bibtex_url
stringlengths 41
50
| proceedings
stringlengths 38
47
| bibtext
stringlengths 709
3.56k
| abstract
stringlengths 17
2.11k
| authors
sequencelengths 1
72
| title
stringlengths 12
207
| id
stringlengths 7
16
| type
stringclasses 2
values | arxiv_id
stringlengths 0
10
| GitHub
sequencelengths 1
1
| paper_page
stringclasses 276
values | n_linked_authors
int64 -1
13
| upvotes
int64 -1
14
| num_comments
int64 -1
11
| n_authors
int64 -1
44
| paper_page_exists_pre_conf
int64 0
1
| Models
sequencelengths 0
100
| Datasets
sequencelengths 0
14
| Spaces
sequencelengths 0
100
|
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
https://aclanthology.org/2023.acl-long.600.bib | https://aclanthology.org/2023.acl-long.600/ | @inproceedings{wu-etal-2023-rethinking,
title = "Rethinking Masked Language Modeling for {C}hinese Spelling Correction",
author = "Wu, Hongqiu and
Zhang, Shaohua and
Zhang, Yuchen and
Zhao, Hai",
editor = "Rogers, Anna and
Boyd-Graber, Jordan and
Okazaki, Naoaki",
booktitle = "Proceedings of the 61st Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers)",
month = jul,
year = "2023",
address = "Toronto, Canada",
publisher = "Association for Computational Linguistics",
url = "https://aclanthology.org/2023.acl-long.600",
doi = "10.18653/v1/2023.acl-long.600",
pages = "10743--10756",
abstract = "In this paper, we study Chinese Spelling Correction (CSC) as a joint decision made by two separate models: a language model and an error model. Through empirical analysis, we find that fine-tuning BERT tends to over-fit the error model while under-fit the language model, resulting in poor generalization to out-of-distribution error patterns. Given that BERT is the backbone of most CSC models, this phenomenon has a significant negative impact. To address this issue, we are releasing a multi-domain benchmark LEMON, with higher quality and diversity than existing benchmarks, to allow a comprehensive assessment of the open domain generalization of CSC models. Then, we demonstrate that a very simple strategy {--} randomly masking 20{\%} non-error tokens from the input sequence during fine-tuning {--} is sufficient for learning a much better language model without sacrificing the error model. This technique can be applied to any model architecture and achieves new state-of-the-art results on SIGHAN, ECSpell, and LEMON.",
}
| In this paper, we study Chinese Spelling Correction (CSC) as a joint decision made by two separate models: a language model and an error model. Through empirical analysis, we find that fine-tuning BERT tends to over-fit the error model while under-fit the language model, resulting in poor generalization to out-of-distribution error patterns. Given that BERT is the backbone of most CSC models, this phenomenon has a significant negative impact. To address this issue, we are releasing a multi-domain benchmark LEMON, with higher quality and diversity than existing benchmarks, to allow a comprehensive assessment of the open domain generalization of CSC models. Then, we demonstrate that a very simple strategy {--} randomly masking 20{\%} non-error tokens from the input sequence during fine-tuning {--} is sufficient for learning a much better language model without sacrificing the error model. This technique can be applied to any model architecture and achieves new state-of-the-art results on SIGHAN, ECSpell, and LEMON. | [
"Wu, Hongqiu",
"Zhang, Shaohua",
"Zhang, Yuchen",
"Zhao, Hai"
] | Rethinking Masked Language Modeling for Chinese Spelling Correction | acl-long.600 | Poster | 2305.17721 | [
"https://github.com/gingasan/lemon"
] | -1 | -1 | -1 | -1 | 0 | [] | [] | [] |
|
https://aclanthology.org/2023.acl-long.601.bib | https://aclanthology.org/2023.acl-long.601/ | @inproceedings{li-etal-2023-multi-modal,
title = "A Multi-Modal Context Reasoning Approach for Conditional Inference on Joint Textual and Visual Clues",
author = "Li, Yunxin and
Hu, Baotian and
Xinyu, Chen and
Ding, Yuxin and
Ma, Lin and
Zhang, Min",
editor = "Rogers, Anna and
Boyd-Graber, Jordan and
Okazaki, Naoaki",
booktitle = "Proceedings of the 61st Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers)",
month = jul,
year = "2023",
address = "Toronto, Canada",
publisher = "Association for Computational Linguistics",
url = "https://aclanthology.org/2023.acl-long.601",
doi = "10.18653/v1/2023.acl-long.601",
pages = "10757--10770",
abstract = "Conditional inference on joint textual and visual clues is a multi-modal reasoning task that textual clues provide prior permutation or external knowledge, which are complementary with visual content and pivotal to deducing the correct option. Previous methods utilizing pretrained vision-language models (VLMs) have achieved impressive performances, yet they show a lack of multimodal context reasoning capability, especially for text-modal information. To address this issue, we propose a Multi-modal Context Reasoning approach, named ModCR. Compared to VLMs performing reasoning via cross modal semantic alignment, it regards the given textual abstract semantic and objective image information as the pre-context information and embeds them into the language model to perform context reasoning. Different from recent vision-aided language models used in natural language processing, ModCR incorporates the multi-view semantic alignment information between language and vision by introducing the learnable alignment prefix between image and text in the pretrained language model. This makes the language model well-suitable for such multi-modal reasoning scenario on joint textual and visual clues. We conduct extensive experiments on two corresponding data sets and experimental results show significantly improved performance (exact gain by 4.8{\%} on PMR test set) compared to previous strong baselines.",
}
| Conditional inference on joint textual and visual clues is a multi-modal reasoning task that textual clues provide prior permutation or external knowledge, which are complementary with visual content and pivotal to deducing the correct option. Previous methods utilizing pretrained vision-language models (VLMs) have achieved impressive performances, yet they show a lack of multimodal context reasoning capability, especially for text-modal information. To address this issue, we propose a Multi-modal Context Reasoning approach, named ModCR. Compared to VLMs performing reasoning via cross modal semantic alignment, it regards the given textual abstract semantic and objective image information as the pre-context information and embeds them into the language model to perform context reasoning. Different from recent vision-aided language models used in natural language processing, ModCR incorporates the multi-view semantic alignment information between language and vision by introducing the learnable alignment prefix between image and text in the pretrained language model. This makes the language model well-suitable for such multi-modal reasoning scenario on joint textual and visual clues. We conduct extensive experiments on two corresponding data sets and experimental results show significantly improved performance (exact gain by 4.8{\%} on PMR test set) compared to previous strong baselines. | [
"Li, Yunxin",
"Hu, Baotian",
"Xinyu, Chen",
"Ding, Yuxin",
"Ma, Lin",
"Zhang, Min"
] | A Multi-Modal Context Reasoning Approach for Conditional Inference on Joint Textual and Visual Clues | acl-long.601 | Poster | 2305.04530 | [
"https://github.com/yunxinli/multimodal-context-reasoning"
] | https://huggingface.co/papers/2305.04530 | 0 | 0 | 0 | 6 | 1 | [
"YunxinLi/ModCR_checkpoints"
] | [] | [] |
https://aclanthology.org/2023.acl-long.602.bib | https://aclanthology.org/2023.acl-long.602/ | @inproceedings{wang-etal-2023-simple,
title = "Simple and Effective Unsupervised Speech Translation",
author = "Wang, Changhan and
Inaguma, Hirofumi and
Chen, Peng-Jen and
Kulikov, Ilia and
Tang, Yun and
Hsu, Wei-Ning and
Auli, Michael and
Pino, Juan",
editor = "Rogers, Anna and
Boyd-Graber, Jordan and
Okazaki, Naoaki",
booktitle = "Proceedings of the 61st Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers)",
month = jul,
year = "2023",
address = "Toronto, Canada",
publisher = "Association for Computational Linguistics",
url = "https://aclanthology.org/2023.acl-long.602",
doi = "10.18653/v1/2023.acl-long.602",
pages = "10771--10784",
abstract = "The amount of labeled data to train models for speech tasks is limited for most languages, however, the data scarcity is exacerbated for speech translation which requires labeled data covering two different languages. To address this issue, we study a simple and effective approach to build speech translation systems without labeled data by leveraging recent advances in unsupervised speech recognition, machine translation and speech synthesis, either in a pipeline approach, or to generate pseudo-labels for training end-to-end speech translation models. Furthermore, we present an unsupervised domain adaptation technique for pre-trained speech models which improves the performance of downstream unsupervised speech recognition, especially for low-resource settings. Experiments show that unsupervised speech-to-text translation outperforms the previous unsupervised state of the art by 3.2 BLEU on the Libri-Trans benchmark, on CoVoST 2, our best systems outperform the best supervised end-to-end models (without pre-training) from only two years ago by an average of 5.0 BLEU over five X-En directions. We also report competitive results on MuST-C and CVSS benchmarks.",
}
| The amount of labeled data to train models for speech tasks is limited for most languages, however, the data scarcity is exacerbated for speech translation which requires labeled data covering two different languages. To address this issue, we study a simple and effective approach to build speech translation systems without labeled data by leveraging recent advances in unsupervised speech recognition, machine translation and speech synthesis, either in a pipeline approach, or to generate pseudo-labels for training end-to-end speech translation models. Furthermore, we present an unsupervised domain adaptation technique for pre-trained speech models which improves the performance of downstream unsupervised speech recognition, especially for low-resource settings. Experiments show that unsupervised speech-to-text translation outperforms the previous unsupervised state of the art by 3.2 BLEU on the Libri-Trans benchmark, on CoVoST 2, our best systems outperform the best supervised end-to-end models (without pre-training) from only two years ago by an average of 5.0 BLEU over five X-En directions. We also report competitive results on MuST-C and CVSS benchmarks. | [
"Wang, Changhan",
"Inaguma, Hirofumi",
"Chen, Peng-Jen",
"Kulikov, Ilia",
"Tang, Yun",
"Hsu, Wei-Ning",
"Auli, Michael",
"Pino, Juan"
] | Simple and Effective Unsupervised Speech Translation | acl-long.602 | Poster | 2210.10191 | [
""
] | -1 | -1 | -1 | -1 | 0 | [] | [] | [] |
|
https://aclanthology.org/2023.acl-long.603.bib | https://aclanthology.org/2023.acl-long.603/ | @inproceedings{do-etal-2023-modeling,
title = "Modeling What-to-ask and How-to-ask for Answer-unaware Conversational Question Generation",
author = "Do, Xuan Long and
Zou, Bowei and
Joty, Shafiq and
Tai, Tran and
Pan, Liangming and
Chen, Nancy and
Aw, Ai Ti",
editor = "Rogers, Anna and
Boyd-Graber, Jordan and
Okazaki, Naoaki",
booktitle = "Proceedings of the 61st Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers)",
month = jul,
year = "2023",
address = "Toronto, Canada",
publisher = "Association for Computational Linguistics",
url = "https://aclanthology.org/2023.acl-long.603",
doi = "10.18653/v1/2023.acl-long.603",
pages = "10785--10803",
abstract = "Conversational Question Generation (CQG) is a critical task for machines to assist humans in fulfilling their information needs through conversations. The task is generally cast into two different settings: answer-aware and answer-unaware. While the former facilitates the models by exposing the expected answer, the latter is more realistic and receiving growing attentions recently. What-to-ask and how-to-ask are the two main challenges in the answer-unaware setting. To address the first challenge, existing methods mainly select sequential sentences in context as the rationales. We argue that the conversation generated using such naive heuristics may not be natural enough as in reality, the interlocutors often talk about the relevant contents that are not necessarily sequential in context. Additionally, previous methods decide the type of question to be generated (boolean/span-based) implicitly. Modeling the question type explicitly is crucial as the answer, which hints the models to generate a boolean or span-based question, is unavailable. To this end, we present SG-CQG, a two-stage CQG framework. For the what-to-ask stage, a sentence is selected as the rationale from a semantic graph that we construct, and extract the answer span from it. For the how-to-ask stage, a classifier determines the target answer type of the question via two explicit control signals before generating and filtering. In addition, we propose Conv-Distinct, a novel evaluation metric for CQG, to evaluate the diversity of the generated conversation from a context. Compared with the existing answer-unaware CQG models, the proposed SG-CQG achieves state-of-the-art performance.",
}
| Conversational Question Generation (CQG) is a critical task for machines to assist humans in fulfilling their information needs through conversations. The task is generally cast into two different settings: answer-aware and answer-unaware. While the former facilitates the models by exposing the expected answer, the latter is more realistic and receiving growing attentions recently. What-to-ask and how-to-ask are the two main challenges in the answer-unaware setting. To address the first challenge, existing methods mainly select sequential sentences in context as the rationales. We argue that the conversation generated using such naive heuristics may not be natural enough as in reality, the interlocutors often talk about the relevant contents that are not necessarily sequential in context. Additionally, previous methods decide the type of question to be generated (boolean/span-based) implicitly. Modeling the question type explicitly is crucial as the answer, which hints the models to generate a boolean or span-based question, is unavailable. To this end, we present SG-CQG, a two-stage CQG framework. For the what-to-ask stage, a sentence is selected as the rationale from a semantic graph that we construct, and extract the answer span from it. For the how-to-ask stage, a classifier determines the target answer type of the question via two explicit control signals before generating and filtering. In addition, we propose Conv-Distinct, a novel evaluation metric for CQG, to evaluate the diversity of the generated conversation from a context. Compared with the existing answer-unaware CQG models, the proposed SG-CQG achieves state-of-the-art performance. | [
"Do, Xuan Long",
"Zou, Bowei",
"Joty, Shafiq",
"Tai, Tran",
"Pan, Liangming",
"Chen, Nancy",
"Aw, Ai Ti"
] | Modeling What-to-ask and How-to-ask for Answer-unaware Conversational Question Generation | acl-long.603 | Poster | 2305.03088 | [
"https://github.com/dxlong2000/sg-cqg"
] | -1 | -1 | -1 | -1 | 0 | [] | [] | [] |
|
https://aclanthology.org/2023.acl-long.604.bib | https://aclanthology.org/2023.acl-long.604/ | @inproceedings{chen-etal-2023-cheer,
title = "{CHEER}: Centrality-aware High-order Event Reasoning Network for Document-level Event Causality Identification",
author = "Chen, Meiqi and
Cao, Yixin and
Zhang, Yan and
Liu, Zhiwei",
editor = "Rogers, Anna and
Boyd-Graber, Jordan and
Okazaki, Naoaki",
booktitle = "Proceedings of the 61st Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers)",
month = jul,
year = "2023",
address = "Toronto, Canada",
publisher = "Association for Computational Linguistics",
url = "https://aclanthology.org/2023.acl-long.604",
doi = "10.18653/v1/2023.acl-long.604",
pages = "10804--10816",
abstract = "Document-level Event Causality Identification (DECI) aims to recognize causal relations between events within a document. Recent studies focus on building a document-level graph for cross-sentence reasoning, but ignore important causal structures {---} there are one or two {``}central{''} events that prevail throughout the document, with most other events serving as either their cause or consequence. In this paper, we manually annotate central events for a systematical investigation and propose a novel DECI model, CHEER, which performs high-order reasoning while considering event centrality. First, we summarize a general GNN-based DECI model and provide a unified view for better understanding. Second, we design an Event Interaction Graph (EIG) involving the interactions among events (e.g., coreference) and event pairs, e.g., causal transitivity, cause(A, B) AND cause(B, C) â cause(A, C). Finally, we incorporate event centrality information into the EIG reasoning network via well-designed features and multi-task learning. We have conducted extensive experiments on two benchmark datasets. The results present great improvements (5.9{\%} F1 gains on average) and demonstrate the effectiveness of each main component.",
}
| Document-level Event Causality Identification (DECI) aims to recognize causal relations between events within a document. Recent studies focus on building a document-level graph for cross-sentence reasoning, but ignore important causal structures {---} there are one or two {``}central{''} events that prevail throughout the document, with most other events serving as either their cause or consequence. In this paper, we manually annotate central events for a systematical investigation and propose a novel DECI model, CHEER, which performs high-order reasoning while considering event centrality. First, we summarize a general GNN-based DECI model and provide a unified view for better understanding. Second, we design an Event Interaction Graph (EIG) involving the interactions among events (e.g., coreference) and event pairs, e.g., causal transitivity, cause(A, B) AND cause(B, C) â cause(A, C). Finally, we incorporate event centrality information into the EIG reasoning network via well-designed features and multi-task learning. We have conducted extensive experiments on two benchmark datasets. The results present great improvements (5.9{\%} F1 gains on average) and demonstrate the effectiveness of each main component. | [
"Chen, Meiqi",
"Cao, Yixin",
"Zhang, Yan",
"Liu, Zhiwei"
] | CHEER: Centrality-aware High-order Event Reasoning Network for Document-level Event Causality Identification | acl-long.604 | Poster | [
""
] | -1 | -1 | -1 | -1 | 0 | [] | [] | [] |
||
https://aclanthology.org/2023.acl-long.605.bib | https://aclanthology.org/2023.acl-long.605/ | @inproceedings{wen-etal-2023-f,
title = "f-Divergence Minimization for Sequence-Level Knowledge Distillation",
author = "Wen, Yuqiao and
Li, Zichao and
Du, Wenyu and
Mou, Lili",
editor = "Rogers, Anna and
Boyd-Graber, Jordan and
Okazaki, Naoaki",
booktitle = "Proceedings of the 61st Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers)",
month = jul,
year = "2023",
address = "Toronto, Canada",
publisher = "Association for Computational Linguistics",
url = "https://aclanthology.org/2023.acl-long.605",
doi = "10.18653/v1/2023.acl-long.605",
pages = "10817--10834",
abstract = "Knowledge distillation (KD) is the process of transferring knowledge from a large model to a small one. It has gained increasing attention in the natural language processing community, driven by the demands of compressing ever-growing language models. In this work, we propose an FDISTILL framework, which formulates sequence-level knowledge distillation as minimizing a generalized f-divergence function. We propose four distilling variants under our framework and show that existing SeqKD and ENGINE approaches are approximations of our FDISTILL methods. We further derive step-wise decomposition for our FDISTILL, reducing intractable sequence-level divergence to word-level losses that can be computed in a tractable manner. Experiments across four datasets show that our methods outperform existing KD approaches, and that our symmetric distilling losses can better force the student to learn from the teacher distribution.",
}
| Knowledge distillation (KD) is the process of transferring knowledge from a large model to a small one. It has gained increasing attention in the natural language processing community, driven by the demands of compressing ever-growing language models. In this work, we propose an FDISTILL framework, which formulates sequence-level knowledge distillation as minimizing a generalized f-divergence function. We propose four distilling variants under our framework and show that existing SeqKD and ENGINE approaches are approximations of our FDISTILL methods. We further derive step-wise decomposition for our FDISTILL, reducing intractable sequence-level divergence to word-level losses that can be computed in a tractable manner. Experiments across four datasets show that our methods outperform existing KD approaches, and that our symmetric distilling losses can better force the student to learn from the teacher distribution. | [
"Wen, Yuqiao",
"Li, Zichao",
"Du, Wenyu",
"Mou, Lili"
] | f-Divergence Minimization for Sequence-Level Knowledge Distillation | acl-long.605 | Oral | 2307.15190 | [
"https://github.com/manga-uofa/fdistill"
] | -1 | -1 | -1 | -1 | 0 | [] | [] | [] |
|
https://aclanthology.org/2023.acl-long.606.bib | https://aclanthology.org/2023.acl-long.606/ | @inproceedings{hu-etal-2023-supervised,
title = "Supervised Adversarial Contrastive Learning for Emotion Recognition in Conversations",
author = "Hu, Dou and
Bao, Yinan and
Wei, Lingwei and
Zhou, Wei and
Hu, Songlin",
editor = "Rogers, Anna and
Boyd-Graber, Jordan and
Okazaki, Naoaki",
booktitle = "Proceedings of the 61st Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers)",
month = jul,
year = "2023",
address = "Toronto, Canada",
publisher = "Association for Computational Linguistics",
url = "https://aclanthology.org/2023.acl-long.606",
doi = "10.18653/v1/2023.acl-long.606",
pages = "10835--10852",
abstract = "Extracting generalized and robust representations is a major challenge in emotion recognition in conversations (ERC). To address this, we propose a supervised adversarial contrastive learning (SACL) framework for learning class-spread structured representations in a supervised manner. SACL applies contrast-aware adversarial training to generate worst-case samples and uses joint class-spread contrastive learning to extract structured representations. It can effectively utilize label-level feature consistency and retain fine-grained intra-class features. To avoid the negative impact of adversarial perturbations on context-dependent data, we design a contextual adversarial training (CAT) strategy to learn more diverse features from context and enhance the model{'}s context robustness. Under the framework with CAT, we develop a sequence-based SACL-LSTM to learn label-consistent and context-robust features for ERC. Experiments on three datasets show that SACL-LSTM achieves state-of-the-art performance on ERC. Extended experiments prove the effectiveness of SACL and CAT.",
}
| Extracting generalized and robust representations is a major challenge in emotion recognition in conversations (ERC). To address this, we propose a supervised adversarial contrastive learning (SACL) framework for learning class-spread structured representations in a supervised manner. SACL applies contrast-aware adversarial training to generate worst-case samples and uses joint class-spread contrastive learning to extract structured representations. It can effectively utilize label-level feature consistency and retain fine-grained intra-class features. To avoid the negative impact of adversarial perturbations on context-dependent data, we design a contextual adversarial training (CAT) strategy to learn more diverse features from context and enhance the model{'}s context robustness. Under the framework with CAT, we develop a sequence-based SACL-LSTM to learn label-consistent and context-robust features for ERC. Experiments on three datasets show that SACL-LSTM achieves state-of-the-art performance on ERC. Extended experiments prove the effectiveness of SACL and CAT. | [
"Hu, Dou",
"Bao, Yinan",
"Wei, Lingwei",
"Zhou, Wei",
"Hu, Songlin"
] | Supervised Adversarial Contrastive Learning for Emotion Recognition in Conversations | acl-long.606 | Poster | 2306.01505 | [
"https://github.com/zerohd4869/sacl"
] | -1 | -1 | -1 | -1 | 0 | [] | [] | [] |
|
https://aclanthology.org/2023.acl-long.607.bib | https://aclanthology.org/2023.acl-long.607/ | @inproceedings{zhang-etal-2023-novel,
title = "A Novel Table-to-Graph Generation Approach for Document-Level Joint Entity and Relation Extraction",
author = "Zhang, Ruoyu and
Li, Yanzeng and
Zou, Lei",
editor = "Rogers, Anna and
Boyd-Graber, Jordan and
Okazaki, Naoaki",
booktitle = "Proceedings of the 61st Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers)",
month = jul,
year = "2023",
address = "Toronto, Canada",
publisher = "Association for Computational Linguistics",
url = "https://aclanthology.org/2023.acl-long.607",
doi = "10.18653/v1/2023.acl-long.607",
pages = "10853--10865",
abstract = "Document-level relation extraction (DocRE) aims to extract relations among entities within a document, which is crucial for applications like knowledge graph construction. Existing methods usually assume that entities and their mentions are identified beforehand, which falls short of real-world applications. To overcome this limitation, we propose TaG, a novel table-to-graph generation model for joint extractionof entities and relations at document-level. To enhance the learning of task dependencies, TaG induces a latent graph among mentions, with different types of edges indicating different task information, which is further broadcast with a relational graph convolutional network. To alleviate the error propagation problem, we adapt the hierarchical agglomerative clustering algorithm to back-propagate task information at decoding stage. Experiments on the benchmark dataset, DocRED, demonstrate that TaG surpasses previous methods by a large margin and achieves state-of-the-art results.",
}
| Document-level relation extraction (DocRE) aims to extract relations among entities within a document, which is crucial for applications like knowledge graph construction. Existing methods usually assume that entities and their mentions are identified beforehand, which falls short of real-world applications. To overcome this limitation, we propose TaG, a novel table-to-graph generation model for joint extractionof entities and relations at document-level. To enhance the learning of task dependencies, TaG induces a latent graph among mentions, with different types of edges indicating different task information, which is further broadcast with a relational graph convolutional network. To alleviate the error propagation problem, we adapt the hierarchical agglomerative clustering algorithm to back-propagate task information at decoding stage. Experiments on the benchmark dataset, DocRED, demonstrate that TaG surpasses previous methods by a large margin and achieves state-of-the-art results. | [
"Zhang, Ruoyu",
"Li, Yanzeng",
"Zou, Lei"
] | A Novel Table-to-Graph Generation Approach for Document-Level Joint Entity and Relation Extraction | acl-long.607 | Poster | [
""
] | -1 | -1 | -1 | -1 | 0 | [] | [] | [] |
||
https://aclanthology.org/2023.acl-long.608.bib | https://aclanthology.org/2023.acl-long.608/ | @inproceedings{bao-etal-2023-synthetic,
title = "A Synthetic Data Generation Framework for Grounded Dialogues",
author = "Bao, Jianzhu and
Wang, Rui and
Wang, Yasheng and
Sun, Aixin and
Li, Yitong and
Mi, Fei and
Xu, Ruifeng",
editor = "Rogers, Anna and
Boyd-Graber, Jordan and
Okazaki, Naoaki",
booktitle = "Proceedings of the 61st Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers)",
month = jul,
year = "2023",
address = "Toronto, Canada",
publisher = "Association for Computational Linguistics",
url = "https://aclanthology.org/2023.acl-long.608",
doi = "10.18653/v1/2023.acl-long.608",
pages = "10866--10882",
abstract = "Training grounded response generation models often requires a large collection of grounded dialogues. However, it is costly to build such dialogues. In this paper, we present a synthetic data generation framework (SynDG) for grounded dialogues. The generation process utilizes large pre-trained language models and freely available knowledge data (e.g., Wikipedia pages, persona profiles, etc.). The key idea of designing SynDG is to consider dialogue flow and coherence in the generation process. Specifically, given knowledge data, we first heuristically determine a dialogue flow, which is a series of knowledge pieces. Then, we employ T5 to incrementally turn the dialogue flow into a dialogue. To ensure coherence of both the dialogue flow and the synthetic dialogue, we design a two-level filtering strategy, at the flow-level and the utterance-level respectively. Experiments on two public benchmarks show that the synthetic grounded dialogue data produced by our framework is able to significantly boost model performance in both full training data and low-resource scenarios.",
}
| Training grounded response generation models often requires a large collection of grounded dialogues. However, it is costly to build such dialogues. In this paper, we present a synthetic data generation framework (SynDG) for grounded dialogues. The generation process utilizes large pre-trained language models and freely available knowledge data (e.g., Wikipedia pages, persona profiles, etc.). The key idea of designing SynDG is to consider dialogue flow and coherence in the generation process. Specifically, given knowledge data, we first heuristically determine a dialogue flow, which is a series of knowledge pieces. Then, we employ T5 to incrementally turn the dialogue flow into a dialogue. To ensure coherence of both the dialogue flow and the synthetic dialogue, we design a two-level filtering strategy, at the flow-level and the utterance-level respectively. Experiments on two public benchmarks show that the synthetic grounded dialogue data produced by our framework is able to significantly boost model performance in both full training data and low-resource scenarios. | [
"Bao, Jianzhu",
"Wang, Rui",
"Wang, Yasheng",
"Sun, Aixin",
"Li, Yitong",
"Mi, Fei",
"Xu, Ruifeng"
] | A Synthetic Data Generation Framework for Grounded Dialogues | acl-long.608 | Poster | [
""
] | -1 | -1 | -1 | -1 | 0 | [] | [] | [] |
||
https://aclanthology.org/2023.acl-long.609.bib | https://aclanthology.org/2023.acl-long.609/ | @inproceedings{dione-etal-2023-masakhapos,
title = "{M}asakha{POS}: Part-of-Speech Tagging for Typologically Diverse {A}frican languages",
author = "Dione, Cheikh M. Bamba and
Adelani, David Ifeoluwa and
Nabende, Peter and
Alabi, Jesujoba and
Sindane, Thapelo and
Buzaaba, Happy and
Muhammad, Shamsuddeen Hassan and
Emezue, Chris Chinenye and
Ogayo, Perez and
Aremu, Anuoluwapo and
Gitau, Catherine and
Mbaye, Derguene and
Mukiibi, Jonathan and
Sibanda, Blessing and
Dossou, Bonaventure F. P. and
Bukula, Andiswa and
Mabuya, Rooweither and
Tapo, Allahsera Auguste and
Munkoh-Buabeng, Edwin and
Memdjokam Koagne, Victoire and
Ouoba Kabore, Fatoumata and
Taylor, Amelia and
Kalipe, Godson and
Macucwa, Tebogo and
Marivate, Vukosi and
Gwadabe, Tajuddeen and
Elvis, Mboning Tchiaze and
Onyenwe, Ikechukwu and
Atindogbe, Gratien and
Adelani, Tolulope and
Akinade, Idris and
Samuel, Olanrewaju and
Nahimana, Marien and
Musabeyezu, Th{\'e}og{\`e}ne and
Niyomutabazi, Emile and
Chimhenga, Ester and
Gotosa, Kudzai and
Mizha, Patrick and
Agbolo, Apelete and
Traore, Seydou and
Uchechukwu, Chinedu and
Yusuf, Aliyu and
Abdullahi, Muhammad and
Klakow, Dietrich",
editor = "Rogers, Anna and
Boyd-Graber, Jordan and
Okazaki, Naoaki",
booktitle = "Proceedings of the 61st Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers)",
month = jul,
year = "2023",
address = "Toronto, Canada",
publisher = "Association for Computational Linguistics",
url = "https://aclanthology.org/2023.acl-long.609",
doi = "10.18653/v1/2023.acl-long.609",
pages = "10883--10900",
abstract = "In this paper, we present AfricaPOS, the largest part-of-speech (POS) dataset for 20 typologically diverse African languages. We discuss the challenges in annotating POS for these languages using the universal dependencies (UD) guidelines. We conducted extensive POS baseline experiments using both conditional random field and several multilingual pre-trained language models. We applied various cross-lingual transfer models trained with data available in the UD. Evaluating on the AfricaPOS dataset, we show that choosing the best transfer language(s) in both single-source and multi-source setups greatly improves the POS tagging performance of the target languages, in particular when combined with parameter-fine-tuning methods. Crucially, transferring knowledge from a language that matches the language family and morphosyntactic properties seems to be more effective for POS tagging in unseen languages.",
}
| In this paper, we present AfricaPOS, the largest part-of-speech (POS) dataset for 20 typologically diverse African languages. We discuss the challenges in annotating POS for these languages using the universal dependencies (UD) guidelines. We conducted extensive POS baseline experiments using both conditional random field and several multilingual pre-trained language models. We applied various cross-lingual transfer models trained with data available in the UD. Evaluating on the AfricaPOS dataset, we show that choosing the best transfer language(s) in both single-source and multi-source setups greatly improves the POS tagging performance of the target languages, in particular when combined with parameter-fine-tuning methods. Crucially, transferring knowledge from a language that matches the language family and morphosyntactic properties seems to be more effective for POS tagging in unseen languages. | [
"Dione, Cheikh M. Bamba",
"Adelani, David Ifeoluwa",
"Nabende, Peter",
"Alabi, Jesujoba",
"Sindane, Thapelo",
"Buzaaba, Happy",
"Muhammad, Shamsuddeen Hassan",
"Emezue, Chris Chinenye",
"Ogayo, Perez",
"Aremu, Anuoluwapo",
"Gitau, Catherine",
"Mbaye, Derguene",
"Mukiibi, Jonathan",
"Sib",
"a, Blessing",
"Dossou, Bonaventure F. P.",
"Bukula, Andiswa",
"Mabuya, Rooweither",
"Tapo, Allahsera Auguste",
"Munkoh-Buabeng, Edwin",
"Memdjokam Koagne, Victoire",
"Ouoba Kabore, Fatoumata",
"Taylor, Amelia",
"Kalipe, Godson",
"Macucwa, Tebogo",
"Marivate, Vukosi",
"Gwadabe, Tajuddeen",
"Elvis, Mboning Tchiaze",
"Onyenwe, Ikechukwu",
"Atindogbe, Gratien",
"Adelani, Tolulope",
"Akinade, Idris",
"Samuel, Olanrewaju",
"Nahimana, Marien",
"Musabeyezu, Th{\\'e}og{\\`e}ne",
"Niyomutabazi, Emile",
"Chimhenga, Ester",
"Gotosa, Kudzai",
"Mizha, Patrick",
"Agbolo, Apelete",
"Traore, Seydou",
"Uchechukwu, Chinedu",
"Yusuf, Aliyu",
"Abdullahi, Muhammad",
"Klakow, Dietrich"
] | MasakhaPOS: Part-of-Speech Tagging for Typologically Diverse African languages | acl-long.609 | Oral | 2305.13989 | [
"https://github.com/masakhane-io/masakhane-pos"
] | https://huggingface.co/papers/2305.13989 | 1 | 0 | 0 | 44 | 1 | [] | [] | [] |
https://aclanthology.org/2023.acl-long.610.bib | https://aclanthology.org/2023.acl-long.610/ | @inproceedings{hu-etal-2023-semantic,
title = "Semantic Structure Enhanced Event Causality Identification",
author = "Hu, Zhilei and
Li, Zixuan and
Jin, Xiaolong and
Bai, Long and
Guan, Saiping and
Guo, Jiafeng and
Cheng, Xueqi",
editor = "Rogers, Anna and
Boyd-Graber, Jordan and
Okazaki, Naoaki",
booktitle = "Proceedings of the 61st Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers)",
month = jul,
year = "2023",
address = "Toronto, Canada",
publisher = "Association for Computational Linguistics",
url = "https://aclanthology.org/2023.acl-long.610",
doi = "10.18653/v1/2023.acl-long.610",
pages = "10901--10913",
abstract = "Event Causality Identification (ECI) aims to identify causal relations between events in unstructured texts. This is a very challenging task, because causal relations are usually expressed by implicit associations between events. Existing methods usually capture such associations by directly modeling the texts with pre-trained language models, which underestimate two kinds of semantic structures vital to the ECI task, namely, event-centric structure and event-associated structure. The former includes important semantic elements related to the events to describe them more precisely, while the latter contains semantic paths between two events to provide possible supports for ECI. In this paper, we study the implicit associations between events by modeling the above explicit semantic structures, and propose a Semantic Structure Integration model (SemSIn).It utilizes a GNN-based event aggregator to integrate the event-centric structure information, and employs an LSTM-based path aggregator to capture the event-associated structure information between two events. Experimental results on three widely used datasets show that SemSIn achieves significant improvements over baseline methods.",
}
| Event Causality Identification (ECI) aims to identify causal relations between events in unstructured texts. This is a very challenging task, because causal relations are usually expressed by implicit associations between events. Existing methods usually capture such associations by directly modeling the texts with pre-trained language models, which underestimate two kinds of semantic structures vital to the ECI task, namely, event-centric structure and event-associated structure. The former includes important semantic elements related to the events to describe them more precisely, while the latter contains semantic paths between two events to provide possible supports for ECI. In this paper, we study the implicit associations between events by modeling the above explicit semantic structures, and propose a Semantic Structure Integration model (SemSIn).It utilizes a GNN-based event aggregator to integrate the event-centric structure information, and employs an LSTM-based path aggregator to capture the event-associated structure information between two events. Experimental results on three widely used datasets show that SemSIn achieves significant improvements over baseline methods. | [
"Hu, Zhilei",
"Li, Zixuan",
"Jin, Xiaolong",
"Bai, Long",
"Guan, Saiping",
"Guo, Jiafeng",
"Cheng, Xueqi"
] | Semantic Structure Enhanced Event Causality Identification | acl-long.610 | Poster | 2305.12792 | [
""
] | -1 | -1 | -1 | -1 | 0 | [] | [] | [] |
|
https://aclanthology.org/2023.acl-long.611.bib | https://aclanthology.org/2023.acl-long.611/ | @inproceedings{wang-etal-2023-weakly,
title = "Weakly-Supervised Spoken Video Grounding via Semantic Interaction Learning",
author = "Wang, Ye and
Lin, Wang and
Zhang, Shengyu and
Jin, Tao and
Li, Linjun and
Cheng, Xize and
Zhao, Zhou",
editor = "Rogers, Anna and
Boyd-Graber, Jordan and
Okazaki, Naoaki",
booktitle = "Proceedings of the 61st Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers)",
month = jul,
year = "2023",
address = "Toronto, Canada",
publisher = "Association for Computational Linguistics",
url = "https://aclanthology.org/2023.acl-long.611",
doi = "10.18653/v1/2023.acl-long.611",
pages = "10914--10932",
abstract = "The task of spoken video grounding aims to localize moments in videos that are relevant to descriptive spoken queries. However, extracting semantic information from speech and modeling the cross-modal correlation pose two critical challenges. Previous studies solve them by representing spoken queries based on the matched video frames, which require tremendous effort for frame-level labeling. In this work, we investigate weakly-supervised spoken video grounding, i.e., learning to localize moments without expensive temporal annotations. To effectively represent the cross-modal semantics, we propose Semantic Interaction Learning (SIL), a novel framework consisting of the acoustic-semantic pre-training (ASP) and acoustic-visual contrastive learning (AVCL). In ASP, we pre-train an effective encoder for the grounding task with three comprehensive tasks, where the robustness task enhances stability by explicitly capturing the invariance between time- and frequency-domain features, the conciseness task avoids over-smooth attention by compressing long sequence into segments, and the semantic task improves spoken language understanding by modeling the precise semantics. In AVCL, we mine pseudo labels with discriminative sampling strategies and directly strengthen the interaction between speech and video by maximizing their mutual information. Extensive experiments demonstrate the effectiveness and superiority of our method.",
}
| The task of spoken video grounding aims to localize moments in videos that are relevant to descriptive spoken queries. However, extracting semantic information from speech and modeling the cross-modal correlation pose two critical challenges. Previous studies solve them by representing spoken queries based on the matched video frames, which require tremendous effort for frame-level labeling. In this work, we investigate weakly-supervised spoken video grounding, i.e., learning to localize moments without expensive temporal annotations. To effectively represent the cross-modal semantics, we propose Semantic Interaction Learning (SIL), a novel framework consisting of the acoustic-semantic pre-training (ASP) and acoustic-visual contrastive learning (AVCL). In ASP, we pre-train an effective encoder for the grounding task with three comprehensive tasks, where the robustness task enhances stability by explicitly capturing the invariance between time- and frequency-domain features, the conciseness task avoids over-smooth attention by compressing long sequence into segments, and the semantic task improves spoken language understanding by modeling the precise semantics. In AVCL, we mine pseudo labels with discriminative sampling strategies and directly strengthen the interaction between speech and video by maximizing their mutual information. Extensive experiments demonstrate the effectiveness and superiority of our method. | [
"Wang, Ye",
"Lin, Wang",
"Zhang, Shengyu",
"Jin, Tao",
"Li, Linjun",
"Cheng, Xize",
"Zhao, Zhou"
] | Weakly-Supervised Spoken Video Grounding via Semantic Interaction Learning | acl-long.611 | Oral | [
""
] | -1 | -1 | -1 | -1 | 0 | [] | [] | [] |
||
https://aclanthology.org/2023.acl-long.612.bib | https://aclanthology.org/2023.acl-long.612/ | @inproceedings{wang-etal-2023-rehearsal,
title = "Rehearsal-free Continual Language Learning via Efficient Parameter Isolation",
author = "Wang, Zhicheng and
Liu, Yufang and
Ji, Tao and
Wang, Xiaoling and
Wu, Yuanbin and
Jiang, Congcong and
Chao, Ye and
Han, Zhencong and
Wang, Ling and
Shao, Xu and
Zeng, Wenqiu",
editor = "Rogers, Anna and
Boyd-Graber, Jordan and
Okazaki, Naoaki",
booktitle = "Proceedings of the 61st Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers)",
month = jul,
year = "2023",
address = "Toronto, Canada",
publisher = "Association for Computational Linguistics",
url = "https://aclanthology.org/2023.acl-long.612",
doi = "10.18653/v1/2023.acl-long.612",
pages = "10933--10946",
abstract = "We study the problem of defying catastrophic forgetting when learning a series of language processing tasks. Compared with previous methods, we emphasize the importance of not caching history tasks{'} data, which makes the problem more challenging. Our proposed method applies the parameter isolation strategy. For each task, it allocates a small portion of private parameters and learns them with a shared pre-trained model. To load correct parameters at testing time, we introduce a simple yet effective non-parametric method. Experiments on continual language learning benchmarks show that our method is significantly better than all existing no-data-cache methods, and is comparable (or even better) than those using historical data.",
}
| We study the problem of defying catastrophic forgetting when learning a series of language processing tasks. Compared with previous methods, we emphasize the importance of not caching history tasks{'} data, which makes the problem more challenging. Our proposed method applies the parameter isolation strategy. For each task, it allocates a small portion of private parameters and learns them with a shared pre-trained model. To load correct parameters at testing time, we introduce a simple yet effective non-parametric method. Experiments on continual language learning benchmarks show that our method is significantly better than all existing no-data-cache methods, and is comparable (or even better) than those using historical data. | [
"Wang, Zhicheng",
"Liu, Yufang",
"Ji, Tao",
"Wang, Xiaoling",
"Wu, Yuanbin",
"Jiang, Congcong",
"Chao, Ye",
"Han, Zhencong",
"Wang, Ling",
"Shao, Xu",
"Zeng, Wenqiu"
] | Rehearsal-free Continual Language Learning via Efficient Parameter Isolation | acl-long.612 | Poster | [
""
] | -1 | -1 | -1 | -1 | 0 | [] | [] | [] |
||
https://aclanthology.org/2023.acl-long.613.bib | https://aclanthology.org/2023.acl-long.613/ | @inproceedings{chen-etal-2023-label,
title = "Label-Aware Hyperbolic Embeddings for Fine-grained Emotion Classification",
author = "Chen, Chih Yao and
Hung, Tun Min and
Hsu, Yi-Li and
Ku, Lun-Wei",
editor = "Rogers, Anna and
Boyd-Graber, Jordan and
Okazaki, Naoaki",
booktitle = "Proceedings of the 61st Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers)",
month = jul,
year = "2023",
address = "Toronto, Canada",
publisher = "Association for Computational Linguistics",
url = "https://aclanthology.org/2023.acl-long.613",
doi = "10.18653/v1/2023.acl-long.613",
pages = "10947--10958",
abstract = "Fine-grained emotion classification (FEC) is a challenging task. Specifically, FEC needs to handle subtle nuance between labels, which can be complex and confusing. Most existing models only address text classification problem in the euclidean space, which we believe may not be the optimal solution as labels of close semantic (e.g., afraid and terrified) may not be differentiated in such space, which harms the performance. In this paper, we propose HypEmo, a novel framework that can integrate hyperbolic embeddings to improve the FEC task. First, we learn label embeddings in the hyperbolic space to better capture their hierarchical structure, and then our model projects contextualized representations to the hyperbolic space to compute the distance between samples and labels. Experimental results show that incorporating such distance to weight cross entropy loss substantially improve the performance on two benchmark datasets, with around 3{\%} improvement compared to previous state-of-the-art, and could even improve up to 8.6{\%} when the labels are hard to distinguish. Code is available at \url{https://github.com/dinobby/HypEmo}.",
}
| Fine-grained emotion classification (FEC) is a challenging task. Specifically, FEC needs to handle subtle nuance between labels, which can be complex and confusing. Most existing models only address text classification problem in the euclidean space, which we believe may not be the optimal solution as labels of close semantic (e.g., afraid and terrified) may not be differentiated in such space, which harms the performance. In this paper, we propose HypEmo, a novel framework that can integrate hyperbolic embeddings to improve the FEC task. First, we learn label embeddings in the hyperbolic space to better capture their hierarchical structure, and then our model projects contextualized representations to the hyperbolic space to compute the distance between samples and labels. Experimental results show that incorporating such distance to weight cross entropy loss substantially improve the performance on two benchmark datasets, with around 3{\%} improvement compared to previous state-of-the-art, and could even improve up to 8.6{\%} when the labels are hard to distinguish. Code is available at \url{https://github.com/dinobby/HypEmo}. | [
"Chen, Chih Yao",
"Hung, Tun Min",
"Hsu, Yi-Li",
"Ku, Lun-Wei"
] | Label-Aware Hyperbolic Embeddings for Fine-grained Emotion Classification | acl-long.613 | Poster | 2306.14822 | [
"https://github.com/dinobby/hypemo"
] | -1 | -1 | -1 | -1 | 0 | [] | [] | [] |
|
https://aclanthology.org/2023.acl-long.614.bib | https://aclanthology.org/2023.acl-long.614/ | @inproceedings{si-etal-2023-combo,
title = "Combo of Thinking and Observing for Outside-Knowledge {VQA}",
author = "Si, Qingyi and
Mo, Yuchen and
Lin, Zheng and
Ji, Huishan and
Wang, Weiping",
editor = "Rogers, Anna and
Boyd-Graber, Jordan and
Okazaki, Naoaki",
booktitle = "Proceedings of the 61st Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers)",
month = jul,
year = "2023",
address = "Toronto, Canada",
publisher = "Association for Computational Linguistics",
url = "https://aclanthology.org/2023.acl-long.614",
doi = "10.18653/v1/2023.acl-long.614",
pages = "10959--10975",
abstract = "Outside-knowledge visual question answering is a challenging task that requires both the acquisition and the use of open-ended real-world knowledge. Some existing solutions draw external knowledge into the cross-modality space which overlooks the much vaster textual knowledge in natural-language space, while others transform the image into a text which further fuses with the textual knowledge into the natural-language space and completely abandons the use of visual features. In this paper, we are inspired to constrain the cross-modality space into the same space of natural-language space which makes the visual features preserved directly, and the model still benefits from the vast knowledge in natural-language space. To this end, we propose a novel framework consisting of a multimodal encoder, a textual encoder and an answer decoder. Such structure allows us to introduce more types of knowledge including explicit and implicit multimodal and textual knowledge. Extensive experiments validate the superiority of the proposed method which outperforms the state-of-the-art by 6.17{\%} accuracy. We also conduct comprehensive ablations of each component, and systematically study the roles of varying types of knowledge. Codes and knowledge data are to be released.",
}
| Outside-knowledge visual question answering is a challenging task that requires both the acquisition and the use of open-ended real-world knowledge. Some existing solutions draw external knowledge into the cross-modality space which overlooks the much vaster textual knowledge in natural-language space, while others transform the image into a text which further fuses with the textual knowledge into the natural-language space and completely abandons the use of visual features. In this paper, we are inspired to constrain the cross-modality space into the same space of natural-language space which makes the visual features preserved directly, and the model still benefits from the vast knowledge in natural-language space. To this end, we propose a novel framework consisting of a multimodal encoder, a textual encoder and an answer decoder. Such structure allows us to introduce more types of knowledge including explicit and implicit multimodal and textual knowledge. Extensive experiments validate the superiority of the proposed method which outperforms the state-of-the-art by 6.17{\%} accuracy. We also conduct comprehensive ablations of each component, and systematically study the roles of varying types of knowledge. Codes and knowledge data are to be released. | [
"Si, Qingyi",
"Mo, Yuchen",
"Lin, Zheng",
"Ji, Huishan",
"Wang, Weiping"
] | Combo of Thinking and Observing for Outside-Knowledge VQA | acl-long.614 | Poster | 2305.06407 | [
"https://github.com/phoebussi/thinking-while-observing"
] | -1 | -1 | -1 | -1 | 0 | [] | [] | [] |
|
https://aclanthology.org/2023.acl-long.615.bib | https://aclanthology.org/2023.acl-long.615/ | @inproceedings{hsu-etal-2023-ampere,
title = "{AMPERE}: {AMR}-Aware Prefix for Generation-Based Event Argument Extraction Model",
author = "Hsu, I-Hung and
Xie, Zhiyu and
Huang, Kuan-Hao and
Natarajan, Prem and
Peng, Nanyun",
editor = "Rogers, Anna and
Boyd-Graber, Jordan and
Okazaki, Naoaki",
booktitle = "Proceedings of the 61st Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers)",
month = jul,
year = "2023",
address = "Toronto, Canada",
publisher = "Association for Computational Linguistics",
url = "https://aclanthology.org/2023.acl-long.615",
doi = "10.18653/v1/2023.acl-long.615",
pages = "10976--10993",
abstract = "Event argument extraction (EAE) identifies event arguments and their specific roles for a given event. Recent advancement in generation-based EAE models has shown great performance and generalizability over classification-based models. However, existing generation-based EAE models mostly focus on problem re-formulation and prompt design, without incorporating additional information that has been shown to be effective for classification-based models, such as the abstract meaning representation (AMR) of the input passages. Incorporating such information into generation-based models is challenging due to the heterogeneous nature of the natural language form prevalently used in generation-based models and the structured form of AMRs. In this work, we study strategies to incorporate AMR into generation-based EAE models. We propose AMPERE, which generates AMR-aware prefixes for every layer of the generation model. Thus, the prefix introduces AMR information to the generation-based EAE model and then improves the generation. We also introduce an adjusted copy mechanism to AMPERE to help overcome potential noises brought by the AMR graph. Comprehensive experiments and analyses on ACE2005 and ERE datasets show that AMPERE can get 4{\%} - 10{\%} absolute F1 score improvements with reduced training data and it is in general powerful across different training sizes.",
}
| Event argument extraction (EAE) identifies event arguments and their specific roles for a given event. Recent advancement in generation-based EAE models has shown great performance and generalizability over classification-based models. However, existing generation-based EAE models mostly focus on problem re-formulation and prompt design, without incorporating additional information that has been shown to be effective for classification-based models, such as the abstract meaning representation (AMR) of the input passages. Incorporating such information into generation-based models is challenging due to the heterogeneous nature of the natural language form prevalently used in generation-based models and the structured form of AMRs. In this work, we study strategies to incorporate AMR into generation-based EAE models. We propose AMPERE, which generates AMR-aware prefixes for every layer of the generation model. Thus, the prefix introduces AMR information to the generation-based EAE model and then improves the generation. We also introduce an adjusted copy mechanism to AMPERE to help overcome potential noises brought by the AMR graph. Comprehensive experiments and analyses on ACE2005 and ERE datasets show that AMPERE can get 4{\%} - 10{\%} absolute F1 score improvements with reduced training data and it is in general powerful across different training sizes. | [
"Hsu, I-Hung",
"Xie, Zhiyu",
"Huang, Kuan-Hao",
"Natarajan, Prem",
"Peng, Nanyun"
] | AMPERE: AMR-Aware Prefix for Generation-Based Event Argument Extraction Model | acl-long.615 | Poster | 2305.16734 | [
"https://github.com/pluslabnlp/ampere"
] | -1 | -1 | -1 | -1 | 0 | [] | [] | [] |
|
https://aclanthology.org/2023.acl-long.616.bib | https://aclanthology.org/2023.acl-long.616/ | @inproceedings{jurgens-etal-2023-spouse,
title = "Your spouse needs professional help: Determining the Contextual Appropriateness of Messages through Modeling Social Relationships",
author = "Jurgens, David and
Seth, Agrima and
Sargent, Jackson and
Aghighi, Athena and
Geraci, Michael",
editor = "Rogers, Anna and
Boyd-Graber, Jordan and
Okazaki, Naoaki",
booktitle = "Proceedings of the 61st Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers)",
month = jul,
year = "2023",
address = "Toronto, Canada",
publisher = "Association for Computational Linguistics",
url = "https://aclanthology.org/2023.acl-long.616",
doi = "10.18653/v1/2023.acl-long.616",
pages = "10994--11013",
abstract = "Understanding interpersonal communication requires, in part, understanding the social context and norms in which a message is said. However, current methods for identifying offensive content in such communication largely operate independent of context, with only a few approaches considering community norms or prior conversation as context. Here, we introduce a new approach to identifying inappropriate communication by explicitly modeling the social relationship between the individuals. We introduce a new dataset of contextually-situated judgments of appropriateness and show that large language models can readily incorporate relationship information to accurately identify appropriateness in a given context. Using data from online conversations and movie dialogues, we provide insight into how the relationships themselves function as implicit norms and quantify the degree to which context-sensitivity is needed in different conversation settings. Further, we also demonstrate that contextual-appropriateness judgments are predictive of other social factors expressed in language such as condescension and politeness.",
}
| Understanding interpersonal communication requires, in part, understanding the social context and norms in which a message is said. However, current methods for identifying offensive content in such communication largely operate independent of context, with only a few approaches considering community norms or prior conversation as context. Here, we introduce a new approach to identifying inappropriate communication by explicitly modeling the social relationship between the individuals. We introduce a new dataset of contextually-situated judgments of appropriateness and show that large language models can readily incorporate relationship information to accurately identify appropriateness in a given context. Using data from online conversations and movie dialogues, we provide insight into how the relationships themselves function as implicit norms and quantify the degree to which context-sensitivity is needed in different conversation settings. Further, we also demonstrate that contextual-appropriateness judgments are predictive of other social factors expressed in language such as condescension and politeness. | [
"Jurgens, David",
"Seth, Agrima",
"Sargent, Jackson",
"Aghighi, Athena",
"Geraci, Michael"
] | Your spouse needs professional help: Determining the Contextual Appropriateness of Messages through Modeling Social Relationships | acl-long.616 | Oral | 2307.02763 | [
"https://github.com/davidjurgens/contextual-appropriateness"
] | -1 | -1 | -1 | -1 | 0 | [] | [] | [] |
|
https://aclanthology.org/2023.acl-long.617.bib | https://aclanthology.org/2023.acl-long.617/ | @inproceedings{lei-etal-2023-tart,
title = "{TART}: Improved Few-shot Text Classification Using Task-Adaptive Reference Transformation",
author = "Lei, Shuo and
Zhang, Xuchao and
He, Jianfeng and
Chen, Fanglan and
Lu, Chang-Tien",
editor = "Rogers, Anna and
Boyd-Graber, Jordan and
Okazaki, Naoaki",
booktitle = "Proceedings of the 61st Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers)",
month = jul,
year = "2023",
address = "Toronto, Canada",
publisher = "Association for Computational Linguistics",
url = "https://aclanthology.org/2023.acl-long.617",
doi = "10.18653/v1/2023.acl-long.617",
pages = "11014--11026",
abstract = "Meta-learning has emerged as a trending technique to tackle few-shot text classification and achieve state-of-the-art performance. However, the performance of existing approaches heavily depends on the inter-class variance of the support set. As a result, it can perform well on tasks when the semantics of sampled classes are distinct while failing to differentiate classes with similar semantics. In this paper, we propose a novel Task-Adaptive Reference Transformation (TART) network, aiming to enhance the generalization by transforming the class prototypes to per-class fixed reference points in task-adaptive metric spaces. To further maximize divergence between transformed prototypes in task-adaptive metric spaces, TART introduces a discriminative reference regularization among transformed prototypes. Extensive experiments are conducted on four benchmark datasets and our method demonstrates clear superiority over the state-of-the-art models in all the datasets. In particular, our model surpasses the state-of-the-art method by 7.4{\%} and 5.4{\%} in 1-shot and 5-shot classification on the 20 Newsgroups dataset, respectively.",
}
| Meta-learning has emerged as a trending technique to tackle few-shot text classification and achieve state-of-the-art performance. However, the performance of existing approaches heavily depends on the inter-class variance of the support set. As a result, it can perform well on tasks when the semantics of sampled classes are distinct while failing to differentiate classes with similar semantics. In this paper, we propose a novel Task-Adaptive Reference Transformation (TART) network, aiming to enhance the generalization by transforming the class prototypes to per-class fixed reference points in task-adaptive metric spaces. To further maximize divergence between transformed prototypes in task-adaptive metric spaces, TART introduces a discriminative reference regularization among transformed prototypes. Extensive experiments are conducted on four benchmark datasets and our method demonstrates clear superiority over the state-of-the-art models in all the datasets. In particular, our model surpasses the state-of-the-art method by 7.4{\%} and 5.4{\%} in 1-shot and 5-shot classification on the 20 Newsgroups dataset, respectively. | [
"Lei, Shuo",
"Zhang, Xuchao",
"He, Jianfeng",
"Chen, Fanglan",
"Lu, Chang-Tien"
] | TART: Improved Few-shot Text Classification Using Task-Adaptive Reference Transformation | acl-long.617 | Poster | 2306.02175 | [
"https://github.com/slei109/tart"
] | -1 | -1 | -1 | -1 | 0 | [] | [] | [] |
|
https://aclanthology.org/2023.acl-long.618.bib | https://aclanthology.org/2023.acl-long.618/ | @inproceedings{an-etal-2023-context,
title = "How Do In-Context Examples Affect Compositional Generalization?",
author = "An, Shengnan and
Lin, Zeqi and
Fu, Qiang and
Chen, Bei and
Zheng, Nanning and
Lou, Jian-Guang and
Zhang, Dongmei",
editor = "Rogers, Anna and
Boyd-Graber, Jordan and
Okazaki, Naoaki",
booktitle = "Proceedings of the 61st Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers)",
month = jul,
year = "2023",
address = "Toronto, Canada",
publisher = "Association for Computational Linguistics",
url = "https://aclanthology.org/2023.acl-long.618",
doi = "10.18653/v1/2023.acl-long.618",
pages = "11027--11052",
abstract = "Compositional generalization{--}understanding unseen combinations of seen primitives{--}is an essential reasoning capability in human intelligence. The AI community mainly studies this capability by fine-tuning neural networks on lots of training samples, while it is still unclear whether and how in-context learning{--}the prevailing few-shot paradigm based on large language models{--}exhibits compositional generalization. In this paper, we present CoFe, a test suite to investigate in-context compositional generalization. We find that the compositional generalization performance can be easily affected by the selection of in-context examples, thus raising the research question what the key factors are to make good in-context examples for compositional generalization. We study three potential factors: similarity, diversity and complexity. Our systematic experiments indicate that in-context examples should be structurally similar to the test case, diverse from each other, and individually simple. Furthermore, two strong limitations are observed: in-context compositional generalization on fictional words is much weaker than that on commonly used ones; it is still critical that the in-context examples should cover required linguistic structures, even though the backbone model has been pre-trained on large corpus. We hope our analysis would facilitate the understanding and utilization of in-context learning paradigm.",
}
| Compositional generalization{--}understanding unseen combinations of seen primitives{--}is an essential reasoning capability in human intelligence. The AI community mainly studies this capability by fine-tuning neural networks on lots of training samples, while it is still unclear whether and how in-context learning{--}the prevailing few-shot paradigm based on large language models{--}exhibits compositional generalization. In this paper, we present CoFe, a test suite to investigate in-context compositional generalization. We find that the compositional generalization performance can be easily affected by the selection of in-context examples, thus raising the research question what the key factors are to make good in-context examples for compositional generalization. We study three potential factors: similarity, diversity and complexity. Our systematic experiments indicate that in-context examples should be structurally similar to the test case, diverse from each other, and individually simple. Furthermore, two strong limitations are observed: in-context compositional generalization on fictional words is much weaker than that on commonly used ones; it is still critical that the in-context examples should cover required linguistic structures, even though the backbone model has been pre-trained on large corpus. We hope our analysis would facilitate the understanding and utilization of in-context learning paradigm. | [
"An, Shengnan",
"Lin, Zeqi",
"Fu, Qiang",
"Chen, Bei",
"Zheng, Nanning",
"Lou, Jian-Guang",
"Zhang, Dongmei"
] | How Do In-Context Examples Affect Compositional Generalization? | acl-long.618 | Poster | 2305.04835 | [
""
] | -1 | -1 | -1 | -1 | 0 | [] | [] | [] |
|
https://aclanthology.org/2023.acl-long.619.bib | https://aclanthology.org/2023.acl-long.619/ | @inproceedings{yang-jin-2023-attractive,
title = "Attractive Storyteller: Stylized Visual Storytelling with Unpaired Text",
author = "Yang, Dingyi and
Jin, Qin",
editor = "Rogers, Anna and
Boyd-Graber, Jordan and
Okazaki, Naoaki",
booktitle = "Proceedings of the 61st Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers)",
month = jul,
year = "2023",
address = "Toronto, Canada",
publisher = "Association for Computational Linguistics",
url = "https://aclanthology.org/2023.acl-long.619",
doi = "10.18653/v1/2023.acl-long.619",
pages = "11053--11066",
abstract = "Most research on stylized image captioning aims to generate style-specific captions using unpaired text, and has achieved impressive performance for simple styles like positive and negative. However, unlike previous single-sentence captions whose style is mostly embodied in distinctive words or phrases, real-world styles are likely to be implied at the syntactic and discourse levels. In this work, we introduce a new task of Stylized Visual Storytelling (SVST), which aims to describe a photo stream with stylized stories that are more expressive and attractive. We propose a multitasking memory-augmented framework called StyleVSG, which is jointly trained on factual visual storytelling data and unpaired style corpus, achieving a trade-off between style accuracy and visual relevance. Particularly for unpaired stylized text, StyleVSG learns to reconstruct the stylistic story from roughly parallel visual inputs mined with the CLIP model, avoiding problems caused by random mapping in previous methods. Furthermore, a memory module is designed to preserve the consistency and coherence of generated stories. Experiments show that our method can generate attractive and coherent stories with different styles such as fairy tale, romance, and humor. The overall performance of our StyleVSG surpasses state-of-the-art methods on both automatic and human evaluation metrics.",
}
| Most research on stylized image captioning aims to generate style-specific captions using unpaired text, and has achieved impressive performance for simple styles like positive and negative. However, unlike previous single-sentence captions whose style is mostly embodied in distinctive words or phrases, real-world styles are likely to be implied at the syntactic and discourse levels. In this work, we introduce a new task of Stylized Visual Storytelling (SVST), which aims to describe a photo stream with stylized stories that are more expressive and attractive. We propose a multitasking memory-augmented framework called StyleVSG, which is jointly trained on factual visual storytelling data and unpaired style corpus, achieving a trade-off between style accuracy and visual relevance. Particularly for unpaired stylized text, StyleVSG learns to reconstruct the stylistic story from roughly parallel visual inputs mined with the CLIP model, avoiding problems caused by random mapping in previous methods. Furthermore, a memory module is designed to preserve the consistency and coherence of generated stories. Experiments show that our method can generate attractive and coherent stories with different styles such as fairy tale, romance, and humor. The overall performance of our StyleVSG surpasses state-of-the-art methods on both automatic and human evaluation metrics. | [
"Yang, Dingyi",
"Jin, Qin"
] | Attractive Storyteller: Stylized Visual Storytelling with Unpaired Text | acl-long.619 | Poster | [
""
] | -1 | -1 | -1 | -1 | 0 | [] | [] | [] |
||
https://aclanthology.org/2023.acl-long.620.bib | https://aclanthology.org/2023.acl-long.620/ | @inproceedings{giaquinto-etal-2023-multitask,
title = "Multitask Pretraining with Structured Knowledge for Text-to-{SQL} Generation",
author = "Giaquinto, Robert and
Zhang, Dejiao and
Kleiner, Benjamin and
Li, Yang and
Tan, Ming and
Bhatia, Parminder and
Nallapati, Ramesh and
Ma, Xiaofei",
editor = "Rogers, Anna and
Boyd-Graber, Jordan and
Okazaki, Naoaki",
booktitle = "Proceedings of the 61st Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers)",
month = jul,
year = "2023",
address = "Toronto, Canada",
publisher = "Association for Computational Linguistics",
url = "https://aclanthology.org/2023.acl-long.620",
doi = "10.18653/v1/2023.acl-long.620",
pages = "11067--11083",
abstract = "Many machine learning-based low-code or no-code applications involve generating code that interacts with structured knowledge. For example, one of the most studied tasks in this area is generating SQL code from a natural language statement. Prior work shows that incorporating context information from the database schema, such as table and column names, is beneficial to model performance on this task. In this work we present a large pretraining dataset and strategy for learning representations of text, tables, and SQL code that leverages the entire context of the problem. Specifically, we build on existing encoder-decoder architecture by introducing a multitask pretraining framework that complements the unique attributes of our diverse pretraining data. Our work represents the first study on large-scale pretraining of encoder-decoder models for interacting with structured knowledge, and offers a new state-of-the-art foundation model in text-to-SQL generation. We validate our approach with experiments on two SQL tasks, showing improvement over existing methods, including a 1.7 and 2.2 percentage point improvement over prior state-of-the-arts on Spider and CoSQL.",
}
| Many machine learning-based low-code or no-code applications involve generating code that interacts with structured knowledge. For example, one of the most studied tasks in this area is generating SQL code from a natural language statement. Prior work shows that incorporating context information from the database schema, such as table and column names, is beneficial to model performance on this task. In this work we present a large pretraining dataset and strategy for learning representations of text, tables, and SQL code that leverages the entire context of the problem. Specifically, we build on existing encoder-decoder architecture by introducing a multitask pretraining framework that complements the unique attributes of our diverse pretraining data. Our work represents the first study on large-scale pretraining of encoder-decoder models for interacting with structured knowledge, and offers a new state-of-the-art foundation model in text-to-SQL generation. We validate our approach with experiments on two SQL tasks, showing improvement over existing methods, including a 1.7 and 2.2 percentage point improvement over prior state-of-the-arts on Spider and CoSQL. | [
"Giaquinto, Robert",
"Zhang, Dejiao",
"Kleiner, Benjamin",
"Li, Yang",
"Tan, Ming",
"Bhatia, Parminder",
"Nallapati, Ramesh",
"Ma, Xiaofei"
] | Multitask Pretraining with Structured Knowledge for Text-to-SQL Generation | acl-long.620 | Poster | [
""
] | -1 | -1 | -1 | -1 | 0 | [] | [] | [] |
||
https://aclanthology.org/2023.acl-long.621.bib | https://aclanthology.org/2023.acl-long.621/ | @inproceedings{wu-etal-2023-wspalign,
title = "{WSPA}lign: Word Alignment Pre-training via Large-Scale Weakly Supervised Span Prediction",
author = "Wu, Qiyu and
Nagata, Masaaki and
Tsuruoka, Yoshimasa",
editor = "Rogers, Anna and
Boyd-Graber, Jordan and
Okazaki, Naoaki",
booktitle = "Proceedings of the 61st Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers)",
month = jul,
year = "2023",
address = "Toronto, Canada",
publisher = "Association for Computational Linguistics",
url = "https://aclanthology.org/2023.acl-long.621",
doi = "10.18653/v1/2023.acl-long.621",
pages = "11084--11099",
abstract = "Most existing word alignment methods rely on manual alignment datasets or parallel corpora, which limits their usefulness. Here, to mitigate the dependence on manual data, we broaden the source of supervision by relaxing the requirement for correct, fully-aligned, and parallel sentences. Specifically, we make noisy, partially aligned, and non-parallel paragraphs in this paper. We then use such a large-scale weakly-supervised dataset for word alignment pre-training via span prediction. Extensive experiments with various settings empirically demonstrate that our approach, which is named WSPAlign, is an effective and scalable way to pre-train word aligners without manual data. When fine-tuned on standard benchmarks, WSPAlign has set a new state of the art by improving upon the best supervised baseline by 3.3 6.1 points in F1 and 1.5 6.1 points in AER. Furthermore, WSPAlign also achieves competitive performance compared with the corresponding baselines in few-shot, zero-shot and cross-lingual tests, which demonstrates that WSPAlign is potentially more practical for low-resource languages than existing methods.",
}
| Most existing word alignment methods rely on manual alignment datasets or parallel corpora, which limits their usefulness. Here, to mitigate the dependence on manual data, we broaden the source of supervision by relaxing the requirement for correct, fully-aligned, and parallel sentences. Specifically, we make noisy, partially aligned, and non-parallel paragraphs in this paper. We then use such a large-scale weakly-supervised dataset for word alignment pre-training via span prediction. Extensive experiments with various settings empirically demonstrate that our approach, which is named WSPAlign, is an effective and scalable way to pre-train word aligners without manual data. When fine-tuned on standard benchmarks, WSPAlign has set a new state of the art by improving upon the best supervised baseline by 3.3 6.1 points in F1 and 1.5 6.1 points in AER. Furthermore, WSPAlign also achieves competitive performance compared with the corresponding baselines in few-shot, zero-shot and cross-lingual tests, which demonstrates that WSPAlign is potentially more practical for low-resource languages than existing methods. | [
"Wu, Qiyu",
"Nagata, Masaaki",
"Tsuruoka, Yoshimasa"
] | WSPAlign: Word Alignment Pre-training via Large-Scale Weakly Supervised Span Prediction | acl-long.621 | Poster | 2306.05644 | [
"https://github.com/qiyuw/wspalign"
] | -1 | -1 | -1 | -1 | 0 | [] | [] | [] |
|
https://aclanthology.org/2023.acl-long.622.bib | https://aclanthology.org/2023.acl-long.622/ | @inproceedings{kang-etal-2023-distill,
title = "Distill or Annotate? Cost-Efficient Fine-Tuning of Compact Models",
author = "Kang, Junmo and
Xu, Wei and
Ritter, Alan",
editor = "Rogers, Anna and
Boyd-Graber, Jordan and
Okazaki, Naoaki",
booktitle = "Proceedings of the 61st Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers)",
month = jul,
year = "2023",
address = "Toronto, Canada",
publisher = "Association for Computational Linguistics",
url = "https://aclanthology.org/2023.acl-long.622",
doi = "10.18653/v1/2023.acl-long.622",
pages = "11100--11119",
abstract = "Fine-tuning large models is highly effective, however, inference can be expensive and produces carbon emissions. Knowledge distillation has been shown to be a practical solution to reduce inference costs, but the distillation process itself requires significant computational resources. Rather than buying or renting GPUs to fine-tune, then distill a large model, an NLP practitioner might instead choose to allocate the available budget to hire annotators and manually label additional fine-tuning data. In this paper, we investigate how to most efficiently use a fixed budget to build a compact model. Through extensive experiments on six diverse tasks, we show that distilling from T5-XXL (11B) to T5-Small (60M) is almost always a cost-efficient strategy compared to annotating more data to directly train a compact model (T5-Small). We further investigate how the optimal budget allocated towards computation varies across scenarios. We will make our code, datasets, annotation cost estimates, and baseline models available as a benchmark to support further work on cost-efficient training of compact models.",
}
| Fine-tuning large models is highly effective, however, inference can be expensive and produces carbon emissions. Knowledge distillation has been shown to be a practical solution to reduce inference costs, but the distillation process itself requires significant computational resources. Rather than buying or renting GPUs to fine-tune, then distill a large model, an NLP practitioner might instead choose to allocate the available budget to hire annotators and manually label additional fine-tuning data. In this paper, we investigate how to most efficiently use a fixed budget to build a compact model. Through extensive experiments on six diverse tasks, we show that distilling from T5-XXL (11B) to T5-Small (60M) is almost always a cost-efficient strategy compared to annotating more data to directly train a compact model (T5-Small). We further investigate how the optimal budget allocated towards computation varies across scenarios. We will make our code, datasets, annotation cost estimates, and baseline models available as a benchmark to support further work on cost-efficient training of compact models. | [
"Kang, Junmo",
"Xu, Wei",
"Ritter, Alan"
] | Distill or Annotate? Cost-Efficient Fine-Tuning of Compact Models | acl-long.622 | Poster | 2305.01645 | [
""
] | -1 | -1 | -1 | -1 | 0 | [] | [] | [] |
|
https://aclanthology.org/2023.acl-long.623.bib | https://aclanthology.org/2023.acl-long.623/ | @inproceedings{ning-etal-2023-od,
title = "{OD}-{RTE}: A One-Stage Object Detection Framework for Relational Triple Extraction",
author = "Ning, Jinzhong and
Yang, Zhihao and
Sun, Yuanyuan and
Wang, Zhizheng and
Lin, Hongfei",
editor = "Rogers, Anna and
Boyd-Graber, Jordan and
Okazaki, Naoaki",
booktitle = "Proceedings of the 61st Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers)",
month = jul,
year = "2023",
address = "Toronto, Canada",
publisher = "Association for Computational Linguistics",
url = "https://aclanthology.org/2023.acl-long.623",
doi = "10.18653/v1/2023.acl-long.623",
pages = "11120--11135",
abstract = "The Relational Triple Extraction (RTE) task is a fundamental and essential information extraction task. Recently, the table-filling RTE methods have received lots of attention. Despite their success, they suffer from some inherent problems such as underutilizing regional information of triple. In this work, we treat the RTE task based on table-filling method as an Object Detection task and propose a one-stage Object Detection framework for Relational Triple Extraction (OD-RTE). In this framework, the vertices-based bounding box detection, coupled with auxiliary global relational triple region detection, ensuring that regional information of triple could be fully utilized. Besides, our proposed decoding scheme could extract all types of triples. In addition, the negative sampling strategy of relations in the training stage improves the training efficiency while alleviating the imbalance of positive and negative relations. The experimental results show that 1) OD-RTE achieves the state-of-the-art performance on two widely used datasets (i.e., NYT and WebNLG). 2) Compared with the best performing table-filling method, OD-RTE achieves faster training and inference speed with lower GPU memory usage. To facilitate future research in this area, the codes are publicly available at \url{https://github.com/NingJinzhong/ODRTE}.",
}
| The Relational Triple Extraction (RTE) task is a fundamental and essential information extraction task. Recently, the table-filling RTE methods have received lots of attention. Despite their success, they suffer from some inherent problems such as underutilizing regional information of triple. In this work, we treat the RTE task based on table-filling method as an Object Detection task and propose a one-stage Object Detection framework for Relational Triple Extraction (OD-RTE). In this framework, the vertices-based bounding box detection, coupled with auxiliary global relational triple region detection, ensuring that regional information of triple could be fully utilized. Besides, our proposed decoding scheme could extract all types of triples. In addition, the negative sampling strategy of relations in the training stage improves the training efficiency while alleviating the imbalance of positive and negative relations. The experimental results show that 1) OD-RTE achieves the state-of-the-art performance on two widely used datasets (i.e., NYT and WebNLG). 2) Compared with the best performing table-filling method, OD-RTE achieves faster training and inference speed with lower GPU memory usage. To facilitate future research in this area, the codes are publicly available at \url{https://github.com/NingJinzhong/ODRTE}. | [
"Ning, Jinzhong",
"Yang, Zhihao",
"Sun, Yuanyuan",
"Wang, Zhizheng",
"Lin, Hongfei"
] | OD-RTE: A One-Stage Object Detection Framework for Relational Triple Extraction | acl-long.623 | Poster | [
""
] | -1 | -1 | -1 | -1 | 0 | [] | [] | [] |
||
https://aclanthology.org/2023.acl-long.624.bib | https://aclanthology.org/2023.acl-long.624/ | @inproceedings{zhou-etal-2023-cast,
title = "{I} Cast Detect Thoughts: Learning to Converse and Guide with Intents and Theory-of-Mind in Dungeons and Dragons",
author = "Zhou, Pei and
Zhu, Andrew and
Hu, Jennifer and
Pujara, Jay and
Ren, Xiang and
Callison-Burch, Chris and
Choi, Yejin and
Ammanabrolu, Prithviraj",
editor = "Rogers, Anna and
Boyd-Graber, Jordan and
Okazaki, Naoaki",
booktitle = "Proceedings of the 61st Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers)",
month = jul,
year = "2023",
address = "Toronto, Canada",
publisher = "Association for Computational Linguistics",
url = "https://aclanthology.org/2023.acl-long.624",
doi = "10.18653/v1/2023.acl-long.624",
pages = "11136--11155",
abstract = "We propose a novel task, G4C, to study teacher-student natural language interactions in a goal-driven and grounded environment. Dungeons and Dragons (D{\&}D), a role-playing game, provides an ideal setting to investigate such interactions. Here, the Dungeon Master (DM), i.e., the teacher, guides the actions of several players{---}students, each with their own personas and abilities{---}to achieve shared goals grounded in a fantasy world. Our approach is to decompose and model these interactions into (1) the DM{'}s intent to guide players toward a given goal; (2) the DM{'}s guidance utterance to the players expressing this intent; and (3) a theory-of-mind (ToM) model that anticipates the players{'} reaction to the guidance one turn into the future. We develop a novel reinforcement learning (RL) method for training a DM that generates guidance for players by rewarding utterances where the intent matches the ToM-anticipated player actions. Human and automated evaluations show that a DM trained to explicitly model intents and incorporate ToM of the players using RL generates better-quality guidance that is 3x more likely to fulfill the DM{'}s intent than a vanilla natural language generation (NLG) approach.",
}
| We propose a novel task, G4C, to study teacher-student natural language interactions in a goal-driven and grounded environment. Dungeons and Dragons (D{\&}D), a role-playing game, provides an ideal setting to investigate such interactions. Here, the Dungeon Master (DM), i.e., the teacher, guides the actions of several players{---}students, each with their own personas and abilities{---}to achieve shared goals grounded in a fantasy world. Our approach is to decompose and model these interactions into (1) the DM{'}s intent to guide players toward a given goal; (2) the DM{'}s guidance utterance to the players expressing this intent; and (3) a theory-of-mind (ToM) model that anticipates the players{'} reaction to the guidance one turn into the future. We develop a novel reinforcement learning (RL) method for training a DM that generates guidance for players by rewarding utterances where the intent matches the ToM-anticipated player actions. Human and automated evaluations show that a DM trained to explicitly model intents and incorporate ToM of the players using RL generates better-quality guidance that is 3x more likely to fulfill the DM{'}s intent than a vanilla natural language generation (NLG) approach. | [
"Zhou, Pei",
"Zhu, Andrew",
"Hu, Jennifer",
"Pujara, Jay",
"Ren, Xiang",
"Callison-Burch, Chris",
"Choi, Yejin",
"Ammanabrolu, Prithviraj"
] | I Cast Detect Thoughts: Learning to Converse and Guide with Intents and Theory-of-Mind in Dungeons and Dragons | acl-long.624 | Poster | 2212.10060 | [
""
] | https://huggingface.co/papers/2212.10060 | 1 | 0 | 0 | 8 | 1 | [] | [] | [] |
https://aclanthology.org/2023.acl-long.625.bib | https://aclanthology.org/2023.acl-long.625/ | @inproceedings{sun-etal-2023-multitask,
title = "Multitask Pre-training of Modular Prompt for {C}hinese Few-Shot Learning",
author = "Sun, Tianxiang and
He, Zhengfu and
Zhu, Qin and
Qiu, Xipeng and
Huang, Xuanjing",
editor = "Rogers, Anna and
Boyd-Graber, Jordan and
Okazaki, Naoaki",
booktitle = "Proceedings of the 61st Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers)",
month = jul,
year = "2023",
address = "Toronto, Canada",
publisher = "Association for Computational Linguistics",
url = "https://aclanthology.org/2023.acl-long.625",
doi = "10.18653/v1/2023.acl-long.625",
pages = "11156--11172",
abstract = "Prompt tuning is a parameter-efficient approach to adapting pre-trained language models to downstream tasks. Although prompt tuning has been shown to match the performance of full model tuning when training data is sufficient, it tends to struggle in few-shot learning settings. In this paper, we present Multi-task Pre-trained Modular Prompt (MP2) to boost prompt tuning for few-shot learning. MP2 is a set of combinable prompts pre-trained on 38 Chinese tasks. On downstream tasks, the pre-trained prompts are selectively activated and combined, leading to strong compositional generalization to unseen tasks. To bridge the gap between pre-training and fine-tuning, we formulate upstream and downstream tasks into a unified machine reading comprehension task. Extensive experiments under two learning paradigms, i.e., gradient descent and black-box tuning, show that MP2 significantly outperforms prompt tuning, full model tuning, and prior prompt pre-training methods in few-shot settings. In addition, we demonstrate that MP2 can achieve surprisingly fast and strong adaptation to downstream tasks by merely learning 8 parameters to combine the pre-trained modular prompts.",
}
| Prompt tuning is a parameter-efficient approach to adapting pre-trained language models to downstream tasks. Although prompt tuning has been shown to match the performance of full model tuning when training data is sufficient, it tends to struggle in few-shot learning settings. In this paper, we present Multi-task Pre-trained Modular Prompt (MP2) to boost prompt tuning for few-shot learning. MP2 is a set of combinable prompts pre-trained on 38 Chinese tasks. On downstream tasks, the pre-trained prompts are selectively activated and combined, leading to strong compositional generalization to unseen tasks. To bridge the gap between pre-training and fine-tuning, we formulate upstream and downstream tasks into a unified machine reading comprehension task. Extensive experiments under two learning paradigms, i.e., gradient descent and black-box tuning, show that MP2 significantly outperforms prompt tuning, full model tuning, and prior prompt pre-training methods in few-shot settings. In addition, we demonstrate that MP2 can achieve surprisingly fast and strong adaptation to downstream tasks by merely learning 8 parameters to combine the pre-trained modular prompts. | [
"Sun, Tianxiang",
"He, Zhengfu",
"Zhu, Qin",
"Qiu, Xipeng",
"Huang, Xuanjing"
] | Multitask Pre-training of Modular Prompt for Chinese Few-Shot Learning | acl-long.625 | Poster | 2210.07565 | [
"https://github.com/Hzfinfdu/MPMP"
] | -1 | -1 | -1 | -1 | 0 | [] | [] | [] |
|
https://aclanthology.org/2023.acl-long.626.bib | https://aclanthology.org/2023.acl-long.626/ | @inproceedings{ding-etal-2023-gpt,
title = "Is {GPT}-3 a Good Data Annotator?",
author = "Ding, Bosheng and
Qin, Chengwei and
Liu, Linlin and
Chia, Yew Ken and
Li, Boyang and
Joty, Shafiq and
Bing, Lidong",
editor = "Rogers, Anna and
Boyd-Graber, Jordan and
Okazaki, Naoaki",
booktitle = "Proceedings of the 61st Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers)",
month = jul,
year = "2023",
address = "Toronto, Canada",
publisher = "Association for Computational Linguistics",
url = "https://aclanthology.org/2023.acl-long.626",
doi = "10.18653/v1/2023.acl-long.626",
pages = "11173--11195",
abstract = "Data annotation is the process of labeling data that could be used to train machine learning models. Having high quality annotation is crucial, as it allows the model to learn the relationship between the input data and the desired output. GPT-3, a large-scale language model developed by OpenAI, has demonstrated im- impressive zero- and few-shot performance on a wide range of NLP tasks. It is therefore natural to wonder whether it can be used to effectively annotate data for NLP tasks. In this paper, we evaluate the performance of GPT-3 as a data annotator by comparing it with traditional data annotation methods and analyzing its output on a range of tasks. Through this analysis, we aim to provide insight into the potential of GPT-3 as a general-purpose data annotator in NLP.",
}
| Data annotation is the process of labeling data that could be used to train machine learning models. Having high quality annotation is crucial, as it allows the model to learn the relationship between the input data and the desired output. GPT-3, a large-scale language model developed by OpenAI, has demonstrated im- impressive zero- and few-shot performance on a wide range of NLP tasks. It is therefore natural to wonder whether it can be used to effectively annotate data for NLP tasks. In this paper, we evaluate the performance of GPT-3 as a data annotator by comparing it with traditional data annotation methods and analyzing its output on a range of tasks. Through this analysis, we aim to provide insight into the potential of GPT-3 as a general-purpose data annotator in NLP. | [
"Ding, Bosheng",
"Qin, Chengwei",
"Liu, Linlin",
"Chia, Yew Ken",
"Li, Boyang",
"Joty, Shafiq",
"Bing, Lidong"
] | Is GPT-3 a Good Data Annotator? | acl-long.626 | Poster | 2212.10450 | [
"https://github.com/damo-nlp-sg/llm-data-annotator"
] | -1 | -1 | -1 | -1 | 0 | [] | [] | [] |
|
https://aclanthology.org/2023.acl-long.627.bib | https://aclanthology.org/2023.acl-long.627/ | @inproceedings{wan-etal-2023-multi,
title = "Multi-Grained Knowledge Retrieval for End-to-End Task-Oriented Dialog",
author = "Wan, Fanqi and
Shen, Weizhou and
Yang, Ke and
Quan, Xiaojun and
Bi, Wei",
editor = "Rogers, Anna and
Boyd-Graber, Jordan and
Okazaki, Naoaki",
booktitle = "Proceedings of the 61st Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers)",
month = jul,
year = "2023",
address = "Toronto, Canada",
publisher = "Association for Computational Linguistics",
url = "https://aclanthology.org/2023.acl-long.627",
doi = "10.18653/v1/2023.acl-long.627",
pages = "11196--11210",
abstract = "Retrieving proper domain knowledge from an external database lies at the heart of end-to-end task-oriented dialog systems to generate informative responses. Most existing systems blend knowledge retrieval with response generation and optimize them with direct supervision from reference responses, leading to suboptimal retrieval performance when the knowledge base becomes large-scale. To address this, we propose to decouple knowledge retrieval from response generation and introduce a multi-grained knowledge retriever (MAKER) that includes an entity selector to search for relevant entities and an attribute selector to filter out irrelevant attributes. To train the retriever, we propose a novel distillation objective that derives supervision signals from the response generator. Experiments conducted on three standard benchmarks with both small and large-scale knowledge bases demonstrate that our retriever performs knowledge retrieval more effectively than existing methods. Our code has been made publicly available at \url{https://github.com/18907305772/MAKER}.",
}
| Retrieving proper domain knowledge from an external database lies at the heart of end-to-end task-oriented dialog systems to generate informative responses. Most existing systems blend knowledge retrieval with response generation and optimize them with direct supervision from reference responses, leading to suboptimal retrieval performance when the knowledge base becomes large-scale. To address this, we propose to decouple knowledge retrieval from response generation and introduce a multi-grained knowledge retriever (MAKER) that includes an entity selector to search for relevant entities and an attribute selector to filter out irrelevant attributes. To train the retriever, we propose a novel distillation objective that derives supervision signals from the response generator. Experiments conducted on three standard benchmarks with both small and large-scale knowledge bases demonstrate that our retriever performs knowledge retrieval more effectively than existing methods. Our code has been made publicly available at \url{https://github.com/18907305772/MAKER}. | [
"Wan, Fanqi",
"Shen, Weizhou",
"Yang, Ke",
"Quan, Xiaojun",
"Bi, Wei"
] | Multi-Grained Knowledge Retrieval for End-to-End Task-Oriented Dialog | acl-long.627 | Poster | 2305.10149 | [
"https://github.com/18907305772/maker"
] | -1 | -1 | -1 | -1 | 0 | [] | [] | [] |
|
https://aclanthology.org/2023.acl-long.628.bib | https://aclanthology.org/2023.acl-long.628/ | @inproceedings{ma-etal-2023-shot,
title = "Few-shot Event Detection: An Empirical Study and a Unified View",
author = "Ma, Yubo and
Wang, Zehao and
Cao, Yixin and
Sun, Aixin",
editor = "Rogers, Anna and
Boyd-Graber, Jordan and
Okazaki, Naoaki",
booktitle = "Proceedings of the 61st Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers)",
month = jul,
year = "2023",
address = "Toronto, Canada",
publisher = "Association for Computational Linguistics",
url = "https://aclanthology.org/2023.acl-long.628",
doi = "10.18653/v1/2023.acl-long.628",
pages = "11211--11236",
abstract = "Few-shot event detection (ED) has been widely studied, while this brings noticeable discrepancies, e.g., various motivations, tasks, and experimental settings, that hinder the understanding of models for future progress. This paper presents a thorough empirical study, a unified view of ED models, and a better unified baseline. For fair evaluation, we compare 12 representative methods on three datasets, which are roughly grouped into prompt-based and prototype-based models for detailed analysis. Experiments consistently demonstrate that prompt-based methods, including ChatGPT, still significantly trail prototype-based methods in terms of overall performance. To investigate their superior performance, we break down their design elements along several dimensions and build a unified framework on prototype-based methods. Under such unified view, each prototype-method can be viewed a combination of different modules from these design elements. We further combine all advantageous modules and propose a simple yet effective baseline, which outperforms existing methods by a large margin (e.g., 2.7{\%} F1 gains under low-resource setting).",
}
| Few-shot event detection (ED) has been widely studied, while this brings noticeable discrepancies, e.g., various motivations, tasks, and experimental settings, that hinder the understanding of models for future progress. This paper presents a thorough empirical study, a unified view of ED models, and a better unified baseline. For fair evaluation, we compare 12 representative methods on three datasets, which are roughly grouped into prompt-based and prototype-based models for detailed analysis. Experiments consistently demonstrate that prompt-based methods, including ChatGPT, still significantly trail prototype-based methods in terms of overall performance. To investigate their superior performance, we break down their design elements along several dimensions and build a unified framework on prototype-based methods. Under such unified view, each prototype-method can be viewed a combination of different modules from these design elements. We further combine all advantageous modules and propose a simple yet effective baseline, which outperforms existing methods by a large margin (e.g., 2.7{\%} F1 gains under low-resource setting). | [
"Ma, Yubo",
"Wang, Zehao",
"Cao, Yixin",
"Sun, Aixin"
] | Few-shot Event Detection: An Empirical Study and a Unified View | acl-long.628 | Poster | 2305.01901 | [
"https://github.com/mayubo2333/fewshot_ed"
] | -1 | -1 | -1 | -1 | 0 | [] | [] | [] |
|
https://aclanthology.org/2023.acl-long.629.bib | https://aclanthology.org/2023.acl-long.629/ | @inproceedings{mueller-linzen-2023-plant,
title = "How to Plant Trees in Language Models: Data and Architectural Effects on the Emergence of Syntactic Inductive Biases",
author = "Mueller, Aaron and
Linzen, Tal",
editor = "Rogers, Anna and
Boyd-Graber, Jordan and
Okazaki, Naoaki",
booktitle = "Proceedings of the 61st Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers)",
month = jul,
year = "2023",
address = "Toronto, Canada",
publisher = "Association for Computational Linguistics",
url = "https://aclanthology.org/2023.acl-long.629",
doi = "10.18653/v1/2023.acl-long.629",
pages = "11237--11252",
abstract = "Accurate syntactic representations are essential for robust generalization in natural language. Recent work has found that pre-training can teach language models to rely on hierarchical syntactic features{---}as opposed to incorrect linear features{---}when performing tasks after fine-tuning. We test what aspects of pre-training are important for endowing encoder-decoder Transformers with an inductive bias that favors hierarchical syntactic generalizations. We focus on architectural features (depth, width, and number of parameters), as well as the genre and size of the pre-training corpus, diagnosing inductive biases using two syntactic transformation tasks: question formation and passivization, both in English. We find that the number of parameters alone does not explain hierarchical generalization: model depth plays greater role than model width. We also find that pre-training on simpler language, such as child-directed speech, induces a hierarchical bias using an order-of-magnitude less data than pre-training on more typical datasets based on web text or Wikipedia; this suggests that in cognitively plausible language acquisition settings, neural language models may be more data-efficient than previously thought.",
}
| Accurate syntactic representations are essential for robust generalization in natural language. Recent work has found that pre-training can teach language models to rely on hierarchical syntactic features{---}as opposed to incorrect linear features{---}when performing tasks after fine-tuning. We test what aspects of pre-training are important for endowing encoder-decoder Transformers with an inductive bias that favors hierarchical syntactic generalizations. We focus on architectural features (depth, width, and number of parameters), as well as the genre and size of the pre-training corpus, diagnosing inductive biases using two syntactic transformation tasks: question formation and passivization, both in English. We find that the number of parameters alone does not explain hierarchical generalization: model depth plays greater role than model width. We also find that pre-training on simpler language, such as child-directed speech, induces a hierarchical bias using an order-of-magnitude less data than pre-training on more typical datasets based on web text or Wikipedia; this suggests that in cognitively plausible language acquisition settings, neural language models may be more data-efficient than previously thought. | [
"Mueller, Aaron",
"Linzen, Tal"
] | How to Plant Trees in Language Models: Data and Architectural Effects on the Emergence of Syntactic Inductive Biases | acl-long.629 | Poster | 2305.19905 | [
"https://github.com/aaronmueller/emergent-syntax"
] | -1 | -1 | -1 | -1 | 0 | [] | [] | [] |
|
https://aclanthology.org/2023.acl-long.630.bib | https://aclanthology.org/2023.acl-long.630/ | @inproceedings{pyatkin-etal-2023-clarifydelphi,
title = "{C}larify{D}elphi: Reinforced Clarification Questions with Defeasibility Rewards for Social and Moral Situations",
author = "Pyatkin, Valentina and
Hwang, Jena D. and
Srikumar, Vivek and
Lu, Ximing and
Jiang, Liwei and
Choi, Yejin and
Bhagavatula, Chandra",
editor = "Rogers, Anna and
Boyd-Graber, Jordan and
Okazaki, Naoaki",
booktitle = "Proceedings of the 61st Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers)",
month = jul,
year = "2023",
address = "Toronto, Canada",
publisher = "Association for Computational Linguistics",
url = "https://aclanthology.org/2023.acl-long.630",
doi = "10.18653/v1/2023.acl-long.630",
pages = "11253--11271",
abstract = "Context is everything, even in commonsense moral reasoning. Changing contexts can flip the moral judgment of an action; Lying to a friend is wrong in general, but may be morally acceptable if it is intended to protect their life. We present ClarifyDelphi, an interactive system that learns to ask clarification questions (e.g., why did you lie to your friend?) in order to elicit additional salient contexts of a social or moral situation. We posit that questions whose potential answers lead to \textit{diverging} moral judgments are the most informative. Thus, we propose a reinforcement learning framework with a defeasibility reward that aims to maximize the divergence between moral judgments of hypothetical answers to a question. Human evaluation demonstrates that our system generates more relevant, informative and defeasible questions compared to competitive baselines. Our work is ultimately inspired by studies in cognitive science that have investigated the flexibility in moral cognition (i.e., the diverse contexts in which moral rules can be bent), and we hope that research in this direction can assist both cognitive and computational investigations of moral judgments.",
}
| Context is everything, even in commonsense moral reasoning. Changing contexts can flip the moral judgment of an action; Lying to a friend is wrong in general, but may be morally acceptable if it is intended to protect their life. We present ClarifyDelphi, an interactive system that learns to ask clarification questions (e.g., why did you lie to your friend?) in order to elicit additional salient contexts of a social or moral situation. We posit that questions whose potential answers lead to \textit{diverging} moral judgments are the most informative. Thus, we propose a reinforcement learning framework with a defeasibility reward that aims to maximize the divergence between moral judgments of hypothetical answers to a question. Human evaluation demonstrates that our system generates more relevant, informative and defeasible questions compared to competitive baselines. Our work is ultimately inspired by studies in cognitive science that have investigated the flexibility in moral cognition (i.e., the diverse contexts in which moral rules can be bent), and we hope that research in this direction can assist both cognitive and computational investigations of moral judgments. | [
"Pyatkin, Valentina",
"Hwang, Jena D.",
"Srikumar, Vivek",
"Lu, Ximing",
"Jiang, Liwei",
"Choi, Yejin",
"Bhagavatula, Ch",
"ra"
] | ClarifyDelphi: Reinforced Clarification Questions with Defeasibility Rewards for Social and Moral Situations | acl-long.630 | Poster | 2212.10409 | [
"https://github.com/valentinapy/clarifyd"
] | https://huggingface.co/papers/2212.10409 | 1 | 0 | 0 | 7 | 1 | [] | [] | [] |
https://aclanthology.org/2023.acl-long.631.bib | https://aclanthology.org/2023.acl-long.631/ | @inproceedings{ivison-etal-2023-hint,
title = "{HINT}: Hypernetwork Instruction Tuning for Efficient Zero- and Few-Shot Generalisation",
author = "Ivison, Hamish and
Bhagia, Akshita and
Wang, Yizhong and
Hajishirzi, Hannaneh and
Peters, Matthew",
editor = "Rogers, Anna and
Boyd-Graber, Jordan and
Okazaki, Naoaki",
booktitle = "Proceedings of the 61st Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers)",
month = jul,
year = "2023",
address = "Toronto, Canada",
publisher = "Association for Computational Linguistics",
url = "https://aclanthology.org/2023.acl-long.631",
doi = "10.18653/v1/2023.acl-long.631",
pages = "11272--11288",
abstract = "Recent NLP models have shown the remarkable ability to effectively generalise {`}zero-shot{'} to new tasks using only natural language instructions as guidance. However, many of these approaches suffer from high computational costs due to their reliance on concatenating lengthy instructions with every input example, resulting in costly reprocessing of the instruction. To avoid this, we introduce Hypernetworks for INstruction Tuning (HINT), which convert task instructions and examples into parameter-efficient modules inserted into an underlying model using a pretrained text encoder, eliminating the need to include instructions in the model input. The hypernetwork in HINT also produces an encoded instruction, which we concatenate with encoded inputs during decoding to further improve performance. HINT models outperform strong state-of-the-art baselines by over 10{\%} when controlling for compute (measured in FLOPs). By converting instructions into modules, HINT models can effectively disregard the length of instructions and few-shot example inputs in terms of compute usage. As a result, HINT can enhance its performance by up to 25{\%} by incorporating additional few-shot data, while utilizing only up to 5{\%} more compute. This combines the strengths of parameter-efficient fine-tuning and in-context learning.",
}
| Recent NLP models have shown the remarkable ability to effectively generalise {`}zero-shot{'} to new tasks using only natural language instructions as guidance. However, many of these approaches suffer from high computational costs due to their reliance on concatenating lengthy instructions with every input example, resulting in costly reprocessing of the instruction. To avoid this, we introduce Hypernetworks for INstruction Tuning (HINT), which convert task instructions and examples into parameter-efficient modules inserted into an underlying model using a pretrained text encoder, eliminating the need to include instructions in the model input. The hypernetwork in HINT also produces an encoded instruction, which we concatenate with encoded inputs during decoding to further improve performance. HINT models outperform strong state-of-the-art baselines by over 10{\%} when controlling for compute (measured in FLOPs). By converting instructions into modules, HINT models can effectively disregard the length of instructions and few-shot example inputs in terms of compute usage. As a result, HINT can enhance its performance by up to 25{\%} by incorporating additional few-shot data, while utilizing only up to 5{\%} more compute. This combines the strengths of parameter-efficient fine-tuning and in-context learning. | [
"Ivison, Hamish",
"Bhagia, Akshita",
"Wang, Yizhong",
"Hajishirzi, Hannaneh",
"Peters, Matthew"
] | HINT: Hypernetwork Instruction Tuning for Efficient Zero- and Few-Shot Generalisation | acl-long.631 | Poster | [
""
] | -1 | -1 | -1 | -1 | 0 | [] | [] | [] |
||
https://aclanthology.org/2023.acl-long.632.bib | https://aclanthology.org/2023.acl-long.632/ | @inproceedings{si-etal-2023-measuring,
title = "Measuring Inductive Biases of In-Context Learning with Underspecified Demonstrations",
author = "Si, Chenglei and
Friedman, Dan and
Joshi, Nitish and
Feng, Shi and
Chen, Danqi and
He, He",
editor = "Rogers, Anna and
Boyd-Graber, Jordan and
Okazaki, Naoaki",
booktitle = "Proceedings of the 61st Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers)",
month = jul,
year = "2023",
address = "Toronto, Canada",
publisher = "Association for Computational Linguistics",
url = "https://aclanthology.org/2023.acl-long.632",
doi = "10.18653/v1/2023.acl-long.632",
pages = "11289--11310",
abstract = "In-context learning (ICL) is an important paradigm for adapting large language models (LLMs) to new tasks, but the generalization behavior of ICL remains poorly understood. We investigate the inductive biases of ICL from the perspective of feature bias: which feature ICL is more likely to use given a set of underspecified demonstrations in which two features are equally predictive of the labels. First, we characterize the feature biases of GPT-3 models by constructing underspecified demonstrations from a range of NLP datasets and feature combinations. We find that LLMs exhibit clear feature biases{---}for example, demonstrating a strong bias to predict labels according to sentiment rather than shallow lexical features, like punctuation. Second, we evaluate the effect of different interventions that are designed to impose an inductive bias in favor of a particular feature, such as adding a natural language instruction or using semantically relevant label words. We find that, while many interventions can influence the learner to prefer a particular feature, it can be difficult to overcome strong prior biases. Overall, our results provide a broader picture of the types of features that ICL may be more likely to exploit and how to impose inductive biases that are better aligned with the intended task.",
}
| In-context learning (ICL) is an important paradigm for adapting large language models (LLMs) to new tasks, but the generalization behavior of ICL remains poorly understood. We investigate the inductive biases of ICL from the perspective of feature bias: which feature ICL is more likely to use given a set of underspecified demonstrations in which two features are equally predictive of the labels. First, we characterize the feature biases of GPT-3 models by constructing underspecified demonstrations from a range of NLP datasets and feature combinations. We find that LLMs exhibit clear feature biases{---}for example, demonstrating a strong bias to predict labels according to sentiment rather than shallow lexical features, like punctuation. Second, we evaluate the effect of different interventions that are designed to impose an inductive bias in favor of a particular feature, such as adding a natural language instruction or using semantically relevant label words. We find that, while many interventions can influence the learner to prefer a particular feature, it can be difficult to overcome strong prior biases. Overall, our results provide a broader picture of the types of features that ICL may be more likely to exploit and how to impose inductive biases that are better aligned with the intended task. | [
"Si, Chenglei",
"Friedman, Dan",
"Joshi, Nitish",
"Feng, Shi",
"Chen, Danqi",
"He, He"
] | Measuring Inductive Biases of In-Context Learning with Underspecified Demonstrations | acl-long.632 | Oral | 2305.13299 | [
"https://github.com/noviscl/ambigprompt"
] | -1 | -1 | -1 | -1 | 0 | [] | [] | [] |
|
https://aclanthology.org/2023.acl-long.633.bib | https://aclanthology.org/2023.acl-long.633/ | @inproceedings{kuznetsov-gurevych-2023-inclusive,
title = "An Inclusive Notion of Text",
author = "Kuznetsov, Ilia and
Gurevych, Iryna",
editor = "Rogers, Anna and
Boyd-Graber, Jordan and
Okazaki, Naoaki",
booktitle = "Proceedings of the 61st Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers)",
month = jul,
year = "2023",
address = "Toronto, Canada",
publisher = "Association for Computational Linguistics",
url = "https://aclanthology.org/2023.acl-long.633",
doi = "10.18653/v1/2023.acl-long.633",
pages = "11311--11327",
abstract = "Natural language processing (NLP) researchers develop models of grammar, meaning and communication based on written text. Due to task and data differences, what is considered text can vary substantially across studies. A conceptual framework for systematically capturing these differences is lacking. We argue that clarity on the notion of text is crucial for reproducible and generalizable NLP. Towards that goal, we propose common terminology to discuss the production and transformation of textual data, and introduce a two-tier taxonomy of linguistic and non-linguistic elements that are available in textual sources and can be used in NLP modeling. We apply this taxonomy to survey existing work that extends the notion of text beyond the conservative language-centered view. We outline key desiderata and challenges of the emerging inclusive approach to text in NLP, and suggest community-level reporting as a crucial next step to consolidate the discussion.",
}
| Natural language processing (NLP) researchers develop models of grammar, meaning and communication based on written text. Due to task and data differences, what is considered text can vary substantially across studies. A conceptual framework for systematically capturing these differences is lacking. We argue that clarity on the notion of text is crucial for reproducible and generalizable NLP. Towards that goal, we propose common terminology to discuss the production and transformation of textual data, and introduce a two-tier taxonomy of linguistic and non-linguistic elements that are available in textual sources and can be used in NLP modeling. We apply this taxonomy to survey existing work that extends the notion of text beyond the conservative language-centered view. We outline key desiderata and challenges of the emerging inclusive approach to text in NLP, and suggest community-level reporting as a crucial next step to consolidate the discussion. | [
"Kuznetsov, Ilia",
"Gurevych, Iryna"
] | An Inclusive Notion of Text | acl-long.633 | Poster | 2211.05604 | [
""
] | -1 | -1 | -1 | -1 | 0 | [] | [] | [] |
|
https://aclanthology.org/2023.acl-long.634.bib | https://aclanthology.org/2023.acl-long.634/ | @inproceedings{zha-etal-2023-alignscore,
title = "{A}lign{S}core: Evaluating Factual Consistency with A Unified Alignment Function",
author = "Zha, Yuheng and
Yang, Yichi and
Li, Ruichen and
Hu, Zhiting",
editor = "Rogers, Anna and
Boyd-Graber, Jordan and
Okazaki, Naoaki",
booktitle = "Proceedings of the 61st Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers)",
month = jul,
year = "2023",
address = "Toronto, Canada",
publisher = "Association for Computational Linguistics",
url = "https://aclanthology.org/2023.acl-long.634",
doi = "10.18653/v1/2023.acl-long.634",
pages = "11328--11348",
abstract = "Many text generation applications require the generated text to be factually consistent with input information. Automatic evaluation of factual consistency is challenging. Previous work has developed various metrics that often depend on specific functions, such as natural language inference (NLI) or question answering (QA), trained on limited data. Those metrics thus can hardly assess diverse factual inconsistencies (e.g., contradictions, hallucinations) that occur in varying inputs/outputs (e.g., sentences, documents) from different tasks. In this paper, we propose AlignScore, a new holistic metric that applies to a variety of factual inconsistency scenarios as above. AlignScore is based on a general function of information alignment between two arbitrary text pieces. Crucially, we develop a unified training framework of the alignment function by integrating a large diversity of data sources, resulting in 4.7M training examples from 7 well-established tasks (NLI, QA, paraphrasing, fact verification, information retrieval, semantic similarity, and summarization). We conduct extensive experiments on large-scale benchmarks including 22 evaluation datasets, where 19 of the datasets were never seen in the alignment training. AlignScore achieves substantial improvement over a wide range of previous metrics. Moreover, AlignScore (355M parameters) matches or even outperforms metrics based on ChatGPT and GPT-4 that are orders of magnitude larger.",
}
| Many text generation applications require the generated text to be factually consistent with input information. Automatic evaluation of factual consistency is challenging. Previous work has developed various metrics that often depend on specific functions, such as natural language inference (NLI) or question answering (QA), trained on limited data. Those metrics thus can hardly assess diverse factual inconsistencies (e.g., contradictions, hallucinations) that occur in varying inputs/outputs (e.g., sentences, documents) from different tasks. In this paper, we propose AlignScore, a new holistic metric that applies to a variety of factual inconsistency scenarios as above. AlignScore is based on a general function of information alignment between two arbitrary text pieces. Crucially, we develop a unified training framework of the alignment function by integrating a large diversity of data sources, resulting in 4.7M training examples from 7 well-established tasks (NLI, QA, paraphrasing, fact verification, information retrieval, semantic similarity, and summarization). We conduct extensive experiments on large-scale benchmarks including 22 evaluation datasets, where 19 of the datasets were never seen in the alignment training. AlignScore achieves substantial improvement over a wide range of previous metrics. Moreover, AlignScore (355M parameters) matches or even outperforms metrics based on ChatGPT and GPT-4 that are orders of magnitude larger. | [
"Zha, Yuheng",
"Yang, Yichi",
"Li, Ruichen",
"Hu, Zhiting"
] | AlignScore: Evaluating Factual Consistency with A Unified Alignment Function | acl-long.634 | Poster | 2305.16739 | [
"https://github.com/yuh-zha/alignscore"
] | https://huggingface.co/papers/2305.16739 | 1 | 0 | 0 | 4 | 1 | [
"yzha/AlignScore",
"krotima1/AlignScoreCS"
] | [] | [] |
https://aclanthology.org/2023.acl-long.635.bib | https://aclanthology.org/2023.acl-long.635/ | @inproceedings{jing-etal-2023-multi,
title = "Multi-source Semantic Graph-based Multimodal Sarcasm Explanation Generation",
author = "Jing, Liqiang and
Song, Xuemeng and
Ouyang, Kun and
Jia, Mengzhao and
Nie, Liqiang",
editor = "Rogers, Anna and
Boyd-Graber, Jordan and
Okazaki, Naoaki",
booktitle = "Proceedings of the 61st Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers)",
month = jul,
year = "2023",
address = "Toronto, Canada",
publisher = "Association for Computational Linguistics",
url = "https://aclanthology.org/2023.acl-long.635",
doi = "10.18653/v1/2023.acl-long.635",
pages = "11349--11361",
abstract = "Multimodal Sarcasm Explanation (MuSE) is a new yet challenging task, which aims to generate a natural language sentence for a multimodal social post (an image as well as its caption) to explain why it contains sarcasm. Although the existing pioneer study has achieved great success with the BART backbone, it overlooks the gap between the visual feature space and the decoder semantic space, the object-level metadata of the image, as well as the potential external knowledge. To solve these limitations, in this work, we propose a novel mulTi-source sEmantic grAph-based Multimodal sarcasm explanation scheme, named TEAM. In particular, TEAM extracts the object-level semantic meta-data instead of the traditional global visual features from the input image. Meanwhile, TEAM resorts to ConceptNet to obtain the external related knowledge concepts for the input text and the extracted object meta-data. Thereafter, TEAM introduces a multi-source semantic graph that comprehensively characterize the multi-source (i.e., caption, object meta-data, external knowledge) semantic relations to facilitate the sarcasm reasoning. Extensive experiments on a public released dataset MORE verify the superiority of our model over cutting-edge methods.",
}
| Multimodal Sarcasm Explanation (MuSE) is a new yet challenging task, which aims to generate a natural language sentence for a multimodal social post (an image as well as its caption) to explain why it contains sarcasm. Although the existing pioneer study has achieved great success with the BART backbone, it overlooks the gap between the visual feature space and the decoder semantic space, the object-level metadata of the image, as well as the potential external knowledge. To solve these limitations, in this work, we propose a novel mulTi-source sEmantic grAph-based Multimodal sarcasm explanation scheme, named TEAM. In particular, TEAM extracts the object-level semantic meta-data instead of the traditional global visual features from the input image. Meanwhile, TEAM resorts to ConceptNet to obtain the external related knowledge concepts for the input text and the extracted object meta-data. Thereafter, TEAM introduces a multi-source semantic graph that comprehensively characterize the multi-source (i.e., caption, object meta-data, external knowledge) semantic relations to facilitate the sarcasm reasoning. Extensive experiments on a public released dataset MORE verify the superiority of our model over cutting-edge methods. | [
"Jing, Liqiang",
"Song, Xuemeng",
"Ouyang, Kun",
"Jia, Mengzhao",
"Nie, Liqiang"
] | Multi-source Semantic Graph-based Multimodal Sarcasm Explanation Generation | acl-long.635 | Poster | 2306.16650 | [
"https://github.com/liqiangjing/team"
] | -1 | -1 | -1 | -1 | 0 | [] | [] | [] |
|
https://aclanthology.org/2023.acl-long.636.bib | https://aclanthology.org/2023.acl-long.636/ | @inproceedings{deng-etal-2023-counterfactual,
title = "Counterfactual Active Learning for Out-of-Distribution Generalization",
author = "Deng, Xun and
Wang, Wenjie and
Feng, Fuli and
Zhang, Hanwang and
He, Xiangnan and
Liao, Yong",
editor = "Rogers, Anna and
Boyd-Graber, Jordan and
Okazaki, Naoaki",
booktitle = "Proceedings of the 61st Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers)",
month = jul,
year = "2023",
address = "Toronto, Canada",
publisher = "Association for Computational Linguistics",
url = "https://aclanthology.org/2023.acl-long.636",
doi = "10.18653/v1/2023.acl-long.636",
pages = "11362--11377",
abstract = "We study the out-of-distribution generalization of active learning that adaptively selects samples for annotation in learning the decision boundary of classification. Our empirical study finds that increasingly annotating seen samples may hardly benefit the generalization. To address the problem, we propose Counterfactual Active Learning (CounterAL) that empowers active learning with counterfactual thinking to bridge the seen samples with unseen cases. In addition to annotating factual samples, CounterAL requires annotators to answer counterfactual questions to construct counterfactual samples for training. To achieve CounterAL, we design a new acquisition strategy that selects the informative factual-counterfactual pairs for annotation; and a new training strategy that pushes the model update to focus on the discrepancy between factual and counterfactual samples. We evaluate CounterAL on multiple public datasets of sentiment analysis and natural language inference. The experiment results show that CounterAL requires fewer acquisition rounds and outperforms existing active learning methods by a large margin in OOD tests with comparable IID performance.",
}
| We study the out-of-distribution generalization of active learning that adaptively selects samples for annotation in learning the decision boundary of classification. Our empirical study finds that increasingly annotating seen samples may hardly benefit the generalization. To address the problem, we propose Counterfactual Active Learning (CounterAL) that empowers active learning with counterfactual thinking to bridge the seen samples with unseen cases. In addition to annotating factual samples, CounterAL requires annotators to answer counterfactual questions to construct counterfactual samples for training. To achieve CounterAL, we design a new acquisition strategy that selects the informative factual-counterfactual pairs for annotation; and a new training strategy that pushes the model update to focus on the discrepancy between factual and counterfactual samples. We evaluate CounterAL on multiple public datasets of sentiment analysis and natural language inference. The experiment results show that CounterAL requires fewer acquisition rounds and outperforms existing active learning methods by a large margin in OOD tests with comparable IID performance. | [
"Deng, Xun",
"Wang, Wenjie",
"Feng, Fuli",
"Zhang, Hanwang",
"He, Xiangnan",
"Liao, Yong"
] | Counterfactual Active Learning for Out-of-Distribution Generalization | acl-long.636 | Poster | [
""
] | -1 | -1 | -1 | -1 | 0 | [] | [] | [] |
||
https://aclanthology.org/2023.acl-long.637.bib | https://aclanthology.org/2023.acl-long.637/ | @inproceedings{chen-etal-2023-multi,
title = "Multi-granularity Temporal Question Answering over Knowledge Graphs",
author = "Chen, Ziyang and
Liao, Jinzhi and
Zhao, Xiang",
editor = "Rogers, Anna and
Boyd-Graber, Jordan and
Okazaki, Naoaki",
booktitle = "Proceedings of the 61st Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers)",
month = jul,
year = "2023",
address = "Toronto, Canada",
publisher = "Association for Computational Linguistics",
url = "https://aclanthology.org/2023.acl-long.637",
doi = "10.18653/v1/2023.acl-long.637",
pages = "11378--11392",
abstract = "Recently, question answering over temporal knowledge graphs (i.e., TKGQA) has been introduced and investigated, in quest of reasoning about dynamic factual knowledge. To foster research on TKGQA, a few datasets have been curated (e.g., CronQuestions and Complex-CronQuestions), and various models have been proposed based on these datasets. Nevertheless, existing efforts overlook the fact that real-life applications of TKGQA also tend to be complex in temporal granularity, i.e., the questions may concern mixed temporal granularities (e.g., both day and month). To overcome the limitation, in this paper, we motivate the notion of multi-granularity temporal question answering over knowledge graphs and present a large scale dataset for multi-granularity TKGQA, namely MultiTQ. To the best of our knowledge, MultiTQis among the first of its kind, and compared with existing datasets on TKGQA, MultiTQfeatures at least two desirable aspects{---}ample relevant facts and multiple temporal granularities. It is expected to better reflect real-world challenges, and serve as a test bed for TKGQA models. In addition, we propose a competing baseline MultiQA over MultiTQ, which is experimentally demonstrated to be effective in dealing with TKGQA. The data and code are released at \url{https://github.com/czy1999/MultiTQ}.",
}
| Recently, question answering over temporal knowledge graphs (i.e., TKGQA) has been introduced and investigated, in quest of reasoning about dynamic factual knowledge. To foster research on TKGQA, a few datasets have been curated (e.g., CronQuestions and Complex-CronQuestions), and various models have been proposed based on these datasets. Nevertheless, existing efforts overlook the fact that real-life applications of TKGQA also tend to be complex in temporal granularity, i.e., the questions may concern mixed temporal granularities (e.g., both day and month). To overcome the limitation, in this paper, we motivate the notion of multi-granularity temporal question answering over knowledge graphs and present a large scale dataset for multi-granularity TKGQA, namely MultiTQ. To the best of our knowledge, MultiTQis among the first of its kind, and compared with existing datasets on TKGQA, MultiTQfeatures at least two desirable aspects{---}ample relevant facts and multiple temporal granularities. It is expected to better reflect real-world challenges, and serve as a test bed for TKGQA models. In addition, we propose a competing baseline MultiQA over MultiTQ, which is experimentally demonstrated to be effective in dealing with TKGQA. The data and code are released at \url{https://github.com/czy1999/MultiTQ}. | [
"Chen, Ziyang",
"Liao, Jinzhi",
"Zhao, Xiang"
] | Multi-granularity Temporal Question Answering over Knowledge Graphs | acl-long.637 | Poster | [
""
] | -1 | -1 | -1 | -1 | 0 | [] | [] | [] |
||
https://aclanthology.org/2023.acl-long.638.bib | https://aclanthology.org/2023.acl-long.638/ | @inproceedings{toborek-etal-2023-new,
title = "A New Aligned Simple {G}erman Corpus",
author = "Toborek, Vanessa and
Busch, Moritz and
Bo{\ss}ert, Malte and
Bauckhage, Christian and
Welke, Pascal",
editor = "Rogers, Anna and
Boyd-Graber, Jordan and
Okazaki, Naoaki",
booktitle = "Proceedings of the 61st Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers)",
month = jul,
year = "2023",
address = "Toronto, Canada",
publisher = "Association for Computational Linguistics",
url = "https://aclanthology.org/2023.acl-long.638",
doi = "10.18653/v1/2023.acl-long.638",
pages = "11393--11412",
abstract = "{``}Leichte Sprache{''}, the German counterpart to Simple English, is a regulated language aiming to facilitate complex written language that would otherwise stay inaccessible to different groups of people. We present a new sentence-aligned monolingual corpus for Simple German {--} German. It contains multiple document-aligned sources which we have aligned using automatic sentence-alignment methods. We evaluate our alignments based on a manually labelled subset of aligned documents. The quality of our sentence alignments, as measured by the F1-score, surpasses previous work. We publish the dataset under CC BY-SA and the accompanying code under MIT license.",
}
| {``}Leichte Sprache{''}, the German counterpart to Simple English, is a regulated language aiming to facilitate complex written language that would otherwise stay inaccessible to different groups of people. We present a new sentence-aligned monolingual corpus for Simple German {--} German. It contains multiple document-aligned sources which we have aligned using automatic sentence-alignment methods. We evaluate our alignments based on a manually labelled subset of aligned documents. The quality of our sentence alignments, as measured by the F1-score, surpasses previous work. We publish the dataset under CC BY-SA and the accompanying code under MIT license. | [
"Toborek, Vanessa",
"Busch, Moritz",
"Bo{\\ss}ert, Malte",
"Bauckhage, Christian",
"Welke, Pascal"
] | A New Aligned Simple German Corpus | acl-long.638 | Poster | 2209.01106 | [
"https://github.com/mlai-bonn/simple-german-corpus"
] | -1 | -1 | -1 | -1 | 0 | [] | [] | [] |
|
https://aclanthology.org/2023.acl-long.639.bib | https://aclanthology.org/2023.acl-long.639/ | @inproceedings{xu-etal-2023-introducing,
title = "Introducing Semantics into Speech Encoders",
author = "Xu, Derek and
Dong, Shuyan and
Wang, Changhan and
Kim, Suyoun and
Lin, Zhaojiang and
Liu, Bing and
Shrivastava, Akshat and
Li, Shang-Wen and
Tseng, Liang-Hsuan and
Lin, Guan-Ting and
Baevski, Alexei and
Lee, Hung-yi and
Sun, Yizhou and
Wang, Wei",
editor = "Rogers, Anna and
Boyd-Graber, Jordan and
Okazaki, Naoaki",
booktitle = "Proceedings of the 61st Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers)",
month = jul,
year = "2023",
address = "Toronto, Canada",
publisher = "Association for Computational Linguistics",
url = "https://aclanthology.org/2023.acl-long.639",
doi = "10.18653/v1/2023.acl-long.639",
pages = "11413--11429",
abstract = "Recent studies find existing self-supervised speech encoders contain primarily acoustic rather than semantic information. As a result, pipelined supervised automatic speech recognition (ASR) to large language model (LLM) systems achieve state-of-the-art results on semantic spoken language tasks by utilizing rich semantic representations from the LLM. These systems come at the cost of labeled audio transcriptions, which is expensive and time-consuming to obtain. We propose a task-agnostic unsupervised way of incorporating semantic information from LLMs into self-supervised speech encoders without labeled audio transcriptions. By introducing semantics, we improve existing speech encoder spoken language understanding (SLU) performance by over 5{\%} on intent classification (IC), with modest gains in named entity resolution (NER) and slot filling (SF), and spoken question answering (SQA) FF1 score by over 2{\%}. Our approach, which uses no ASR data, achieves similar performance as methods trained on over 100 hours of labeled audio transcripts, demonstrating the feasibility of unsupervised semantic augmentations to existing speech encoders.",
}
| Recent studies find existing self-supervised speech encoders contain primarily acoustic rather than semantic information. As a result, pipelined supervised automatic speech recognition (ASR) to large language model (LLM) systems achieve state-of-the-art results on semantic spoken language tasks by utilizing rich semantic representations from the LLM. These systems come at the cost of labeled audio transcriptions, which is expensive and time-consuming to obtain. We propose a task-agnostic unsupervised way of incorporating semantic information from LLMs into self-supervised speech encoders without labeled audio transcriptions. By introducing semantics, we improve existing speech encoder spoken language understanding (SLU) performance by over 5{\%} on intent classification (IC), with modest gains in named entity resolution (NER) and slot filling (SF), and spoken question answering (SQA) FF1 score by over 2{\%}. Our approach, which uses no ASR data, achieves similar performance as methods trained on over 100 hours of labeled audio transcripts, demonstrating the feasibility of unsupervised semantic augmentations to existing speech encoders. | [
"Xu, Derek",
"Dong, Shuyan",
"Wang, Changhan",
"Kim, Suyoun",
"Lin, Zhaojiang",
"Liu, Bing",
"Shrivastava, Akshat",
"Li, Shang-Wen",
"Tseng, Liang-Hsuan",
"Lin, Guan-Ting",
"Baevski, Alexei",
"Lee, Hung-yi",
"Sun, Yizhou",
"Wang, Wei"
] | Introducing Semantics into Speech Encoders | acl-long.639 | Poster | 2211.08402 | [
""
] | -1 | -1 | -1 | -1 | 0 | [] | [] | [] |
|
https://aclanthology.org/2023.acl-long.640.bib | https://aclanthology.org/2023.acl-long.640/ | @inproceedings{xue-etal-2023-constrained,
title = "Constrained Tuple Extraction with Interaction-Aware Network",
author = "Xue, Xiaojun and
Zhang, Chunxia and
Xu, Tianxiang and
Niu, Zhendong",
editor = "Rogers, Anna and
Boyd-Graber, Jordan and
Okazaki, Naoaki",
booktitle = "Proceedings of the 61st Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers)",
month = jul,
year = "2023",
address = "Toronto, Canada",
publisher = "Association for Computational Linguistics",
url = "https://aclanthology.org/2023.acl-long.640",
doi = "10.18653/v1/2023.acl-long.640",
pages = "11430--11444",
abstract = "Tuples extraction is a fundamental task for information extraction and knowledge graph construction. The extracted tuples are usually represented as knowledge triples consisting of subject, relation, and object. In practice, however, the validity of knowledge triples is associated with and changes with the spatial, temporal, or other kinds of constraints. Motivated by this observation, this paper proposes a constrained tuple extraction (CTE) task to guarantee the validity of knowledge tuples. Formally, the CTE task is to extract constrained tuples from unstructured text, which adds constraints to conventional triples. To this end, we propose an interaction-aware network. Combinatorial interactions among context-specific external features and distinct-granularity internal features are exploited to effectively mine the potential constraints. Moreover, we have built a new dataset containing totally 1,748,826 constrained tuples for training and 3656 ones for evaluation. Experiments on our dataset and the public CaRB dataset demonstrate the superiority of the proposed model. The constructed dataset and the codes are publicly available.",
}
| Tuples extraction is a fundamental task for information extraction and knowledge graph construction. The extracted tuples are usually represented as knowledge triples consisting of subject, relation, and object. In practice, however, the validity of knowledge triples is associated with and changes with the spatial, temporal, or other kinds of constraints. Motivated by this observation, this paper proposes a constrained tuple extraction (CTE) task to guarantee the validity of knowledge tuples. Formally, the CTE task is to extract constrained tuples from unstructured text, which adds constraints to conventional triples. To this end, we propose an interaction-aware network. Combinatorial interactions among context-specific external features and distinct-granularity internal features are exploited to effectively mine the potential constraints. Moreover, we have built a new dataset containing totally 1,748,826 constrained tuples for training and 3656 ones for evaluation. Experiments on our dataset and the public CaRB dataset demonstrate the superiority of the proposed model. The constructed dataset and the codes are publicly available. | [
"Xue, Xiaojun",
"Zhang, Chunxia",
"Xu, Tianxiang",
"Niu, Zhendong"
] | Constrained Tuple Extraction with Interaction-Aware Network | acl-long.640 | Poster | [
""
] | -1 | -1 | -1 | -1 | 0 | [] | [] | [] |
||
https://aclanthology.org/2023.acl-long.641.bib | https://aclanthology.org/2023.acl-long.641/ | @inproceedings{xu-etal-2023-multiinstruct,
title = "{M}ulti{I}nstruct: Improving Multi-Modal Zero-Shot Learning via Instruction Tuning",
author = "Xu, Zhiyang and
Shen, Ying and
Huang, Lifu",
editor = "Rogers, Anna and
Boyd-Graber, Jordan and
Okazaki, Naoaki",
booktitle = "Proceedings of the 61st Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers)",
month = jul,
year = "2023",
address = "Toronto, Canada",
publisher = "Association for Computational Linguistics",
url = "https://aclanthology.org/2023.acl-long.641",
doi = "10.18653/v1/2023.acl-long.641",
pages = "11445--11465",
abstract = "Instruction tuning, a new learning paradigm that fine-tunes pre-trained language models on tasks specified through instructions, has shown promising zero-shot performance on various natural language processing tasks. However, it has yet to be explored for vision and multimodal tasks. In this work, we introduce MultiInstruct, the first multimodal instruction tuning benchmark dataset that consists of 62 diverse multimodal tasks in a unified seq-to-seq format covering 10 broad categories. The tasks are derived from 21 existing open-source datasets and each task is equipped with 5 expert-written instructions. We take OFA as the base pre-trained model for multimodal instruction tuning, and to further improve its zero-shot performance, we explore multiple transfer learning strategies to leverage the large-scale Natural Instructions dataset. Experimental results demonstrate strong zero-shot performance on various unseen multimodal tasks and the benefit of transfer learning from a text-only instruction dataset. We also design a new evaluation metric {--} Sensitivity, to evaluate how sensitive the model is to the variety of instructions. Our results indicate that fine-tuning the model on a diverse set of tasks and instructions leads to a reduced sensitivity to variations in instructions for each task.",
}
| Instruction tuning, a new learning paradigm that fine-tunes pre-trained language models on tasks specified through instructions, has shown promising zero-shot performance on various natural language processing tasks. However, it has yet to be explored for vision and multimodal tasks. In this work, we introduce MultiInstruct, the first multimodal instruction tuning benchmark dataset that consists of 62 diverse multimodal tasks in a unified seq-to-seq format covering 10 broad categories. The tasks are derived from 21 existing open-source datasets and each task is equipped with 5 expert-written instructions. We take OFA as the base pre-trained model for multimodal instruction tuning, and to further improve its zero-shot performance, we explore multiple transfer learning strategies to leverage the large-scale Natural Instructions dataset. Experimental results demonstrate strong zero-shot performance on various unseen multimodal tasks and the benefit of transfer learning from a text-only instruction dataset. We also design a new evaluation metric {--} Sensitivity, to evaluate how sensitive the model is to the variety of instructions. Our results indicate that fine-tuning the model on a diverse set of tasks and instructions leads to a reduced sensitivity to variations in instructions for each task. | [
"Xu, Zhiyang",
"Shen, Ying",
"Huang, Lifu"
] | MultiInstruct: Improving Multi-Modal Zero-Shot Learning via Instruction Tuning | acl-long.641 | Poster | 2212.10773 | [
"https://github.com/vt-nlp/multiinstruct"
] | -1 | -1 | -1 | -1 | 0 | [] | [] | [] |
|
https://aclanthology.org/2023.acl-long.642.bib | https://aclanthology.org/2023.acl-long.642/ | @inproceedings{ramesh-etal-2023-single,
title = "Single Sequence Prediction over Reasoning Graphs for Multi-hop {QA}",
author = "Ramesh, Gowtham and
Sreedhar, Makesh Narsimhan and
Hu, Junjie",
editor = "Rogers, Anna and
Boyd-Graber, Jordan and
Okazaki, Naoaki",
booktitle = "Proceedings of the 61st Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers)",
month = jul,
year = "2023",
address = "Toronto, Canada",
publisher = "Association for Computational Linguistics",
url = "https://aclanthology.org/2023.acl-long.642",
doi = "10.18653/v1/2023.acl-long.642",
pages = "11466--11481",
abstract = "Recent generative approaches for multi-hop question answering (QA) utilize the fusion-in-decoder method to generate a single sequence output which includes both a final answer and a reasoning path taken to arrive at that answer, such as passage titles and key facts from those passages. While such models can lead to better interpretability and high quantitative scores, they often have difficulty accurately identifying the passages corresponding to key entities in the context, resulting in incorrect passage hops and a lack of faithfulness in the reasoning path. To address this, we propose a single-sequence prediction method over a local reasoning graph that integrates a graph structure connecting key entities in each context passage to relevant subsequent passages for each question. We use a graph neural network to encode this graph structure and fuse the resulting representations into the entity representations of the model. Our experiments show significant improvements in answer exact-match/F1 scores and faithfulness of grounding in the reasoning path on the HotpotQA dataset and achieve state-of-the-art numbers on the Musique dataset with only up to a 4{\%} increase in model parameters.",
}
| Recent generative approaches for multi-hop question answering (QA) utilize the fusion-in-decoder method to generate a single sequence output which includes both a final answer and a reasoning path taken to arrive at that answer, such as passage titles and key facts from those passages. While such models can lead to better interpretability and high quantitative scores, they often have difficulty accurately identifying the passages corresponding to key entities in the context, resulting in incorrect passage hops and a lack of faithfulness in the reasoning path. To address this, we propose a single-sequence prediction method over a local reasoning graph that integrates a graph structure connecting key entities in each context passage to relevant subsequent passages for each question. We use a graph neural network to encode this graph structure and fuse the resulting representations into the entity representations of the model. Our experiments show significant improvements in answer exact-match/F1 scores and faithfulness of grounding in the reasoning path on the HotpotQA dataset and achieve state-of-the-art numbers on the Musique dataset with only up to a 4{\%} increase in model parameters. | [
"Ramesh, Gowtham",
"Sreedhar, Makesh Narsimhan",
"Hu, Junjie"
] | Single Sequence Prediction over Reasoning Graphs for Multi-hop QA | acl-long.642 | Poster | 2307.00335 | [
""
] | -1 | -1 | -1 | -1 | 0 | [] | [] | [] |
|
https://aclanthology.org/2023.acl-long.643.bib | https://aclanthology.org/2023.acl-long.643/ | @inproceedings{ladhak-etal-2023-contrastive,
title = "Contrastive Error Attribution for Finetuned Language Models",
author = "Ladhak, Faisal and
Durmus, Esin and
Hashimoto, Tatsunori",
editor = "Rogers, Anna and
Boyd-Graber, Jordan and
Okazaki, Naoaki",
booktitle = "Proceedings of the 61st Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers)",
month = jul,
year = "2023",
address = "Toronto, Canada",
publisher = "Association for Computational Linguistics",
url = "https://aclanthology.org/2023.acl-long.643",
doi = "10.18653/v1/2023.acl-long.643",
pages = "11482--11498",
abstract = "Recent work has identified noisy and misannotated data as a core cause of hallucinations and unfaithful outputs in Natural Language Generation (NLG) tasks. Consequently, identifying and removing these examples is a key open challenge in creating reliable NLG systems. In this work, we introduce a framework to identify and remove low-quality training instances that lead to undesirable outputs, such as faithfulness errors in text summarization. We show that existing approaches for error tracing, such as gradient-based influence measures, do not perform reliably for detecting faithfulness errors in NLG datasets. We overcome the drawbacks of existing error tracing methods through a new, contrast-based estimate that compares undesired generations to human-corrected outputs. Our proposed method can achieve a mean average precision of 0.93 at detecting known data errors across synthetic tasks with known ground truth, substantially outperforming existing approaches. Using this approach and re-training models on cleaned data leads to a 70{\%} reduction in entity hallucinations on the NYT dataset and a 55{\%} reduction in semantic errors on the E2E dataset.",
}
| Recent work has identified noisy and misannotated data as a core cause of hallucinations and unfaithful outputs in Natural Language Generation (NLG) tasks. Consequently, identifying and removing these examples is a key open challenge in creating reliable NLG systems. In this work, we introduce a framework to identify and remove low-quality training instances that lead to undesirable outputs, such as faithfulness errors in text summarization. We show that existing approaches for error tracing, such as gradient-based influence measures, do not perform reliably for detecting faithfulness errors in NLG datasets. We overcome the drawbacks of existing error tracing methods through a new, contrast-based estimate that compares undesired generations to human-corrected outputs. Our proposed method can achieve a mean average precision of 0.93 at detecting known data errors across synthetic tasks with known ground truth, substantially outperforming existing approaches. Using this approach and re-training models on cleaned data leads to a 70{\%} reduction in entity hallucinations on the NYT dataset and a 55{\%} reduction in semantic errors on the E2E dataset. | [
"Ladhak, Faisal",
"Durmus, Esin",
"Hashimoto, Tatsunori"
] | Contrastive Error Attribution for Finetuned Language Models | acl-long.643 | Poster | 2212.10722 | [
"https://github.com/fladhak/contrastive_error_attribution"
] | -1 | -1 | -1 | -1 | 0 | [] | [] | [] |
|
https://aclanthology.org/2023.acl-long.644.bib | https://aclanthology.org/2023.acl-long.644/ | @inproceedings{ivankay-etal-2023-dare,
title = "{DARE}: Towards Robust Text Explanations in Biomedical and Healthcare Applications",
author = "Ivankay, Adam and
Rigotti, Mattia and
Frossard, Pascal",
editor = "Rogers, Anna and
Boyd-Graber, Jordan and
Okazaki, Naoaki",
booktitle = "Proceedings of the 61st Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers)",
month = jul,
year = "2023",
address = "Toronto, Canada",
publisher = "Association for Computational Linguistics",
url = "https://aclanthology.org/2023.acl-long.644",
doi = "10.18653/v1/2023.acl-long.644",
pages = "11499--11533",
abstract = "Along with the successful deployment of deep neural networks in several application domains, the need to unravel the black-box nature of these networks has seen a significant increase recently. Several methods have been introduced to provide insight into the inference process of deep neural networks. However, most of these explainability methods have been shown to be brittle in the face of adversarial perturbations of their inputs in the image and generic textual domain. In this work we show that this phenomenon extends to specific and important high stakes domains like biomedical datasets. In particular, we observe that the robustness of explanations should be characterized in terms of the accuracy of the explanation in linking a model{'}s inputs and its decisions - faithfulness - and its relevance from the perspective of domain experts - plausibility. This is crucial to prevent explanations that are inaccurate but still look convincing in the context of the domain at hand. To this end, we show how to adapt current attribution robustness estimation methods to a given domain, so as to take into account domain-specific plausibility. This results in our DomainAdaptiveAREstimator (DARE) attribution robustness estimator, allowing us to properly characterize the domain-specific robustness of faithful explanations. Next, we provide two methods, adversarial training and FAR training, to mitigate the brittleness characterized by DARE, allowing us to train networks that display robust attributions. Finally, we empirically validate our methods with extensive experiments on three established biomedical benchmarks.",
}
| Along with the successful deployment of deep neural networks in several application domains, the need to unravel the black-box nature of these networks has seen a significant increase recently. Several methods have been introduced to provide insight into the inference process of deep neural networks. However, most of these explainability methods have been shown to be brittle in the face of adversarial perturbations of their inputs in the image and generic textual domain. In this work we show that this phenomenon extends to specific and important high stakes domains like biomedical datasets. In particular, we observe that the robustness of explanations should be characterized in terms of the accuracy of the explanation in linking a model{'}s inputs and its decisions - faithfulness - and its relevance from the perspective of domain experts - plausibility. This is crucial to prevent explanations that are inaccurate but still look convincing in the context of the domain at hand. To this end, we show how to adapt current attribution robustness estimation methods to a given domain, so as to take into account domain-specific plausibility. This results in our DomainAdaptiveAREstimator (DARE) attribution robustness estimator, allowing us to properly characterize the domain-specific robustness of faithful explanations. Next, we provide two methods, adversarial training and FAR training, to mitigate the brittleness characterized by DARE, allowing us to train networks that display robust attributions. Finally, we empirically validate our methods with extensive experiments on three established biomedical benchmarks. | [
"Ivankay, Adam",
"Rigotti, Mattia",
"Frossard, Pascal"
] | DARE: Towards Robust Text Explanations in Biomedical and Healthcare Applications | acl-long.644 | Oral | 2307.02094 | [
"https://github.com/ibm/domain-adaptive-attribution-robustness"
] | -1 | -1 | -1 | -1 | 0 | [] | [] | [] |
|
https://aclanthology.org/2023.acl-long.645.bib | https://aclanthology.org/2023.acl-long.645/ | @inproceedings{petersen-etal-2023-neural,
title = "Neural Machine Translation for Mathematical Formulae",
author = "Petersen, Felix and
Schubotz, Moritz and
Greiner-Petter, Andre and
Gipp, Bela",
editor = "Rogers, Anna and
Boyd-Graber, Jordan and
Okazaki, Naoaki",
booktitle = "Proceedings of the 61st Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers)",
month = jul,
year = "2023",
address = "Toronto, Canada",
publisher = "Association for Computational Linguistics",
url = "https://aclanthology.org/2023.acl-long.645",
doi = "10.18653/v1/2023.acl-long.645",
pages = "11534--11550",
abstract = "We tackle the problem of neural machine translation of mathematical formulae between ambiguous presentation languages and unambiguous content languages. Compared to neural machine translation on natural language, mathematical formulae have a much smaller vocabulary and much longer sequences of symbols, while their translation requires extreme precision to satisfy mathematical information needs. In this work, we perform the tasks of translating from LaTeX to Mathematica as well as from LaTeX to semantic LaTeX. While recurrent, recursive, and transformer networks struggle with preserving all contained information, we find that convolutional sequence-to-sequence networks achieve 95.1{\%} and 90.7{\%} exact matches, respectively.",
}
| We tackle the problem of neural machine translation of mathematical formulae between ambiguous presentation languages and unambiguous content languages. Compared to neural machine translation on natural language, mathematical formulae have a much smaller vocabulary and much longer sequences of symbols, while their translation requires extreme precision to satisfy mathematical information needs. In this work, we perform the tasks of translating from LaTeX to Mathematica as well as from LaTeX to semantic LaTeX. While recurrent, recursive, and transformer networks struggle with preserving all contained information, we find that convolutional sequence-to-sequence networks achieve 95.1{\%} and 90.7{\%} exact matches, respectively. | [
"Petersen, Felix",
"Schubotz, Moritz",
"Greiner-Petter, Andre",
"Gipp, Bela"
] | Neural Machine Translation for Mathematical Formulae | acl-long.645 | Poster | 2305.16433 | [
""
] | -1 | -1 | -1 | -1 | 0 | [] | [] | [] |
|
https://aclanthology.org/2023.acl-long.646.bib | https://aclanthology.org/2023.acl-long.646/ | @inproceedings{lee-etal-2023-query,
title = "Query-Efficient Black-Box Red Teaming via {B}ayesian Optimization",
author = "Lee, Deokjae and
Lee, JunYeong and
Ha, Jung-Woo and
Kim, Jin-Hwa and
Lee, Sang-Woo and
Lee, Hwaran and
Song, Hyun Oh",
editor = "Rogers, Anna and
Boyd-Graber, Jordan and
Okazaki, Naoaki",
booktitle = "Proceedings of the 61st Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers)",
month = jul,
year = "2023",
address = "Toronto, Canada",
publisher = "Association for Computational Linguistics",
url = "https://aclanthology.org/2023.acl-long.646",
doi = "10.18653/v1/2023.acl-long.646",
pages = "11551--11574",
abstract = "The deployment of large-scale generative models is often restricted by their potential risk of causing harm to users in unpredictable ways. We focus on the problem of black-box red teaming, where a red team generates test cases and interacts with the victim model to discover a diverse set of failures with limited query access. Existing red teaming methods construct test cases based on human supervision or language model (LM) and query all test cases in a brute-force manner without incorporating any information from past evaluations, resulting in a prohibitively large number of queries. To this end, we propose \textit{Bayesian red teaming} (BRT), novel query-efficient black-box red teaming methods based on Bayesian optimization, which iteratively identify diverse positive test cases leading to model failures by utilizing the pre-defined user input pool and the past evaluations. Experimental results on various user input pools demonstrate that our method consistently finds a significantly larger number of diverse positive test cases under the limited query budget than the baseline methods.The source code is available at \url{https://github.com/snu-mllab/Bayesian-Red-Teaming}.",
}
| The deployment of large-scale generative models is often restricted by their potential risk of causing harm to users in unpredictable ways. We focus on the problem of black-box red teaming, where a red team generates test cases and interacts with the victim model to discover a diverse set of failures with limited query access. Existing red teaming methods construct test cases based on human supervision or language model (LM) and query all test cases in a brute-force manner without incorporating any information from past evaluations, resulting in a prohibitively large number of queries. To this end, we propose \textit{Bayesian red teaming} (BRT), novel query-efficient black-box red teaming methods based on Bayesian optimization, which iteratively identify diverse positive test cases leading to model failures by utilizing the pre-defined user input pool and the past evaluations. Experimental results on various user input pools demonstrate that our method consistently finds a significantly larger number of diverse positive test cases under the limited query budget than the baseline methods.The source code is available at \url{https://github.com/snu-mllab/Bayesian-Red-Teaming}. | [
"Lee, Deokjae",
"Lee, JunYeong",
"Ha, Jung-Woo",
"Kim, Jin-Hwa",
"Lee, Sang-Woo",
"Lee, Hwaran",
"Song, Hyun Oh"
] | Query-Efficient Black-Box Red Teaming via Bayesian Optimization | acl-long.646 | Poster | 2305.17444 | [
"https://github.com/snu-mllab/bayesian-red-teaming"
] | -1 | -1 | -1 | -1 | 0 | [] | [] | [] |
|
https://aclanthology.org/2023.acl-long.647.bib | https://aclanthology.org/2023.acl-long.647/ | @inproceedings{han-etal-2023-ssd,
title = "{SSD}-{LM}: Semi-autoregressive Simplex-based Diffusion Language Model for Text Generation and Modular Control",
author = "Han, Xiaochuang and
Kumar, Sachin and
Tsvetkov, Yulia",
editor = "Rogers, Anna and
Boyd-Graber, Jordan and
Okazaki, Naoaki",
booktitle = "Proceedings of the 61st Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers)",
month = jul,
year = "2023",
address = "Toronto, Canada",
publisher = "Association for Computational Linguistics",
url = "https://aclanthology.org/2023.acl-long.647",
doi = "10.18653/v1/2023.acl-long.647",
pages = "11575--11596",
abstract = "Despite the growing success of diffusion models in continuous-valued domains (e.g., images), similar efforts for discrete domains such as text have yet to match the performance of autoregressive language models. In this work, we present SSD-LM{---}a diffusion-based language model with two key design choices. First, SSD-LM is semi-autoregressive, iteratively generating blocks of text, allowing for flexible output length at decoding time while enabling local bidirectional context updates. Second, it is simplex-based, performing diffusion on the natural vocabulary space rather than a learned latent space, allowing us to incorporate classifier guidance and modular control using off-the-shelf classifiers without any adaptation. We evaluate SSD-LM on unconstrained text generation benchmarks, and show that it matches or outperforms strong autoregressive GPT-2 models across standard quality and diversity metrics, while vastly outperforming diffusion-based baselines. On controlled text generation, SSD-LM also outperforms competitive baselines, with an extra advantage in modularity.",
}
| Despite the growing success of diffusion models in continuous-valued domains (e.g., images), similar efforts for discrete domains such as text have yet to match the performance of autoregressive language models. In this work, we present SSD-LM{---}a diffusion-based language model with two key design choices. First, SSD-LM is semi-autoregressive, iteratively generating blocks of text, allowing for flexible output length at decoding time while enabling local bidirectional context updates. Second, it is simplex-based, performing diffusion on the natural vocabulary space rather than a learned latent space, allowing us to incorporate classifier guidance and modular control using off-the-shelf classifiers without any adaptation. We evaluate SSD-LM on unconstrained text generation benchmarks, and show that it matches or outperforms strong autoregressive GPT-2 models across standard quality and diversity metrics, while vastly outperforming diffusion-based baselines. On controlled text generation, SSD-LM also outperforms competitive baselines, with an extra advantage in modularity. | [
"Han, Xiaochuang",
"Kumar, Sachin",
"Tsvetkov, Yulia"
] | SSD-LM: Semi-autoregressive Simplex-based Diffusion Language Model for Text Generation and Modular Control | acl-long.647 | Poster | 2210.17432 | [
"https://github.com/xhan77/ssd-lm"
] | https://huggingface.co/papers/2210.17432 | 2 | 1 | 0 | 3 | 1 | [] | [] | [] |
https://aclanthology.org/2023.acl-long.648.bib | https://aclanthology.org/2023.acl-long.648/ | @inproceedings{jiang-etal-2023-recall,
title = "Recall, Expand, and Multi-Candidate Cross-Encode: Fast and Accurate Ultra-Fine Entity Typing",
author = "Jiang, Chengyue and
Hui, Wenyang and
Jiang, Yong and
Wang, Xiaobin and
Xie, Pengjun and
Tu, Kewei",
editor = "Rogers, Anna and
Boyd-Graber, Jordan and
Okazaki, Naoaki",
booktitle = "Proceedings of the 61st Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers)",
month = jul,
year = "2023",
address = "Toronto, Canada",
publisher = "Association for Computational Linguistics",
url = "https://aclanthology.org/2023.acl-long.648",
doi = "10.18653/v1/2023.acl-long.648",
pages = "11597--11609",
abstract = "Ultra-fine entity typing (UFET) predicts extremely free-formed types (e.g., \textit{president, politician}) of a given entity mention (e.g., \textit{Joe Biden}) in context. State-of-the-art (SOTA) methods use the cross-encoder (CE) based architecture. CE concatenates a mention (and its context) with each type and feeds the pair into a pretrained language model (PLM) to score their relevance. It brings deeper interaction between the mention and the type to reach better performance but has to perform $N$ (the type set size) forward passes to infer all the types of a single mention. CE is therefore very slow in inference when the type set is large (e.g., $N=10k$ for UFET). {\%} Cross-encoder also ignores the correlation between different types.To this end, we propose to perform entity typing in a recall-expand-filter manner. The recall and expansion stages prune the large type set and generate $K$ (typically much smaller than $N$) most relevant type candidates for each mention. At the filter stage, we use a novel model called {pasted macro {`}NAME{'}} to concurrently encode and score all these $K$ candidates in only one forward pass to obtain the final type prediction. We investigate different model options for each stage and conduct extensive experiments to compare each option, experiments show that our method reaches SOTA performance on UFET and is thousands of times faster than the CE-based architecture. We also found our method is very effective in fine-grained (130 types) and coarse-grained (9 types) entity typing. Our code is available at {pasted macro {`}CODE{'}}.",
}
| Ultra-fine entity typing (UFET) predicts extremely free-formed types (e.g., \textit{president, politician}) of a given entity mention (e.g., \textit{Joe Biden}) in context. State-of-the-art (SOTA) methods use the cross-encoder (CE) based architecture. CE concatenates a mention (and its context) with each type and feeds the pair into a pretrained language model (PLM) to score their relevance. It brings deeper interaction between the mention and the type to reach better performance but has to perform $N$ (the type set size) forward passes to infer all the types of a single mention. CE is therefore very slow in inference when the type set is large (e.g., $N=10k$ for UFET). {\%} Cross-encoder also ignores the correlation between different types.To this end, we propose to perform entity typing in a recall-expand-filter manner. The recall and expansion stages prune the large type set and generate $K$ (typically much smaller than $N$) most relevant type candidates for each mention. At the filter stage, we use a novel model called {pasted macro {`}NAME{'}} to concurrently encode and score all these $K$ candidates in only one forward pass to obtain the final type prediction. We investigate different model options for each stage and conduct extensive experiments to compare each option, experiments show that our method reaches SOTA performance on UFET and is thousands of times faster than the CE-based architecture. We also found our method is very effective in fine-grained (130 types) and coarse-grained (9 types) entity typing. Our code is available at {pasted macro {`}CODE{'}}. | [
"Jiang, Chengyue",
"Hui, Wenyang",
"Jiang, Yong",
"Wang, Xiaobin",
"Xie, Pengjun",
"Tu, Kewei"
] | Recall, Expand, and Multi-Candidate Cross-Encode: Fast and Accurate Ultra-Fine Entity Typing | acl-long.648 | Poster | [
"https://github.com/modelscope/adaseq"
] | -1 | -1 | -1 | -1 | 0 | [] | [] | [] |
||
https://aclanthology.org/2023.acl-long.649.bib | https://aclanthology.org/2023.acl-long.649/ | @inproceedings{hu-etal-2023-mir,
title = "{MIR}-{GAN}: Refining Frame-Level Modality-Invariant Representations with Adversarial Network for Audio-Visual Speech Recognition",
author = "Hu, Yuchen and
Chen, Chen and
Li, Ruizhe and
Zou, Heqing and
Chng, Eng Siong",
editor = "Rogers, Anna and
Boyd-Graber, Jordan and
Okazaki, Naoaki",
booktitle = "Proceedings of the 61st Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers)",
month = jul,
year = "2023",
address = "Toronto, Canada",
publisher = "Association for Computational Linguistics",
url = "https://aclanthology.org/2023.acl-long.649",
doi = "10.18653/v1/2023.acl-long.649",
pages = "11610--11625",
abstract = "Audio-visual speech recognition (AVSR) attracts a surge of research interest recently by leveraging multimodal signals to understand human speech. Mainstream approaches addressing this task have developed sophisticated architectures and techniques for multi-modality fusion and representation learning. However, the natural heterogeneity of different modalities causes distribution gap between their representations, making it challenging to fuse them. In this paper, we aim to learn the shared representations across modalities to bridge their gap. Different from existing similar methods on other multimodal tasks like sentiment analysis, we focus on the temporal contextual dependencies considering the sequence-to-sequence task setting of AVSR. In particular, we propose an adversarial network to refine frame-level modality-invariant representations (MIR-GAN), which captures the commonality across modalities to ease the subsequent multimodal fusion process. Extensive experiments on public benchmarks LRS3 and LRS2 show that our approach outperforms the state-of-the-arts.",
}
| Audio-visual speech recognition (AVSR) attracts a surge of research interest recently by leveraging multimodal signals to understand human speech. Mainstream approaches addressing this task have developed sophisticated architectures and techniques for multi-modality fusion and representation learning. However, the natural heterogeneity of different modalities causes distribution gap between their representations, making it challenging to fuse them. In this paper, we aim to learn the shared representations across modalities to bridge their gap. Different from existing similar methods on other multimodal tasks like sentiment analysis, we focus on the temporal contextual dependencies considering the sequence-to-sequence task setting of AVSR. In particular, we propose an adversarial network to refine frame-level modality-invariant representations (MIR-GAN), which captures the commonality across modalities to ease the subsequent multimodal fusion process. Extensive experiments on public benchmarks LRS3 and LRS2 show that our approach outperforms the state-of-the-arts. | [
"Hu, Yuchen",
"Chen, Chen",
"Li, Ruizhe",
"Zou, Heqing",
"Chng, Eng Siong"
] | MIR-GAN: Refining Frame-Level Modality-Invariant Representations with Adversarial Network for Audio-Visual Speech Recognition | acl-long.649 | Oral | 2306.10567 | [
"https://github.com/yuchen005/mir-gan"
] | -1 | -1 | -1 | -1 | 0 | [] | [] | [] |
|
https://aclanthology.org/2023.acl-long.650.bib | https://aclanthology.org/2023.acl-long.650/ | @inproceedings{tang-etal-2023-understanding,
title = "Understanding Factual Errors in Summarization: Errors, Summarizers, Datasets, Error Detectors",
author = "Tang, Liyan and
Goyal, Tanya and
Fabbri, Alex and
Laban, Philippe and
Xu, Jiacheng and
Yavuz, Semih and
Kryscinski, Wojciech and
Rousseau, Justin and
Durrett, Greg",
editor = "Rogers, Anna and
Boyd-Graber, Jordan and
Okazaki, Naoaki",
booktitle = "Proceedings of the 61st Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers)",
month = jul,
year = "2023",
address = "Toronto, Canada",
publisher = "Association for Computational Linguistics",
url = "https://aclanthology.org/2023.acl-long.650",
doi = "10.18653/v1/2023.acl-long.650",
pages = "11626--11644",
abstract = "The propensity of abstractive summarization models to make factual errors has been studied extensively, including design of metrics to detect factual errors and annotation of errors in current systems{'} outputs. However, the ever-evolving nature of summarization systems, metrics, and annotated benchmarks makes factuality evaluation a moving target, and drawing clear comparisons among metrics has become increasingly difficult. In this work, we aggregate factuality error annotations from nine existing datasets and stratify them according to the underlying summarization model. We compare performance of state-of-the-art factuality metrics, including recent ChatGPT-based metrics, on this stratified benchmark and show that their performance varies significantly across different types of summarization models. Critically, our analysis shows that much of the recent improvement in the factuality detection space has been on summaries from older (pre-Transformer) models instead of more relevant recent summarization models. We further perform a finer-grained analysis per error-type and find similar performance variance across error types for different factuality metrics. Our results show that no one metric is superior in all settings or for all error types, and we provide recommendations for best practices given these insights.",
}
| The propensity of abstractive summarization models to make factual errors has been studied extensively, including design of metrics to detect factual errors and annotation of errors in current systems{'} outputs. However, the ever-evolving nature of summarization systems, metrics, and annotated benchmarks makes factuality evaluation a moving target, and drawing clear comparisons among metrics has become increasingly difficult. In this work, we aggregate factuality error annotations from nine existing datasets and stratify them according to the underlying summarization model. We compare performance of state-of-the-art factuality metrics, including recent ChatGPT-based metrics, on this stratified benchmark and show that their performance varies significantly across different types of summarization models. Critically, our analysis shows that much of the recent improvement in the factuality detection space has been on summaries from older (pre-Transformer) models instead of more relevant recent summarization models. We further perform a finer-grained analysis per error-type and find similar performance variance across error types for different factuality metrics. Our results show that no one metric is superior in all settings or for all error types, and we provide recommendations for best practices given these insights. | [
"Tang, Liyan",
"Goyal, Tanya",
"Fabbri, Alex",
"Laban, Philippe",
"Xu, Jiacheng",
"Yavuz, Semih",
"Kryscinski, Wojciech",
"Rousseau, Justin",
"Durrett, Greg"
] | Understanding Factual Errors in Summarization: Errors, Summarizers, Datasets, Error Detectors | acl-long.650 | Poster | 2205.12854 | [
"https://github.com/liyan06/aggrefact"
] | https://huggingface.co/papers/2205.12854 | 0 | 0 | 0 | 9 | 1 | [
"vectara/hallucination_evaluation_model"
] | [] | [
"vectara/leaderboard",
"TeamTonic/MultiMed",
"jayash391/RAG_MedMind",
"Tonic1/hallucination-test",
"itsJB/Fact-Checked",
"Tonic/MultiMedTulu",
"girgis/Cloudilic-Demo",
"eaglelandsonce/Breaking-Free-Hackathon",
"jimshadow666/vectara-hallucination_evaluation_model",
"TeamTonic/TruEraMultiMed",
"subhanliaqat/hhem",
"eaglelandsonce/hhem",
"ahmadtalha/hhem",
"pyresearch/KitchenCreators",
"Tonic/SureRAG",
"Johan713/MedMind01",
"abidlabs/HHEM",
"ranavikas/NEXUS"
] |
https://aclanthology.org/2023.acl-long.651.bib | https://aclanthology.org/2023.acl-long.651/ | @inproceedings{gu-etal-2023-gift,
title = "{GIFT}: Graph-Induced Fine-Tuning for Multi-Party Conversation Understanding",
author = "Gu, Jia-Chen and
Ling, Zhenhua and
Liu, Quan and
Liu, Cong and
Hu, Guoping",
editor = "Rogers, Anna and
Boyd-Graber, Jordan and
Okazaki, Naoaki",
booktitle = "Proceedings of the 61st Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers)",
month = jul,
year = "2023",
address = "Toronto, Canada",
publisher = "Association for Computational Linguistics",
url = "https://aclanthology.org/2023.acl-long.651",
doi = "10.18653/v1/2023.acl-long.651",
pages = "11645--11658",
abstract = "Addressing the issues of who saying what to whom in multi-party conversations (MPCs) has recently attracted a lot of research attention. However, existing methods on MPC understanding typically embed interlocutors and utterances into sequential information flows, or utilize only the superficial of inherent graph structures in MPCs. To this end, we present a plug-and-play and lightweight method named graph-induced fine-tuning (GIFT) which can adapt various Transformer-based pre-trained language models (PLMs) for universal MPC understanding. In detail, the full and equivalent connections among utterances in regular Transformer ignore the sparse but distinctive dependency of an utterance on another in MPCs. To distinguish different relationships between utterances, four types of edges are designed to integrate graph-induced signals into attention mechanisms to refine PLMs originally designed for processing sequential texts. We evaluate GIFT by implementing it into three PLMs, and test the performance on three downstream tasks including addressee recognition, speaker identification and response selection. Experimental results show that GIFT can significantly improve the performance of three PLMs on three downstream tasks and two benchmarks with only 4 additional parameters per encoding layer, achieving new state-of-the-art performance on MPC understanding.",
}
| Addressing the issues of who saying what to whom in multi-party conversations (MPCs) has recently attracted a lot of research attention. However, existing methods on MPC understanding typically embed interlocutors and utterances into sequential information flows, or utilize only the superficial of inherent graph structures in MPCs. To this end, we present a plug-and-play and lightweight method named graph-induced fine-tuning (GIFT) which can adapt various Transformer-based pre-trained language models (PLMs) for universal MPC understanding. In detail, the full and equivalent connections among utterances in regular Transformer ignore the sparse but distinctive dependency of an utterance on another in MPCs. To distinguish different relationships between utterances, four types of edges are designed to integrate graph-induced signals into attention mechanisms to refine PLMs originally designed for processing sequential texts. We evaluate GIFT by implementing it into three PLMs, and test the performance on three downstream tasks including addressee recognition, speaker identification and response selection. Experimental results show that GIFT can significantly improve the performance of three PLMs on three downstream tasks and two benchmarks with only 4 additional parameters per encoding layer, achieving new state-of-the-art performance on MPC understanding. | [
"Gu, Jia-Chen",
"Ling, Zhenhua",
"Liu, Quan",
"Liu, Cong",
"Hu, Guoping"
] | GIFT: Graph-Induced Fine-Tuning for Multi-Party Conversation Understanding | acl-long.651 | Oral | 2305.09360 | [
"https://github.com/JasonForJoy/MPC-BERT"
] | -1 | -1 | -1 | -1 | 0 | [] | [] | [] |
|
https://aclanthology.org/2023.acl-long.652.bib | https://aclanthology.org/2023.acl-long.652/ | @inproceedings{vazhentsev-etal-2023-hybrid,
title = "Hybrid Uncertainty Quantification for Selective Text Classification in Ambiguous Tasks",
author = "Vazhentsev, Artem and
Kuzmin, Gleb and
Tsvigun, Akim and
Panchenko, Alexander and
Panov, Maxim and
Burtsev, Mikhail and
Shelmanov, Artem",
editor = "Rogers, Anna and
Boyd-Graber, Jordan and
Okazaki, Naoaki",
booktitle = "Proceedings of the 61st Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers)",
month = jul,
year = "2023",
address = "Toronto, Canada",
publisher = "Association for Computational Linguistics",
url = "https://aclanthology.org/2023.acl-long.652",
doi = "10.18653/v1/2023.acl-long.652",
pages = "11659--11681",
abstract = "Many text classification tasks are inherently ambiguous, which results in automatic systems having a high risk of making mistakes, in spite of using advanced machine learning models. For example, toxicity detection in user-generated content is a subjective task, and notions of toxicity can be annotated according to a variety of definitions that can be in conflict with one another. Instead of relying solely on automatic solutions, moderation of the most difficult and ambiguous cases can be delegated to human workers. Potential mistakes in automated classification can be identified by using uncertainty estimation (UE) techniques. Although UE is a rapidly growing field within natural language processing, we find that state-of-the-art UE methods estimate only epistemic uncertainty and show poor performance, or under-perform trivial methods for ambiguous tasks such as toxicity detection. We argue that in order to create robust uncertainty estimation methods for ambiguous tasks it is necessary to account also for aleatoric uncertainty. In this paper, we propose a new uncertainty estimation method that combines epistemic and aleatoric UE methods. We show that by using our hybrid method, we can outperform state-of-the-art UE methods for toxicity detection and other ambiguous text classification tasks.",
}
| Many text classification tasks are inherently ambiguous, which results in automatic systems having a high risk of making mistakes, in spite of using advanced machine learning models. For example, toxicity detection in user-generated content is a subjective task, and notions of toxicity can be annotated according to a variety of definitions that can be in conflict with one another. Instead of relying solely on automatic solutions, moderation of the most difficult and ambiguous cases can be delegated to human workers. Potential mistakes in automated classification can be identified by using uncertainty estimation (UE) techniques. Although UE is a rapidly growing field within natural language processing, we find that state-of-the-art UE methods estimate only epistemic uncertainty and show poor performance, or under-perform trivial methods for ambiguous tasks such as toxicity detection. We argue that in order to create robust uncertainty estimation methods for ambiguous tasks it is necessary to account also for aleatoric uncertainty. In this paper, we propose a new uncertainty estimation method that combines epistemic and aleatoric UE methods. We show that by using our hybrid method, we can outperform state-of-the-art UE methods for toxicity detection and other ambiguous text classification tasks. | [
"Vazhentsev, Artem",
"Kuzmin, Gleb",
"Tsvigun, Akim",
"Panchenko, Alex",
"er",
"Panov, Maxim",
"Burtsev, Mikhail",
"Shelmanov, Artem"
] | Hybrid Uncertainty Quantification for Selective Text Classification in Ambiguous Tasks | acl-long.652 | Poster | [
""
] | -1 | -1 | -1 | -1 | 0 | [] | [] | [] |
||
https://aclanthology.org/2023.acl-long.653.bib | https://aclanthology.org/2023.acl-long.653/ | @inproceedings{yong-etal-2023-bloom,
title = "{BLOOM}+1: Adding Language Support to {BLOOM} for Zero-Shot Prompting",
author = "Yong, Zheng Xin and
Schoelkopf, Hailey and
Muennighoff, Niklas and
Aji, Alham Fikri and
Adelani, David Ifeoluwa and
Almubarak, Khalid and
Bari, M Saiful and
Sutawika, Lintang and
Kasai, Jungo and
Baruwa, Ahmed and
Winata, Genta and
Biderman, Stella and
Raff, Edward and
Radev, Dragomir and
Nikoulina, Vassilina",
editor = "Rogers, Anna and
Boyd-Graber, Jordan and
Okazaki, Naoaki",
booktitle = "Proceedings of the 61st Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers)",
month = jul,
year = "2023",
address = "Toronto, Canada",
publisher = "Association for Computational Linguistics",
url = "https://aclanthology.org/2023.acl-long.653",
doi = "10.18653/v1/2023.acl-long.653",
pages = "11682--11703",
abstract = "The BLOOM model is a large publicly available multilingual language model, but its pretraining was limited to 46 languages. To extend the benefits of BLOOM to other languages without incurring prohibitively large costs, it is desirable to adapt BLOOM to new languages not seen during pretraining. In this work, we apply existing language adaptation strategies to BLOOM and benchmark its zero-shot prompting performance on eight new languages in a resource-constrained setting. We find language adaptation to be effective at improving zero-shot performance in new languages. Surprisingly, we find that adapter-based finetuning is more effective than continued pretraining for large models. In addition, we discover that prompting performance is not significantly affected by language specifics, such as the writing system. It is primarily determined by the size of the language adaptation data. We also add new languages to BLOOMZ, which is a multitask finetuned version of BLOOM capable of following task instructions zero-shot. We find including a new language in the multitask fine-tuning mixture to be the most effective method to teach BLOOMZ a new language. We conclude that with sufficient training data language adaptation can generalize well to diverse languages. Our code is available at \url{https://github.com/bigscience-workshop/multilingual-modeling}.",
}
| The BLOOM model is a large publicly available multilingual language model, but its pretraining was limited to 46 languages. To extend the benefits of BLOOM to other languages without incurring prohibitively large costs, it is desirable to adapt BLOOM to new languages not seen during pretraining. In this work, we apply existing language adaptation strategies to BLOOM and benchmark its zero-shot prompting performance on eight new languages in a resource-constrained setting. We find language adaptation to be effective at improving zero-shot performance in new languages. Surprisingly, we find that adapter-based finetuning is more effective than continued pretraining for large models. In addition, we discover that prompting performance is not significantly affected by language specifics, such as the writing system. It is primarily determined by the size of the language adaptation data. We also add new languages to BLOOMZ, which is a multitask finetuned version of BLOOM capable of following task instructions zero-shot. We find including a new language in the multitask fine-tuning mixture to be the most effective method to teach BLOOMZ a new language. We conclude that with sufficient training data language adaptation can generalize well to diverse languages. Our code is available at \url{https://github.com/bigscience-workshop/multilingual-modeling}. | [
"Yong, Zheng Xin",
"Schoelkopf, Hailey",
"Muennighoff, Niklas",
"Aji, Alham Fikri",
"Adelani, David Ifeoluwa",
"Almubarak, Khalid",
"Bari, M Saiful",
"Sutawika, Lintang",
"Kasai, Jungo",
"Baruwa, Ahmed",
"Winata, Genta",
"Biderman, Stella",
"Raff, Edward",
"Radev, Dragomir",
"Nikoulina, Vassilina"
] | BLOOM+1: Adding Language Support to BLOOM for Zero-Shot Prompting | acl-long.653 | Poster | 2212.09535 | [
"https://github.com/bigscience-workshop/multilingual-modeling"
] | https://huggingface.co/papers/2212.09535 | 6 | 1 | 0 | 14 | 1 | [
"basilepp19/bloom-1b7_it",
"basilepp19/bloom-1b7-it-dolly-evalita",
"basilepp19/bloom-1b7-it-dolly",
"basilepp19/bloom-1b7-it-evalita"
] | [
"desik98/UniversallyJailbreakingLLMInputOutputSafetyFilters"
] | [] |
https://aclanthology.org/2023.acl-long.654.bib | https://aclanthology.org/2023.acl-long.654/ | @inproceedings{medina-grespan-etal-2023-logic,
title = "Logic-driven Indirect Supervision: An Application to Crisis Counseling",
author = "Medina Grespan, Mattia and
Broadbent, Meghan and
Zhang, Xinyao and
Axford, Katherine and
Kious, Brent and
Imel, Zac and
Srikumar, Vivek",
editor = "Rogers, Anna and
Boyd-Graber, Jordan and
Okazaki, Naoaki",
booktitle = "Proceedings of the 61st Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers)",
month = jul,
year = "2023",
address = "Toronto, Canada",
publisher = "Association for Computational Linguistics",
url = "https://aclanthology.org/2023.acl-long.654",
doi = "10.18653/v1/2023.acl-long.654",
pages = "11704--11722",
abstract = "Ensuring the effectiveness of text-based crisis counseling requires observing ongoing conversations and providing feedback, both labor-intensive tasks. Automatic analysis of conversations{---}at the full chat and utterance levels{---}may help support counselors and provide better care. While some session-level training data (e.g., rating of patient risk) is often available from counselors, labeling utterances requires expensive post hoc annotation. But the latter can not only provide insights about conversation dynamics, but can also serve to support quality assurance efforts for counselors. In this paper, we examine if inexpensive{---}and potentially noisy{---}session-level annotation can help improve label utterances. To this end, we propose a logic-based indirect supervision approach that exploits declaratively stated structural dependencies between both levels of annotation to improve utterance modeling. We show that adding these rules gives an improvement of 3.5{\%} f-score over a strong multi-task baseline for utterance-level predictions. We demonstrate via ablation studies how indirect supervision via logic rules also improves the consistency and robustness of the system.",
}
| Ensuring the effectiveness of text-based crisis counseling requires observing ongoing conversations and providing feedback, both labor-intensive tasks. Automatic analysis of conversations{---}at the full chat and utterance levels{---}may help support counselors and provide better care. While some session-level training data (e.g., rating of patient risk) is often available from counselors, labeling utterances requires expensive post hoc annotation. But the latter can not only provide insights about conversation dynamics, but can also serve to support quality assurance efforts for counselors. In this paper, we examine if inexpensive{---}and potentially noisy{---}session-level annotation can help improve label utterances. To this end, we propose a logic-based indirect supervision approach that exploits declaratively stated structural dependencies between both levels of annotation to improve utterance modeling. We show that adding these rules gives an improvement of 3.5{\%} f-score over a strong multi-task baseline for utterance-level predictions. We demonstrate via ablation studies how indirect supervision via logic rules also improves the consistency and robustness of the system. | [
"Medina Grespan, Mattia",
"Broadbent, Meghan",
"Zhang, Xinyao",
"Axford, Katherine",
"Kious, Brent",
"Imel, Zac",
"Srikumar, Vivek"
] | Logic-driven Indirect Supervision: An Application to Crisis Counseling | acl-long.654 | Poster | [
""
] | -1 | -1 | -1 | -1 | 0 | [] | [] | [] |
||
https://aclanthology.org/2023.acl-long.655.bib | https://aclanthology.org/2023.acl-long.655/ | @inproceedings{soni-etal-2023-grounding,
title = "Grounding Characters and Places in Narrative Text",
author = "Soni, Sandeep and
Sihra, Amanpreet and
Evans, Elizabeth and
Wilkens, Matthew and
Bamman, David",
editor = "Rogers, Anna and
Boyd-Graber, Jordan and
Okazaki, Naoaki",
booktitle = "Proceedings of the 61st Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers)",
month = jul,
year = "2023",
address = "Toronto, Canada",
publisher = "Association for Computational Linguistics",
url = "https://aclanthology.org/2023.acl-long.655",
doi = "10.18653/v1/2023.acl-long.655",
pages = "11723--11736",
abstract = "Tracking characters and locations throughout a story can help improve the understanding of its plot structure. Prior research has analyzed characters and locations from text independently without grounding characters to their locations in narrative time. Here, we address this gap by proposing a new spatial relationship categorization task. The objective of the task is to assign a spatial relationship category for every character and location co-mention within a window of text, taking into consideration linguistic context, narrative tense, and temporal scope. To this end, we annotate spatial relationships in approximately 2500 book excerpts and train a model using contextual embeddings as features to predict these relationships. When applied to a set of books, this model allows us to test several hypotheses on mobility and domestic space, revealing that protagonists are more mobile than non-central characters and that women as characters tend to occupy more interior space than men. Overall, our work is the first step towards joint modeling and analysis of characters and places in narrative text.",
}
| Tracking characters and locations throughout a story can help improve the understanding of its plot structure. Prior research has analyzed characters and locations from text independently without grounding characters to their locations in narrative time. Here, we address this gap by proposing a new spatial relationship categorization task. The objective of the task is to assign a spatial relationship category for every character and location co-mention within a window of text, taking into consideration linguistic context, narrative tense, and temporal scope. To this end, we annotate spatial relationships in approximately 2500 book excerpts and train a model using contextual embeddings as features to predict these relationships. When applied to a set of books, this model allows us to test several hypotheses on mobility and domestic space, revealing that protagonists are more mobile than non-central characters and that women as characters tend to occupy more interior space than men. Overall, our work is the first step towards joint modeling and analysis of characters and places in narrative text. | [
"Soni, S",
"eep",
"Sihra, Amanpreet",
"Evans, Elizabeth",
"Wilkens, Matthew",
"Bamman, David"
] | Grounding Characters and Places in Narrative Text | acl-long.655 | Oral | [
"https://github.com/sandeepsoni/mobility-books"
] | -1 | -1 | -1 | -1 | 0 | [] | [] | [] |
||
https://aclanthology.org/2023.acl-long.656.bib | https://aclanthology.org/2023.acl-long.656/ | @inproceedings{feng-etal-2023-pretraining,
title = "From Pretraining Data to Language Models to Downstream Tasks: Tracking the Trails of Political Biases Leading to Unfair {NLP} Models",
author = "Feng, Shangbin and
Park, Chan Young and
Liu, Yuhan and
Tsvetkov, Yulia",
editor = "Rogers, Anna and
Boyd-Graber, Jordan and
Okazaki, Naoaki",
booktitle = "Proceedings of the 61st Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers)",
month = jul,
year = "2023",
address = "Toronto, Canada",
publisher = "Association for Computational Linguistics",
url = "https://aclanthology.org/2023.acl-long.656",
doi = "10.18653/v1/2023.acl-long.656",
pages = "11737--11762",
abstract = "Language models (LMs) are pretrained on diverse data sources{---}news, discussion forums, books, online encyclopedias. A significant portion of this data includes facts and opinions which, on one hand, celebrate democracy and diversity of ideas, and on the other hand are inherently socially biased. Our work develops new methods to (1) measure media biases in LMs trained on such corpora, along social and economic axes, and (2) measure the fairness of downstream NLP models trained on top of politically biased LMs. We focus on hate speech and misinformation detection, aiming to empirically quantify the effects of political (social, economic) biases in pretraining data on the fairness of high-stakes social-oriented tasks. Our findings reveal that pretrained LMs do have political leanings which reinforce the polarization present in pretraining corpora, propagating social biases into hate speech predictions and media biases into misinformation detectors. We discuss the implications of our findings for NLP research and propose future directions to mitigate unfairness.",
}
| Language models (LMs) are pretrained on diverse data sources{---}news, discussion forums, books, online encyclopedias. A significant portion of this data includes facts and opinions which, on one hand, celebrate democracy and diversity of ideas, and on the other hand are inherently socially biased. Our work develops new methods to (1) measure media biases in LMs trained on such corpora, along social and economic axes, and (2) measure the fairness of downstream NLP models trained on top of politically biased LMs. We focus on hate speech and misinformation detection, aiming to empirically quantify the effects of political (social, economic) biases in pretraining data on the fairness of high-stakes social-oriented tasks. Our findings reveal that pretrained LMs do have political leanings which reinforce the polarization present in pretraining corpora, propagating social biases into hate speech predictions and media biases into misinformation detectors. We discuss the implications of our findings for NLP research and propose future directions to mitigate unfairness. | [
"Feng, Shangbin",
"Park, Chan Young",
"Liu, Yuhan",
"Tsvetkov, Yulia"
] | From Pretraining Data to Language Models to Downstream Tasks: Tracking the Trails of Political Biases Leading to Unfair NLP Models | acl-long.656 | Poster | 2305.08283 | [
"https://github.com/bunsenfeng/polilean"
] | -1 | -1 | -1 | -1 | 0 | [] | [] | [] |
|
https://aclanthology.org/2023.acl-long.657.bib | https://aclanthology.org/2023.acl-long.657/ | @inproceedings{yadavalli-etal-2023-slabert,
title = "{SLABERT} Talk Pretty One Day: Modeling Second Language Acquisition with {BERT}",
author = "Yadavalli, Aditya and
Yadavalli, Alekhya and
Tobin, Vera",
editor = "Rogers, Anna and
Boyd-Graber, Jordan and
Okazaki, Naoaki",
booktitle = "Proceedings of the 61st Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers)",
month = jul,
year = "2023",
address = "Toronto, Canada",
publisher = "Association for Computational Linguistics",
url = "https://aclanthology.org/2023.acl-long.657",
doi = "10.18653/v1/2023.acl-long.657",
pages = "11763--11777",
abstract = "Second language acquisition (SLA) research has extensively studied cross-linguistic transfer, the influence of linguistic structure of a speaker{'}s native language [L1] on the successful acquisition of a foreign language [L2]. Effects of such transfer can be positive (facilitating acquisition) or negative (impeding acquisition). We find that NLP literature has not given enough attention to the phenomenon of negative transfer. To understand patterns of both positive and negative transfer between L1 and L2, we model sequential second language acquisition in LMs. Further, we build a Mutlilingual Age Ordered CHILDES (MAO-CHILDES){---}a dataset consisting of 5 typologically diverse languages, i.e., German, French, Polish, Indonesian, and Japanese{---}to understand the degree to which native Child-Directed Speech (CDS) [L1] can help or conflict with English language acquisition [L2]. To examine the impact of native CDS, we use the TILT-based cross lingual transfer learning approach established by Papadimitriou and Jurafsky (2020) and find that, as in human SLA, language family distance predicts more negative transfer. Additionally, we find that conversational speech data shows greater facilitation for language acquisition than scripted speech data. Our findings call for further research using our novel Transformer-based SLA models and we would like to encourage it by releasing our code, data, and models.",
}
| Second language acquisition (SLA) research has extensively studied cross-linguistic transfer, the influence of linguistic structure of a speaker{'}s native language [L1] on the successful acquisition of a foreign language [L2]. Effects of such transfer can be positive (facilitating acquisition) or negative (impeding acquisition). We find that NLP literature has not given enough attention to the phenomenon of negative transfer. To understand patterns of both positive and negative transfer between L1 and L2, we model sequential second language acquisition in LMs. Further, we build a Mutlilingual Age Ordered CHILDES (MAO-CHILDES){---}a dataset consisting of 5 typologically diverse languages, i.e., German, French, Polish, Indonesian, and Japanese{---}to understand the degree to which native Child-Directed Speech (CDS) [L1] can help or conflict with English language acquisition [L2]. To examine the impact of native CDS, we use the TILT-based cross lingual transfer learning approach established by Papadimitriou and Jurafsky (2020) and find that, as in human SLA, language family distance predicts more negative transfer. Additionally, we find that conversational speech data shows greater facilitation for language acquisition than scripted speech data. Our findings call for further research using our novel Transformer-based SLA models and we would like to encourage it by releasing our code, data, and models. | [
"Yadavalli, Aditya",
"Yadavalli, Alekhya",
"Tobin, Vera"
] | SLABERT Talk Pretty One Day: Modeling Second Language Acquisition with BERT | acl-long.657 | Poster | 2305.19589 | [
""
] | -1 | -1 | -1 | -1 | 0 | [] | [] | [] |
|
https://aclanthology.org/2023.acl-long.658.bib | https://aclanthology.org/2023.acl-long.658/ | @inproceedings{xu-etal-2023-contrastive,
title = "Contrastive Novelty-Augmented Learning: Anticipating Outliers with Large Language Models",
author = "Xu, Albert and
Ren, Xiang and
Jia, Robin",
editor = "Rogers, Anna and
Boyd-Graber, Jordan and
Okazaki, Naoaki",
booktitle = "Proceedings of the 61st Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers)",
month = jul,
year = "2023",
address = "Toronto, Canada",
publisher = "Association for Computational Linguistics",
url = "https://aclanthology.org/2023.acl-long.658",
doi = "10.18653/v1/2023.acl-long.658",
pages = "11778--11801",
abstract = "In many task settings, text classification models are likely to encounter examples from novel classes on which they cannot predict correctly. Selective prediction, in which models abstain on low-confidence examples, provides a possible solution, but existing models are often overly confident on unseen classes. To remedy this overconfidence, we introduce Contrastive Novelty-Augmented Learning (CoNAL), a two-step method that generates OOD examples representative of novel classes, then trains to decrease confidence on them. First, we generate OOD examples by prompting a large language model twice: we prompt it to enumerate relevant novel classes, then generate examples from each novel class matching the task format. Second, we train a classifier with a novel contrastive objective that encourages lower confidence on generated OOD examples than training examples. When trained with CoNAL, classifiers improve in their ability to detect and abstain on novel class examples over prior methods by an average of 2.3{\%} in terms of accuracy under the accuracy-coverage curve (AUAC) and 5.5{\%} AUROC across 4 NLP datasets, with no cost to in-distribution accuracy.",
}
| In many task settings, text classification models are likely to encounter examples from novel classes on which they cannot predict correctly. Selective prediction, in which models abstain on low-confidence examples, provides a possible solution, but existing models are often overly confident on unseen classes. To remedy this overconfidence, we introduce Contrastive Novelty-Augmented Learning (CoNAL), a two-step method that generates OOD examples representative of novel classes, then trains to decrease confidence on them. First, we generate OOD examples by prompting a large language model twice: we prompt it to enumerate relevant novel classes, then generate examples from each novel class matching the task format. Second, we train a classifier with a novel contrastive objective that encourages lower confidence on generated OOD examples than training examples. When trained with CoNAL, classifiers improve in their ability to detect and abstain on novel class examples over prior methods by an average of 2.3{\%} in terms of accuracy under the accuracy-coverage curve (AUAC) and 5.5{\%} AUROC across 4 NLP datasets, with no cost to in-distribution accuracy. | [
"Xu, Albert",
"Ren, Xiang",
"Jia, Robin"
] | Contrastive Novelty-Augmented Learning: Anticipating Outliers with Large Language Models | acl-long.658 | Poster | 2211.15718 | [
"https://github.com/albertkx/conal"
] | -1 | -1 | -1 | -1 | 0 | [] | [] | [] |
|
https://aclanthology.org/2023.acl-long.659.bib | https://aclanthology.org/2023.acl-long.659/ | @inproceedings{qin-etal-2023-learning,
title = "Learning to Initialize: Can Meta Learning Improve Cross-task Generalization in Prompt Tuning?",
author = "Qin, Chengwei and
Joty, Shafiq and
Li, Qian and
Zhao, Ruochen",
editor = "Rogers, Anna and
Boyd-Graber, Jordan and
Okazaki, Naoaki",
booktitle = "Proceedings of the 61st Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers)",
month = jul,
year = "2023",
address = "Toronto, Canada",
publisher = "Association for Computational Linguistics",
url = "https://aclanthology.org/2023.acl-long.659",
doi = "10.18653/v1/2023.acl-long.659",
pages = "11802--11832",
abstract = "Prompt tuning (PT) which only tunes the embeddings of an additional sequence of tokens per task, keeping the pre-trained language model (PLM) frozen, has shown remarkable performance in few-shot learning. Despite this, PT has been shown to rely heavily on good initialization of the prompt embeddings. In this work, we study meta prompt tuning (MPT) to systematically explore how meta-learning can help improve (if it can) cross-task generalization in PT through learning to initialize the prompt embeddings from other relevant tasks. We empirically analyze a representative set of meta learning algorithms in a wide range of adaptation settings with different source/target task configurations on a large set of few-shot tasks. With extensive experiments and analysis, we demonstrate the effectiveness of MPT. We find the improvement to be significant particularly on classification tasks. For other kinds of tasks such as question answering, we observe that while MPT can outperform PT in most cases, it does not always outperform multi-task learning. We further provide an in-depth analysis from the perspective of task similarity.",
}
| Prompt tuning (PT) which only tunes the embeddings of an additional sequence of tokens per task, keeping the pre-trained language model (PLM) frozen, has shown remarkable performance in few-shot learning. Despite this, PT has been shown to rely heavily on good initialization of the prompt embeddings. In this work, we study meta prompt tuning (MPT) to systematically explore how meta-learning can help improve (if it can) cross-task generalization in PT through learning to initialize the prompt embeddings from other relevant tasks. We empirically analyze a representative set of meta learning algorithms in a wide range of adaptation settings with different source/target task configurations on a large set of few-shot tasks. With extensive experiments and analysis, we demonstrate the effectiveness of MPT. We find the improvement to be significant particularly on classification tasks. For other kinds of tasks such as question answering, we observe that while MPT can outperform PT in most cases, it does not always outperform multi-task learning. We further provide an in-depth analysis from the perspective of task similarity. | [
"Qin, Chengwei",
"Joty, Shafiq",
"Li, Qian",
"Zhao, Ruochen"
] | Learning to Initialize: Can Meta Learning Improve Cross-task Generalization in Prompt Tuning? | acl-long.659 | Poster | 2302.08143 | [
""
] | -1 | -1 | -1 | -1 | 0 | [] | [] | [] |
|
https://aclanthology.org/2023.acl-long.660.bib | https://aclanthology.org/2023.acl-long.660/ | @inproceedings{bansal-etal-2023-rethinking,
title = "Rethinking the Role of Scale for In-Context Learning: An Interpretability-based Case Study at 66 Billion Scale",
author = "Bansal, Hritik and
Gopalakrishnan, Karthik and
Dingliwal, Saket and
Bodapati, Sravan and
Kirchhoff, Katrin and
Roth, Dan",
editor = "Rogers, Anna and
Boyd-Graber, Jordan and
Okazaki, Naoaki",
booktitle = "Proceedings of the 61st Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers)",
month = jul,
year = "2023",
address = "Toronto, Canada",
publisher = "Association for Computational Linguistics",
url = "https://aclanthology.org/2023.acl-long.660",
doi = "10.18653/v1/2023.acl-long.660",
pages = "11833--11856",
abstract = "Language models have been shown to perform better with an increase in scale on a wide variety of tasks via the in-context learning paradigm. In this paper, we investigate the hypothesis that the ability of a large language model to in-context learn-perform a task is not uniformly spread across all of its underlying components. Using a 66 billion parameter language model (OPT-66B) across a diverse set of 14 downstream tasks, we find this is indeed the case: {\textasciitilde}70{\%} of the attention heads and {\textasciitilde}20{\%} of the feed forward networks can be removed with minimal decline in task performance. We find substantial overlap in the set of attention heads (un)important for in-context learning across tasks and number of in-context examples. We also address our hypothesis through a task-agnostic lens, finding that a small set of attention heads in OPT-66B score highly on their ability to perform primitive induction operations associated with in-context learning, namely, prefix matching and copying. These induction heads overlap with task-specific important heads, reinforcing arguments by Olsson et al. (2022) regarding induction head generality to more sophisticated behaviors associated with in-context learning. Overall, our study provides several insights that indicate large language models may be under-trained for in-context learning and opens up questions on how to pre-train language models to more effectively perform in-context learning.",
}
| Language models have been shown to perform better with an increase in scale on a wide variety of tasks via the in-context learning paradigm. In this paper, we investigate the hypothesis that the ability of a large language model to in-context learn-perform a task is not uniformly spread across all of its underlying components. Using a 66 billion parameter language model (OPT-66B) across a diverse set of 14 downstream tasks, we find this is indeed the case: {\textasciitilde}70{\%} of the attention heads and {\textasciitilde}20{\%} of the feed forward networks can be removed with minimal decline in task performance. We find substantial overlap in the set of attention heads (un)important for in-context learning across tasks and number of in-context examples. We also address our hypothesis through a task-agnostic lens, finding that a small set of attention heads in OPT-66B score highly on their ability to perform primitive induction operations associated with in-context learning, namely, prefix matching and copying. These induction heads overlap with task-specific important heads, reinforcing arguments by Olsson et al. (2022) regarding induction head generality to more sophisticated behaviors associated with in-context learning. Overall, our study provides several insights that indicate large language models may be under-trained for in-context learning and opens up questions on how to pre-train language models to more effectively perform in-context learning. | [
"Bansal, Hritik",
"Gopalakrishnan, Karthik",
"Dingliwal, Saket",
"Bodapati, Sravan",
"Kirchhoff, Katrin",
"Roth, Dan"
] | Rethinking the Role of Scale for In-Context Learning: An Interpretability-based Case Study at 66 Billion Scale | acl-long.660 | Poster | 2212.09095 | [
"https://github.com/amazon-science/llm-interpret"
] | -1 | -1 | -1 | -1 | 0 | [] | [] | [] |
|
https://aclanthology.org/2023.acl-long.661.bib | https://aclanthology.org/2023.acl-long.661/ | @inproceedings{gaim-etal-2023-question,
title = "Question-Answering in a Low-resourced Language: Benchmark Dataset and Models for {T}igrinya",
author = "Gaim, Fitsum and
Yang, Wonsuk and
Park, Hancheol and
Park, Jong",
editor = "Rogers, Anna and
Boyd-Graber, Jordan and
Okazaki, Naoaki",
booktitle = "Proceedings of the 61st Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers)",
month = jul,
year = "2023",
address = "Toronto, Canada",
publisher = "Association for Computational Linguistics",
url = "https://aclanthology.org/2023.acl-long.661",
doi = "10.18653/v1/2023.acl-long.661",
pages = "11857--11870",
abstract = "Question-Answering (QA) has seen significant advances recently, achieving near human-level performance over some benchmarks. However, these advances focus on high-resourced languages such as English, while the task remains unexplored for most other languages, mainly due to the lack of annotated datasets. This work presents a native QA dataset for an East African language, Tigrinya. The dataset contains 10.6K question-answer pairs spanning 572 paragraphs extracted from 290 news articles on various topics. The dataset construction method is discussed, which is applicable to constructing similar resources for related languages. We present comprehensive experiments and analyses of several resource-efficient approaches to QA, including monolingual, cross-lingual, and multilingual setups, along with comparisons against machine-translated silver data. Our strong baseline models reach 76{\%} in the F1 score, while the estimated human performance is 92{\%}, indicating that the benchmark presents a good challenge for future work. We make the dataset, models, and leaderboard publicly available.",
}
| Question-Answering (QA) has seen significant advances recently, achieving near human-level performance over some benchmarks. However, these advances focus on high-resourced languages such as English, while the task remains unexplored for most other languages, mainly due to the lack of annotated datasets. This work presents a native QA dataset for an East African language, Tigrinya. The dataset contains 10.6K question-answer pairs spanning 572 paragraphs extracted from 290 news articles on various topics. The dataset construction method is discussed, which is applicable to constructing similar resources for related languages. We present comprehensive experiments and analyses of several resource-efficient approaches to QA, including monolingual, cross-lingual, and multilingual setups, along with comparisons against machine-translated silver data. Our strong baseline models reach 76{\%} in the F1 score, while the estimated human performance is 92{\%}, indicating that the benchmark presents a good challenge for future work. We make the dataset, models, and leaderboard publicly available. | [
"Gaim, Fitsum",
"Yang, Wonsuk",
"Park, Hancheol",
"Park, Jong"
] | Question-Answering in a Low-resourced Language: Benchmark Dataset and Models for Tigrinya | acl-long.661 | Oral | [
""
] | -1 | -1 | -1 | -1 | 0 | [] | [] | [] |
||
https://aclanthology.org/2023.acl-long.662.bib | https://aclanthology.org/2023.acl-long.662/ | @inproceedings{zhang-etal-2023-escoxlm,
title = "{ESCOXLM}-{R}: Multilingual Taxonomy-driven Pre-training for the Job Market Domain",
author = "Zhang, Mike and
van der Goot, Rob and
Plank, Barbara",
editor = "Rogers, Anna and
Boyd-Graber, Jordan and
Okazaki, Naoaki",
booktitle = "Proceedings of the 61st Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers)",
month = jul,
year = "2023",
address = "Toronto, Canada",
publisher = "Association for Computational Linguistics",
url = "https://aclanthology.org/2023.acl-long.662",
doi = "10.18653/v1/2023.acl-long.662",
pages = "11871--11890",
abstract = "The increasing number of benchmarks for Natural Language Processing (NLP) tasks in the computational job market domain highlights the demand for methods that can handle job-related tasks such as skill extraction, skill classification, job title classification, and de-identification. While some approaches have been developed that are specific to the job market domain, there is a lack of generalized, multilingual models and benchmarks for these tasks. In this study, we introduce a language model called ESCOXLM-R, based on XLM-R-large, which uses domain-adaptive pre-training on the European Skills, Competences, Qualifications and Occupations (ESCO) taxonomy, covering 27 languages. The pre-training objectives for ESCOXLM-R include dynamic masked language modeling and a novel additional objective for inducing multilingual taxonomical ESCO relations. We comprehensively evaluate the performance of ESCOXLM-R on 6 sequence labeling and 3 classification tasks in 4 languages and find that it achieves state-of-the-art results on 6 out of 9 datasets. Our analysis reveals that ESCOXLM-R performs better on short spans and outperforms XLM-R-large on entity-level and surface-level span-F1, likely due to ESCO containing short skill and occupation titles, and encoding information on the entity-level.",
}
| The increasing number of benchmarks for Natural Language Processing (NLP) tasks in the computational job market domain highlights the demand for methods that can handle job-related tasks such as skill extraction, skill classification, job title classification, and de-identification. While some approaches have been developed that are specific to the job market domain, there is a lack of generalized, multilingual models and benchmarks for these tasks. In this study, we introduce a language model called ESCOXLM-R, based on XLM-R-large, which uses domain-adaptive pre-training on the European Skills, Competences, Qualifications and Occupations (ESCO) taxonomy, covering 27 languages. The pre-training objectives for ESCOXLM-R include dynamic masked language modeling and a novel additional objective for inducing multilingual taxonomical ESCO relations. We comprehensively evaluate the performance of ESCOXLM-R on 6 sequence labeling and 3 classification tasks in 4 languages and find that it achieves state-of-the-art results on 6 out of 9 datasets. Our analysis reveals that ESCOXLM-R performs better on short spans and outperforms XLM-R-large on entity-level and surface-level span-F1, likely due to ESCO containing short skill and occupation titles, and encoding information on the entity-level. | [
"Zhang, Mike",
"van der Goot, Rob",
"Plank, Barbara"
] | ESCOXLM-R: Multilingual Taxonomy-driven Pre-training for the Job Market Domain | acl-long.662 | Poster | 2305.12092 | [
"https://github.com/mainlp/escoxlmr"
] | -1 | -1 | -1 | -1 | 0 | [] | [] | [] |
|
https://aclanthology.org/2023.acl-long.663.bib | https://aclanthology.org/2023.acl-long.663/ | @inproceedings{li-etal-2023-citadel,
title = "{CITADEL}: Conditional Token Interaction via Dynamic Lexical Routing for Efficient and Effective Multi-Vector Retrieval",
author = "Li, Minghan and
Lin, Sheng-Chieh and
Oguz, Barlas and
Ghoshal, Asish and
Lin, Jimmy and
Mehdad, Yashar and
Yih, Wen-tau and
Chen, Xilun",
editor = "Rogers, Anna and
Boyd-Graber, Jordan and
Okazaki, Naoaki",
booktitle = "Proceedings of the 61st Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers)",
month = jul,
year = "2023",
address = "Toronto, Canada",
publisher = "Association for Computational Linguistics",
url = "https://aclanthology.org/2023.acl-long.663",
doi = "10.18653/v1/2023.acl-long.663",
pages = "11891--11907",
abstract = "Multi-vector retrieval methods combine the merits of sparse (e.g. BM25) and dense (e.g. DPR) retrievers and have achieved state-of-the-art performance on various retrieval tasks. These methods, however, are orders of magnitude slower and need much more space to store their indices compared to their single-vector counterparts. In this paper, we unify different multi-vector retrieval models from a token routing viewpoint and propose conditional token interaction via dynamic lexical routing, namely CITADEL, for efficient and effective multi-vector retrieval.CITADEL learns to route different token vectors to the predicted lexical keys such that a query token vector only interacts with document token vectors routed to the same key. This design significantly reduces the computation cost while maintaining high accuracy. Notably, CITADEL achieves the same or slightly better performance than the previous state of the art, ColBERT-v2, on both in-domain (MS MARCO) and out-of-domain (BEIR) evaluations, while being nearly 40 times faster. Source code and data are available at \url{https://github.com/facebookresearch/dpr-scale/tree/citadel}.",
}
| Multi-vector retrieval methods combine the merits of sparse (e.g. BM25) and dense (e.g. DPR) retrievers and have achieved state-of-the-art performance on various retrieval tasks. These methods, however, are orders of magnitude slower and need much more space to store their indices compared to their single-vector counterparts. In this paper, we unify different multi-vector retrieval models from a token routing viewpoint and propose conditional token interaction via dynamic lexical routing, namely CITADEL, for efficient and effective multi-vector retrieval.CITADEL learns to route different token vectors to the predicted lexical keys such that a query token vector only interacts with document token vectors routed to the same key. This design significantly reduces the computation cost while maintaining high accuracy. Notably, CITADEL achieves the same or slightly better performance than the previous state of the art, ColBERT-v2, on both in-domain (MS MARCO) and out-of-domain (BEIR) evaluations, while being nearly 40 times faster. Source code and data are available at \url{https://github.com/facebookresearch/dpr-scale/tree/citadel}. | [
"Li, Minghan",
"Lin, Sheng-Chieh",
"Oguz, Barlas",
"Ghoshal, Asish",
"Lin, Jimmy",
"Mehdad, Yashar",
"Yih, Wen-tau",
"Chen, Xilun"
] | CITADEL: Conditional Token Interaction via Dynamic Lexical Routing for Efficient and Effective Multi-Vector Retrieval | acl-long.663 | Oral | 2211.10411 | [
"https://github.com/facebookresearch/dpr-scale"
] | -1 | -1 | -1 | -1 | 0 | [] | [] | [] |
|
https://aclanthology.org/2023.acl-long.664.bib | https://aclanthology.org/2023.acl-long.664/ | @inproceedings{yang-etal-2023-multicapclip,
title = "{M}ulti{C}ap{CLIP}: Auto-Encoding Prompts for Zero-Shot Multilingual Visual Captioning",
author = "Yang, Bang and
Liu, Fenglin and
Wu, Xian and
Wang, Yaowei and
Sun, Xu and
Zou, Yuexian",
editor = "Rogers, Anna and
Boyd-Graber, Jordan and
Okazaki, Naoaki",
booktitle = "Proceedings of the 61st Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers)",
month = jul,
year = "2023",
address = "Toronto, Canada",
publisher = "Association for Computational Linguistics",
url = "https://aclanthology.org/2023.acl-long.664",
doi = "10.18653/v1/2023.acl-long.664",
pages = "11908--11922",
abstract = "Supervised visual captioning models typically require a large scale of images or videos paired with descriptions in a specific language (i.e., the vision-caption pairs) for training. However, collecting and labeling large-scale datasets is time-consuming and expensive for many scenarios and languages. Therefore, sufficient labeled pairs are usually not available. To deal with the label shortage problem, we present a simple yet effective zero-shot approach MultiCapCLIP that can generate visual captions for different scenarios and languages without any labeled vision-caption pairs of downstream datasets. In the training stage, MultiCapCLIP only requires text data for input. Then it conducts two main steps: 1) retrieving concept prompts that preserve the corresponding domain knowledge of new scenarios; 2) auto-encoding the prompts to learn writing styles to output captions in a desired language. In the testing stage, MultiCapCLIP instead takes visual data as input directly to retrieve the concept prompts to generate the final visual descriptions. The extensive experiments on image and video captioning across four benchmarks and four languages (i.e., English, Chinese, German, and French) confirm the effectiveness of our approach. Compared with state-of-the-art zero-shot and weakly-supervised methods, our method achieves 4.8{\%} and 21.5{\%} absolute improvements in terms of BLEU@4 and CIDEr metrics. Our code is available at \url{https://github.com/yangbang18/MultiCapCLIP}.",
}
| Supervised visual captioning models typically require a large scale of images or videos paired with descriptions in a specific language (i.e., the vision-caption pairs) for training. However, collecting and labeling large-scale datasets is time-consuming and expensive for many scenarios and languages. Therefore, sufficient labeled pairs are usually not available. To deal with the label shortage problem, we present a simple yet effective zero-shot approach MultiCapCLIP that can generate visual captions for different scenarios and languages without any labeled vision-caption pairs of downstream datasets. In the training stage, MultiCapCLIP only requires text data for input. Then it conducts two main steps: 1) retrieving concept prompts that preserve the corresponding domain knowledge of new scenarios; 2) auto-encoding the prompts to learn writing styles to output captions in a desired language. In the testing stage, MultiCapCLIP instead takes visual data as input directly to retrieve the concept prompts to generate the final visual descriptions. The extensive experiments on image and video captioning across four benchmarks and four languages (i.e., English, Chinese, German, and French) confirm the effectiveness of our approach. Compared with state-of-the-art zero-shot and weakly-supervised methods, our method achieves 4.8{\%} and 21.5{\%} absolute improvements in terms of BLEU@4 and CIDEr metrics. Our code is available at \url{https://github.com/yangbang18/MultiCapCLIP}. | [
"Yang, Bang",
"Liu, Fenglin",
"Wu, Xian",
"Wang, Yaowei",
"Sun, Xu",
"Zou, Yuexian"
] | MultiCapCLIP: Auto-Encoding Prompts for Zero-Shot Multilingual Visual Captioning | acl-long.664 | Poster | 2308.13218 | [
"https://github.com/yangbang18/multicapclip"
] | -1 | -1 | -1 | -1 | 0 | [] | [] | [] |
|
https://aclanthology.org/2023.acl-long.665.bib | https://aclanthology.org/2023.acl-long.665/ | @inproceedings{varadarajan-etal-2023-transfer,
title = "Transfer and Active Learning for Dissonance Detection: Addressing the Rare-Class Challenge",
author = "Varadarajan, Vasudha and
Juhng, Swanie and
Mahwish, Syeda and
Liu, Xiaoran and
Luby, Jonah and
Luhmann, Christian and
Schwartz, H. Andrew",
editor = "Rogers, Anna and
Boyd-Graber, Jordan and
Okazaki, Naoaki",
booktitle = "Proceedings of the 61st Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers)",
month = jul,
year = "2023",
address = "Toronto, Canada",
publisher = "Association for Computational Linguistics",
url = "https://aclanthology.org/2023.acl-long.665",
doi = "10.18653/v1/2023.acl-long.665",
pages = "11923--11936",
abstract = "While transformer-based systems have enabled greater accuracies with fewer training examples, data acquisition obstacles still persist for rare-class tasks {--} when the class label is very infrequent (e.g. {\textless} 5{\%} of samples). Active learning has in general been proposed to alleviate such challenges, but choice of selection strategy, the criteria by which rare-class examples are chosen, has not been systematically evaluated. Further, transformers enable iterative transfer-learning approaches. We propose and investigate transfer- and active learning solutions to the rare class problem of dissonance detection through utilizing models trained on closely related tasks and the evaluation of acquisition strategies, including a proposed probability-of-rare-class (PRC) approach. We perform these experiments for a specific rare-class problem: collecting language samples of cognitive dissonance from social media. We find that PRC is a simple and effective strategy to guide annotations and ultimately improve model accuracy while transfer-learning in a specific order can improve the cold-start performance of the learner but does not benefit iterations of active learning.",
}
| While transformer-based systems have enabled greater accuracies with fewer training examples, data acquisition obstacles still persist for rare-class tasks {--} when the class label is very infrequent (e.g. {\textless} 5{\%} of samples). Active learning has in general been proposed to alleviate such challenges, but choice of selection strategy, the criteria by which rare-class examples are chosen, has not been systematically evaluated. Further, transformers enable iterative transfer-learning approaches. We propose and investigate transfer- and active learning solutions to the rare class problem of dissonance detection through utilizing models trained on closely related tasks and the evaluation of acquisition strategies, including a proposed probability-of-rare-class (PRC) approach. We perform these experiments for a specific rare-class problem: collecting language samples of cognitive dissonance from social media. We find that PRC is a simple and effective strategy to guide annotations and ultimately improve model accuracy while transfer-learning in a specific order can improve the cold-start performance of the learner but does not benefit iterations of active learning. | [
"Varadarajan, Vasudha",
"Juhng, Swanie",
"Mahwish, Syeda",
"Liu, Xiaoran",
"Luby, Jonah",
"Luhmann, Christian",
"Schwartz, H. Andrew"
] | Transfer and Active Learning for Dissonance Detection: Addressing the Rare-Class Challenge | acl-long.665 | Poster | 2305.02459 | [
"https://github.com/humanlab/rare-class-AL"
] | https://huggingface.co/papers/2305.02459 | 0 | 0 | 0 | 7 | 1 | [
"vasevarad/roberta_dissonance_detector"
] | [] | [] |
https://aclanthology.org/2023.acl-long.666.bib | https://aclanthology.org/2023.acl-long.666/ | @inproceedings{jia-etal-2023-sample,
title = "In-sample Curriculum Learning by Sequence Completion for Natural Language Generation",
author = "Jia, Qi and
Liu, Yizhu and
Tang, Haifeng and
Zhu, Kenny",
editor = "Rogers, Anna and
Boyd-Graber, Jordan and
Okazaki, Naoaki",
booktitle = "Proceedings of the 61st Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers)",
month = jul,
year = "2023",
address = "Toronto, Canada",
publisher = "Association for Computational Linguistics",
url = "https://aclanthology.org/2023.acl-long.666",
doi = "10.18653/v1/2023.acl-long.666",
pages = "11937--11950",
abstract = "Curriculum learning has shown promising improvements in multiple domains by training machine learning models from easy samples to hard ones. Previous works which either design rules or train models for scoring the difficulty highly rely on task-specific expertise, and cannot generalize. Inspired by the {``}easy-to-hard{''} intuition, we propose to do in-sample curriculum learning for natural language generation tasks. Our learning strategy starts training the model to generate the last few words, i.e., do sequence completion, and gradually extends to generate the whole output sequence. Comprehensive experiments show that it generalizes well to different tasks and achieves significant improvements over strong baselines.",
}
| Curriculum learning has shown promising improvements in multiple domains by training machine learning models from easy samples to hard ones. Previous works which either design rules or train models for scoring the difficulty highly rely on task-specific expertise, and cannot generalize. Inspired by the {``}easy-to-hard{''} intuition, we propose to do in-sample curriculum learning for natural language generation tasks. Our learning strategy starts training the model to generate the last few words, i.e., do sequence completion, and gradually extends to generate the whole output sequence. Comprehensive experiments show that it generalizes well to different tasks and achieves significant improvements over strong baselines. | [
"Jia, Qi",
"Liu, Yizhu",
"Tang, Haifeng",
"Zhu, Kenny"
] | In-sample Curriculum Learning by Sequence Completion for Natural Language Generation | acl-long.666 | Poster | 2211.11297 | [
"https://github.com/jiaqisjtu/insamplecurriculumlearning"
] | -1 | -1 | -1 | -1 | 0 | [] | [] | [] |
|
https://aclanthology.org/2023.acl-long.667.bib | https://aclanthology.org/2023.acl-long.667/ | @inproceedings{deng-etal-2023-product,
title = "Product Question Answering in {E}-Commerce: A Survey",
author = "Deng, Yang and
Zhang, Wenxuan and
Yu, Qian and
Lam, Wai",
editor = "Rogers, Anna and
Boyd-Graber, Jordan and
Okazaki, Naoaki",
booktitle = "Proceedings of the 61st Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers)",
month = jul,
year = "2023",
address = "Toronto, Canada",
publisher = "Association for Computational Linguistics",
url = "https://aclanthology.org/2023.acl-long.667",
doi = "10.18653/v1/2023.acl-long.667",
pages = "11951--11964",
abstract = "Product question answering (PQA), aiming to automatically provide instant responses to customer{'}s questions in E-Commerce platforms, has drawn increasing attention in recent years. Compared with typical QA problems, PQA exhibits unique challenges such as the subjectivity and reliability of user-generated contents in E-commerce platforms. Therefore, various problem settings and novel methods have been proposed to capture these special characteristics. In this paper, we aim to systematically review existing research efforts on PQA. Specifically, we categorize PQA studies into four problem settings in terms of the form of provided answers. We analyze the pros and cons, as well as present existing datasets and evaluation protocols for each setting. We further summarize the most significant challenges that characterize PQA from general QA applications and discuss their corresponding solutions. Finally, we conclude this paper by providing the prospect on several future directions.",
}
| Product question answering (PQA), aiming to automatically provide instant responses to customer{'}s questions in E-Commerce platforms, has drawn increasing attention in recent years. Compared with typical QA problems, PQA exhibits unique challenges such as the subjectivity and reliability of user-generated contents in E-commerce platforms. Therefore, various problem settings and novel methods have been proposed to capture these special characteristics. In this paper, we aim to systematically review existing research efforts on PQA. Specifically, we categorize PQA studies into four problem settings in terms of the form of provided answers. We analyze the pros and cons, as well as present existing datasets and evaluation protocols for each setting. We further summarize the most significant challenges that characterize PQA from general QA applications and discuss their corresponding solutions. Finally, we conclude this paper by providing the prospect on several future directions. | [
"Deng, Yang",
"Zhang, Wenxuan",
"Yu, Qian",
"Lam, Wai"
] | Product Question Answering in E-Commerce: A Survey | acl-long.667 | Poster | 2302.08092 | [
""
] | -1 | -1 | -1 | -1 | 0 | [] | [] | [] |
|
https://aclanthology.org/2023.acl-long.668.bib | https://aclanthology.org/2023.acl-long.668/ | @inproceedings{farzana-parde-2023-towards,
title = "Towards Domain-Agnostic and Domain-Adaptive Dementia Detection from Spoken Language",
author = "Farzana, Shahla and
Parde, Natalie",
editor = "Rogers, Anna and
Boyd-Graber, Jordan and
Okazaki, Naoaki",
booktitle = "Proceedings of the 61st Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers)",
month = jul,
year = "2023",
address = "Toronto, Canada",
publisher = "Association for Computational Linguistics",
url = "https://aclanthology.org/2023.acl-long.668",
doi = "10.18653/v1/2023.acl-long.668",
pages = "11965--11978",
abstract = "Health-related speech datasets are often small and varied in focus. This makes it difficult to leverage them to effectively support healthcare goals. Robust transfer of linguistic features across different datasets orbiting the same goal carries potential to address this concern. To test this hypothesis, we experiment with domain adaptation (DA) techniques on heterogeneous spoken language data to evaluate generalizability across diverse datasets for a common task: dementia detection. We find that adapted models exhibit better performance across conversational and task-oriented datasets. The feature-augmented DA method achieves a 22{\%} increase in accuracy adapting from a conversational to task-specific dataset compared to a jointly trained baseline. This suggests promising capacity of these techniques to allow for productive use of disparate data for a complex spoken language healthcare task.",
}
| Health-related speech datasets are often small and varied in focus. This makes it difficult to leverage them to effectively support healthcare goals. Robust transfer of linguistic features across different datasets orbiting the same goal carries potential to address this concern. To test this hypothesis, we experiment with domain adaptation (DA) techniques on heterogeneous spoken language data to evaluate generalizability across diverse datasets for a common task: dementia detection. We find that adapted models exhibit better performance across conversational and task-oriented datasets. The feature-augmented DA method achieves a 22{\%} increase in accuracy adapting from a conversational to task-specific dataset compared to a jointly trained baseline. This suggests promising capacity of these techniques to allow for productive use of disparate data for a complex spoken language healthcare task. | [
"Farzana, Shahla",
"Parde, Natalie"
] | Towards Domain-Agnostic and Domain-Adaptive Dementia Detection from Spoken Language | acl-long.668 | Oral | [
""
] | -1 | -1 | -1 | -1 | 0 | [] | [] | [] |
||
https://aclanthology.org/2023.acl-long.669.bib | https://aclanthology.org/2023.acl-long.669/ | @inproceedings{du-etal-2023-generalizing,
title = "Generalizing Backpropagation for Gradient-Based Interpretability",
author = "Du, Kevin and
Torroba Hennigen, Lucas and
Stoehr, Niklas and
Warstadt, Alex and
Cotterell, Ryan",
editor = "Rogers, Anna and
Boyd-Graber, Jordan and
Okazaki, Naoaki",
booktitle = "Proceedings of the 61st Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers)",
month = jul,
year = "2023",
address = "Toronto, Canada",
publisher = "Association for Computational Linguistics",
url = "https://aclanthology.org/2023.acl-long.669",
doi = "10.18653/v1/2023.acl-long.669",
pages = "11979--11995",
abstract = "Many popular feature-attribution methods for interpreting deep neural networks rely on computing the gradients of a model{'}s output with respect to its inputs. While these methods can indicate which input features may be important for the model{'}s prediction, they reveal little about the inner workings of the model itself. In this paper, we observe that the gradient computation of a model is a special case of a more general formulation using semirings. This observation allows us to generalize the backpropagation algorithm to efficiently compute other interpretable statistics about the gradient graph of a neural network, such as the highest-weighted path and entropy. We implement this generalized algorithm, evaluate it on synthetic datasets to better understand the statistics it computes, and apply it to study BERT{'}s behavior on the subject{--}verb number agreement task (SVA). With this method, we (a) validate that the amount of gradient flow through a component of a model reflects its importance to a prediction and (b) for SVA, identify which pathways of the self-attention mechanism are most important.",
}
| Many popular feature-attribution methods for interpreting deep neural networks rely on computing the gradients of a model{'}s output with respect to its inputs. While these methods can indicate which input features may be important for the model{'}s prediction, they reveal little about the inner workings of the model itself. In this paper, we observe that the gradient computation of a model is a special case of a more general formulation using semirings. This observation allows us to generalize the backpropagation algorithm to efficiently compute other interpretable statistics about the gradient graph of a neural network, such as the highest-weighted path and entropy. We implement this generalized algorithm, evaluate it on synthetic datasets to better understand the statistics it computes, and apply it to study BERT{'}s behavior on the subject{--}verb number agreement task (SVA). With this method, we (a) validate that the amount of gradient flow through a component of a model reflects its importance to a prediction and (b) for SVA, identify which pathways of the self-attention mechanism are most important. | [
"Du, Kevin",
"Torroba Hennigen, Lucas",
"Stoehr, Niklas",
"Warstadt, Alex",
"Cotterell, Ryan"
] | Generalizing Backpropagation for Gradient-Based Interpretability | acl-long.669 | Oral | 2307.03056 | [
"https://github.com/kdu4108/semiring-backprop-exps"
] | -1 | -1 | -1 | -1 | 0 | [] | [] | [] |
|
https://aclanthology.org/2023.acl-long.670.bib | https://aclanthology.org/2023.acl-long.670/ | @inproceedings{mou-etal-2023-uppam,
title = "{UPPAM}: A Unified Pre-training Architecture for Political Actor Modeling based on Language",
author = "Mou, Xinyi and
Wei, Zhongyu and
Zhang, Qi and
Huang, Xuanjing",
editor = "Rogers, Anna and
Boyd-Graber, Jordan and
Okazaki, Naoaki",
booktitle = "Proceedings of the 61st Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers)",
month = jul,
year = "2023",
address = "Toronto, Canada",
publisher = "Association for Computational Linguistics",
url = "https://aclanthology.org/2023.acl-long.670",
doi = "10.18653/v1/2023.acl-long.670",
pages = "11996--12012",
abstract = "Modeling political actors is at the core of quantitative political science. Existing works have incorporated contextual information to better learn the representation of political actors for specific tasks through graph models. However, they are limited to the structure and objective of training settings and can not be generalized to all politicians and other tasks. In this paper, we propose a Unified Pre-training Architecture for Political Actor Modeling based on language (UPPAM). In UPPAM, we aggregate statements to represent political actors and learn the mapping from languages to representation, instead of learning the representation of particular persons. We further design structure-aware contrastive learning and behavior-driven contrastive learning tasks, to inject multidimensional information in the political context into the mapping. In this framework, we can profile political actors from different aspects and solve various downstream tasks. Experimental results demonstrate the effectiveness and capability of generalization of our method.",
}
| Modeling political actors is at the core of quantitative political science. Existing works have incorporated contextual information to better learn the representation of political actors for specific tasks through graph models. However, they are limited to the structure and objective of training settings and can not be generalized to all politicians and other tasks. In this paper, we propose a Unified Pre-training Architecture for Political Actor Modeling based on language (UPPAM). In UPPAM, we aggregate statements to represent political actors and learn the mapping from languages to representation, instead of learning the representation of particular persons. We further design structure-aware contrastive learning and behavior-driven contrastive learning tasks, to inject multidimensional information in the political context into the mapping. In this framework, we can profile political actors from different aspects and solve various downstream tasks. Experimental results demonstrate the effectiveness and capability of generalization of our method. | [
"Mou, Xinyi",
"Wei, Zhongyu",
"Zhang, Qi",
"Huang, Xuanjing"
] | UPPAM: A Unified Pre-training Architecture for Political Actor Modeling based on Language | acl-long.670 | Poster | [
""
] | -1 | -1 | -1 | -1 | 0 | [] | [] | [] |
||
https://aclanthology.org/2023.acl-long.671.bib | https://aclanthology.org/2023.acl-long.671/ | @inproceedings{feng-etal-2023-generic,
title = "Generic Temporal Reasoning with Differential Analysis and Explanation",
author = "Feng, Yu and
Zhou, Ben and
Wang, Haoyu and
Jin, Helen and
Roth, Dan",
editor = "Rogers, Anna and
Boyd-Graber, Jordan and
Okazaki, Naoaki",
booktitle = "Proceedings of the 61st Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers)",
month = jul,
year = "2023",
address = "Toronto, Canada",
publisher = "Association for Computational Linguistics",
url = "https://aclanthology.org/2023.acl-long.671",
doi = "10.18653/v1/2023.acl-long.671",
pages = "12013--12029",
abstract = "Temporal reasoning is the task of predicting temporal relations of event pairs. While temporal reasoning models can perform reasonably well on in-domain benchmarks, we have little idea of these systems{'} generalizability due to existing datasets{'} limitations. In this work, we introduce a novel task named TODAY that bridges this gap with temporal differential analysis, which as the name suggests, evaluates whether systems can correctly understand the effect of incremental changes. Specifically, TODAY introduces slight contextual changes for given event pairs, and systems are asked to tell how this subtle contextual change would affect relevant temporal relation distributions. To facilitate learning, TODAY also annotates human explanations. We show that existing models, including GPT-3.5, drop to random guessing on TODAY, suggesting that they heavily rely on spurious information rather than proper reasoning for temporal predictions. On the other hand, we show that TODAY{'}s supervision style and explanation annotations can be used in joint learning, encouraging models to use more appropriate signals during training and thus outperform across several benchmarks. TODAY can also be used to train models to solicit incidental supervision from noisy sources such as GPT-3.5, thus moving us more toward the goal of generic temporal reasoning systems.",
}
| Temporal reasoning is the task of predicting temporal relations of event pairs. While temporal reasoning models can perform reasonably well on in-domain benchmarks, we have little idea of these systems{'} generalizability due to existing datasets{'} limitations. In this work, we introduce a novel task named TODAY that bridges this gap with temporal differential analysis, which as the name suggests, evaluates whether systems can correctly understand the effect of incremental changes. Specifically, TODAY introduces slight contextual changes for given event pairs, and systems are asked to tell how this subtle contextual change would affect relevant temporal relation distributions. To facilitate learning, TODAY also annotates human explanations. We show that existing models, including GPT-3.5, drop to random guessing on TODAY, suggesting that they heavily rely on spurious information rather than proper reasoning for temporal predictions. On the other hand, we show that TODAY{'}s supervision style and explanation annotations can be used in joint learning, encouraging models to use more appropriate signals during training and thus outperform across several benchmarks. TODAY can also be used to train models to solicit incidental supervision from noisy sources such as GPT-3.5, thus moving us more toward the goal of generic temporal reasoning systems. | [
"Feng, Yu",
"Zhou, Ben",
"Wang, Haoyu",
"Jin, Helen",
"Roth, Dan"
] | Generic Temporal Reasoning with Differential Analysis and Explanation | acl-long.671 | Poster | 2212.10467 | [
""
] | -1 | -1 | -1 | -1 | 0 | [] | [] | [] |
|
https://aclanthology.org/2023.acl-long.672.bib | https://aclanthology.org/2023.acl-long.672/ | @inproceedings{towle-zhou-2023-model,
title = "Model-Based Simulation for Optimising Smart Reply",
author = "Towle, Benjamin and
Zhou, Ke",
editor = "Rogers, Anna and
Boyd-Graber, Jordan and
Okazaki, Naoaki",
booktitle = "Proceedings of the 61st Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers)",
month = jul,
year = "2023",
address = "Toronto, Canada",
publisher = "Association for Computational Linguistics",
url = "https://aclanthology.org/2023.acl-long.672",
doi = "10.18653/v1/2023.acl-long.672",
pages = "12030--12043",
abstract = "Smart Reply (SR) systems present a user with a set of replies, of which one can be selected in place of having to type out a response. To perform well at this task, a system should be able to effectively present the user with a diverse set of options, to maximise the chance that at least one of them conveys the user{'}s desired response. This is a significant challenge, due to the lack of datasets containing sets of responses to learn from. Resultantly, previous work has focused largely on post-hoc diversification, rather than explicitly learning to predict sets of responses. Motivated by this problem, we present a novel method SimSR, that employs model-based simulation to discover high-value response sets, through simulating possible user responses with a learned world model. Unlike previous approaches, this allows our method to directly optimise the end-goal of SR{--}maximising the relevance of at least one of the predicted replies. Empirically on two public datasets, when compared to SoTA baselines, our method achieves up to 21{\%} and 18{\%} improvement in ROUGE score and Self-ROUGE score respectively.",
}
| Smart Reply (SR) systems present a user with a set of replies, of which one can be selected in place of having to type out a response. To perform well at this task, a system should be able to effectively present the user with a diverse set of options, to maximise the chance that at least one of them conveys the user{'}s desired response. This is a significant challenge, due to the lack of datasets containing sets of responses to learn from. Resultantly, previous work has focused largely on post-hoc diversification, rather than explicitly learning to predict sets of responses. Motivated by this problem, we present a novel method SimSR, that employs model-based simulation to discover high-value response sets, through simulating possible user responses with a learned world model. Unlike previous approaches, this allows our method to directly optimise the end-goal of SR{--}maximising the relevance of at least one of the predicted replies. Empirically on two public datasets, when compared to SoTA baselines, our method achieves up to 21{\%} and 18{\%} improvement in ROUGE score and Self-ROUGE score respectively. | [
"Towle, Benjamin",
"Zhou, Ke"
] | Model-Based Simulation for Optimising Smart Reply | acl-long.672 | Poster | 2305.16852 | [
"https://github.com/benjamintowle/simsr"
] | -1 | -1 | -1 | -1 | 0 | [] | [] | [] |
|
https://aclanthology.org/2023.acl-long.673.bib | https://aclanthology.org/2023.acl-long.673/ | @inproceedings{wieting-etal-2023-beyond,
title = "Beyond Contrastive Learning: A Variational Generative Model for Multilingual Retrieval",
author = "Wieting, John and
Clark, Jonathan and
Cohen, William and
Neubig, Graham and
Berg-Kirkpatrick, Taylor",
editor = "Rogers, Anna and
Boyd-Graber, Jordan and
Okazaki, Naoaki",
booktitle = "Proceedings of the 61st Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers)",
month = jul,
year = "2023",
address = "Toronto, Canada",
publisher = "Association for Computational Linguistics",
url = "https://aclanthology.org/2023.acl-long.673",
doi = "10.18653/v1/2023.acl-long.673",
pages = "12044--12066",
abstract = "Contrastive learning has been successfully used for retrieval of semantically aligned sentences, but it often requires large batch sizes or careful engineering to work well. In this paper, we instead propose a generative model for learning multilingual text embeddings which can be used to retrieve or score sentence pairs. Our model operates on parallel data in N languages and, through an approximation we introduce, efficiently encourages source separation in this multilingual setting, separating semantic information that is shared between translations from stylistic or language-specific variation. We show careful large-scale comparisons between contrastive and generation-based approaches for learning multilingual text embeddings, a comparison that has not been done to the best of our knowledge despite the popularity of these approaches. We evaluate this method on a suite of tasks including semantic similarity, bitext mining, and cross-lingual question retrieval - the last of which we introduce in this paper. Overall, our model outperforms both a strong contrastive and generative baseline on these tasks.",
}
| Contrastive learning has been successfully used for retrieval of semantically aligned sentences, but it often requires large batch sizes or careful engineering to work well. In this paper, we instead propose a generative model for learning multilingual text embeddings which can be used to retrieve or score sentence pairs. Our model operates on parallel data in N languages and, through an approximation we introduce, efficiently encourages source separation in this multilingual setting, separating semantic information that is shared between translations from stylistic or language-specific variation. We show careful large-scale comparisons between contrastive and generation-based approaches for learning multilingual text embeddings, a comparison that has not been done to the best of our knowledge despite the popularity of these approaches. We evaluate this method on a suite of tasks including semantic similarity, bitext mining, and cross-lingual question retrieval - the last of which we introduce in this paper. Overall, our model outperforms both a strong contrastive and generative baseline on these tasks. | [
"Wieting, John",
"Clark, Jonathan",
"Cohen, William",
"Neubig, Graham",
"Berg-Kirkpatrick, Taylor"
] | Beyond Contrastive Learning: A Variational Generative Model for Multilingual Retrieval | acl-long.673 | Poster | 2212.10726 | [
"https://github.com/google-research/google-research"
] | -1 | -1 | -1 | -1 | 0 | [] | [] | [] |
|
https://aclanthology.org/2023.acl-long.674.bib | https://aclanthology.org/2023.acl-long.674/ | @inproceedings{he-etal-2023-blind,
title = "On the Blind Spots of Model-Based Evaluation Metrics for Text Generation",
author = "He, Tianxing and
Zhang, Jingyu and
Wang, Tianle and
Kumar, Sachin and
Cho, Kyunghyun and
Glass, James and
Tsvetkov, Yulia",
editor = "Rogers, Anna and
Boyd-Graber, Jordan and
Okazaki, Naoaki",
booktitle = "Proceedings of the 61st Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers)",
month = jul,
year = "2023",
address = "Toronto, Canada",
publisher = "Association for Computational Linguistics",
url = "https://aclanthology.org/2023.acl-long.674",
doi = "10.18653/v1/2023.acl-long.674",
pages = "12067--12097",
abstract = "In this work, we explore a useful but often neglected methodology for robustness analysis of text generation evaluation metrics: stress tests with synthetic data. Basically, we design and synthesize a wide range of potential errors and check whether they result in a commensurate drop in the metric scores. We examine a range of recently proposed evaluation metrics based on pretrained language models, for the tasks of open-ended generation, translation, and summarization. Our experiments reveal interesting insensitivities, biases, or even loopholes in existing metrics. For example, we find that BERTScore is confused by truncation errors in summarization, and MAUVE (built on top of GPT-2) is insensitive to errors at the beginning or middle of generations. Further, we investigate the reasons behind these blind spots and suggest practical workarounds for a more reliable evaluation of text generation. We have released our code and data at \url{https://github.com/cloudygoose/blindspot_nlg}.",
}
| In this work, we explore a useful but often neglected methodology for robustness analysis of text generation evaluation metrics: stress tests with synthetic data. Basically, we design and synthesize a wide range of potential errors and check whether they result in a commensurate drop in the metric scores. We examine a range of recently proposed evaluation metrics based on pretrained language models, for the tasks of open-ended generation, translation, and summarization. Our experiments reveal interesting insensitivities, biases, or even loopholes in existing metrics. For example, we find that BERTScore is confused by truncation errors in summarization, and MAUVE (built on top of GPT-2) is insensitive to errors at the beginning or middle of generations. Further, we investigate the reasons behind these blind spots and suggest practical workarounds for a more reliable evaluation of text generation. We have released our code and data at \url{https://github.com/cloudygoose/blindspot_nlg}. | [
"He, Tianxing",
"Zhang, Jingyu",
"Wang, Tianle",
"Kumar, Sachin",
"Cho, Kyunghyun",
"Glass, James",
"Tsvetkov, Yulia"
] | On the Blind Spots of Model-Based Evaluation Metrics for Text Generation | acl-long.674 | Oral | 2212.10020 | [
"https://github.com/cloudygoose/blindspot_nlg"
] | -1 | -1 | -1 | -1 | 0 | [] | [] | [] |
|
https://aclanthology.org/2023.acl-long.675.bib | https://aclanthology.org/2023.acl-long.675/ | @inproceedings{pezzelle-2023-dealing,
title = "Dealing with Semantic Underspecification in Multimodal {NLP}",
author = "Pezzelle, Sandro",
editor = "Rogers, Anna and
Boyd-Graber, Jordan and
Okazaki, Naoaki",
booktitle = "Proceedings of the 61st Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers)",
month = jul,
year = "2023",
address = "Toronto, Canada",
publisher = "Association for Computational Linguistics",
url = "https://aclanthology.org/2023.acl-long.675",
doi = "10.18653/v1/2023.acl-long.675",
pages = "12098--12112",
abstract = "Intelligent systems that aim at mastering language as humans do must deal with its semantic underspecification, namely, the possibility for a linguistic signal to convey only part of the information needed for communication to succeed. Consider the usages of the pronoun they, which can leave the gender and number of its referent(s) underspecified. Semantic underspecification is not a bug but a crucial language feature that boosts its storage and processing efficiency. Indeed, human speakers can quickly and effortlessly integrate semantically-underspecified linguistic signals with a wide range of non-linguistic information, e.g., the multimodal context, social or cultural conventions, and shared knowledge. Standard NLP models have, in principle, no or limited access to such extra information, while multimodal systems grounding language into other modalities, such as vision, are naturally equipped to account for this phenomenon. However, we show that they struggle with it, which could negatively affect their performance and lead to harmful consequences when used for applications. In this position paper, we argue that our community should be aware of semantic underspecification if it aims to develop language technology that can successfully interact with human users. We discuss some applications where mastering it is crucial and outline a few directions toward achieving this goal.",
}
| Intelligent systems that aim at mastering language as humans do must deal with its semantic underspecification, namely, the possibility for a linguistic signal to convey only part of the information needed for communication to succeed. Consider the usages of the pronoun they, which can leave the gender and number of its referent(s) underspecified. Semantic underspecification is not a bug but a crucial language feature that boosts its storage and processing efficiency. Indeed, human speakers can quickly and effortlessly integrate semantically-underspecified linguistic signals with a wide range of non-linguistic information, e.g., the multimodal context, social or cultural conventions, and shared knowledge. Standard NLP models have, in principle, no or limited access to such extra information, while multimodal systems grounding language into other modalities, such as vision, are naturally equipped to account for this phenomenon. However, we show that they struggle with it, which could negatively affect their performance and lead to harmful consequences when used for applications. In this position paper, we argue that our community should be aware of semantic underspecification if it aims to develop language technology that can successfully interact with human users. We discuss some applications where mastering it is crucial and outline a few directions toward achieving this goal. | [
"Pezzelle, S",
"ro"
] | Dealing with Semantic Underspecification in Multimodal NLP | acl-long.675 | Poster | 2306.05240 | [
"https://github.com/sandropezzelle/sunglass"
] | -1 | -1 | -1 | -1 | 0 | [] | [] | [] |
|
https://aclanthology.org/2023.acl-long.676.bib | https://aclanthology.org/2023.acl-long.676/ | @inproceedings{wiegmann-etal-2023-trigger,
title = "Trigger Warning Assignment as a Multi-Label Document Classification Problem",
author = {Wiegmann, Matti and
Wolska, Magdalena and
Schr{\"o}der, Christopher and
Borchardt, Ole and
Stein, Benno and
Potthast, Martin},
editor = "Rogers, Anna and
Boyd-Graber, Jordan and
Okazaki, Naoaki",
booktitle = "Proceedings of the 61st Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers)",
month = jul,
year = "2023",
address = "Toronto, Canada",
publisher = "Association for Computational Linguistics",
url = "https://aclanthology.org/2023.acl-long.676",
doi = "10.18653/v1/2023.acl-long.676",
pages = "12113--12134",
abstract = "A trigger warning is used to warn people about potentially disturbing content. We introduce trigger warning assignment as a multi-label classification task, create the Webis Trigger Warning Corpus 2022, and with it the first dataset of 1 million fanfiction works from Archive of our Own with up to 36 different warnings per document. To provide a reliable catalog of trigger warnings, we organized 41 million of free-form tags assigned by fanfiction authors into the first comprehensive taxonomy of trigger warnings by mapping them to the 36 institutionally recommended warnings. To determine the best operationalization of trigger warnings, we explore state-of-the-art multi-label models, examining the trade-off between assigning coarse- and fine-grained warnings, open- and closed-set classification, document length, and label confidence. Our models achieve micro-F1 scores of about 0.5, which reveals the difficulty of the task. Tailored representations, long input sequences, and a higher recall on rare warnings would help.",
}
| A trigger warning is used to warn people about potentially disturbing content. We introduce trigger warning assignment as a multi-label classification task, create the Webis Trigger Warning Corpus 2022, and with it the first dataset of 1 million fanfiction works from Archive of our Own with up to 36 different warnings per document. To provide a reliable catalog of trigger warnings, we organized 41 million of free-form tags assigned by fanfiction authors into the first comprehensive taxonomy of trigger warnings by mapping them to the 36 institutionally recommended warnings. To determine the best operationalization of trigger warnings, we explore state-of-the-art multi-label models, examining the trade-off between assigning coarse- and fine-grained warnings, open- and closed-set classification, document length, and label confidence. Our models achieve micro-F1 scores of about 0.5, which reveals the difficulty of the task. Tailored representations, long input sequences, and a higher recall on rare warnings would help. | [
"Wiegmann, Matti",
"Wolska, Magdalena",
"Schr{\\\"o}der, Christopher",
"Borchardt, Ole",
"Stein, Benno",
"Potthast, Martin"
] | Trigger Warning Assignment as a Multi-Label Document Classification Problem | acl-long.676 | Poster | [
""
] | -1 | -1 | -1 | -1 | 0 | [] | [] | [] |
||
https://aclanthology.org/2023.acl-long.677.bib | https://aclanthology.org/2023.acl-long.677/ | @inproceedings{zhuo-etal-2023-whitenedcse,
title = "{W}hitened{CSE}: Whitening-based Contrastive Learning of Sentence Embeddings",
author = "Zhuo, Wenjie and
Sun, Yifan and
Wang, Xiaohan and
Zhu, Linchao and
Yang, Yi",
editor = "Rogers, Anna and
Boyd-Graber, Jordan and
Okazaki, Naoaki",
booktitle = "Proceedings of the 61st Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers)",
month = jul,
year = "2023",
address = "Toronto, Canada",
publisher = "Association for Computational Linguistics",
url = "https://aclanthology.org/2023.acl-long.677",
doi = "10.18653/v1/2023.acl-long.677",
pages = "12135--12148",
abstract = "This paper presents a whitening-based contrastive learning method for sentence embedding learning (WhitenedCSE), which combines contrastive learning with a novel shuffled group whitening. Generally, contrastive learning pulls distortions of a single sample (i.e., positive samples) close and push negative samples far away, correspondingly facilitating the alignment and uniformity in the feature space. A popular alternative to the {``}pushing{''} operation is whitening the feature space, which scatters all the samples for uniformity. Since the whitening and the contrastive learning have large redundancy w.r.t. the uniformity, they are usually used separately and do not easily work together. For the first time, this paper integrates whitening into the contrastive learning scheme and facilitates two benefits. 1) Better uniformity. We find that these two approaches are not totally redundant but actually have some complementarity due to different uniformity mechanism. 2) Better alignment. We randomly divide the feature into multiple groups along the channel axis and perform whitening independently within each group. By shuffling the group division, we derive multiple distortions of a single sample and thus increase the positive sample diversity. Consequently, using multiple positive samples with enhanced diversity further improves contrastive learning due to better alignment. Extensive experiments on seven semantic textual similarity tasks show our method achieves consistent improvement over the contrastive learning baseline and sets new states of the art, e.g., 78.78{\%} (+2.53{\%} based on BERT{pasted macro {`}BA{'}}) Spearman correlation on STS tasks.",
}
| This paper presents a whitening-based contrastive learning method for sentence embedding learning (WhitenedCSE), which combines contrastive learning with a novel shuffled group whitening. Generally, contrastive learning pulls distortions of a single sample (i.e., positive samples) close and push negative samples far away, correspondingly facilitating the alignment and uniformity in the feature space. A popular alternative to the {``}pushing{''} operation is whitening the feature space, which scatters all the samples for uniformity. Since the whitening and the contrastive learning have large redundancy w.r.t. the uniformity, they are usually used separately and do not easily work together. For the first time, this paper integrates whitening into the contrastive learning scheme and facilitates two benefits. 1) Better uniformity. We find that these two approaches are not totally redundant but actually have some complementarity due to different uniformity mechanism. 2) Better alignment. We randomly divide the feature into multiple groups along the channel axis and perform whitening independently within each group. By shuffling the group division, we derive multiple distortions of a single sample and thus increase the positive sample diversity. Consequently, using multiple positive samples with enhanced diversity further improves contrastive learning due to better alignment. Extensive experiments on seven semantic textual similarity tasks show our method achieves consistent improvement over the contrastive learning baseline and sets new states of the art, e.g., 78.78{\%} (+2.53{\%} based on BERT{pasted macro {`}BA{'}}) Spearman correlation on STS tasks. | [
"Zhuo, Wenjie",
"Sun, Yifan",
"Wang, Xiaohan",
"Zhu, Linchao",
"Yang, Yi"
] | WhitenedCSE: Whitening-based Contrastive Learning of Sentence Embeddings | acl-long.677 | Oral | [
""
] | -1 | -1 | -1 | -1 | 0 | [] | [] | [] |
||
https://aclanthology.org/2023.acl-long.678.bib | https://aclanthology.org/2023.acl-long.678/ | @inproceedings{zhang-etal-2023-federated,
title = "Federated Learning for Semantic Parsing: Task Formulation, Evaluation Setup, New Algorithms",
author = "Zhang, Tianshu and
Liu, Changchang and
Lee, Wei-Han and
Su, Yu and
Sun, Huan",
editor = "Rogers, Anna and
Boyd-Graber, Jordan and
Okazaki, Naoaki",
booktitle = "Proceedings of the 61st Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers)",
month = jul,
year = "2023",
address = "Toronto, Canada",
publisher = "Association for Computational Linguistics",
url = "https://aclanthology.org/2023.acl-long.678",
doi = "10.18653/v1/2023.acl-long.678",
pages = "12149--12163",
abstract = "This paper studies a new task of federated learning (FL) for semantic parsing, where multiple clients collaboratively train one global model without sharing their semantic parsing data. By leveraging data from multiple clients, the FL paradigm can be especially beneficial for clients that have little training data to develop a data-hungry neural semantic parser on their own. We propose an evaluation setup to study this task, where we re-purpose widely-used single-domain text-to-SQL datasets as clients to form a realistic heterogeneous FL setting and collaboratively train a global model. As standard FL algorithms suffer from the high client heterogeneity in our realistic setup, we further propose a novel LOss Reduction Adjusted Re-weighting (Lorar) mechanism, which adjusts each client{'}s contribution to the global model update based on its training loss reduction during each round. Our intuition is that the larger the loss reduction, the further away the current global model is from the client{'}s local optimum, and the larger weight the client should get. By applying Lorar to three widely adopted FL algorithms (FedAvg, FedOPT and FedProx), we observe that their performance can be improved substantially on average (4{\%}-20{\%} absolute gain under MacroAvg) and that clients with smaller datasets enjoy larger performance gains. In addition, the global model converges faster for almost all the clients.",
}
| This paper studies a new task of federated learning (FL) for semantic parsing, where multiple clients collaboratively train one global model without sharing their semantic parsing data. By leveraging data from multiple clients, the FL paradigm can be especially beneficial for clients that have little training data to develop a data-hungry neural semantic parser on their own. We propose an evaluation setup to study this task, where we re-purpose widely-used single-domain text-to-SQL datasets as clients to form a realistic heterogeneous FL setting and collaboratively train a global model. As standard FL algorithms suffer from the high client heterogeneity in our realistic setup, we further propose a novel LOss Reduction Adjusted Re-weighting (Lorar) mechanism, which adjusts each client{'}s contribution to the global model update based on its training loss reduction during each round. Our intuition is that the larger the loss reduction, the further away the current global model is from the client{'}s local optimum, and the larger weight the client should get. By applying Lorar to three widely adopted FL algorithms (FedAvg, FedOPT and FedProx), we observe that their performance can be improved substantially on average (4{\%}-20{\%} absolute gain under MacroAvg) and that clients with smaller datasets enjoy larger performance gains. In addition, the global model converges faster for almost all the clients. | [
"Zhang, Tianshu",
"Liu, Changchang",
"Lee, Wei-Han",
"Su, Yu",
"Sun, Huan"
] | Federated Learning for Semantic Parsing: Task Formulation, Evaluation Setup, New Algorithms | acl-long.678 | Poster | 2305.17221 | [
"https://github.com/osu-nlp-group/fl4semanticparsing"
] | -1 | -1 | -1 | -1 | 0 | [] | [] | [] |
|
https://aclanthology.org/2023.acl-long.679.bib | https://aclanthology.org/2023.acl-long.679/ | @inproceedings{luo-etal-2023-causality,
title = "Causality-Guided Multi-Memory Interaction Network for Multivariate Stock Price Movement Prediction",
author = "Luo, Di and
Liao, Weiheng and
Li, Shuqi and
Cheng, Xin and
Yan, Rui",
editor = "Rogers, Anna and
Boyd-Graber, Jordan and
Okazaki, Naoaki",
booktitle = "Proceedings of the 61st Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers)",
month = jul,
year = "2023",
address = "Toronto, Canada",
publisher = "Association for Computational Linguistics",
url = "https://aclanthology.org/2023.acl-long.679",
doi = "10.18653/v1/2023.acl-long.679",
pages = "12164--12176",
abstract = "Over the past few years, we{'}ve witnessed an enormous interest in stock price movement prediction using AI techniques. In recent literature, auxiliary data has been used to improve prediction accuracy, such as textual news. When predicting a particular stock, we assume that information from other stocks should also be utilized as auxiliary data to enhance performance. In this paper, we propose the Causality-guided Multi-memory Interaction Network (CMIN), a novel end-to-end deep neural network for stock movement prediction which, for the first time, models the multi-modality between financial text data and causality-enhanced stock correlations to achieve higher prediction accuracy. CMIN transforms the basic attention mechanism into Causal Attention by calculating transfer entropy between multivariate stocks in order to avoid attention on spurious correlations. Furthermore, we introduce a fusion mechanism to model the multi-directional interactions through which CMIN learns not only the self-influence but also the interactive influence in information flows representing the interrelationship between text and stock correlations. The effectiveness of the proposed approach is demonstrated by experiments on three real-world datasets collected from the U.S. and Chinese markets, where CMIN outperforms existing models to establish a new state-of-the-art prediction accuracy.",
}
| Over the past few years, we{'}ve witnessed an enormous interest in stock price movement prediction using AI techniques. In recent literature, auxiliary data has been used to improve prediction accuracy, such as textual news. When predicting a particular stock, we assume that information from other stocks should also be utilized as auxiliary data to enhance performance. In this paper, we propose the Causality-guided Multi-memory Interaction Network (CMIN), a novel end-to-end deep neural network for stock movement prediction which, for the first time, models the multi-modality between financial text data and causality-enhanced stock correlations to achieve higher prediction accuracy. CMIN transforms the basic attention mechanism into Causal Attention by calculating transfer entropy between multivariate stocks in order to avoid attention on spurious correlations. Furthermore, we introduce a fusion mechanism to model the multi-directional interactions through which CMIN learns not only the self-influence but also the interactive influence in information flows representing the interrelationship between text and stock correlations. The effectiveness of the proposed approach is demonstrated by experiments on three real-world datasets collected from the U.S. and Chinese markets, where CMIN outperforms existing models to establish a new state-of-the-art prediction accuracy. | [
"Luo, Di",
"Liao, Weiheng",
"Li, Shuqi",
"Cheng, Xin",
"Yan, Rui"
] | Causality-Guided Multi-Memory Interaction Network for Multivariate Stock Price Movement Prediction | acl-long.679 | Poster | [
""
] | -1 | -1 | -1 | -1 | 0 | [] | [] | [] |
||
https://aclanthology.org/2023.acl-long.680.bib | https://aclanthology.org/2023.acl-long.680/ | @inproceedings{gao-etal-2023-dsrm,
title = "{DSRM}: Boost Textual Adversarial Training with Distribution Shift Risk Minimization",
author = "Gao, SongYang and
Dou, Shihan and
Liu, Yan and
Wang, Xiao and
Zhang, Qi and
Wei, Zhongyu and
Ma, Jin and
Shan, Ying",
editor = "Rogers, Anna and
Boyd-Graber, Jordan and
Okazaki, Naoaki",
booktitle = "Proceedings of the 61st Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers)",
month = jul,
year = "2023",
address = "Toronto, Canada",
publisher = "Association for Computational Linguistics",
url = "https://aclanthology.org/2023.acl-long.680",
doi = "10.18653/v1/2023.acl-long.680",
pages = "12177--12189",
abstract = "Adversarial training is one of the best-performing methods in improving the robustness of deep language models. However, robust models come at the cost of high time consumption, as they require multi-step gradient ascents or word substitutions to obtain adversarial samples. In addition, these generated samples are deficient in grammatical quality and semantic consistency, which impairs the effectiveness of adversarial training. To address these problems, we introduce a novel, effective procedure for instead adversarial training with only clean data. Our procedure, distribution shift risk minimization (DSRM), estimates the adversarial loss by perturbing the input data{'}s probability distribution rather than their embeddings. This formulation results in a robust model that minimizes the expected global loss under adversarial attacks. Our approach requires zero adversarial samples for training and reduces time consumption by up to 70{\%} compared to current best-performing adversarial training methods. Experiments demonstrate that DSRM considerably improves BERT{'}s resistance to textual adversarial attacks and achieves state-of-the-art robust accuracy on various benchmarks.",
}
| Adversarial training is one of the best-performing methods in improving the robustness of deep language models. However, robust models come at the cost of high time consumption, as they require multi-step gradient ascents or word substitutions to obtain adversarial samples. In addition, these generated samples are deficient in grammatical quality and semantic consistency, which impairs the effectiveness of adversarial training. To address these problems, we introduce a novel, effective procedure for instead adversarial training with only clean data. Our procedure, distribution shift risk minimization (DSRM), estimates the adversarial loss by perturbing the input data{'}s probability distribution rather than their embeddings. This formulation results in a robust model that minimizes the expected global loss under adversarial attacks. Our approach requires zero adversarial samples for training and reduces time consumption by up to 70{\%} compared to current best-performing adversarial training methods. Experiments demonstrate that DSRM considerably improves BERT{'}s resistance to textual adversarial attacks and achieves state-of-the-art robust accuracy on various benchmarks. | [
"Gao, SongYang",
"Dou, Shihan",
"Liu, Yan",
"Wang, Xiao",
"Zhang, Qi",
"Wei, Zhongyu",
"Ma, Jin",
"Shan, Ying"
] | DSRM: Boost Textual Adversarial Training with Distribution Shift Risk Minimization | acl-long.680 | Poster | 2306.15164 | [
"https://github.com/sleepthroughdifficulties/dsrm"
] | -1 | -1 | -1 | -1 | 0 | [] | [] | [] |
|
https://aclanthology.org/2023.acl-long.681.bib | https://aclanthology.org/2023.acl-long.681/ | @inproceedings{song-etal-2023-simple,
title = "A Simple and Flexible Modeling for Mental Disorder Detection by Learning from Clinical Questionnaires",
author = "Song, Hoyun and
Shin, Jisu and
Lee, Huije and
Park, Jong",
editor = "Rogers, Anna and
Boyd-Graber, Jordan and
Okazaki, Naoaki",
booktitle = "Proceedings of the 61st Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers)",
month = jul,
year = "2023",
address = "Toronto, Canada",
publisher = "Association for Computational Linguistics",
url = "https://aclanthology.org/2023.acl-long.681",
doi = "10.18653/v1/2023.acl-long.681",
pages = "12190--12206",
abstract = "Social media is one of the most highly sought resources for analyzing characteristics of the language by its users. In particular, many researchers utilized various linguistic features of mental health problems from social media. However, existing approaches to detecting mental disorders face critical challenges, such as the scarcity of high-quality data or the trade-off between addressing the complexity of models and presenting interpretable results grounded in expert domain knowledge. To address these challenges, we design a simple but flexible model that preserves domain-based interpretability. We propose a novel approach that captures the semantic meanings directly from the text and compares them to symptom-related descriptions. Experimental results demonstrate that our model outperforms relevant baselines on various mental disorder detection tasks. Our detailed analysis shows that the proposed model is effective at leveraging domain knowledge, transferable to other mental disorders, and providing interpretable detection results.",
}
| Social media is one of the most highly sought resources for analyzing characteristics of the language by its users. In particular, many researchers utilized various linguistic features of mental health problems from social media. However, existing approaches to detecting mental disorders face critical challenges, such as the scarcity of high-quality data or the trade-off between addressing the complexity of models and presenting interpretable results grounded in expert domain knowledge. To address these challenges, we design a simple but flexible model that preserves domain-based interpretability. We propose a novel approach that captures the semantic meanings directly from the text and compares them to symptom-related descriptions. Experimental results demonstrate that our model outperforms relevant baselines on various mental disorder detection tasks. Our detailed analysis shows that the proposed model is effective at leveraging domain knowledge, transferable to other mental disorders, and providing interpretable detection results. | [
"Song, Hoyun",
"Shin, Jisu",
"Lee, Huije",
"Park, Jong"
] | A Simple and Flexible Modeling for Mental Disorder Detection by Learning from Clinical Questionnaires | acl-long.681 | Poster | 2306.02955 | [
"https://github.com/HoyunSong/acl23-multi-head-siamese-mental-illness"
] | -1 | -1 | -1 | -1 | 0 | [] | [] | [] |
|
https://aclanthology.org/2023.acl-long.682.bib | https://aclanthology.org/2023.acl-long.682/ | @inproceedings{krishna-etal-2023-downstream,
title = "Downstream Datasets Make Surprisingly Good Pretraining Corpora",
author = "Krishna, Kundan and
Garg, Saurabh and
Bigham, Jeffrey and
Lipton, Zachary",
editor = "Rogers, Anna and
Boyd-Graber, Jordan and
Okazaki, Naoaki",
booktitle = "Proceedings of the 61st Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers)",
month = jul,
year = "2023",
address = "Toronto, Canada",
publisher = "Association for Computational Linguistics",
url = "https://aclanthology.org/2023.acl-long.682",
doi = "10.18653/v1/2023.acl-long.682",
pages = "12207--12222",
abstract = "For most natural language processing tasks, the dominant practice is to finetune large pretrained transformer models (e.g., BERT) using smaller downstream datasets. Despite the success of this approach, it remains unclear to what extent these gainsare attributable to the massive background corpora employed for pretraining versus to the pretraining objectives themselves. This paper introduces a large-scale study of self-pretraining, where the same (downstream) training data is used for both pretraining and finetuning.In experiments addressing both ELECTRA and RoBERTa models and 10 distinct downstream classification datasets, we observe that self-pretraining rivals standard pretraining on the BookWiki corpus (despite using around 10x{--}500x less data), outperforming the latter on 7 and 5 datasets, respectively. Surprisingly, these task-specific pretrained models often perform well on other tasks,including the GLUE benchmark. Besides classification tasks, self-pretraining also provides benefits on structured output prediction tasks such as span based question answering and commonsense inference, often providing more than 50{\%} of the performance boosts provided by pretraining on the BookWiki corpus. Our results hint that in many scenarios, performance gains attributable to pretraining are driven primarily by the pretraining objective itself and are not always attributable to the use of external pretraining data in massive amounts. These findings are especially relevant in light of concerns about intellectual property and offensive content in web-scale pretraining data.",
}
| For most natural language processing tasks, the dominant practice is to finetune large pretrained transformer models (e.g., BERT) using smaller downstream datasets. Despite the success of this approach, it remains unclear to what extent these gainsare attributable to the massive background corpora employed for pretraining versus to the pretraining objectives themselves. This paper introduces a large-scale study of self-pretraining, where the same (downstream) training data is used for both pretraining and finetuning.In experiments addressing both ELECTRA and RoBERTa models and 10 distinct downstream classification datasets, we observe that self-pretraining rivals standard pretraining on the BookWiki corpus (despite using around 10x{--}500x less data), outperforming the latter on 7 and 5 datasets, respectively. Surprisingly, these task-specific pretrained models often perform well on other tasks,including the GLUE benchmark. Besides classification tasks, self-pretraining also provides benefits on structured output prediction tasks such as span based question answering and commonsense inference, often providing more than 50{\%} of the performance boosts provided by pretraining on the BookWiki corpus. Our results hint that in many scenarios, performance gains attributable to pretraining are driven primarily by the pretraining objective itself and are not always attributable to the use of external pretraining data in massive amounts. These findings are especially relevant in light of concerns about intellectual property and offensive content in web-scale pretraining data. | [
"Krishna, Kundan",
"Garg, Saurabh",
"Bigham, Jeffrey",
"Lipton, Zachary"
] | Downstream Datasets Make Surprisingly Good Pretraining Corpora | acl-long.682 | Poster | 2209.14389 | [
"https://github.com/acmi-lab/self-pretrain"
] | -1 | -1 | -1 | -1 | 0 | [] | [] | [] |
|
https://aclanthology.org/2023.acl-long.683.bib | https://aclanthology.org/2023.acl-long.683/ | @inproceedings{xu-etal-2023-towards,
title = "Towards Open-World Product Attribute Mining: A Lightly-Supervised Approach",
author = "Xu, Liyan and
Zhang, Chenwei and
Li, Xian and
Shang, Jingbo and
Choi, Jinho D.",
editor = "Rogers, Anna and
Boyd-Graber, Jordan and
Okazaki, Naoaki",
booktitle = "Proceedings of the 61st Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers)",
month = jul,
year = "2023",
address = "Toronto, Canada",
publisher = "Association for Computational Linguistics",
url = "https://aclanthology.org/2023.acl-long.683",
doi = "10.18653/v1/2023.acl-long.683",
pages = "12223--12239",
abstract = "We present a new task setting for attribute mining on e-commerce products, serving as a practical solution to extract open-world attributes without extensive human intervention. Our supervision comes from a high-quality seed attribute set bootstrapped from existing resources, and we aim to expand the attribute vocabulary of existing seed types, and also to discover any new attribute types automatically. A new dataset is created to support our setting, and our approach Amacer is proposed specifically to tackle the limited supervision. Especially, given that no direct supervision is available for those unseen new attributes, our novel formulation exploits self-supervised heuristic and unsupervised latent attributes, which attains implicit semantic signals as additional supervision by leveraging product context. Experiments suggest that our approach surpasses various baselines by 12 F1, expanding attributes of existing types significantly by up to 12 times, and discovering values from 39{\%} new types.",
}
| We present a new task setting for attribute mining on e-commerce products, serving as a practical solution to extract open-world attributes without extensive human intervention. Our supervision comes from a high-quality seed attribute set bootstrapped from existing resources, and we aim to expand the attribute vocabulary of existing seed types, and also to discover any new attribute types automatically. A new dataset is created to support our setting, and our approach Amacer is proposed specifically to tackle the limited supervision. Especially, given that no direct supervision is available for those unseen new attributes, our novel formulation exploits self-supervised heuristic and unsupervised latent attributes, which attains implicit semantic signals as additional supervision by leveraging product context. Experiments suggest that our approach surpasses various baselines by 12 F1, expanding attributes of existing types significantly by up to 12 times, and discovering values from 39{\%} new types. | [
"Xu, Liyan",
"Zhang, Chenwei",
"Li, Xian",
"Shang, Jingbo",
"Choi, Jinho D."
] | Towards Open-World Product Attribute Mining: A Lightly-Supervised Approach | acl-long.683 | Poster | 2305.18350 | [
"https://github.com/lxucs/woam"
] | -1 | -1 | -1 | -1 | 0 | [] | [] | [] |
|
https://aclanthology.org/2023.acl-long.684.bib | https://aclanthology.org/2023.acl-long.684/ | @inproceedings{liu-etal-2023-xdailydialog,
title = "{XD}aily{D}ialog: A Multilingual Parallel Dialogue Corpus",
author = "Liu, Zeming and
Nie, Ping and
Cai, Jie and
Wang, Haifeng and
Niu, Zheng-Yu and
Zhang, Peng and
Sachan, Mrinmaya and
Peng, Kaiping",
editor = "Rogers, Anna and
Boyd-Graber, Jordan and
Okazaki, Naoaki",
booktitle = "Proceedings of the 61st Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers)",
month = jul,
year = "2023",
address = "Toronto, Canada",
publisher = "Association for Computational Linguistics",
url = "https://aclanthology.org/2023.acl-long.684",
doi = "10.18653/v1/2023.acl-long.684",
pages = "12240--12253",
abstract = "High-quality datasets are significant to the development of dialogue models. However, most existing datasets for open-domain dialogue modeling are limited to a single language. The absence of multilingual open-domain dialog datasets not only limits the research on multilingual or cross-lingual transfer learning, but also hinders the development of robust open-domain dialog systems that can be deployed in other parts of the world. In this paper, we provide a multilingual parallel open-domain dialog dataset, XDailyDialog, to enable researchers to explore the challenging task of multilingual and cross-lingual open-domain dialog. XDailyDialog includes 13K dialogues aligned across 4 languages (52K dialogues and 410K utterances in total). We then propose a dialog generation model, kNN-Chat, which has a novel kNN-search mechanism to support unified response retrieval for monolingual, multilingual, and cross-lingual dialogue. Experiment results show the effectiveness of this framework. We will make XDailyDialog and kNN-Chat publicly available soon.",
}
| High-quality datasets are significant to the development of dialogue models. However, most existing datasets for open-domain dialogue modeling are limited to a single language. The absence of multilingual open-domain dialog datasets not only limits the research on multilingual or cross-lingual transfer learning, but also hinders the development of robust open-domain dialog systems that can be deployed in other parts of the world. In this paper, we provide a multilingual parallel open-domain dialog dataset, XDailyDialog, to enable researchers to explore the challenging task of multilingual and cross-lingual open-domain dialog. XDailyDialog includes 13K dialogues aligned across 4 languages (52K dialogues and 410K utterances in total). We then propose a dialog generation model, kNN-Chat, which has a novel kNN-search mechanism to support unified response retrieval for monolingual, multilingual, and cross-lingual dialogue. Experiment results show the effectiveness of this framework. We will make XDailyDialog and kNN-Chat publicly available soon. | [
"Liu, Zeming",
"Nie, Ping",
"Cai, Jie",
"Wang, Haifeng",
"Niu, Zheng-Yu",
"Zhang, Peng",
"Sachan, Mrinmaya",
"Peng, Kaiping"
] | XDailyDialog: A Multilingual Parallel Dialogue Corpus | acl-long.684 | Poster | [
""
] | -1 | -1 | -1 | -1 | 0 | [] | [] | [] |
||
https://aclanthology.org/2023.acl-long.685.bib | https://aclanthology.org/2023.acl-long.685/ | @inproceedings{mishra-etal-2023-pal,
title = "{PAL} to Lend a Helping Hand: Towards Building an Emotion Adaptive Polite and Empathetic Counseling Conversational Agent",
author = "Mishra, Kshitij and
Priya, Priyanshu and
Ekbal, Asif",
editor = "Rogers, Anna and
Boyd-Graber, Jordan and
Okazaki, Naoaki",
booktitle = "Proceedings of the 61st Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers)",
month = jul,
year = "2023",
address = "Toronto, Canada",
publisher = "Association for Computational Linguistics",
url = "https://aclanthology.org/2023.acl-long.685",
doi = "10.18653/v1/2023.acl-long.685",
pages = "12254--12271",
abstract = "The World Health Organization (WHO) has significantly emphasized the need for mental health care. The social stigma associated with mental illness prevents individuals from addressing their issues and getting assistance. In such a scenario, the relevance of online counseling has increased dramatically. The feelings and attitudes that a client and a counselor express towards each other result in a higher or lower counseling experience. A counselor should be friendly and gain clients{'} trust to make them share their problems comfortably. Thus, it is essential for the counselor to adequately comprehend the client{'}s emotions and ensure client{'}s welfare, i.e. s/he should adapt and deal with the clients politely and empathetically to provide a pleasant, cordial and personalized experience. Motivated by this, in this work, we attempt to build a novel Polite and empAthetic counseLing conversational agent PAL to lay down the counseling support to substance addict and crime victims. To have client{'}s emotion-based polite and empathetic responses, two counseling datasets laying down the counseling support to substance addicts and crime victims are annotated. These annotated datasets are used to build PAL in a reinforcement learning framework. A novel reward function is formulated to ensure correct politeness and empathy preferences as per client{'}s emotions with naturalness and non-repetitiveness in responses. Thorough automatic and human evaluation showcase the usefulness and strength of the designed novel reward function. Our proposed system is scalable and can be easily modified with different modules of preference models as per need.",
}
| The World Health Organization (WHO) has significantly emphasized the need for mental health care. The social stigma associated with mental illness prevents individuals from addressing their issues and getting assistance. In such a scenario, the relevance of online counseling has increased dramatically. The feelings and attitudes that a client and a counselor express towards each other result in a higher or lower counseling experience. A counselor should be friendly and gain clients{'} trust to make them share their problems comfortably. Thus, it is essential for the counselor to adequately comprehend the client{'}s emotions and ensure client{'}s welfare, i.e. s/he should adapt and deal with the clients politely and empathetically to provide a pleasant, cordial and personalized experience. Motivated by this, in this work, we attempt to build a novel Polite and empAthetic counseLing conversational agent PAL to lay down the counseling support to substance addict and crime victims. To have client{'}s emotion-based polite and empathetic responses, two counseling datasets laying down the counseling support to substance addicts and crime victims are annotated. These annotated datasets are used to build PAL in a reinforcement learning framework. A novel reward function is formulated to ensure correct politeness and empathy preferences as per client{'}s emotions with naturalness and non-repetitiveness in responses. Thorough automatic and human evaluation showcase the usefulness and strength of the designed novel reward function. Our proposed system is scalable and can be easily modified with different modules of preference models as per need. | [
"Mishra, Kshitij",
"Priya, Priyanshu",
"Ekbal, Asif"
] | PAL to Lend a Helping Hand: Towards Building an Emotion Adaptive Polite and Empathetic Counseling Conversational Agent | acl-long.685 | Poster | [
""
] | -1 | -1 | -1 | -1 | 0 | [] | [] | [] |
||
https://aclanthology.org/2023.acl-long.686.bib | https://aclanthology.org/2023.acl-long.686/ | @inproceedings{deng-etal-2023-bidirectional,
title = "Bidirectional Generative Framework for Cross-domain Aspect-based Sentiment Analysis",
author = "Deng, Yue and
Zhang, Wenxuan and
Pan, Sinno Jialin and
Bing, Lidong",
editor = "Rogers, Anna and
Boyd-Graber, Jordan and
Okazaki, Naoaki",
booktitle = "Proceedings of the 61st Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers)",
month = jul,
year = "2023",
address = "Toronto, Canada",
publisher = "Association for Computational Linguistics",
url = "https://aclanthology.org/2023.acl-long.686",
doi = "10.18653/v1/2023.acl-long.686",
pages = "12272--12285",
abstract = "Cross-domain aspect-based sentiment analysis (ABSA) aims to perform various fine-grained sentiment analysis tasks on a target domain by transferring knowledge from a source domain. Since labeled data only exists in the source domain, a model is expected to bridge the domain gap for tackling cross-domain ABSA. Though domain adaptation methods have proven to be effective, most of them are based on a discriminative model, which needs to be specifically designed for different ABSA tasks. To offer a more general solution, we propose a unified bidirectional generative framework to tackle various cross-domain ABSA tasks. Specifically, our framework trains a generative model in both text-to-label and label-to-text directions. The former transforms each task into a unified format to learn domain-agnostic features, and the latter generates natural sentences from noisy labels for data augmentation, with which a more accurate model can be trained. To investigate the effectiveness and generality of our framework, we conduct extensive experiments on four cross-domain ABSA tasks and present new state-of-the-art results on all tasks. Our data and code are publicly available at \url{https://github.com/DAMO-NLP-SG/BGCA}.",
}
| Cross-domain aspect-based sentiment analysis (ABSA) aims to perform various fine-grained sentiment analysis tasks on a target domain by transferring knowledge from a source domain. Since labeled data only exists in the source domain, a model is expected to bridge the domain gap for tackling cross-domain ABSA. Though domain adaptation methods have proven to be effective, most of them are based on a discriminative model, which needs to be specifically designed for different ABSA tasks. To offer a more general solution, we propose a unified bidirectional generative framework to tackle various cross-domain ABSA tasks. Specifically, our framework trains a generative model in both text-to-label and label-to-text directions. The former transforms each task into a unified format to learn domain-agnostic features, and the latter generates natural sentences from noisy labels for data augmentation, with which a more accurate model can be trained. To investigate the effectiveness and generality of our framework, we conduct extensive experiments on four cross-domain ABSA tasks and present new state-of-the-art results on all tasks. Our data and code are publicly available at \url{https://github.com/DAMO-NLP-SG/BGCA}. | [
"Deng, Yue",
"Zhang, Wenxuan",
"Pan, Sinno Jialin",
"Bing, Lidong"
] | Bidirectional Generative Framework for Cross-domain Aspect-based Sentiment Analysis | acl-long.686 | Poster | 2305.09509 | [
"https://github.com/damo-nlp-sg/bgca"
] | -1 | -1 | -1 | -1 | 0 | [] | [] | [] |
|
https://aclanthology.org/2023.acl-long.687.bib | https://aclanthology.org/2023.acl-long.687/ | @inproceedings{li-etal-2023-contrastive,
title = "Contrastive Decoding: Open-ended Text Generation as Optimization",
author = "Li, Xiang Lisa and
Holtzman, Ari and
Fried, Daniel and
Liang, Percy and
Eisner, Jason and
Hashimoto, Tatsunori and
Zettlemoyer, Luke and
Lewis, Mike",
editor = "Rogers, Anna and
Boyd-Graber, Jordan and
Okazaki, Naoaki",
booktitle = "Proceedings of the 61st Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers)",
month = jul,
year = "2023",
address = "Toronto, Canada",
publisher = "Association for Computational Linguistics",
url = "https://aclanthology.org/2023.acl-long.687",
doi = "10.18653/v1/2023.acl-long.687",
pages = "12286--12312",
abstract = "Given a language model (LM), maximum probability is a poor decoding objective for open-ended generation, because it produces short and repetitive text. On the other hand, sampling can often produce incoherent text that drifts from the original topics. We propose contrastive decoding (CD), a reliable decoding approach that optimizes a contrastive objective subject to a plausibility constraint. The contrastive objective returns the difference between the likelihood under a large LM (called the expert, e.g. OPT-13B) and a small LM (called the amateur, e.g. OPT-125M), and the constraint ensures that the outputs are plausible. CD is inspired by the fact that the failures of larger LMs (e.g., repetition, inco- herence) are even more prevalent in smaller LMs, and that this difference signals which texts should be preferred. CD requires zero additional training, and produces higher quality text than decoding from the larger LM alone. It also works across model scales (OPT-13B and GPT2-1.5B) and significantly outperforms four strong decoding algorithms (e.g., nucleus, top-k) in automatic and human evaluations across wikipedia, news and story domains.",
}
| Given a language model (LM), maximum probability is a poor decoding objective for open-ended generation, because it produces short and repetitive text. On the other hand, sampling can often produce incoherent text that drifts from the original topics. We propose contrastive decoding (CD), a reliable decoding approach that optimizes a contrastive objective subject to a plausibility constraint. The contrastive objective returns the difference between the likelihood under a large LM (called the expert, e.g. OPT-13B) and a small LM (called the amateur, e.g. OPT-125M), and the constraint ensures that the outputs are plausible. CD is inspired by the fact that the failures of larger LMs (e.g., repetition, inco- herence) are even more prevalent in smaller LMs, and that this difference signals which texts should be preferred. CD requires zero additional training, and produces higher quality text than decoding from the larger LM alone. It also works across model scales (OPT-13B and GPT2-1.5B) and significantly outperforms four strong decoding algorithms (e.g., nucleus, top-k) in automatic and human evaluations across wikipedia, news and story domains. | [
"Li, Xiang Lisa",
"Holtzman, Ari",
"Fried, Daniel",
"Liang, Percy",
"Eisner, Jason",
"Hashimoto, Tatsunori",
"Zettlemoyer, Luke",
"Lewis, Mike"
] | Contrastive Decoding: Open-ended Text Generation as Optimization | acl-long.687 | Poster | 2210.15097 | [
"https://github.com/xiangli1999/contrastivedecoding"
] | -1 | -1 | -1 | -1 | 0 | [] | [] | [] |
|
https://aclanthology.org/2023.acl-long.688.bib | https://aclanthology.org/2023.acl-long.688/ | @inproceedings{hosseini-etal-2023-resolving,
title = "Resolving Indirect Referring Expressions for Entity Selection",
author = "Hosseini, Mohammad Javad and
Radlinski, Filip and
Pareti, Silvia and
Louis, Annie",
editor = "Rogers, Anna and
Boyd-Graber, Jordan and
Okazaki, Naoaki",
booktitle = "Proceedings of the 61st Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers)",
month = jul,
year = "2023",
address = "Toronto, Canada",
publisher = "Association for Computational Linguistics",
url = "https://aclanthology.org/2023.acl-long.688",
doi = "10.18653/v1/2023.acl-long.688",
pages = "12313--12335",
abstract = "Recent advances in language modeling have enabled new conversational systems. In particular, it is often desirable for people to make choices among specified options when using such systems. We address the problem of reference resolution, when people use natural expressions to choose between real world entities. For example, given the choice {`}Should we make a Simnel cake or a Pandan cake?` a natural response from a non-expert may be indirect: {`}let{'}s make the green one{`}. Reference resolution has been little studied with natural expressions, thus robustly understanding such language has large potential for improving naturalness in dialog, recommendation, and search systems. We create AltEntities (Alternative Entities), a new public dataset of entity pairs and utterances, and develop models for the disambiguation problem. Consisting of 42K indirect referring expressions across three domains, it enables for the first time the study of how large language models can be adapted to this task. We find they achieve 82{\%}-87{\%} accuracy in realistic settings, which while reasonable also invites further advances.",
}
| Recent advances in language modeling have enabled new conversational systems. In particular, it is often desirable for people to make choices among specified options when using such systems. We address the problem of reference resolution, when people use natural expressions to choose between real world entities. For example, given the choice {`}Should we make a Simnel cake or a Pandan cake?` a natural response from a non-expert may be indirect: {`}let{'}s make the green one{`}. Reference resolution has been little studied with natural expressions, thus robustly understanding such language has large potential for improving naturalness in dialog, recommendation, and search systems. We create AltEntities (Alternative Entities), a new public dataset of entity pairs and utterances, and develop models for the disambiguation problem. Consisting of 42K indirect referring expressions across three domains, it enables for the first time the study of how large language models can be adapted to this task. We find they achieve 82{\%}-87{\%} accuracy in realistic settings, which while reasonable also invites further advances. | [
"Hosseini, Mohammad Javad",
"Radlinski, Filip",
"Pareti, Silvia",
"Louis, Annie"
] | Resolving Indirect Referring Expressions for Entity Selection | acl-long.688 | Poster | 2212.10933 | [
"https://github.com/google-research-datasets/altentities"
] | -1 | -1 | -1 | -1 | 0 | [] | [] | [] |
|
https://aclanthology.org/2023.acl-long.689.bib | https://aclanthology.org/2023.acl-long.689/ | @inproceedings{santilli-etal-2023-accelerating,
title = "Accelerating Transformer Inference for Translation via Parallel Decoding",
author = "Santilli, Andrea and
Severino, Silvio and
Postolache, Emilian and
Maiorca, Valentino and
Mancusi, Michele and
Marin, Riccardo and
Rodola, Emanuele",
editor = "Rogers, Anna and
Boyd-Graber, Jordan and
Okazaki, Naoaki",
booktitle = "Proceedings of the 61st Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers)",
month = jul,
year = "2023",
address = "Toronto, Canada",
publisher = "Association for Computational Linguistics",
url = "https://aclanthology.org/2023.acl-long.689",
doi = "10.18653/v1/2023.acl-long.689",
pages = "12336--12355",
abstract = "Autoregressive decoding limits the efficiency of transformers for Machine Translation (MT). The community proposed specific network architectures and learning-based methods to solve this issue, which are expensive and require changes to the MT model, trading inference speed at the cost of the translation quality. In this paper, we propose to address the problem from the point of view of decoding algorithms, as a less explored but rather compelling direction. We propose to reframe the standard greedy autoregressive decoding of MT with a parallel formulation leveraging Jacobi and Gauss-Seidel fixed-point iteration methods for fast inference. This formulation allows to speed up existing models without training or modifications while retaining translation quality. We present three parallel decoding algorithms and test them on different languages and models showing how the parallelization introduces a speedup up to 38{\%} w.r.t. the standard autoregressive decoding and nearly 2x when scaling the method on parallel resources. Finally, we introduce a decoding dependency graph visualizer (DDGviz) that let us see how the model has learned the conditional dependence between tokens and inspect the decoding procedure.",
}
| Autoregressive decoding limits the efficiency of transformers for Machine Translation (MT). The community proposed specific network architectures and learning-based methods to solve this issue, which are expensive and require changes to the MT model, trading inference speed at the cost of the translation quality. In this paper, we propose to address the problem from the point of view of decoding algorithms, as a less explored but rather compelling direction. We propose to reframe the standard greedy autoregressive decoding of MT with a parallel formulation leveraging Jacobi and Gauss-Seidel fixed-point iteration methods for fast inference. This formulation allows to speed up existing models without training or modifications while retaining translation quality. We present three parallel decoding algorithms and test them on different languages and models showing how the parallelization introduces a speedup up to 38{\%} w.r.t. the standard autoregressive decoding and nearly 2x when scaling the method on parallel resources. Finally, we introduce a decoding dependency graph visualizer (DDGviz) that let us see how the model has learned the conditional dependence between tokens and inspect the decoding procedure. | [
"Santilli, Andrea",
"Severino, Silvio",
"Postolache, Emilian",
"Maiorca, Valentino",
"Mancusi, Michele",
"Marin, Riccardo",
"Rodola, Emanuele"
] | Accelerating Transformer Inference for Translation via Parallel Decoding | acl-long.689 | Poster | 2305.10427 | [
"https://github.com/teelinsan/parallel-decoding"
] | https://huggingface.co/papers/2305.10427 | 2 | 0 | 0 | 7 | 1 | [] | [] | [] |
https://aclanthology.org/2023.acl-long.690.bib | https://aclanthology.org/2023.acl-long.690/ | @inproceedings{xu-etal-2023-hard,
title = "Hard Sample Aware Prompt-Tuning",
author = "Xu, Yuanjian and
An, Qi and
Zhang, Jiahuan and
Li, Peng and
Nie, Zaiqing",
editor = "Rogers, Anna and
Boyd-Graber, Jordan and
Okazaki, Naoaki",
booktitle = "Proceedings of the 61st Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers)",
month = jul,
year = "2023",
address = "Toronto, Canada",
publisher = "Association for Computational Linguistics",
url = "https://aclanthology.org/2023.acl-long.690",
doi = "10.18653/v1/2023.acl-long.690",
pages = "12356--12369",
abstract = "Prompt-tuning based few-shot learning has garnered increasing attention in recent years due to its efficiency and promising capability. To achieve the best performance for NLP tasks with just a few samples, it is vital to include as many informative samples as possible and to avoid misleading ones. However, there is no work in prompt-tuning literature addressing the problem of differentiating informative hard samples from misleading ones in model training, which is challenging due to the lack of supervision signals about the quality of the samples to train a well-performed model. We propose a Hard Sample Aware Prompt-Tuning framework (i.e. HardPT) to solve the non-differentiable problem in hard sample identification with reinforcement learning, and to strengthen the discrimination of the feature space without changing the original data distribution via an adaptive contrastive learning method. An extensive empirical study on a series of NLP tasks demonstrates the capability of HardPT in few-shot scenarios. HardPT obtains new SOTA results on all evaluated NLP tasks, including pushing the SST-5 accuracy to 49.5{\%} (1.1{\%} point absolute improvement), QNLI accuracy to 74.6{\%} (1.9{\%} absolute improvement), NMLI accuracy to 71.5 (0.7{\%} absolute improvement), TACREV $F_1$-score to 28.2 (1.0 absolute improvement), and i2b2/VA $F_1$-score to 41.2 (1.3 absolute improvement).",
}
| Prompt-tuning based few-shot learning has garnered increasing attention in recent years due to its efficiency and promising capability. To achieve the best performance for NLP tasks with just a few samples, it is vital to include as many informative samples as possible and to avoid misleading ones. However, there is no work in prompt-tuning literature addressing the problem of differentiating informative hard samples from misleading ones in model training, which is challenging due to the lack of supervision signals about the quality of the samples to train a well-performed model. We propose a Hard Sample Aware Prompt-Tuning framework (i.e. HardPT) to solve the non-differentiable problem in hard sample identification with reinforcement learning, and to strengthen the discrimination of the feature space without changing the original data distribution via an adaptive contrastive learning method. An extensive empirical study on a series of NLP tasks demonstrates the capability of HardPT in few-shot scenarios. HardPT obtains new SOTA results on all evaluated NLP tasks, including pushing the SST-5 accuracy to 49.5{\%} (1.1{\%} point absolute improvement), QNLI accuracy to 74.6{\%} (1.9{\%} absolute improvement), NMLI accuracy to 71.5 (0.7{\%} absolute improvement), TACREV $F_1$-score to 28.2 (1.0 absolute improvement), and i2b2/VA $F_1$-score to 41.2 (1.3 absolute improvement). | [
"Xu, Yuanjian",
"An, Qi",
"Zhang, Jiahuan",
"Li, Peng",
"Nie, Zaiqing"
] | Hard Sample Aware Prompt-Tuning | acl-long.690 | Poster | [
""
] | -1 | -1 | -1 | -1 | 0 | [] | [] | [] |
||
https://aclanthology.org/2023.acl-long.691.bib | https://aclanthology.org/2023.acl-long.691/ | @inproceedings{stranisci-etal-2023-wikibio,
title = "{W}iki{B}io: a Semantic Resource for the Intersectional Analysis of Biographical Events",
author = "Stranisci, Marco Antonio and
Damiano, Rossana and
Mensa, Enrico and
Patti, Viviana and
Radicioni, Daniele and
Caselli, Tommaso",
editor = "Rogers, Anna and
Boyd-Graber, Jordan and
Okazaki, Naoaki",
booktitle = "Proceedings of the 61st Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers)",
month = jul,
year = "2023",
address = "Toronto, Canada",
publisher = "Association for Computational Linguistics",
url = "https://aclanthology.org/2023.acl-long.691",
doi = "10.18653/v1/2023.acl-long.691",
pages = "12370--12384",
abstract = "Biographical event detection is a relevant task that allows for the exploration and comparison of the ways in which people{'}s lives are told and represented. This may support several real-life applications in digital humanities and in works aimed at exploring bias about minoritized groups. Despite that, there are no corpora and models specifically designed for this task. In this paper we fill this gap by presenting a new corpus annotated for biographical event detection. The corpus, which includes 20 Wikipedia biographies, was aligned with 5 existing corpora in order to train a model for the biographical event detection task. The model was able to detect all mentions of the target-entity in a biography with an F-score of 0.808 and the entity-related events with an F-score of 0.859. Finally, the model was used for performing an analysis of biases about women and non-Western people in Wikipedia biographies.",
}
| Biographical event detection is a relevant task that allows for the exploration and comparison of the ways in which people{'}s lives are told and represented. This may support several real-life applications in digital humanities and in works aimed at exploring bias about minoritized groups. Despite that, there are no corpora and models specifically designed for this task. In this paper we fill this gap by presenting a new corpus annotated for biographical event detection. The corpus, which includes 20 Wikipedia biographies, was aligned with 5 existing corpora in order to train a model for the biographical event detection task. The model was able to detect all mentions of the target-entity in a biography with an F-score of 0.808 and the entity-related events with an F-score of 0.859. Finally, the model was used for performing an analysis of biases about women and non-Western people in Wikipedia biographies. | [
"Stranisci, Marco Antonio",
"Damiano, Rossana",
"Mensa, Enrico",
"Patti, Viviana",
"Radicioni, Daniele",
"Caselli, Tommaso"
] | WikiBio: a Semantic Resource for the Intersectional Analysis of Biographical Events | acl-long.691 | Poster | 2306.09505 | [
"https://github.com/marcostranisci/wikibio"
] | -1 | -1 | -1 | -1 | 0 | [] | [] | [] |
|
https://aclanthology.org/2023.acl-long.692.bib | https://aclanthology.org/2023.acl-long.692/ | @inproceedings{xu-etal-2023-best,
title = "Best-k Search Algorithm for Neural Text Generation",
author = "Xu, Jiacheng and
Xiong, Caiming and
Savarese, Silvio and
Zhou, Yingbo",
editor = "Rogers, Anna and
Boyd-Graber, Jordan and
Okazaki, Naoaki",
booktitle = "Proceedings of the 61st Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers)",
month = jul,
year = "2023",
address = "Toronto, Canada",
publisher = "Association for Computational Linguistics",
url = "https://aclanthology.org/2023.acl-long.692",
doi = "10.18653/v1/2023.acl-long.692",
pages = "12385--12401",
abstract = "Modern natural language generation paradigms require a decoding strategy to obtain quality sequences out of the model. Beam search yields high-quality but low diversity outputs; stochastic approaches suffer from high variance and sometimes low quality. In this work, we propose a deterministic search algorithm balancing both quality and diversity. We first investigate the vanilla best-first search (BFS) algorithm and then propose the best-k search algorithm. Inspired by BFS, we greedily expand the top k nodes, instead of the first node, to boost efficiency and diversity. Upweighting recently discovered nodes accompanied by heap pruning ensures the completeness of the search procedure. Experiments on four NLG tasks show that best-k search yields more diverse and natural outputs compared to strong baselines, while our approach maintains high text quality. The proposed algorithm is parameter-free, lightweight, efficient, and easy-to-use.",
}
| Modern natural language generation paradigms require a decoding strategy to obtain quality sequences out of the model. Beam search yields high-quality but low diversity outputs; stochastic approaches suffer from high variance and sometimes low quality. In this work, we propose a deterministic search algorithm balancing both quality and diversity. We first investigate the vanilla best-first search (BFS) algorithm and then propose the best-k search algorithm. Inspired by BFS, we greedily expand the top k nodes, instead of the first node, to boost efficiency and diversity. Upweighting recently discovered nodes accompanied by heap pruning ensures the completeness of the search procedure. Experiments on four NLG tasks show that best-k search yields more diverse and natural outputs compared to strong baselines, while our approach maintains high text quality. The proposed algorithm is parameter-free, lightweight, efficient, and easy-to-use. | [
"Xu, Jiacheng",
"Xiong, Caiming",
"Savarese, Silvio",
"Zhou, Yingbo"
] | Best-k Search Algorithm for Neural Text Generation | acl-long.692 | Poster | [
""
] | -1 | -1 | -1 | -1 | 0 | [] | [] | [] |
||
https://aclanthology.org/2023.acl-long.693.bib | https://aclanthology.org/2023.acl-long.693/ | @inproceedings{doddapaneni-etal-2023-towards,
title = "Towards Leaving No {I}ndic Language Behind: Building Monolingual Corpora, Benchmark and Models for {I}ndic Languages",
author = "Doddapaneni, Sumanth and
Aralikatte, Rahul and
Ramesh, Gowtham and
Goyal, Shreya and
Khapra, Mitesh M. and
Kunchukuttan, Anoop and
Kumar, Pratyush",
editor = "Rogers, Anna and
Boyd-Graber, Jordan and
Okazaki, Naoaki",
booktitle = "Proceedings of the 61st Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers)",
month = jul,
year = "2023",
address = "Toronto, Canada",
publisher = "Association for Computational Linguistics",
url = "https://aclanthology.org/2023.acl-long.693",
doi = "10.18653/v1/2023.acl-long.693",
pages = "12402--12426",
abstract = "Building Natural Language Understanding (NLU) capabilities for Indic languages, which have a collective speaker base of more than one billion speakers is absolutely crucial. In this work, we aim to improve the NLU capabilities of Indic languages by making contributions along 3 important axes (i) monolingual corpora (ii) NLU testsets (iii) multilingual LLMs focusing on Indic languages. Specifically, we curate the largest monolingual corpora, IndicCorp, with 20.9B tokens covering 24 languages from 4 language families - a 2.3x increase over prior work, while supporting 12 additional languages. Next, we create a human-supervised benchmark, IndicXTREME, consisting of nine diverse NLU tasks covering 20 languages. Across languages and tasks, IndicXTREME contains a total of 105 evaluation sets, of which 52 are new contributions to the literature. To the best of our knowledge, this is the first effort towards creating a standard benchmark for Indic languages that aims to test the multilingual zero-shot capabilities of pretrained language models. Finally, we train IndicBERT v2, a state-of-the-art model supporting all the languages. Averaged across languages and tasks, the model achieves an absolute improvement of 2 points over a strong baseline. The data and models are available at \url{https://github.com/AI4Bharat/IndicBERT}.",
}
| Building Natural Language Understanding (NLU) capabilities for Indic languages, which have a collective speaker base of more than one billion speakers is absolutely crucial. In this work, we aim to improve the NLU capabilities of Indic languages by making contributions along 3 important axes (i) monolingual corpora (ii) NLU testsets (iii) multilingual LLMs focusing on Indic languages. Specifically, we curate the largest monolingual corpora, IndicCorp, with 20.9B tokens covering 24 languages from 4 language families - a 2.3x increase over prior work, while supporting 12 additional languages. Next, we create a human-supervised benchmark, IndicXTREME, consisting of nine diverse NLU tasks covering 20 languages. Across languages and tasks, IndicXTREME contains a total of 105 evaluation sets, of which 52 are new contributions to the literature. To the best of our knowledge, this is the first effort towards creating a standard benchmark for Indic languages that aims to test the multilingual zero-shot capabilities of pretrained language models. Finally, we train IndicBERT v2, a state-of-the-art model supporting all the languages. Averaged across languages and tasks, the model achieves an absolute improvement of 2 points over a strong baseline. The data and models are available at \url{https://github.com/AI4Bharat/IndicBERT}. | [
"Doddapaneni, Sumanth",
"Aralikatte, Rahul",
"Ramesh, Gowtham",
"Goyal, Shreya",
"Khapra, Mitesh M.",
"Kunchukuttan, Anoop",
"Kumar, Pratyush"
] | Towards Leaving No Indic Language Behind: Building Monolingual Corpora, Benchmark and Models for Indic Languages | acl-long.693 | Poster | 2212.05409 | [
"https://github.com/ai4bharat/indicbert"
] | https://huggingface.co/papers/2212.05409 | 0 | 0 | 0 | 7 | 1 | [] | [
"satpalsr/indicCorpv2"
] | [] |
https://aclanthology.org/2023.acl-long.694.bib | https://aclanthology.org/2023.acl-long.694/ | @inproceedings{yang-etal-2023-transforming,
title = "Transforming Visual Scene Graphs to Image Captions",
author = "Yang, Xu and
Peng, Jiawei and
Wang, Zihua and
Xu, Haiyang and
Ye, Qinghao and
Li, Chenliang and
Huang, Songfang and
Huang, Fei and
Li, Zhangzikang and
Zhang, Yu",
editor = "Rogers, Anna and
Boyd-Graber, Jordan and
Okazaki, Naoaki",
booktitle = "Proceedings of the 61st Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers)",
month = jul,
year = "2023",
address = "Toronto, Canada",
publisher = "Association for Computational Linguistics",
url = "https://aclanthology.org/2023.acl-long.694",
doi = "10.18653/v1/2023.acl-long.694",
pages = "12427--12440",
abstract = "We propose to TransForm Scene Graphs into more descriptive Captions (TFSGC). In TFSGC, we apply multi-head attention (MHA) to design the Graph Neural Network (GNN) for embedding scene graphs. After embedding, different graph embeddings contain diverse specific knowledge for generating the words with different part-of-speech, e.g., object/attribute embedding is good for generating nouns/adjectives. Motivated by this, we design a Mixture-of-Expert (MOE)-based decoder, where each expert is built on MHA, for discriminating the graph embeddings to generate different kinds of words. Since both the encoder and decoder are built based on the MHA, as a result, we construct a simple and homogeneous encoder-decoder unlike the previous heterogeneous ones which usually apply Fully-Connected-based GNN and LSTM-based decoder. The homogeneous architecture enables us to unify the training configuration of the whole model instead of specifying different training strategies for diverse sub-networks as in the heterogeneous pipeline, which releases the training difficulty. Extensive experiments on the MS-COCO captioning benchmark validate the effectiveness of our TFSGC. The code is in: \url{https://anonymous.4open.science/r/ACL23_TFSGC}.",
}
| We propose to TransForm Scene Graphs into more descriptive Captions (TFSGC). In TFSGC, we apply multi-head attention (MHA) to design the Graph Neural Network (GNN) for embedding scene graphs. After embedding, different graph embeddings contain diverse specific knowledge for generating the words with different part-of-speech, e.g., object/attribute embedding is good for generating nouns/adjectives. Motivated by this, we design a Mixture-of-Expert (MOE)-based decoder, where each expert is built on MHA, for discriminating the graph embeddings to generate different kinds of words. Since both the encoder and decoder are built based on the MHA, as a result, we construct a simple and homogeneous encoder-decoder unlike the previous heterogeneous ones which usually apply Fully-Connected-based GNN and LSTM-based decoder. The homogeneous architecture enables us to unify the training configuration of the whole model instead of specifying different training strategies for diverse sub-networks as in the heterogeneous pipeline, which releases the training difficulty. Extensive experiments on the MS-COCO captioning benchmark validate the effectiveness of our TFSGC. The code is in: \url{https://anonymous.4open.science/r/ACL23_TFSGC}. | [
"Yang, Xu",
"Peng, Jiawei",
"Wang, Zihua",
"Xu, Haiyang",
"Ye, Qinghao",
"Li, Chenliang",
"Huang, Songfang",
"Huang, Fei",
"Li, Zhangzikang",
"Zhang, Yu"
] | Transforming Visual Scene Graphs to Image Captions | acl-long.694 | Poster | 2305.02177 | [
"https://github.com/garyjiajia/tsg"
] | -1 | -1 | -1 | -1 | 0 | [] | [] | [] |
|
https://aclanthology.org/2023.acl-long.695.bib | https://aclanthology.org/2023.acl-long.695/ | @inproceedings{tang-etal-2023-hybrid,
title = "Hybrid Transducer and Attention based Encoder-Decoder Modeling for Speech-to-Text Tasks",
author = "Tang, Yun and
Sun, Anna and
Inaguma, Hirofumi and
Chen, Xinyue and
Dong, Ning and
Ma, Xutai and
Tomasello, Paden and
Pino, Juan",
editor = "Rogers, Anna and
Boyd-Graber, Jordan and
Okazaki, Naoaki",
booktitle = "Proceedings of the 61st Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers)",
month = jul,
year = "2023",
address = "Toronto, Canada",
publisher = "Association for Computational Linguistics",
url = "https://aclanthology.org/2023.acl-long.695",
doi = "10.18653/v1/2023.acl-long.695",
pages = "12441--12455",
abstract = "Transducer and Attention based Encoder-Decoder (AED) are two widely used frameworks for speech-to-text tasks. They are designed for different purposes and each has its own benefits and drawbacks for speech-to-text tasks. In order to leverage strengths of both modeling methods, we propose a solution by combining Transducer and Attention based Encoder-Decoder (TAED) for speech-to-text tasks. The new method leverages AED{'}s strength in non-monotonic sequence to sequence learning while retaining Transducer{'}s streaming property. In the proposed framework, Transducer and AED share the same speech encoder. The predictor in Transducer is replaced by the decoder in the AED model, and the outputs of the decoder are conditioned on the speech inputs instead of outputs from an unconditioned language model. The proposed solution ensures that the model is optimized by covering all possible read/write scenarios and creates a matched environment for streaming applications. We evaluate the proposed approach on the MuST-C dataset and the findings demonstrate that TAED performs significantly better than Transducer for offline automatic speech recognition (ASR) and speech-to-text translation (ST) tasks. In the streaming case, TAED outperforms Transducer in the ASR task and one ST direction while comparable results are achieved in another translation direction.",
}
| Transducer and Attention based Encoder-Decoder (AED) are two widely used frameworks for speech-to-text tasks. They are designed for different purposes and each has its own benefits and drawbacks for speech-to-text tasks. In order to leverage strengths of both modeling methods, we propose a solution by combining Transducer and Attention based Encoder-Decoder (TAED) for speech-to-text tasks. The new method leverages AED{'}s strength in non-monotonic sequence to sequence learning while retaining Transducer{'}s streaming property. In the proposed framework, Transducer and AED share the same speech encoder. The predictor in Transducer is replaced by the decoder in the AED model, and the outputs of the decoder are conditioned on the speech inputs instead of outputs from an unconditioned language model. The proposed solution ensures that the model is optimized by covering all possible read/write scenarios and creates a matched environment for streaming applications. We evaluate the proposed approach on the MuST-C dataset and the findings demonstrate that TAED performs significantly better than Transducer for offline automatic speech recognition (ASR) and speech-to-text translation (ST) tasks. In the streaming case, TAED outperforms Transducer in the ASR task and one ST direction while comparable results are achieved in another translation direction. | [
"Tang, Yun",
"Sun, Anna",
"Inaguma, Hirofumi",
"Chen, Xinyue",
"Dong, Ning",
"Ma, Xutai",
"Tomasello, Paden",
"Pino, Juan"
] | Hybrid Transducer and Attention based Encoder-Decoder Modeling for Speech-to-Text Tasks | acl-long.695 | Poster | 2305.03101 | [
""
] | https://huggingface.co/papers/2305.03101 | 1 | 0 | 0 | 8 | 1 | [] | [] | [] |
https://aclanthology.org/2023.acl-long.696.bib | https://aclanthology.org/2023.acl-long.696/ | @inproceedings{jiang-etal-2023-improving,
title = "Improving Domain Generalization for Prompt-Aware Essay Scoring via Disentangled Representation Learning",
author = "Jiang, Zhiwei and
Gao, Tianyi and
Yin, Yafeng and
Liu, Meng and
Yu, Hua and
Cheng, Zifeng and
Gu, Qing",
editor = "Rogers, Anna and
Boyd-Graber, Jordan and
Okazaki, Naoaki",
booktitle = "Proceedings of the 61st Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers)",
month = jul,
year = "2023",
address = "Toronto, Canada",
publisher = "Association for Computational Linguistics",
url = "https://aclanthology.org/2023.acl-long.696",
doi = "10.18653/v1/2023.acl-long.696",
pages = "12456--12470",
abstract = "Automated Essay Scoring (AES) aims to score essays written in response to specific prompts. Many AES models have been proposed, but most of them are either prompt-specific or prompt-adaptive and cannot generalize well on {``}unseen{''} prompts. This work focuses on improving the generalization ability of AES models from the perspective of domain generalization, where the data of target prompts cannot be accessed during training. Specifically, we propose a prompt-aware neural AES model to extract comprehensive representation for essay scoring, including both prompt-invariant and prompt-specific features. To improve the generalization of representation, we further propose a novel disentangled representation learning framework. In this framework, a contrastive norm-angular alignment strategy and a counterfactual self-training strategy are designed to disentangle the prompt-invariant information and prompt-specific information in representation. Extensive experimental results on datasets of both ASAP and TOEFL11 demonstrate the effectiveness of our method under the domain generalization setting.",
}
| Automated Essay Scoring (AES) aims to score essays written in response to specific prompts. Many AES models have been proposed, but most of them are either prompt-specific or prompt-adaptive and cannot generalize well on {``}unseen{''} prompts. This work focuses on improving the generalization ability of AES models from the perspective of domain generalization, where the data of target prompts cannot be accessed during training. Specifically, we propose a prompt-aware neural AES model to extract comprehensive representation for essay scoring, including both prompt-invariant and prompt-specific features. To improve the generalization of representation, we further propose a novel disentangled representation learning framework. In this framework, a contrastive norm-angular alignment strategy and a counterfactual self-training strategy are designed to disentangle the prompt-invariant information and prompt-specific information in representation. Extensive experimental results on datasets of both ASAP and TOEFL11 demonstrate the effectiveness of our method under the domain generalization setting. | [
"Jiang, Zhiwei",
"Gao, Tianyi",
"Yin, Yafeng",
"Liu, Meng",
"Yu, Hua",
"Cheng, Zifeng",
"Gu, Qing"
] | Improving Domain Generalization for Prompt-Aware Essay Scoring via Disentangled Representation Learning | acl-long.696 | Poster | [
""
] | -1 | -1 | -1 | -1 | 0 | [] | [] | [] |
||
https://aclanthology.org/2023.acl-long.697.bib | https://aclanthology.org/2023.acl-long.697/ | @inproceedings{tedeschi-etal-2023-whats,
title = "What{'}s the Meaning of Superhuman Performance in Today{'}s {NLU}?",
author = "Tedeschi, Simone and
Bos, Johan and
Declerck, Thierry and
Haji{\v{c}}, Jan and
Hershcovich, Daniel and
Hovy, Eduard and
Koller, Alexander and
Krek, Simon and
Schockaert, Steven and
Sennrich, Rico and
Shutova, Ekaterina and
Navigli, Roberto",
editor = "Rogers, Anna and
Boyd-Graber, Jordan and
Okazaki, Naoaki",
booktitle = "Proceedings of the 61st Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers)",
month = jul,
year = "2023",
address = "Toronto, Canada",
publisher = "Association for Computational Linguistics",
url = "https://aclanthology.org/2023.acl-long.697",
doi = "10.18653/v1/2023.acl-long.697",
pages = "12471--12491",
abstract = "In the last five years, there has been a significant focus in Natural Language Processing (NLP) on developing larger Pretrained Language Models (PLMs) and introducing benchmarks such as SuperGLUE and SQuAD to measure their abilities in language understanding, reasoning, and reading comprehension. These PLMs have achieved impressive results on these benchmarks, even surpassing human performance in some cases. This has led to claims of superhuman capabilities and the provocative idea that certain tasks have been solved. In this position paper, we take a critical look at these claims and ask whether PLMs truly have superhuman abilities and what the current benchmarks are really evaluating. We show that these benchmarks have serious limitations affecting the comparison between humans and PLMs and provide recommendations for fairer and more transparent benchmarks.",
}
| In the last five years, there has been a significant focus in Natural Language Processing (NLP) on developing larger Pretrained Language Models (PLMs) and introducing benchmarks such as SuperGLUE and SQuAD to measure their abilities in language understanding, reasoning, and reading comprehension. These PLMs have achieved impressive results on these benchmarks, even surpassing human performance in some cases. This has led to claims of superhuman capabilities and the provocative idea that certain tasks have been solved. In this position paper, we take a critical look at these claims and ask whether PLMs truly have superhuman abilities and what the current benchmarks are really evaluating. We show that these benchmarks have serious limitations affecting the comparison between humans and PLMs and provide recommendations for fairer and more transparent benchmarks. | [
"Tedeschi, Simone",
"Bos, Johan",
"Declerck, Thierry",
"Haji{\\v{c}}, Jan",
"Hershcovich, Daniel",
"Hovy, Eduard",
"Koller, Alex",
"er",
"Krek, Simon",
"Schockaert, Steven",
"Sennrich, Rico",
"Shutova, Ekaterina",
"Navigli, Roberto"
] | What's the Meaning of Superhuman Performance in Today's NLU? | acl-long.697 | Oral | 2305.08414 | [
""
] | https://huggingface.co/papers/2305.08414 | 2 | 1 | 0 | 12 | 1 | [] | [] | [] |
https://aclanthology.org/2023.acl-long.698.bib | https://aclanthology.org/2023.acl-long.698/ | @inproceedings{shen-etal-2023-promptner,
title = "{P}rompt{NER}: Prompt Locating and Typing for Named Entity Recognition",
author = "Shen, Yongliang and
Tan, Zeqi and
Wu, Shuhui and
Zhang, Wenqi and
Zhang, Rongsheng and
Xi, Yadong and
Lu, Weiming and
Zhuang, Yueting",
editor = "Rogers, Anna and
Boyd-Graber, Jordan and
Okazaki, Naoaki",
booktitle = "Proceedings of the 61st Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers)",
month = jul,
year = "2023",
address = "Toronto, Canada",
publisher = "Association for Computational Linguistics",
url = "https://aclanthology.org/2023.acl-long.698",
doi = "10.18653/v1/2023.acl-long.698",
pages = "12492--12507",
abstract = "Prompt learning is a new paradigm for utilizing pre-trained language models and has achieved great success in many tasks. To adopt prompt learning in the NER task, two kinds of methods have been explored from a pair of symmetric perspectives, populating the template by enumerating spans to predict their entity types or constructing type-specific prompts to locate entities. However, these methods not only require a multi-round prompting manner with a high time overhead and computational cost, but also require elaborate prompt templates, that are difficult to apply in practical scenarios. In this paper, we unify entity locating and entity typing into prompt learning, and design a dual-slot multi-prompt template with the position slot and type slot to prompt locating and typing respectively. Multiple prompts can be input to the model simultaneously, and then the model extracts all entities by parallel predictions on the slots. To assign labels for the slots during training, we design a dynamic template filling mechanism that uses the extended bipartite graph matching between prompts and the ground-truth entities. We conduct experiments in various settings, including resource-rich flat and nested NER datasets and low-resource in-domain and cross-domain datasets. Experimental results show that the proposed model achieves a significant performance improvement, especially in the cross-domain few-shot setting, which outperforms the state-of-the-art model by +7.7{\%} on average.",
}
| Prompt learning is a new paradigm for utilizing pre-trained language models and has achieved great success in many tasks. To adopt prompt learning in the NER task, two kinds of methods have been explored from a pair of symmetric perspectives, populating the template by enumerating spans to predict their entity types or constructing type-specific prompts to locate entities. However, these methods not only require a multi-round prompting manner with a high time overhead and computational cost, but also require elaborate prompt templates, that are difficult to apply in practical scenarios. In this paper, we unify entity locating and entity typing into prompt learning, and design a dual-slot multi-prompt template with the position slot and type slot to prompt locating and typing respectively. Multiple prompts can be input to the model simultaneously, and then the model extracts all entities by parallel predictions on the slots. To assign labels for the slots during training, we design a dynamic template filling mechanism that uses the extended bipartite graph matching between prompts and the ground-truth entities. We conduct experiments in various settings, including resource-rich flat and nested NER datasets and low-resource in-domain and cross-domain datasets. Experimental results show that the proposed model achieves a significant performance improvement, especially in the cross-domain few-shot setting, which outperforms the state-of-the-art model by +7.7{\%} on average. | [
"Shen, Yongliang",
"Tan, Zeqi",
"Wu, Shuhui",
"Zhang, Wenqi",
"Zhang, Rongsheng",
"Xi, Yadong",
"Lu, Weiming",
"Zhuang, Yueting"
] | PromptNER: Prompt Locating and Typing for Named Entity Recognition | acl-long.698 | Poster | 2305.17104 | [
"https://github.com/tricktreat/promptner"
] | -1 | -1 | -1 | -1 | 0 | [] | [] | [] |
|
https://aclanthology.org/2023.acl-long.699.bib | https://aclanthology.org/2023.acl-long.699/ | @inproceedings{zevallos-bel-2023-hints,
title = "Hints on the data for language modeling of synthetic languages with transformers",
author = "Zevallos, Rodolfo and
Bel, Nuria",
editor = "Rogers, Anna and
Boyd-Graber, Jordan and
Okazaki, Naoaki",
booktitle = "Proceedings of the 61st Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers)",
month = jul,
year = "2023",
address = "Toronto, Canada",
publisher = "Association for Computational Linguistics",
url = "https://aclanthology.org/2023.acl-long.699",
doi = "10.18653/v1/2023.acl-long.699",
pages = "12508--12522",
abstract = "Language Models (LM) are becoming more and more useful for providing representations upon which to train Natural Language Processing applications. However, there is now clear evidence that attention-based transformers require a critical amount of language data to produce good enough LMs. The question we have addressed in this paper is to what extent the critical amount of data varies for languages of different morphological typology, in particular those that have a rich inflectional morphology, and whether the tokenization method to preprocess the data can make a difference. These details can be important for low-resourced languages that need to plan the production of datasets. We evaluated intrinsically and extrinsically the differences of five different languages with different pretraining dataset sizes and three different tokenization methods for each. The results confirm that the size of the vocabulary due to morphological characteristics is directly correlated with both the LM perplexity and the performance of two typical downstream tasks such as NER identification and POS labeling. The experiments also provide new evidence that a canonical tokenizer can reduce perplexity by more than a half for a polysynthetic language like Quechua as well as raising F1 from 0.8 to more than 0.9 in both downstream tasks with a LM trained with only 6M tokens.",
}
| Language Models (LM) are becoming more and more useful for providing representations upon which to train Natural Language Processing applications. However, there is now clear evidence that attention-based transformers require a critical amount of language data to produce good enough LMs. The question we have addressed in this paper is to what extent the critical amount of data varies for languages of different morphological typology, in particular those that have a rich inflectional morphology, and whether the tokenization method to preprocess the data can make a difference. These details can be important for low-resourced languages that need to plan the production of datasets. We evaluated intrinsically and extrinsically the differences of five different languages with different pretraining dataset sizes and three different tokenization methods for each. The results confirm that the size of the vocabulary due to morphological characteristics is directly correlated with both the LM perplexity and the performance of two typical downstream tasks such as NER identification and POS labeling. The experiments also provide new evidence that a canonical tokenizer can reduce perplexity by more than a half for a polysynthetic language like Quechua as well as raising F1 from 0.8 to more than 0.9 in both downstream tasks with a LM trained with only 6M tokens. | [
"Zevallos, Rodolfo",
"Bel, Nuria"
] | Hints on the data for language modeling of synthetic languages with transformers | acl-long.699 | Poster | [
""
] | -1 | -1 | -1 | -1 | 0 | [] | [] | [] |
Subsets and Splits