publication_date
stringlengths 20
20
| doc_url
stringlengths 33
34
| id
stringlengths 33
34
| authors
stringlengths 5
20.1k
| update_date
stringlengths 20
20
| category_all
stringlengths 5
138
| abstract
stringlengths 123
3.09k
| category_primary
stringclasses 78
values | title
stringlengths 10
211
|
|---|---|---|---|---|---|---|---|---|
2024-05-22T20:08:48Z | http://arxiv.org/pdf/2405.13967v3 | http://arxiv.org/abs/2405.13967v3 | Rheeya Uppaal, Apratim Dey, Yiting He, Yiqiao Zhong, Junjie Hu | 2024-05-28T19:55:16Z | cs.CL | Recent alignment algorithms such as direct preference optimization (DPO) have
been developed to improve the safety of large language models (LLMs) by
training these models to match human behaviors exemplified by preference data.
However, these methods are both computationally intensive and lacking in
controllability and transparency, making them prone to jailbreaking and
inhibiting their widespread use. Furthermore, these tuning-based methods
require large-scale preference data for training and are susceptible to noisy
preference data. In this paper, we introduce a tuning-free alignment
alternative (DeTox) and demonstrate its effectiveness under the use case of
toxicity reduction. Grounded on theory from factor analysis, DeTox is a
sample-efficient model editing approach that identifies a toxic subspace in the
model parameter space and reduces model toxicity by projecting away the
detected subspace. The toxic sub-space is identified by extracting preference
data embeddings from the language model, and removing non-toxic information
from these embeddings. We show that DeTox is more sample-efficient than DPO,
further showcasing greater robustness to noisy data. Finally, we establish both
theoretical and empirical connections between DeTox and DPO, showing that DeTox
can be interpreted as a denoised version of a single DPO step. | cs.CL | DeTox: Toxic Subspace Projection for Model Editing |
2024-05-22T20:05:49Z | http://arxiv.org/pdf/2405.13966v1 | http://arxiv.org/abs/2405.13966v1 | Mudit Verma, Siddhant Bhambri, Subbarao Kambhampati | 2024-05-22T20:05:49Z | cs.AI, cs.CL | The reasoning abilities of Large Language Models (LLMs) remain a topic of
debate. Some methods such as ReAct-based prompting, have gained popularity for
claiming to enhance sequential decision-making abilities of agentic LLMs.
However, it is unclear what is the source of improvement in LLM reasoning with
ReAct based prompting. In this paper we examine these claims of ReAct based
prompting in improving agentic LLMs for sequential decision-making. By
introducing systematic variations to the input prompt we perform a sensitivity
analysis along the claims of ReAct and find that the performance is minimally
influenced by the "interleaving reasoning trace with action execution" or the
content of the generated reasoning traces in ReAct, contrary to original claims
and common usage. Instead, the performance of LLMs is driven by the similarity
between input example tasks and queries, implicitly forcing the prompt designer
to provide instance-specific examples which significantly increases the
cognitive burden on the human. Our investigation shows that the perceived
reasoning abilities of LLMs stem from the exemplar-query similarity and
approximate retrieval rather than any inherent reasoning abilities. | cs.AI | On the Brittle Foundations of ReAct Prompting for Agentic Large Language Models |
2024-05-22T19:39:05Z | http://arxiv.org/pdf/2405.13954v1 | http://arxiv.org/abs/2405.13954v1 | Sang Keun Choe, Hwijeen Ahn, Juhan Bae, Kewen Zhao, Minsoo Kang, Youngseog Chung, Adithya Pratapa, Willie Neiswanger, Emma Strubell, Teruko Mitamura, Jeff Schneider, Eduard Hovy, Roger Grosse, Eric Xing | 2024-05-22T19:39:05Z | cs.LG, cs.AI, cs.CL | Large language models (LLMs) are trained on a vast amount of human-written
data, but data providers often remain uncredited. In response to this issue,
data valuation (or data attribution), which quantifies the contribution or
value of each data to the model output, has been discussed as a potential
solution. Nevertheless, applying existing data valuation methods to recent LLMs
and their vast training datasets has been largely limited by prohibitive
compute and memory costs. In this work, we focus on influence functions, a
popular gradient-based data valuation method, and significantly improve its
scalability with an efficient gradient projection strategy called LoGra that
leverages the gradient structure in backpropagation. We then provide a
theoretical motivation of gradient projection approaches to influence functions
to promote trust in the data valuation process. Lastly, we lower the barrier to
implementing data valuation systems by introducing LogIX, a software package
that can transform existing training code into data valuation code with minimal
effort. In our data valuation experiments, LoGra achieves competitive accuracy
against more expensive baselines while showing up to 6,500x improvement in
throughput and 5x reduction in GPU memory usage when applied to
Llama3-8B-Instruct and the 1B-token dataset. | cs.LG | What is Your Data Worth to GPT? LLM-Scale Data Valuation with Influence Functions |
2024-05-22T18:58:58Z | http://arxiv.org/pdf/2405.13929v2 | http://arxiv.org/abs/2405.13929v2 | Aleksandr Nikolich, Konstantin Korolev, Artem Shelmanov, Igor Kiselev | 2024-06-19T17:32:23Z | cs.CL, cs.AI | There has been a surge in the development of various Large Language Models
(LLMs). However, text generation for languages other than English often faces
significant challenges, including poor generation quality and the reduced
computational performance due to the disproportionate representation of tokens
in model's vocabulary. In this work, we address these issues and introduce
Vikhr, a new state-of-the-art open-source instruction-tuned LLM designed
specifically for the Russian language. Unlike previous efforts for Russian that
utilize computationally inexpensive LoRA adapters on top of English-oriented
models, Vikhr features an adapted tokenizer vocabulary and undergoes the
continued pre-training and instruction tuning of all weights. This approach not
only enhances the model's performance but also significantly improves its
computational and contextual efficiency. The remarkable performance of Vikhr
across various Russian-language benchmarks can also be attributed to our
efforts in expanding instruction datasets and corpora for continued
pre-training. Vikhr not only sets the new state of the art among open-source
LLMs for Russian, but even outperforms some proprietary closed-source models on
certain benchmarks. The model weights, instruction sets, and code are publicly
available | cs.CL | Vikhr: The Family of Open-Source Instruction-Tuned Large Language Models for Russian |
2024-05-22T18:53:25Z | http://arxiv.org/pdf/2405.13923v2 | http://arxiv.org/abs/2405.13923v2 | Xiang Geng, Ming Zhu, Jiahuan Li, Zhejian Lai, Wei Zou, Shuaijie She, Jiaxin Guo, Xiaofeng Zhao, Yinglu Li, Yuang Li, Chang Su, Yanqing Zhao, Xinglin Lyu, Min Zhang, Jiajun Chen, Hao Yang, Shujian Huang | 2024-05-31T07:22:45Z | cs.CL | The scarcity of non-English data limits the development of non-English large
language models (LLMs). Transforming English-centric LLMs to non-English has
been identified as an effective and resource-efficient method. Previous works
start from base LLMs and perform knowledge distillation (KD) with data
generated by stronger LLMs, e.g. GPT-4. Compared to base LLMs, chat LLMs are
further optimized for advanced abilities, e.g. multi-turn conversation and
human preference alignment, and thus more powerful in both helpfulness and
safety. However, transforming a chat LLM involves two critical issues: (1) How
can we effectively transfer advanced abilities without their supervised data?
(2) How can we prevent the original knowledge from catastrophic forgetting
during transformation? We target these issues by introducing a simple framework
called TransLLM. For the first issue, TransLLM divides the transfer problem
into some common sub-tasks with the translation chain-of-thought, which uses
the translation as the bridge between English and non-English step-by-step. We
further enhance the performance of sub-tasks with publicly available data. For
the second issue, we propose a method comprising two synergistic components:
low-rank adaptation for training to maintain the original LLM parameters, and
recovery KD, which utilizes data generated by the chat LLM itself to recover
the original knowledge from the frozen parameters. In the experiments, we
transform the LLaMA-2-chat-7B to the Thai language. Our method, using only
single-turn data, outperforms strong baselines and ChatGPT on multi-turn
benchmark MT-bench. Furthermore, our method, without safety data, rejects more
harmful queries of safety benchmark AdvBench than both ChatGPT and GPT-4. | cs.CL | Why Not Transform Chat Large Language Models to Non-English? |
2024-05-22T18:35:10Z | http://arxiv.org/pdf/2405.13911v1 | http://arxiv.org/abs/2405.13911v1 | Wei Li, Hehe Fan, Yongkang Wong, Mohan Kankanhalli, Yi Yang | 2024-05-22T18:35:10Z | cs.CV, cs.AI, cs.CL | Recent advancements in image understanding have benefited from the extensive
use of web image-text pairs. However, video understanding remains a challenge
despite the availability of substantial web video-text data. This difficulty
primarily arises from the inherent complexity of videos and the inefficient
language supervision in recent web-collected video-text datasets. In this
paper, we introduce Text-Only Pre-Alignment (TOPA), a novel approach to extend
large language models (LLMs) for video understanding, without the need for
pre-training on real video data. Specifically, we first employ an advanced LLM
to automatically generate Textual Videos comprising continuous textual frames,
along with corresponding annotations to simulate real video-text data. Then,
these annotated textual videos are used to pre-align a language-only LLM with
the video modality. To bridge the gap between textual and real videos, we
employ the CLIP model as the feature extractor to align image and text
modalities. During text-only pre-alignment, the continuous textual frames,
encoded as a sequence of CLIP text features, are analogous to continuous CLIP
image features, thus aligning the LLM with real video representation. Extensive
experiments, including zero-shot evaluation and finetuning on various video
understanding tasks, demonstrate that TOPA is an effective and efficient
framework for aligning video content with LLMs. In particular, without training
on any video data, the TOPA-Llama2-13B model achieves a Top-1 accuracy of 51.0%
on the challenging long-form video understanding benchmark, Egoschema. This
performance surpasses previous video-text pre-training approaches and proves
competitive with recent GPT-3.5-based video agents. | cs.CV | TOPA: Extend Large Language Models for Video Understanding via Text-Only Pre-Alignment |
2024-05-22T18:28:26Z | http://arxiv.org/pdf/2405.13907v2 | http://arxiv.org/abs/2405.13907v2 | Adam Yang, Chen Chen, Konstantinos Pitas | 2024-06-16T13:49:53Z | cs.CL, cs.AI | State-of-the-art large language models are sometimes distributed as
open-source software but are also increasingly provided as a closed-source
service. These closed-source large-language models typically see the widest
usage by the public, however, they often do not provide an estimate of their
uncertainty when responding to queries. As even the best models are prone to
``hallucinating" false information with high confidence, a lack of a reliable
estimate of uncertainty limits the applicability of these models in critical
settings. We explore estimating the uncertainty of closed-source LLMs via
multiple rephrasings of an original base query. Specifically, we ask the model,
multiple rephrased questions, and use the similarity of the answers as an
estimate of uncertainty. We diverge from previous work in i) providing rules
for rephrasing that are simple to memorize and use in practice ii) proposing a
theoretical framework for why multiple rephrased queries obtain calibrated
uncertainty estimates. Our method demonstrates significant improvements in the
calibration of uncertainty estimates compared to the baseline and provides
intuition as to how query strategies should be designed for optimal test
calibration. | cs.CL | Just rephrase it! Uncertainty estimation in closed-source language models via multiple rephrased queries |
2024-05-22T17:56:53Z | http://arxiv.org/pdf/2405.13873v1 | http://arxiv.org/abs/2405.13873v1 | Yuan Sui, Yufei He, Nian Liu, Xiaoxin He, Kun Wang, Bryan Hooi | 2024-05-22T17:56:53Z | cs.AI, cs.CL | While large language models (LLMs) have achieved significant success in
various applications, they often struggle with hallucinations, especially in
scenarios that require deep and responsible reasoning. These issues could be
partially mitigate by integrating external knowledge graphs (KG) in LLM
reasoning. However, the method of their incorporation is still largely
unexplored. In this paper, we propose a retrieval-exploration interactive
method, FiDelis to handle intermediate steps of reasoning grounded by KGs.
Specifically, we propose Path-RAG module for recalling useful intermediate
knowledge from KG for LLM reasoning. We incorporate the logic and common-sense
reasoning of LLMs and topological connectivity of KGs into the knowledge
retrieval process, which provides more accurate recalling performance.
Furthermore, we propose to leverage deductive reasoning capabilities of LLMs as
a better criterion to automatically guide the reasoning process in a stepwise
and generalizable manner. Deductive verification serve as precise indicators
for when to cease further reasoning, thus avoiding misleading the chains of
reasoning and unnecessary computation. Extensive experiments show that our
method, as a training-free method with lower computational cost and better
generality outperforms the existing strong baselines in three benchmarks. | cs.AI | FiDeLiS: Faithful Reasoning in Large Language Model for Knowledge Graph Question Answering |
2024-05-22T17:56:51Z | http://arxiv.org/pdf/2405.13872v2 | http://arxiv.org/abs/2405.13872v2 | Qiji Zhou, Ruochen Zhou, Zike Hu, Panzhong Lu, Siyang Gao, Yue Zhang | 2024-05-29T02:24:36Z | cs.AI, cs.CL, cs.CV | Recent advancements in Chain-of-Thought (CoT) and related rationale-based
works have significantly improved the performance of Large Language Models
(LLMs) in complex reasoning tasks. With the evolution of Multimodal Large
Language Models (MLLMs), enhancing their capability to tackle complex
multimodal reasoning problems is a crucial frontier. However, incorporating
multimodal rationales in CoT has yet to be thoroughly investigated. We propose
the Image-of-Thought (IoT) prompting method, which helps MLLMs to extract
visual rationales step-by-step. Specifically, IoT prompting can automatically
design critical visual information extraction operations based on the input
images and questions. Each step of visual information refinement identifies
specific visual rationales that support answers to complex visual reasoning
questions. Beyond the textual CoT, IoT simultaneously utilizes visual and
textual rationales to help MLLMs understand complex multimodal information. IoT
prompting has improved zero-shot visual reasoning performance across various
visual understanding tasks in different MLLMs. Moreover, the step-by-step
visual feature explanations generated by IoT prompting elucidate the visual
reasoning process, aiding in analyzing the cognitive processes of large
multimodal models | cs.AI | Image-of-Thought Prompting for Visual Reasoning Refinement in Multimodal Large Language Models |
2024-05-22T17:50:04Z | http://arxiv.org/pdf/2405.13868v1 | http://arxiv.org/abs/2405.13868v1 | Xuyang Ge, Fukang Zhu, Wentao Shu, Junxuan Wang, Zhengfu He, Xipeng Qiu | 2024-05-22T17:50:04Z | cs.LG, cs.CL | Circuit analysis of any certain model behavior is a central task in
mechanistic interpretability. We introduce our circuit discovery pipeline with
sparse autoencoders (SAEs) and a variant called skip SAEs. With these two
modules inserted into the model, the model's computation graph with respect to
OV and MLP circuits becomes strictly linear. Our methods do not require linear
approximation to compute the causal effect of each node. This fine-grained
graph enables identifying both end-to-end and local circuits accounting for
either logits or intermediate features. We can scalably apply this pipeline
with a technique called Hierarchical Attribution. We analyze three kind of
circuits in GPT2-Small, namely bracket, induction and Indirect Object
Identification circuits. Our results reveal new findings underlying existing
discoveries. | cs.LG | Automatically Identifying Local and Global Circuits with Linear Computation Graphs |
2024-05-22T17:13:49Z | http://arxiv.org/pdf/2405.13845v2 | http://arxiv.org/abs/2405.13845v2 | Xin Qiu, Risto Miikkulainen | 2024-05-25T07:20:46Z | cs.CL, cs.AI | With the widespread application of Large Language Models (LLMs) to various
domains, concerns regarding the trustworthiness of LLMs in safety-critical
scenarios have been raised, due to their unpredictable tendency to hallucinate
and generate misinformation. Existing LLMs do not have an inherent
functionality to provide the users with an uncertainty metric for each response
it generates, making it difficult to evaluate trustworthiness. Although a
number of works aim to develop uncertainty quantification methods for LLMs,
they have fundamental limitations, such as being restricted to classification
tasks, requiring additional training and data, considering only lexical instead
of semantic information, and being prompt-wise but not response-wise. A new
framework is proposed in this paper to address these issues. Semantic density
extracts uncertainty information for each response from a probability
distribution perspective in semantic space. It has no restriction on task types
and is "off-the-shelf" for new models and tasks. Experiments on seven
state-of-the-art LLMs, including the latest Llama 3 and Mixtral-8x22B models,
on four free-form question-answering benchmarks demonstrate the superior
performance and robustness of semantic density compared to prior approaches. | cs.CL | Semantic Density: Uncertainty Quantification in Semantic Space for Large Language Models |
2024-05-22T16:57:02Z | http://arxiv.org/pdf/2405.13828v1 | http://arxiv.org/abs/2405.13828v1 | Ziqiao Ma, Zekun Wang, Joyce Chai | 2024-05-22T16:57:02Z | cs.CL, cs.AI | Humans are efficient language learners and inherently social creatures. Our
language development is largely shaped by our social interactions, for example,
the demonstration and feedback from caregivers. Contrary to human language
learning, recent advancements in large language models have primarily adopted a
non-interactive training paradigm, and refined pre-trained models through
feedback afterward. In this work, we aim to examine how corrective feedback
from interactions influences neural language acquisition from the ground up
through systematically controlled experiments, assessing whether it contributes
to learning efficiency in language models. We introduce a
trial-and-demonstration (TnD) learning framework that incorporates three
components: student trials, teacher demonstrations, and a reward conditioned on
language competence at various developmental stages. Our experiments reveal
that the TnD approach accelerates word acquisition for student models of equal
and smaller numbers of parameters, and we highlight the significance of both
trials and demonstrations. We further show that the teacher's choices of words
influence students' word-specific learning efficiency, and a
practice-makes-perfect effect is evident by a strong correlation between the
frequency of words in trials and their respective learning curves. Our findings
suggest that interactive language learning, with teacher demonstrations and
student trials, can facilitate efficient word learning in language models. | cs.CL | Babysit A Language Model From Scratch: Interactive Language Learning by Trials and Demonstrations |
2024-05-22T16:51:07Z | http://arxiv.org/pdf/2405.13820v1 | http://arxiv.org/abs/2405.13820v1 | Weixiang Zhao, Yulin Hu, Zhuojun Li, Yang Deng, Yanyan Zhao, Bing Qin, Tat-Seng Chua | 2024-05-22T16:51:07Z | cs.CL | Safety alignment of large language models (LLMs) has been gaining increasing
attention. However, current safety-aligned LLMs suffer from the fragile and
imbalanced safety mechanisms, which can still be induced to generate unsafe
responses, exhibit over-safety by rejecting safe user inputs, and fail to
preserve general utility after safety alignment. To this end, we propose a
novel post safety alignment (PSA) method to address these inherent and emerging
safety challenges, including safety enhancement, over-safety mitigation, and
utility preservation. In specific, we introduce \textsc{SafePatching}, a novel
framework for comprehensive and efficient PSA, where two distinct safety
patches are developed on the harmful data to enhance safety and mitigate
over-safety concerns, and then seamlessly integrated into the target LLM
backbone without compromising its utility. Extensive experiments show that
\textsc{SafePatching} achieves a more comprehensive and efficient PSA than
baseline methods. It even enhances the utility of the backbone, further
optimizing the balance between being helpful and harmless in current aligned
LLMs. Also, \textsc{SafePatching} demonstrates its superiority in continual PSA
scenarios. | cs.CL | Towards Comprehensive and Efficient Post Safety Alignment of Large Language Models via Safety Patching |
2024-05-22T16:46:19Z | http://arxiv.org/pdf/2405.13816v2 | http://arxiv.org/abs/2405.13816v2 | Shimao Zhang, Changjiang Gao, Wenhao Zhu, Jiajun Chen, Xin Huang, Xue Han, Junlan Feng, Chao Deng, Shujian Huang | 2024-06-18T16:30:01Z | cs.CL | Recently, Large Language Models (LLMs) have shown impressive language
capabilities. While most of the existing LLMs have very unbalanced performance
across different languages, multilingual alignment based on translation
parallel data is an effective method to enhance the LLMs' multilingual
capabilities. In this work, we discover and comprehensively investigate the
spontaneous multilingual alignment improvement of LLMs. We find that LLMs
instruction-tuned on the question translation data (i.e. without annotated
answers) are able to encourage the alignment between English and a wide range
of languages, even including those unseen during instruction-tuning.
Additionally, we utilize different settings and mechanistic interpretability
methods to analyze the LLM's performance in the multilingual scenario
comprehensively. Our work suggests that LLMs have enormous potential for
improving multilingual alignment efficiently with great language and task
generalization. | cs.CL | Getting More from Less: Large Language Models are Good Spontaneous Multilingual Learners |
2024-05-22T16:30:24Z | http://arxiv.org/pdf/2405.13803v2 | http://arxiv.org/abs/2405.13803v2 | Siyi Wu, Feixue Han, Bingsheng Yao, Tianyi Xie, Xuan Zhao, Dakuo Wang | 2024-06-14T03:54:31Z | cs.HC, cs.CL | A longstanding challenge in mental well-being support is the reluctance of
people to adopt psychologically beneficial activities, often due to lack of
motivation, low perceived trustworthiness, and limited personalization of
recommendations. Chatbots have shown promise in promoting positive mental
health practices, yet their rigid interaction flows and less human-like
conversational experiences present significant limitations. In this work, we
explore whether the anthropomorphic design (both LLM's persona design and
conversational experience design) can enhance users' perception of the system
and their willingness to adopt mental well-being activity recommendations. To
this end, we introduce Sunnie, an anthropomorphic LLM-based conversational
agent designed to offer personalized well-being support through multi-turn
conversation and recommend practical actions grounded in positive psychology
and social psychology. An empirical user study comparing the user experience
with Sunnie and with a traditional survey-based activity recommendation system
suggests that the anthropomorphic characteristics of Sunnie significantly
enhance users' perception of the system and the overall usability;
nevertheless, users' willingness to adopt activity recommendations did not
change significantly. | cs.HC | Sunnie: An Anthropomorphic LLM-Based Conversational Agent for Mental Well-Being Activity Recommendation |
2024-05-22T16:23:40Z | http://arxiv.org/pdf/2405.13798v1 | http://arxiv.org/abs/2405.13798v1 | Raghu Mudumbai, Tyler Bell | 2024-05-22T16:23:40Z | cs.CL, cs.AI, cs.IT, math.IT | We propose a new asymptotic equipartition property for the perplexity of a
large piece of text generated by a language model and present theoretical
arguments for this property. Perplexity, defined as a inverse likelihood
function, is widely used as a performance metric for training language models.
Our main result states that the logarithmic perplexity of any large text
produced by a language model must asymptotically converge to the average
entropy of its token distributions. This means that language models are
constrained to only produce outputs from a ``typical set", which we show, is a
vanishingly small subset of all possible grammatically correct outputs. We
present preliminary experimental results from an open-source language model to
support our theoretical claims. This work has possible practical applications
for understanding and improving ``AI detection" tools and theoretical
implications for the uniqueness, predictability and creative potential of
generative models. | cs.CL | Slaves to the Law of Large Numbers: An Asymptotic Equipartition Property for Perplexity in Generative Language Models |
2024-05-22T16:15:17Z | http://arxiv.org/pdf/2405.13792v1 | http://arxiv.org/abs/2405.13792v1 | Xin Cheng, Xun Wang, Xingxing Zhang, Tao Ge, Si-Qing Chen, Furu Wei, Huishuai Zhang, Dongyan Zhao | 2024-05-22T16:15:17Z | cs.CL, cs.AI, cs.IR | This paper introduces xRAG, an innovative context compression method tailored
for retrieval-augmented generation. xRAG reinterprets document embeddings in
dense retrieval--traditionally used solely for retrieval--as features from the
retrieval modality. By employing a modality fusion methodology, xRAG seamlessly
integrates these embeddings into the language model representation space,
effectively eliminating the need for their textual counterparts and achieving
an extreme compression rate. In xRAG, the only trainable component is the
modality bridge, while both the retriever and the language model remain frozen.
This design choice allows for the reuse of offline-constructed document
embeddings and preserves the plug-and-play nature of retrieval augmentation.
Experimental results demonstrate that xRAG achieves an average improvement of
over 10% across six knowledge-intensive tasks, adaptable to various language
model backbones, ranging from a dense 7B model to an 8x7B Mixture of Experts
configuration. xRAG not only significantly outperforms previous context
compression methods but also matches the performance of uncompressed models on
several datasets, while reducing overall FLOPs by a factor of 3.53. Our work
pioneers new directions in retrieval-augmented generation from the perspective
of multimodality fusion, and we hope it lays the foundation for future
efficient and scalable retrieval-augmented systems | cs.CL | xRAG: Extreme Context Compression for Retrieval-augmented Generation with One Token |
2024-05-22T15:56:52Z | http://arxiv.org/pdf/2405.13769v1 | http://arxiv.org/abs/2405.13769v1 | Cyril Chhun, Fabian M. Suchanek, Chloé Clavel | 2024-05-22T15:56:52Z | cs.CL | Storytelling is an integral part of human experience and plays a crucial role
in social interactions. Thus, Automatic Story Evaluation (ASE) and Generation
(ASG) could benefit society in multiple ways, but they are challenging tasks
which require high-level human abilities such as creativity, reasoning and deep
understanding. Meanwhile, Large Language Models (LLM) now achieve
state-of-the-art performance on many NLP tasks. In this paper, we study whether
LLMs can be used as substitutes for human annotators for ASE. We perform an
extensive analysis of the correlations between LLM ratings, other automatic
measures, and human annotations, and we explore the influence of prompting on
the results and the explainability of LLM behaviour. Most notably, we find that
LLMs outperform current automatic measures for system-level evaluation but
still struggle at providing satisfactory explanations for their answers. | cs.CL | Do Language Models Enjoy Their Own Stories? Prompting Large Language Models for Automatic Story Evaluation |
2024-05-22T15:52:52Z | http://arxiv.org/pdf/2405.14899v1 | http://arxiv.org/abs/2405.14899v1 | Zijian Zhou, Xiaoqiang Lin, Xinyi Xu, Alok Prakash, Daniela Rus, Bryan Kian Hsiang Low | 2024-05-22T15:52:52Z | cs.CL, cs.AI, cs.LG | In-context learning (ICL) allows transformer-based language models that are
pre-trained on general text to quickly learn a specific task with a few "task
demonstrations" without updating their parameters, significantly boosting their
flexibility and generality. ICL possesses many distinct characteristics from
conventional machine learning, thereby requiring new approaches to interpret
this learning paradigm. Taking the viewpoint of recent works showing that
transformers learn in context by formulating an internal optimizer, we propose
an influence function-based attribution technique, DETAIL, that addresses the
specific characteristics of ICL. We empirically verify the effectiveness of our
approach for demonstration attribution while being computationally efficient.
Leveraging the results, we then show how DETAIL can help improve model
performance in real-world scenarios through demonstration reordering and
curation. Finally, we experimentally prove the wide applicability of DETAIL by
showing our attribution scores obtained on white-box models are transferable to
black-box models in improving model performance. | cs.CL | DETAIL: Task DEmonsTration Attribution for Interpretable In-context Learning |
2024-05-22T15:38:53Z | http://arxiv.org/pdf/2405.13754v1 | http://arxiv.org/abs/2405.13754v1 | Wondimagegnhue Tsegaye Tufa, Ilia Markov, Piek Vossen | 2024-05-22T15:38:53Z | cs.CL | Social media conversations frequently suffer from toxicity, creating
significant issues for users, moderators, and entire communities. Events in the
real world, like elections or conflicts, can initiate and escalate toxic
behavior online. Our study investigates how real-world events influence the
origin and spread of toxicity in online discussions across various languages
and regions. We gathered Reddit data comprising 4.5 million comments from 31
thousand posts in six different languages (Dutch, English, German, Arabic,
Turkish and Spanish). We target fifteen major social and political world events
that occurred between 2020 and 2023. We observe significant variations in
toxicity, negative sentiment, and emotion expressions across different events
and language communities, showing that toxicity is a complex phenomenon in
which many different factors interact and still need to be investigated. We
will release the data for further research along with our code. | cs.CL | Grounding Toxicity in Real-World Events across Languages |
2024-05-22T14:25:41Z | http://arxiv.org/pdf/2405.13684v1 | http://arxiv.org/abs/2405.13684v1 | Guangzhi Sun, Potsawee Manakul, Adian Liusie, Kunat Pipatanakul, Chao Zhang, Phil Woodland, Mark Gales | 2024-05-22T14:25:41Z | cs.CL | Multimodal foundation models are prone to hallucination, generating outputs
that either contradict the input or are not grounded by factual information.
Given the diversity in architectures, training data and instruction tuning
techniques, there can be large variations in systems' susceptibility to
hallucinations. To assess system hallucination robustness, hallucination
ranking approaches have been developed for specific tasks such as image
captioning, question answering, summarization, or biography generation.
However, these approaches typically compare model outputs to gold-standard
references or labels, limiting hallucination benchmarking for new domains. This
work proposes "CrossCheckGPT", a reference-free universal hallucination ranking
for multimodal foundation models. The core idea of CrossCheckGPT is that the
same hallucinated content is unlikely to be generated by different independent
systems, hence cross-system consistency can provide meaningful and accurate
hallucination assessment scores. CrossCheckGPT can be applied to any model or
task, provided that the information consistency between outputs can be measured
through an appropriate distance metric. Focusing on multimodal large language
models that generate text, we explore two information consistency measures:
CrossCheck-explicit and CrossCheck-implicit. We showcase the applicability of
our method for hallucination ranking across various modalities, namely the
text, image, and audio-visual domains. Further, we propose the first
audio-visual hallucination benchmark, "AVHalluBench", and illustrate the
effectiveness of CrossCheckGPT, achieving correlations of 98% and 89% with
human judgements on MHaluBench and AVHalluBench, respectively. | cs.CL | CrossCheckGPT: Universal Hallucination Ranking for Multimodal Foundation Models |
2024-05-22T13:39:33Z | http://arxiv.org/pdf/2405.13640v1 | http://arxiv.org/abs/2405.13640v1 | Ying Ma, Owen Burns, Mingqiu Wang, Gang Li, Nan Du, Laurent El Shafey, Liqiang Wang, Izhak Shafran, Hagen Soltau | 2024-05-22T13:39:33Z | cs.CL, cs.AI, cs.LG | Reinforcement learning (RL) is an effective method of finding reasoning
pathways in incomplete knowledge graphs (KGs). To overcome the challenges of a
large action space, a self-supervised pre-training method is proposed to warm
up the policy network before the RL training stage. To alleviate the
distributional mismatch issue in general self-supervised RL (SSRL), in our
supervised learning (SL) stage, the agent selects actions based on the policy
network and learns from generated labels; this self-generation of labels is the
intuition behind the name self-supervised. With this training framework, the
information density of our SL objective is increased and the agent is prevented
from getting stuck with the early rewarded paths. Our self-supervised RL (SSRL)
method improves the performance of RL by pairing it with the wide coverage
achieved by SL during pretraining, since the breadth of the SL objective makes
it infeasible to train an agent with that alone. We show that our SSRL model
meets or exceeds current state-of-the-art results on all Hits@k and mean
reciprocal rank (MRR) metrics on four large benchmark KG datasets. This SSRL
method can be used as a plug-in for any RL architecture for a KGR task. We
adopt two RL architectures, i.e., MINERVA and MultiHopKG as our baseline RL
models and experimentally show that our SSRL model consistently outperforms
both baselines on all of these four KG reasoning tasks. Full code for the paper
available at
https://github.com/owenonline/Knowledge-Graph-Reasoning-with-Self-supervised-Reinforcement-Learning. | cs.CL | Knowledge Graph Reasoning with Self-supervised Reinforcement Learning |
2024-05-22T13:14:11Z | http://arxiv.org/pdf/2405.13622v1 | http://arxiv.org/abs/2405.13622v1 | Gauthier Guinet, Behrooz Omidvar-Tehrani, Anoop Deoras, Laurent Callot | 2024-05-22T13:14:11Z | cs.CL, cs.IR | We propose a new method to measure the task-specific accuracy of
Retrieval-Augmented Large Language Models (RAG). Evaluation is performed by
scoring the RAG on an automatically-generated synthetic exam composed of
multiple choice questions based on the corpus of documents associated with the
task. Our method is an automated, cost-efficient, interpretable, and robust
strategy to select the optimal components for a RAG system. We leverage Item
Response Theory (IRT) to estimate the quality of an exam and its
informativeness on task-specific accuracy. IRT also provides a natural way to
iteratively improve the exam by eliminating the exam questions that are not
sufficiently informative about a model's ability. We demonstrate our approach
on four new open-ended Question-Answering tasks based on Arxiv abstracts,
StackExchange questions, AWS DevOps troubleshooting guides, and SEC filings. In
addition, our experiments reveal more general insights into factors impacting
RAG performance like size, retrieval mechanism, prompting and fine-tuning. Most
notably, our findings show that choosing the right retrieval algorithms often
leads to bigger performance gains than simply using a larger language model. | cs.CL | Automated Evaluation of Retrieval-Augmented Language Models with Task-Specific Exam Generation |
2024-05-22T12:53:12Z | http://arxiv.org/pdf/2405.13602v1 | http://arxiv.org/abs/2405.13602v1 | Zhiwei Hu, Víctor Gutiérrez-Basulto, Zhiliang Xiang, Ru Li, Jeff Z. Pan | 2024-05-22T12:53:12Z | cs.AI, cs.CL, cs.LG | Knowledge graph entity typing (KGET) aims to infer missing entity type
instances in knowledge graphs. Previous research has predominantly centered
around leveraging contextual information associated with entities, which
provides valuable clues for inference. However, they have long ignored the dual
nature of information inherent in entities, encompassing both high-level
coarse-grained cluster knowledge and fine-grained type knowledge. This paper
introduces Cross-view Optimal Transport for knowledge graph Entity Typing
(COTET), a method that effectively incorporates the information on how types
are clustered into the representation of entities and types. COTET comprises
three modules: i) Multi-view Generation and Encoder, which captures structured
knowledge at different levels of granularity through entity-type,
entity-cluster, and type-cluster-type perspectives; ii) Cross-view Optimal
Transport, transporting view-specific embeddings to a unified space by
minimizing the Wasserstein distance from a distributional alignment
perspective; iii) Pooling-based Entity Typing Prediction, employing a mixture
pooling mechanism to aggregate prediction scores from diverse neighbors of an
entity. Additionally, we introduce a distribution-based loss function to
mitigate the occurrence of false negatives during training. Extensive
experiments demonstrate the effectiveness of COTET when compared to existing
baselines. | cs.AI | COTET: Cross-view Optimal Transport for Knowledge Graph Entity Typing |
2024-05-22T12:15:52Z | http://arxiv.org/pdf/2405.13578v1 | http://arxiv.org/abs/2405.13578v1 | Weilong Dong, Xinwei Wu, Renren Jin, Shaoyang Xu, Deyi Xiong | 2024-05-22T12:15:52Z | cs.CL | Ensuring large language models (LLM) behave consistently with human goals,
values, and intentions is crucial for their safety but yet computationally
expensive. To reduce the computational cost of alignment training of LLMs,
especially for those with a huge number of parameters, and to reutilize learned
value alignment, we propose ConTrans, a novel framework that enables
weak-to-strong alignment transfer via concept transplantation. From the
perspective of representation engineering, ConTrans refines concept vectors in
value alignment from a source LLM (usually a weak yet aligned LLM). The refined
concept vectors are then reformulated to adapt to the target LLM (usually a
strong yet unaligned base LLM) via affine transformation. In the third step,
ConTrans transplants the reformulated concept vectors into the residual stream
of the target LLM. Experiments demonstrate the successful transplantation of a
wide range of aligned concepts from 7B models to 13B and 70B models across
multiple LLMs and LLM families. Remarkably, ConTrans even surpasses
instruction-tuned models in terms of truthfulness. Experiment results validate
the effectiveness of both inter-LLM-family and intra-LLM-family concept
transplantation. Our work successfully demonstrates an alternative way to
achieve weak-to-strong alignment generalization and control. | cs.CL | ConTrans: Weak-to-Strong Alignment Engineering via Concept Transplantation |
2024-05-22T12:12:40Z | http://arxiv.org/pdf/2405.13576v1 | http://arxiv.org/abs/2405.13576v1 | Jiajie Jin, Yutao Zhu, Xinyu Yang, Chenghao Zhang, Zhicheng Dou | 2024-05-22T12:12:40Z | cs.CL, cs.IR | With the advent of Large Language Models (LLMs), the potential of Retrieval
Augmented Generation (RAG) techniques have garnered considerable research
attention. Numerous novel algorithms and models have been introduced to enhance
various aspects of RAG systems. However, the absence of a standardized
framework for implementation, coupled with the inherently intricate RAG
process, makes it challenging and time-consuming for researchers to compare and
evaluate these approaches in a consistent environment. Existing RAG toolkits
like LangChain and LlamaIndex, while available, are often heavy and unwieldy,
failing to meet the personalized needs of researchers. In response to this
challenge, we propose FlashRAG, an efficient and modular open-source toolkit
designed to assist researchers in reproducing existing RAG methods and in
developing their own RAG algorithms within a unified framework. Our toolkit
implements 12 advanced RAG methods and has gathered and organized 32 benchmark
datasets. Our toolkit has various features, including customizable modular
framework, rich collection of pre-implemented RAG works, comprehensive
datasets, efficient auxiliary pre-processing scripts, and extensive and
standard evaluation metrics. Our toolkit and resources are available at
https://github.com/RUC-NLPIR/FlashRAG. | cs.CL | FlashRAG: A Modular Toolkit for Efficient Retrieval-Augmented Generation Research |
2024-05-22T12:05:17Z | http://arxiv.org/pdf/2405.13568v1 | http://arxiv.org/abs/2405.13568v1 | Wanyu Hu, Vrizlynn L. L. Thing | 2024-05-22T12:05:17Z | cs.CR, cs.AI, cs.CL, cs.LG | With the drastic increase in the number of new vulnerabilities in the
National Vulnerability Database (NVD) every year, the workload for NVD analysts
to associate the Common Platform Enumeration (CPE) with the Common
Vulnerabilities and Exposures (CVE) summaries becomes increasingly laborious
and slow. The delay causes organisations, which depend on NVD for vulnerability
management and security measurement, to be more vulnerable to zero-day attacks.
Thus, it is essential to come out with a technique and tool to extract the CPEs
in the CVE summaries accurately and quickly. In this work, we propose the
CPE-Identifier system, an automated CPE annotating and extracting system, from
the CVE summaries. The system can be used as a tool to identify CPE entities
from new CVE text inputs. Moreover, we also automate the data generating and
labeling processes using deep learning models. Due to the complexity of the CVE
texts, new technical terminologies appear frequently. To identify novel words
in future CVE texts, we apply Natural Language Processing (NLP) Named Entity
Recognition (NER), to identify new technical jargons in the text. Our proposed
model achieves an F1 score of 95.48%, an accuracy score of 99.13%, a precision
of 94.83%, and a recall of 96.14%. We show that it outperforms prior works on
automated CVE-CPE labeling by more than 9% on all metrics. | cs.CR | CPE-Identifier: Automated CPE identification and CVE summaries annotation with Deep Learning and NLP |
2024-05-22T11:33:29Z | http://arxiv.org/pdf/2405.13548v2 | http://arxiv.org/abs/2405.13548v2 | Wei Zhang, Xianfu Cheng, Yi Zhang, Jian Yang, Hongcheng Guo, Zhoujun Li, Xiaolin Yin, Xiangyuan Guan, Xu Shi, Liangfan Zheng, Bo Zhang | 2024-05-24T05:39:24Z | cs.SE, cs.CL | Log parsing, a vital task for interpreting the vast and complex data produced
within software architectures faces significant challenges in the transition
from academic benchmarks to the industrial domain. Existing log parsers, while
highly effective on standardized public datasets, struggle to maintain
performance and efficiency when confronted with the sheer scale and diversity
of real-world industrial logs. These challenges are two-fold: 1) massive log
templates: The performance and efficiency of most existing parsers will be
significantly reduced when logs of growing quantities and different lengths; 2)
Complex and changeable semantics: Traditional template-matching algorithms
cannot accurately match the log templates of complicated industrial logs
because they cannot utilize cross-language logs with similar semantics. To
address these issues, we propose ECLIPSE, Enhanced Cross-Lingual Industrial log
Parsing with Semantic Entropy-LCS, since cross-language logs can robustly parse
industrial logs. On the one hand, it integrates two efficient data-driven
template-matching algorithms and Faiss indexing. On the other hand, driven by
the powerful semantic understanding ability of the Large Language Model (LLM),
the semantics of log keywords were accurately extracted, and the retrieval
space was effectively reduced. Notably, we launch a Chinese and English
cross-platform industrial log parsing benchmark ECLIPSE- BENCH to evaluate the
performance of mainstream parsers in industrial scenarios. Our experimental
results across public benchmarks and ECLIPSE- BENCH underscore the superior
performance and robustness of our proposed ECLIPSE. Notably, ECLIPSE both
delivers state-of-the-art performance when compared to strong baselines and
preserves a significant edge in processing efficiency. | cs.SE | ECLIPSE: Semantic Entropy-LCS for Cross-Lingual Industrial Log Parsing |
2024-05-22T11:30:59Z | http://arxiv.org/pdf/2405.13546v2 | http://arxiv.org/abs/2405.13546v2 | Monika Jain, Raghava Mutharaju, Kuldeep Singh, Ramakanth Kavuluru | 2024-06-18T08:20:19Z | cs.CL, cs.IR | Relation extraction (RE) is a well-known NLP application often treated as a
sentence- or document-level task. However, a handful of recent efforts explore
it across documents or in the cross-document setting (CrossDocRE). This is
distinct from the single document case because different documents often focus
on disparate themes, while text within a document tends to have a single goal.
Linking findings from disparate documents to identify new relationships is at
the core of the popular literature-based knowledge discovery paradigm in
biomedicine and other domains. Current CrossDocRE efforts do not consider
domain knowledge, which are often assumed to be known to the reader when
documents are authored. Here, we propose a novel approach, KXDocRE, that embed
domain knowledge of entities with input text for cross-document RE. Our
proposed framework has three main benefits over baselines: 1) it incorporates
domain knowledge of entities along with documents' text; 2) it offers
interpretability by producing explanatory text for predicted relations between
entities 3) it improves performance over the prior methods. | cs.CL | Knowledge-Driven Cross-Document Relation Extraction |
2024-05-22T11:23:03Z | http://arxiv.org/pdf/2405.13541v1 | http://arxiv.org/abs/2405.13541v1 | Yuu Jinnai, Ukyo Honda | 2024-05-22T11:23:03Z | cs.CL, cs.AI, cs.LG | Preference optimization is a standard approach to fine-tuning large language
models to align with human preferences. The quality, diversity, and quantity of
the preference dataset are critical to the effectiveness of preference
optimization. However, obtaining a large amount of high-quality and diverse
preference annotations is difficult in many applications. This raises the
question of how to use the limited annotation budget to create an effective
preference dataset. To this end, we propose Annotation-Efficient Preference
Optimization (AEPO). Instead of exhaustively annotating preference over all
available response texts, AEPO selects a subset of responses that maximizes
quality and diversity from the available responses, and then annotates
preference over the selected ones. In this way, AEPO focuses the annotation
budget on labeling preference over a smaller subset of responses with diversity
and of high quality. We evaluate the performance of Direct Preference
Optimization (DPO) using AEPO and show that it outperforms models trained using
a standard DPO with the same annotation budget. Our code is available at
https://github.com/CyberAgentAILab/annotation-efficient-po | cs.CL | Annotation-Efficient Preference Optimization for Language Model Alignment |
2024-05-22T11:14:00Z | http://arxiv.org/pdf/2405.13536v1 | http://arxiv.org/abs/2405.13536v1 | Tobias Leemann, Alina Fastowski, Felix Pfeiffer, Gjergji Kasneci | 2024-05-22T11:14:00Z | cs.LG, cs.AI, cs.CL | We address the critical challenge of applying feature attribution methods to
the transformer architecture, which dominates current applications in natural
language processing and beyond. Traditional attribution methods to explainable
AI (XAI) explicitly or implicitly rely on linear or additive surrogate models
to quantify the impact of input features on a model's output. In this work, we
formally prove an alarming incompatibility: transformers are structurally
incapable to align with popular surrogate models for feature attribution,
undermining the grounding of these conventional explanation methodologies. To
address this discrepancy, we introduce the Softmax-Linked Additive Log-Odds
Model (SLALOM), a novel surrogate model specifically designed to align with the
transformer framework. Unlike existing methods, SLALOM demonstrates the
capacity to deliver a range of faithful and insightful explanations across both
synthetic and real-world datasets. Showing that diverse explanations computed
from SLALOM outperform common surrogate explanations on different tasks, we
highlight the need for task-specific feature attributions rather than a
one-size-fits-all approach. | cs.LG | Attention Mechanisms Don't Learn Additive Models: Rethinking Feature Importance for Transformers |
2024-05-22T10:55:26Z | http://arxiv.org/pdf/2405.13529v1 | http://arxiv.org/abs/2405.13529v1 | Huasheng Zhang | 2024-05-22T10:55:26Z | cs.CL | In native speakers' lexical choices, a concept can be more readily expressed
by one expression over another grammatical one, a phenomenon known as
nativelike selection (NLS). In previous research, arbitrary chunks such as
collocations have been considered crucial for this phenomenon. However, this
study examines the possibility of analyzing the semantic motivation and
deducibility behind some NLSs by exploring the correlation between NLS and
prototypicality, specifically the onomasiological hypothesis of Grondelaers and
Geeraerts (2003, Towards a pragmatic model of cognitive onomasiology. In Hubert
Cuyckens, Ren\'e Dirven & John R. Taylor (eds.), Cognitive approaches to
lexical semantics, 67-92. Berlin: De Gruyter Mouton). They hypothesized that
"[a] referent is more readily named by a lexical item if it is a salient member
of the category denoted by that item". To provide a preliminary investigation
of this important but rarely explored phenomenon, a series of innovative
methods and procedures, including the use of semantic embedding and
interlingual comparisons, is designed. Specifically, potential NLSs are
efficiently discovered through an automatic exploratory analysis using topic
modeling techniques, and then confirmed by manual inspection through frame
semantics. Finally, to account for the NLS in question, cluster analysis and
behavioral profile analysis are conducted to uncover a language-specific
prototype for the Chinese verb shang 'harm', providing supporting evidence for
the correlation between NLS and prototypicality. | cs.CL | The correlation between nativelike selection and prototypicality: a multilingual onomasiological case study using semantic embedding |
2024-05-22T10:45:50Z | http://arxiv.org/pdf/2405.13522v2 | http://arxiv.org/abs/2405.13522v2 | Zhijian Xu, Yuxuan Bian, Jianyuan Zhong, Xiangyu Wen, Qiang Xu | 2024-05-24T15:10:27Z | cs.LG, cs.AI, cs.CL | This work introduces a novel Text-Guided Time Series Forecasting (TGTSF)
task. By integrating textual cues, such as channel descriptions and dynamic
news, TGTSF addresses the critical limitations of traditional methods that rely
purely on historical data. To support this task, we propose TGForecaster, a
robust baseline model that fuses textual cues and time series data using
cross-attention mechanisms. We then present four meticulously curated benchmark
datasets to validate the proposed framework, ranging from simple periodic data
to complex, event-driven fluctuations. Our comprehensive evaluations
demonstrate that TGForecaster consistently achieves state-of-the-art
performance, highlighting the transformative potential of incorporating textual
information into time series forecasting. This work not only pioneers a novel
forecasting task but also establishes a new benchmark for future research,
driving advancements in multimodal data integration for time series models. | cs.LG | Beyond Trend and Periodicity: Guiding Time Series Forecasting with Textual Cues |
2024-05-22T10:22:20Z | http://arxiv.org/pdf/2405.13517v1 | http://arxiv.org/abs/2405.13517v1 | Baizhou Huang, Xiaojun Wan | 2024-05-22T10:22:20Z | cs.CR, cs.CL | With the increasing use of large language models (LLMs) in daily life,
concerns have emerged regarding their potential misuse and societal impact.
Watermarking is proposed to trace the usage of specific models by injecting
patterns into their generated texts. An ideal watermark should produce outputs
that are nearly indistinguishable from those of the original LLM
(imperceptibility), while ensuring a high detection rate (efficacy), even when
the text is partially altered (robustness). Despite many methods having been
proposed, none have simultaneously achieved all three properties, revealing an
inherent trade-off. This paper utilizes a key-centered scheme to unify existing
watermarking techniques by decomposing a watermark into two distinct modules: a
key module and a mark module. Through this decomposition, we demonstrate for
the first time that the key module significantly contributes to the trade-off
issues observed in prior methods. Specifically, this reflects the conflict
between the scale of the key sampling space during generation and the
complexity of key restoration during detection. To this end, we introduce
\textbf{WaterPool}, a simple yet effective key module that preserves a complete
key sampling space required by imperceptibility while utilizing semantics-based
search to improve the key restoration process. WaterPool can integrate with
most watermarks, acting as a plug-in. Our experiments with three well-known
watermarking techniques show that WaterPool significantly enhances their
performance, achieving near-optimal imperceptibility and markedly improving
efficacy and robustness (+12.73\% for KGW, +20.27\% for EXP, +7.27\% for ITS). | cs.CR | WaterPool: A Watermark Mitigating Trade-offs among Imperceptibility, Efficacy and Robustness |
2024-05-22T10:21:50Z | http://arxiv.org/pdf/2405.13516v2 | http://arxiv.org/abs/2405.13516v2 | Mingye Zhu, Yi Liu, Lei Zhang, Junbo Guo, Zhendong Mao | 2024-06-04T08:21:05Z | cs.CL, cs.LG | Recently, tremendous strides have been made to align the generation of Large
Language Models (LLMs) with human values to mitigate toxic or unhelpful
content. Leveraging Reinforcement Learning from Human Feedback (RLHF) proves
effective and is widely adopted by researchers. However, implementing RLHF is
complex, and its sensitivity to hyperparameters renders achieving stable
performance and scalability challenging. Furthermore, prevailing approaches to
preference alignment primarily concentrate on pairwise comparisons, with
limited exploration into multi-response scenarios, thereby overlooking the
potential richness within the candidate pool. For the above reasons, we propose
a new approach: Listwise Reward Enhancement for Preference Alignment (LIRE), a
gradient-based reward optimization approach that incorporates the offline
rewards of multiple responses into a streamlined listwise framework, thus
eliminating the need for online sampling during training. LIRE is
straightforward to implement, requiring minimal parameter tuning, and
seamlessly aligns with the pairwise paradigm while naturally extending to
multi-response scenarios. Moreover, we introduce a self-enhancement algorithm
aimed at iteratively refining the reward during training. Our experiments
demonstrate that LIRE consistently outperforms existing methods across several
benchmarks on dialogue and summarization tasks, with good transferability to
out-of-distribution data, assessed using proxy reward models and human
annotators. | cs.CL | LIRE: listwise reward enhancement for preference alignment |
2024-05-22T10:17:30Z | http://arxiv.org/pdf/2405.13514v1 | http://arxiv.org/abs/2405.13514v1 | Muhammad Shakeel, Yui Sudo, Yifan Peng, Shinji Watanabe | 2024-05-22T10:17:30Z | eess.AS, cs.CL, cs.SD | End-to-end (E2E) automatic speech recognition (ASR) can operate in two modes:
streaming and non-streaming, each with its pros and cons. Streaming ASR
processes the speech frames in real-time as it is being received, while
non-streaming ASR waits for the entire speech utterance; thus, professionals
may have to operate in either mode to satisfy their application. In this work,
we present joint optimization of streaming and non-streaming ASR based on
multi-decoder and knowledge distillation. Primarily, we study 1) the encoder
integration of these ASR modules, followed by 2) separate decoders to make the
switching mode flexible, and enhancing performance by 3) incorporating
similarity-preserving knowledge distillation between the two modular encoders
and decoders. Evaluation results show 2.6%-5.3% relative character error rate
reductions (CERR) on CSJ for streaming ASR, and 8.3%-9.7% relative CERRs for
non-streaming ASR within a single model compared to multiple standalone
modules. | eess.AS | Joint Optimization of Streaming and Non-Streaming Automatic Speech Recognition with Multi-Decoder and Knowledge Distillation |
2024-05-22T10:12:32Z | http://arxiv.org/pdf/2405.13511v2 | http://arxiv.org/abs/2405.13511v2 | Tomás Hüttebräucker, Mohamed Sana, Emilio Calvanese Strinati | 2024-06-04T10:13:13Z | cs.LG, cs.CL, cs.IT, math.IT | We relax the constraint of a shared language between agents in a semantic and
goal-oriented communication system to explore the effect of language mismatch
in distributed task solving. We propose a mathematical framework, which
provides a modelling and a measure of the semantic distortion introduced in the
communication when agents use distinct languages. We then propose a new
approach to semantic channel equalization with proven effectiveness through
numerical evaluations. | cs.LG | Latent Space Alignment for Semantic Channel Equalization |
2024-05-22T08:38:26Z | http://arxiv.org/pdf/2405.13448v1 | http://arxiv.org/abs/2405.13448v1 | Yuanhao Yue, Chengyu Wang, Jun Huang, Peng Wang | 2024-05-22T08:38:26Z | cs.CL | The process of instruction tuning aligns pre-trained large language models
(LLMs) with open-domain instructions and human-preferred responses. While
several studies have explored autonomous approaches to distilling and
annotating instructions from more powerful proprietary LLMs, such as ChatGPT,
they often neglect the impact of task distributions and the varying difficulty
of instructions of the training sets. This oversight can lead to imbalanced
knowledge capabilities and poor generalization powers of small student LLMs. To
address this challenge, we introduce Task-Aware Curriculum Planning for
Instruction Refinement (TAPIR), a multi-round distillation framework with
balanced task distributions and dynamic difficulty adjustment. This approach
utilizes an oracle LLM to select instructions that are difficult for a student
LLM to follow and distill instructions with balanced task distributions. By
incorporating curriculum planning, our approach systematically escalates the
difficulty levels, progressively enhancing the student LLM's capabilities. We
rigorously evaluate TAPIR using two widely recognized benchmarks, including
AlpacaEval 2.0 and MT-Bench. The empirical results demonstrate that the student
LLMs, trained with our method and less training data, outperform larger
instruction-tuned models and strong distillation baselines. The improvement is
particularly notable in complex tasks, such as logical reasoning and code
generation. | cs.CL | Distilling Instruction-following Abilities of Large Language Models with Task-aware Curriculum Planning |
2024-05-22T08:18:19Z | http://arxiv.org/pdf/2405.13432v1 | http://arxiv.org/abs/2405.13432v1 | Tingchen Fu, Deng Cai, Lemao Liu, Shuming Shi, Rui Yan | 2024-05-22T08:18:19Z | cs.CL, cs.AI | Supervised fine-tuning (SFT) on instruction-following corpus is a crucial
approach toward the alignment of large language models (LLMs). However, the
performance of LLMs on standard knowledge and reasoning benchmarks tends to
suffer from deterioration at the latter stage of the SFT process, echoing the
phenomenon of alignment tax. Through our pilot study, we put a hypothesis that
the data biases are probably one cause behind the phenomenon. To address the
issue, we introduce a simple disperse-then-merge framework. To be concrete, we
disperse the instruction-following data into portions and train multiple
sub-models using different data portions. Then we merge multiple models into a
single one via model merging techniques. Despite its simplicity, our framework
outperforms various sophisticated methods such as data curation and training
regularization on a series of standard knowledge and reasoning benchmarks. | cs.CL | Disperse-Then-Merge: Pushing the Limits of Instruction Tuning via Alignment Tax Reduction |
2024-05-22T07:21:32Z | http://arxiv.org/pdf/2405.13401v3 | http://arxiv.org/abs/2405.13401v3 | Pengzhou Cheng, Yidong Ding, Tianjie Ju, Zongru Wu, Wei Du, Ping Yi, Zhuosheng Zhang, Gongshen Liu | 2024-05-31T16:59:17Z | cs.CR, cs.CL | Large language models (LLMs) have raised concerns about potential security
threats despite performing significantly in Natural Language Processing (NLP).
Backdoor attacks initially verified that LLM is doing substantial harm at all
stages, but the cost and robustness have been criticized. Attacking LLMs is
inherently risky in security review, while prohibitively expensive. Besides,
the continuous iteration of LLMs will degrade the robustness of backdoors. In
this paper, we propose TrojanRAG, which employs a joint backdoor attack in the
Retrieval-Augmented Generation, thereby manipulating LLMs in universal attack
scenarios. Specifically, the adversary constructs elaborate target contexts and
trigger sets. Multiple pairs of backdoor shortcuts are orthogonally optimized
by contrastive learning, thus constraining the triggering conditions to a
parameter subspace to improve the matching. To improve the recall of the RAG
for the target contexts, we introduce a knowledge graph to construct structured
data to achieve hard matching at a fine-grained level. Moreover, we normalize
the backdoor scenarios in LLMs to analyze the real harm caused by backdoors
from both attackers' and users' perspectives and further verify whether the
context is a favorable tool for jailbreaking models. Extensive experimental
results on truthfulness, language understanding, and harmfulness show that
TrojanRAG exhibits versatility threats while maintaining retrieval capabilities
on normal queries. | cs.CR | TrojanRAG: Retrieval-Augmented Generation Can Be Backdoor Driver in Large Language Models |
2024-05-22T06:45:38Z | http://arxiv.org/pdf/2405.13386v1 | http://arxiv.org/abs/2405.13386v1 | 360Zhinao Team | 2024-05-22T06:45:38Z | cs.CL, cs.AI | We present 360Zhinao models with 7B parameter size and context lengths
spanning 4K, 32K and 360K, all available at
https://github.com/Qihoo360/360zhinao. For rapid development in pretraining, we
establish a stable and sensitive ablation environment to evaluate and compare
experiment runs with minimal model size. Under such guidance, we perfect our
data cleaning and composition strategies to pretrain
$\texttt{360Zhinao-7B-Base}$ on 3.4T tokens. We also mainly emphasize data
during alignment, where we strive to balance quantity and quality with
filtering and reformatting. With tailored data, 360Zhinao-7B's context window
is easily extended to 32K and 360K. RMs and RLHF are trained following SFT and
credibly applied to specific tasks. All together these contributions lead to
360Zhinao-7B's competitive performance among models of similar size. | cs.CL | 360Zhinao Technical Report |
2024-05-22T06:24:55Z | http://arxiv.org/pdf/2405.13379v1 | http://arxiv.org/abs/2405.13379v1 | Ronald Cumbal, Birger Moell, Jose Lopes, Olof Engwall | 2024-05-22T06:24:55Z | cs.CL, cs.SD, eess.AS | The performance of Automatic Speech Recognition (ASR) systems has constantly
increased in state-of-the-art development. However, performance tends to
decrease considerably in more challenging conditions (e.g., background noise,
multiple speaker social conversations) and with more atypical speakers (e.g.,
children, non-native speakers or people with speech disorders), which signifies
that general improvements do not necessarily transfer to applications that rely
on ASR, e.g., educational software for younger students or language learners.
In this study, we focus on the gap in performance between recognition results
for native and non-native, read and spontaneous, Swedish utterances transcribed
by different ASR services. We compare the recognition results using Word Error
Rate and analyze the linguistic factors that may generate the observed
transcription errors. | cs.CL | You don't understand me!: Comparing ASR results for L1 and L2 speakers of Swedish |
2024-05-22T05:32:11Z | http://arxiv.org/pdf/2405.13358v1 | http://arxiv.org/abs/2405.13358v1 | Alireza Ghaffari, Sharareh Younesian, Vahid Partovi Nia, Boxing Chen, Masoud Asgharian | 2024-05-22T05:32:11Z | cs.CL | The ever-growing computational complexity of Large Language Models (LLMs)
necessitates efficient deployment strategies. The current state-of-the-art
approaches for Post-training Quantization (PTQ) often require calibration to
achieve the desired accuracy. This paper presents AdpQ, a novel zero-shot
adaptive PTQ method for LLMs that achieves the state-of-the-art performance in
low-precision quantization (e.g. 3-bit) without requiring any calibration data.
Inspired by Adaptive LASSO regression model, our proposed approach tackles the
challenge of outlier activations by separating salient weights using an
adaptive soft-thresholding method. Guided by Adaptive LASSO, this method
ensures that the quantized weights distribution closely follows the originally
trained weights and eliminates the need for calibration data entirely, setting
our method apart from popular approaches such as SpQR and AWQ. Furthermore, our
method offers an additional benefit in terms of privacy preservation by
eliminating any calibration or training data. We also delve deeper into the
information-theoretic underpinnings of the proposed method. We demonstrate that
it leverages the Adaptive LASSO to minimize the Kullback-Leibler divergence
between the quantized weights and the originally trained weights. This
minimization ensures the quantized model retains the Shannon information
content of the original model to a great extent, guaranteeing efficient
deployment without sacrificing accuracy or information. Our results achieve the
same accuracy as the existing methods on various LLM benchmarks while the
quantization time is reduced by at least 10x, solidifying our contribution to
efficient and privacy-preserving LLM deployment. | cs.CL | AdpQ: A Zero-shot Calibration Free Adaptive Post Training Quantization Method for LLMs |
2024-05-22T05:12:35Z | http://arxiv.org/pdf/2405.13350v2 | http://arxiv.org/abs/2405.13350v2 | Corinne Aars, Lauren Adams, Xiaokan Tian, Zhaoyu Wang, Colton Wismer, Jason Wu, Pablo Rivas, Korn Sooksatra, Matthew Fendt | 2024-05-30T18:42:45Z | cs.CL, cs.LG, I.2.7 | This study presents the development and evaluation of a ByT5-based
multilingual translation model tailored for translating the Bible into
underrepresented languages. Utilizing the comprehensive Johns Hopkins
University Bible Corpus, we trained the model to capture the intricate nuances
of character-based and morphologically rich languages. Our results, measured by
the BLEU score and supplemented with sample translations, suggest the model can
improve accessibility to sacred texts. It effectively handles the distinctive
biblical lexicon and structure, thus bridging the linguistic divide. The study
also discusses the model's limitations and suggests pathways for future
enhancements, focusing on expanding access to sacred literature across
linguistic boundaries. | cs.CL | Efficacy of ByT5 in Multilingual Translation of Biblical Texts for Underrepresented Languages |
2024-05-22T05:03:39Z | http://arxiv.org/pdf/2405.13344v1 | http://arxiv.org/abs/2405.13344v1 | Yui Sudo, Yosuke Fukumoto, Muhammad Shakeel, Yifan Peng, Shinji Watanabe | 2024-05-22T05:03:39Z | eess.AS, cs.CL, cs.SD | Deep biasing (DB) improves the performance of end-to-end automatic speech
recognition (E2E-ASR) for rare words or contextual phrases using a bias list.
However, most existing methods treat bias phrases as sequences of subwords in a
predefined static vocabulary, which can result in ineffective learning of the
dependencies between subwords. More advanced techniques address this problem by
incorporating additional text data, which increases the overall workload. This
paper proposes a dynamic vocabulary where phrase-level bias tokens can be added
during the inference phase. Each bias token represents an entire bias phrase
within a single token, thereby eliminating the need to learn the dependencies
between the subwords within the bias phrases. This method can be applied to
various architectures because it only extends the embedding and output layers
in common E2E-ASR architectures. Experimental results demonstrate that the
proposed method improves the performance of bias phrases on English and
Japanese datasets. | eess.AS | Contextualized Automatic Speech Recognition with Dynamic Vocabulary |
2024-05-22T04:15:41Z | http://arxiv.org/pdf/2405.13329v1 | http://arxiv.org/abs/2405.13329v1 | Nithin Parthasarathy, James Soetedjo, Saarang Panchavati, Nitya Parthasarathy, Corey Arnold, Nader Pouratian, William Speier | 2024-05-22T04:15:41Z | cs.CL, cs.HC, cs.SY, eess.SP, eess.SY | Amyotrophic lateral sclerosis (ALS) severely impairs patients' ability to
communicate, often leading to a decline in their quality of life within a few
years of diagnosis. The P300 speller brain-computer interface (BCI) offers an
alternative communication method by interpreting a subject's EEG response to
characters presented on a grid interface.
This paper addresses the common speed limitations encountered in training
efficient P300-based multi-subject classifiers by introducing innovative
"across-subject" classifiers. We leverage a combination of the
second-generation Generative Pre-Trained Transformer (GPT2) and Dijkstra's
algorithm to optimize stimuli and suggest word completion choices based on
typing history. Additionally, we employ a multi-layered smoothing technique to
accommodate out-of-vocabulary (OOV) words.
Through extensive simulations involving random sampling of EEG data from
subjects, we demonstrate significant speed enhancements in typing passages
containing rare and OOV words. These optimizations result in approximately 10%
improvement in character-level typing speed and up to 40% improvement in
multi-word prediction. We demonstrate that augmenting standard row/column
highlighting techniques with layered word prediction yields close-to-optimal
performance.
Furthermore, we explore both "within-subject" and "across-subject" training
techniques, showing that speed improvements are consistent across both
approaches. | cs.CL | High Performance P300 Spellers Using GPT2 Word Prediction With Cross-Subject Training |
2024-05-22T04:08:20Z | http://arxiv.org/pdf/2405.13326v1 | http://arxiv.org/abs/2405.13326v1 | Ming Li, Pei Chen, Chenguang Wang, Hongyu Zhao, Yijun Liang, Yupeng Hou, Fuxiao Liu, Tianyi Zhou | 2024-05-22T04:08:20Z | cs.CL | Finetuning large language models with a variety of instruction-response pairs
has enhanced their capability to understand and follow instructions. Current
instruction tuning primarily relies on teacher models or human intervention to
generate and refine the instructions and responses, which are costly,
non-sustainable, and may lack diversity. In this paper, we introduce Mosaic
Instruction Tuning (Mosaic-IT), a human/model-free method that can efficiently
create rich and diverse augmentations from existing instruction tuning data to
enhance the finetuned LLM.Mosaic-IT randomly concatenates multiple instruction
data into one and trains the model to produce the corresponding responses with
predefined higher-level meta-instructions to strengthen its multi-step
instruction-following and format-following skills. Our extensive evaluations
demonstrate a superior performance and training efficiency of Mosaic-IT, which
achieves consistent performance improvements over various benchmarks and an 80%
reduction in training costs compared with original instruction tuning. Our
codes and data are available at https://github.com/tianyi-lab/Mosaic-IT. | cs.CL | Mosaic IT: Enhancing Instruction Tuning with Data Mosaics |
2024-05-22T03:56:55Z | http://arxiv.org/pdf/2405.13325v2 | http://arxiv.org/abs/2405.13325v2 | Guanghui Wang, Dexi Liu, Jian-Yun Nie, Qizhi Wan, Rong Hu, Xiping Liu, Wanlong Liu, Jiaming Liu | 2024-06-15T15:40:49Z | cs.CL, cs.AI, cs.IR | Recent advancements in event argument extraction (EAE) involve incorporating
useful auxiliary information into models during training and inference, such as
retrieved instances and event templates. These methods face two challenges: (1)
the retrieval results may be irrelevant and (2) templates are developed
independently for each event without considering their possible relationship.
In this work, we propose DEGAP to address these challenges through a simple yet
effective components: dual prefixes, i.e. learnable prompt vectors, where the
instance-oriented prefix and template-oriented prefix are trained to learn
information from different event instances and templates. Additionally, we
propose an event-guided adaptive gating mechanism, which can adaptively
leverage possible connections between different events and thus capture
relevant information from the prefix. Finally, these event-guided prefixes
provide relevant information as cues to EAE model without retrieval. Extensive
experiments demonstrate that our method achieves new state-of-the-art
performance on four datasets (ACE05, RAMS, WIKIEVENTS, and MLEE). Further
analysis shows the impact of different components. | cs.CL | DEGAP: Dual Event-Guided Adaptive Prefixes for Templated-Based Event Argument Extraction with Slot Querying |
2024-05-22T03:25:35Z | http://arxiv.org/pdf/2405.13319v1 | http://arxiv.org/abs/2405.13319v1 | Denys Katerenchuk, Rivka Levitan | 2024-05-22T03:25:35Z | cs.CL, cs.LG | Humans express ideas, beliefs, and statements through language. The manner of
expression can carry information indicating the author's degree of confidence
in their statement. Understanding the certainty level of a claim is crucial in
areas such as medicine, finance, engineering, and many others where errors can
lead to disastrous results. In this work, we apply a joint model that leverages
words and part-of-speech tags to improve hedge detection in text and achieve a
new top score on the CoNLL-2010 Wikipedia corpus. | cs.CL | ''You should probably read this'': Hedge Detection in Text |
2024-05-22T02:19:13Z | http://arxiv.org/pdf/2405.13292v1 | http://arxiv.org/abs/2405.13292v1 | Co Van Dinh, Son T. Luu | 2024-05-22T02:19:13Z | cs.CL | The problem of detecting spam reviews (opinions) has received significant
attention in recent years, especially with the rapid development of e-commerce.
Spam reviews are often classified based on comment content, but in some cases,
it is insufficient for models to accurately determine the review label. In this
work, we introduce the ViSpamReviews v2 dataset, which includes metadata of
reviews with the objective of integrating supplementary attributes for spam
review classification. We propose a novel approach to simultaneously integrate
both textual and categorical attributes into the classification model. In our
experiments, the product category proved effective when combined with deep
neural network (DNN) models, while text features performed well on both DNN
models and the model achieved state-of-the-art performance in the problem of
detecting spam reviews on Vietnamese e-commerce websites, namely PhoBERT.
Specifically, the PhoBERT model achieves the highest accuracy when combined
with product description features generated from the SPhoBert model, which is
the combination of PhoBERT and SentenceBERT. Using the macro-averaged F1 score,
the task of classifying spam reviews achieved 87.22% (an increase of 1.64%
compared to the baseline), while the task of identifying the type of spam
reviews achieved an accuracy of 73.49% (an increase of 1.93% compared to the
baseline). | cs.CL | Metadata Integration for Spam Reviews Detection on Vietnamese E-commerce Websites |
2024-05-22T02:15:57Z | http://arxiv.org/pdf/2406.00014v2 | http://arxiv.org/abs/2406.00014v2 | Hajung Kim, Chanhwi Kim, Hoonick Lee, Kyochul Jang, Jiwoo Lee, Kyungjae Lee, Gangwoo Kim, Jaewoo Kang | 2024-06-19T16:21:46Z | cs.DB, cs.AI, cs.CL, cs.IR | Transforming natural language questions into SQL queries is crucial for
precise data retrieval from electronic health record (EHR) databases. A
significant challenge in this process is detecting and rejecting unanswerable
questions that request information beyond the database's scope or exceed the
system's capabilities. In this paper, we introduce a novel text-to-SQL
framework that robustly handles out-of-domain questions and verifies the
generated queries with query execution.Our framework begins by standardizing
the structure of questions into a templated format. We use a powerful large
language model (LLM), fine-tuned GPT-3.5 with detailed prompts involving the
table schemas of the EHR database system. Our experimental results demonstrate
the effectiveness of our framework on the EHRSQL-2024 benchmark benchmark, a
shared task in the ClinicalNLP workshop. Although a straightforward fine-tuning
of GPT shows promising results on the development set, it struggled with the
out-of-domain questions in the test set. With our framework, we improve our
system's adaptability and achieve competitive performances in the official
leaderboard of the EHRSQL-2024 challenge. | cs.DB | KU-DMIS at EHRSQL 2024:Generating SQL query via question templatization in EHR |
2024-05-22T01:10:39Z | http://arxiv.org/pdf/2405.13274v1 | http://arxiv.org/abs/2405.13274v1 | Weiting Tan, Jingyu Zhang, Lingfeng Shen, Daniel Khashabi, Philipp Koehn | 2024-05-22T01:10:39Z | cs.CL | Non-autoregressive Transformers (NATs) are recently applied in direct
speech-to-speech translation systems, which convert speech across different
languages without intermediate text data. Although NATs generate high-quality
outputs and offer faster inference than autoregressive models, they tend to
produce incoherent and repetitive results due to complex data distribution
(e.g., acoustic and linguistic variations in speech). In this work, we
introduce DiffNorm, a diffusion-based normalization strategy that simplifies
data distributions for training NAT models. After training with a
self-supervised noise estimation objective, DiffNorm constructs normalized
target data by denoising synthetically corrupted speech features. Additionally,
we propose to regularize NATs with classifier-free guidance, improving model
robustness and translation quality by randomly dropping out source information
during training. Our strategies result in a notable improvement of about +7
ASR-BLEU for English-Spanish (En-Es) and +2 ASR-BLEU for English-French (En-Fr)
translations on the CVSS benchmark, while attaining over 14x speedup for En-Es
and 5x speedup for En-Fr translations compared to autoregressive baselines. | cs.CL | DiffNorm: Self-Supervised Normalization for Non-autoregressive Speech-to-speech Translation |
2024-05-22T01:01:04Z | http://arxiv.org/pdf/2405.13272v1 | http://arxiv.org/abs/2405.13272v1 | Xi Chen, Mattia Samory, Scott Hale, David Jurgens, Przemyslaw A. Grabowicz | 2024-05-22T01:01:04Z | cs.CL, cs.CY | Understanding the writing frame of news articles is vital for addressing
social issues, and thus has attracted notable attention in the fields of
communication studies. Yet, assessing such news article frames remains a
challenge due to the absence of a concrete and unified standard dataset that
considers the comprehensive nuances within news content.
To address this gap, we introduce an extended version of a large labeled news
article dataset with 16,687 new labeled pairs. Leveraging the pairwise
comparison of news articles, our method frees the work of manual identification
of frame classes in traditional news frame analysis studies. Overall we
introduce the most extensive cross-lingual news article similarity dataset
available to date with 26,555 labeled news article pairs across 10 languages.
Each data point has been meticulously annotated according to a codebook
detailing eight critical aspects of news content, under a human-in-the-loop
framework. Application examples demonstrate its potential in unearthing country
communities within global news coverage, exposing media bias among news
outlets, and quantifying the factors related to news creation. We envision that
this news similarity dataset will broaden our understanding of the media
ecosystem in terms of news coverage of events and perspectives across
countries, locations, languages, and other social constructs. By doing so, it
can catalyze advancements in social science research and applied methodologies,
thereby exerting a profound impact on our society. | cs.CL | A Multilingual Similarity Dataset for News Article Frame |
2024-05-21T23:12:03Z | http://arxiv.org/pdf/2405.13245v2 | http://arxiv.org/abs/2405.13245v2 | Vanya Cohen, Jason Xinyu Liu, Raymond Mooney, Stefanie Tellex, David Watkins | 2024-06-22T13:03:53Z | cs.RO, cs.AI, cs.CL | With large language models, robots can understand language more flexibly and
more capable than ever before. This survey reviews and situates recent
literature into a spectrum with two poles: 1) mapping between language and some
manually defined formal representation of meaning, and 2) mapping between
language and high-dimensional vector spaces that translate directly to
low-level robot policy. Using a formal representation allows the meaning of the
language to be precisely represented, limits the size of the learning problem,
and leads to a framework for interpretability and formal safety guarantees.
Methods that embed language and perceptual data into high-dimensional spaces
avoid this manually specified symbolic structure and thus have the potential to
be more general when fed enough data but require more data and computing to
train. We discuss the benefits and tradeoffs of each approach and finish by
providing directions for future work that achieves the best of both worlds. | cs.RO | A Survey of Robotic Language Grounding: Tradeoffs between Symbols and Embeddings |
2024-05-21T22:40:38Z | http://arxiv.org/pdf/2405.13233v1 | http://arxiv.org/abs/2405.13233v1 | Sirou Chen, Sakiko Yahata, Shuichiro Shimizu, Zhengdong Yang, Yihang Li, Chenhui Chu, Sadao Kurohashi | 2024-05-21T22:40:38Z | cs.CL | Emotion plays a crucial role in human conversation. This paper underscores
the significance of considering emotion in speech translation. We present the
MELD-ST dataset for the emotion-aware speech translation task, comprising
English-to-Japanese and English-to-German language pairs. Each language pair
includes about 10,000 utterances annotated with emotion labels from the MELD
dataset. Baseline experiments using the SeamlessM4T model on the dataset
indicate that fine-tuning with emotion labels can enhance translation
performance in some settings, highlighting the need for further research in
emotion-aware speech translation systems. | cs.CL | MELD-ST: An Emotion-aware Speech Translation Dataset |
2024-05-21T22:26:01Z | http://arxiv.org/pdf/2405.13226v1 | http://arxiv.org/abs/2405.13226v1 | Hadi Pouransari, Chun-Liang Li, Jen-Hao Rick Chang, Pavan Kumar Anasosalu Vasu, Cem Koc, Vaishaal Shankar, Oncel Tuzel | 2024-05-21T22:26:01Z | cs.CL, cs.LG | Large language models (LLMs) are commonly trained on datasets consisting of
fixed-length token sequences. These datasets are created by randomly
concatenating documents of various lengths and then chunking them into
sequences of a predetermined target length. However, this method of
concatenation can lead to cross-document attention within a sequence, which is
neither a desirable learning signal nor computationally efficient.
Additionally, training on long sequences becomes computationally prohibitive
due to the quadratic cost of attention. In this study, we introduce dataset
decomposition, a novel variable sequence length training technique, to tackle
these challenges. We decompose a dataset into a union of buckets, each
containing sequences of the same size extracted from a unique document. During
training, we use variable sequence length and batch size, sampling
simultaneously from all buckets with a curriculum. In contrast to the
concat-and-chunk baseline, which incurs a fixed attention cost at every step of
training, our proposed method incurs a penalty proportional to the actual
document lengths at each step, resulting in significant savings in training
time. We train an 8k context-length 1B model at the same cost as a 2k
context-length model trained with the baseline approach. Experiments on a
web-scale corpus demonstrate that our approach significantly enhances
performance on standard language evaluations and long-context benchmarks,
reaching target accuracy 3x faster compared to the baseline. Our method not
only enables efficient pretraining on long sequences but also scales
effectively with dataset size. Lastly, we shed light on a critical yet less
studied aspect of training large language models: the distribution and
curriculum of sequence lengths, which results in a non-negligible difference in
performance. | cs.CL | Dataset Decomposition: Faster LLM Training with Variable Sequence Length Curriculum |
2024-05-21T22:00:13Z | http://arxiv.org/pdf/2405.13219v1 | http://arxiv.org/abs/2405.13219v1 | Ayesha Siddika Nipu, K M Sajjadul Islam, Praveen Madiraju | 2024-05-21T22:00:13Z | cs.AI, cs.CL | Artificial Intelligence (AI) chatbots leveraging Large Language Models (LLMs)
are gaining traction in healthcare for their potential to automate patient
interactions and aid clinical decision-making. This study examines the
reliability of AI chatbots, specifically GPT 4.0, Claude 3 Opus, and Gemini
Ultra 1.0, in predicting diseases from patient complaints in the emergency
department. The methodology includes few-shot learning techniques to evaluate
the chatbots' effectiveness in disease prediction. We also fine-tune the
transformer-based model BERT and compare its performance with the AI chatbots.
Results suggest that GPT 4.0 achieves high accuracy with increased few-shot
data, while Gemini Ultra 1.0 performs well with fewer examples, and Claude 3
Opus maintains consistent performance. BERT's performance, however, is lower
than all the chatbots, indicating limitations due to limited labeled data.
Despite the chatbots' varying accuracy, none of them are sufficiently reliable
for critical medical decision-making, underscoring the need for rigorous
validation and human oversight. This study reflects that while AI chatbots have
potential in healthcare, they should complement, not replace, human expertise
to ensure patient safety. Further refinement and research are needed to improve
AI-based healthcare applications' reliability for disease prediction. | cs.AI | How Reliable AI Chatbots are for Disease Prediction from Patient Complaints? |
2024-05-21T21:41:07Z | http://arxiv.org/pdf/2405.13216v1 | http://arxiv.org/abs/2405.13216v1 | Chenghao Yang, Zi Yang, Nan Hua | 2024-05-21T21:41:07Z | cs.CL | Long-context modeling presents a significant challenge for transformer-based
large language models (LLMs) due to the quadratic complexity of the
self-attention mechanism and issues with length extrapolation caused by
pretraining exclusively on short inputs. Existing methods address computational
complexity through techniques such as text chunking, the kernel approach, and
structured attention, and tackle length extrapolation problems through
positional encoding, continued pretraining, and data engineering. These
approaches typically require $\textbf{sequential access}$ to the document,
necessitating reading from the first to the last token. We contend that for
goal-oriented reading of long documents, such sequential access is not
necessary, and a proficiently trained model can learn to omit hundreds of less
pertinent tokens. Inspired by human reading behaviors and existing empirical
observations, we propose $\textbf{random access}$, a novel reading strategy
that enables transformers to efficiently process long documents without
examining every token. Experimental results from pretraining, fine-tuning, and
inference phases validate the efficacy of our method. | cs.CL | Equipping Transformer with Random-Access Reading for Long-Context Understanding |
2024-05-21T21:24:34Z | http://arxiv.org/pdf/2405.13209v1 | http://arxiv.org/abs/2405.13209v1 | Neisarg Dave, Daniel Kifer, C. Lee Giles, Ankur Mali | 2024-05-21T21:24:34Z | cs.CL, cs.LG | Prompting techniques have significantly enhanced the capabilities of Large
Language Models (LLMs) across various complex tasks, including reasoning,
planning, and solving math word problems. However, most research has
predominantly focused on language-based reasoning and word problems, often
overlooking the potential of LLMs in handling symbol-based calculations and
reasoning. This study aims to bridge this gap by rigorously evaluating LLMs on
a series of symbolic tasks, such as addition, multiplication, modulus
arithmetic, numerical precision, and symbolic counting. Our analysis
encompasses eight LLMs, including four enterprise-grade and four open-source
models, of which three have been pre-trained on mathematical tasks. The
assessment framework is anchored in Chomsky's Hierarchy, providing a robust
measure of the computational abilities of these models. The evaluation employs
minimally explained prompts alongside the zero-shot Chain of Thoughts
technique, allowing models to navigate the solution process autonomously. The
findings reveal a significant decline in LLMs' performance on context-free and
context-sensitive symbolic tasks as the complexity, represented by the number
of symbols, increases. Notably, even the fine-tuned GPT3.5 exhibits only
marginal improvements, mirroring the performance trends observed in other
models. Across the board, all models demonstrated a limited generalization
ability on these symbol-intensive tasks. This research underscores LLMs'
challenges with increasing symbolic complexity and highlights the need for
specialized training, memory and architectural adjustments to enhance their
proficiency in symbol-based reasoning tasks. | cs.CL | Investigating Symbolic Capabilities of Large Language Models |
2024-05-21T21:14:31Z | http://arxiv.org/pdf/2405.13203v1 | http://arxiv.org/abs/2405.13203v1 | Garrett Tanzer, Gustaf Ahdritz, Luke Melas-Kyriazi | 2024-05-21T21:14:31Z | cs.LG, cs.CL | Chatbots built upon language models have exploded in popularity, but they
have largely been limited to synchronous, turn-by-turn dialogues. In this paper
we present a simple yet general method to simulate real-time interactive
conversations using pretrained text-only language models, by modeling timed
diarized transcripts and decoding them with causal rejection sampling. We
demonstrate the promise of this method with two case studies: instant messenger
dialogues and spoken conversations, which require generation at about 30 tok/s
and 20 tok/s respectively to maintain real-time interactivity. These
capabilities can be added into language models using relatively little data and
run on commodity hardware. | cs.LG | Modeling Real-Time Interactive Conversations as Timed Diarized Transcripts |
2024-05-21T20:08:52Z | http://arxiv.org/pdf/2405.13181v1 | http://arxiv.org/abs/2405.13181v1 | Krishna Prasad Varadarajan Srinivasan, Prasanth Gumpena, Madhusudhana Yattapu, Vishal H. Brahmbhatt | 2024-05-21T20:08:52Z | cs.CL, cs.LG | In the domain of large language models (LLMs), arXiv:2305.16938 showed that
few-shot full-model fine-tuning -- namely Vanilla Fine Tuning (FT) and
Pattern-Based Fine Tuning (PBFT) --, and In-Context Learning (ICL) generalize
similarly on Out-Of-Domain (OOD) datasets, but vary in terms of task
adaptation. However, they both pose challenges, especially in term of memory
requirements. In this paper, we further try to push the understanding of
different fine-tuning strategies for LLM and aim to bring a myriad of these on
the same pedestal for an elaborate comparison with full-model fine-tuning on
two diverse datasets. To that end, we conducted a series of experiments,
beginning with state-of-the-art methods like vanilla fine-tuning and
Pattern-Based Fine-Tuning (PBFT) on pre-trained models across two datasets,
COLA and MNLI. We then investigate adaptive fine-tuning and the efficiency of
LoRA adapters in a few-shot setting. Finally, we also compare an alternative
approach that has gained recent popularity -- context distillation -- with the
vanilla FT and PBFT with and without few-shot setup.
Our findings suggest that these alternative strategies that we explored can
exhibit out-of-domain generalization comparable to that of vanilla FT and PBFT.
PBFT under-performs Vanilla FT on out-of-domain (OOD) data, emphasizing the
need for effective prompts. Further, our adaptive-fine tuning and LoRA
experiments perform comparable or slightly worse than the standard fine-tunings
as anticipated, since standard fine-tunings involve tuning the entire model.
Finally, our context distillation experiments out-perform the standard
fine-tuning methods. These findings underscore that eventually the choice of an
appropriate fine-tuning method depends on the available resources (memory,
compute, data) and task adaptability. | cs.CL | Comparative Analysis of Different Efficient Fine Tuning Methods of Large Language Models (LLMs) in Low-Resource Setting |
2024-05-21T20:03:40Z | http://arxiv.org/pdf/2405.13179v4 | http://arxiv.org/abs/2405.13179v4 | Yuelyu Ji, Zhuochun Li, Rui Meng, Sonish Sivarajkumar, Yanshan Wang, Zeshui Yu, Hui Ji, Yushui Han, Hanyu Zeng, Daqing He | 2024-06-24T18:13:16Z | cs.CL | This paper introduces the RAG-RLRC-LaySum framework, designed to make complex
biomedical research understandable to laymen through advanced Natural Language
Processing (NLP) techniques. Our Retrieval Augmented Generation (RAG) solution,
enhanced by a reranking method, utilizes multiple knowledge sources to ensure
the precision and pertinence of lay summaries. Additionally, our Reinforcement
Learning for Readability Control (RLRC) strategy improves readability, making
scientific content comprehensible to non-specialists. Evaluations using the
publicly accessible PLOS and eLife datasets show that our methods surpass Plain
Gemini model, demonstrating a 20% increase in readability scores, a 15%
improvement in ROUGE-2 relevance scores, and a 10% enhancement in factual
accuracy. The RAG-RLRC-LaySum framework effectively democratizes scientific
knowledge, enhancing public engagement with biomedical discoveries. | cs.CL | RAG-RLRC-LaySum at BioLaySumm: Integrating Retrieval-Augmented Generation and Readability Control for Layman Summarization of Biomedical Texts |
2024-05-21T18:29:54Z | http://arxiv.org/pdf/2405.13144v2 | http://arxiv.org/abs/2405.13144v2 | Xuhan Huang, Qingning Shen, Yan Hu, Anningzhe Gao, Benyou Wang | 2024-06-30T05:42:24Z | cs.AI, cs.CL | Mathematical modeling involves representing real-world phenomena, systems, or
problems using mathematical expressions and equations to analyze, understand,
and predict their behavior. Given that this process typically requires
experienced experts, there is an interest in exploring whether Large Language
Models (LLMs) can undertake mathematical modeling to potentially decrease human
labor. To evaluate of LLMs in mathematical modeling, we introduce a new
benchmark, Mamo, that transcends traditional result-oriented assessments.
Unlike conventional methods that primarily assess LLMs based on the accuracy of
solutions to mathematical problems, our approach offers deeper insight into the
modeling process itself. By focusing on the processes LLMs undertake rather
than the correctness of their final solutions, Mamo pioneers a novel evaluation
paradigm. This shift underscores the importance of understanding the inherent
modeling capabilities of LLMs, paving the way for a more nuanced and
comprehensive analysis of their problem-solving strategies. Our work marks a
significant advancement in the field, suggesting a new direction for future
research by emphasizing the evaluation of LLMs' modeling processes over the
mere correctness of answers. This benchmark not only facilitates a better
understanding of LLMs' mathematical modeling capabilities but also sets a new
standard for evaluating their performance in complex problem-solving scenarios. | cs.AI | Mamo: a Mathematical Modeling Benchmark with Solvers |
2024-05-21T18:12:37Z | http://arxiv.org/pdf/2405.13135v1 | http://arxiv.org/abs/2405.13135v1 | Tong Zeng, Daniel Acuna | 2024-05-21T18:12:37Z | cs.CL, cs.LG | Datasets are critical for scientific research, playing an important role in
replication, reproducibility, and efficiency. Researchers have recently shown
that datasets are becoming more important for science to function properly,
even serving as artifacts of study themselves. However, citing datasets is not
a common or standard practice in spite of recent efforts by data repositories
and funding agencies. This greatly affects our ability to track their usage and
importance. A potential solution to this problem is to automatically extract
dataset mentions from scientific articles. In this work, we propose to achieve
such extraction by using a neural network based on a Bi-LSTM-CRF architecture.
Our method achieves F1 = 0.885 in social science articles released as part of
the Rich Context Dataset. We discuss the limitations of the current datasets
and propose modifications to the model to be done in the future. | cs.CL | Dataset Mention Extraction in Scientific Articles Using Bi-LSTM-CRF Model |
2024-05-21T18:05:44Z | http://arxiv.org/pdf/2405.13131v1 | http://arxiv.org/abs/2405.13131v1 | Raghuveer Thirukovalluru, Yukun Huang, Bhuwan Dhingra | 2024-05-21T18:05:44Z | cs.CL | Recent work has aimed to improve LLM generations by filtering out
hallucinations, thereby improving the precision of the information in
responses. Correctness of a long-form response, however, also depends on the
recall of multiple pieces of information relevant to the question. In this
paper, we introduce Atomic Self-Consistency (ASC), a technique for improving
the recall of relevant information in an LLM response. ASC follows recent work,
Universal Self-Consistency (USC) in using multiple stochastic samples from an
LLM to improve the long-form response. Unlike USC which only focuses on
selecting the best single generation, ASC picks authentic subparts from the
samples and merges them into a superior composite answer. Through extensive
experiments and ablations, we show that merging relevant subparts of multiple
samples performs significantly better than picking a single sample. ASC
demonstrates significant gains over USC on multiple factoids and open-ended QA
datasets - ASQA, QAMPARI, QUEST, ELI5 with ChatGPT and Llama2. Our analysis
also reveals untapped potential for enhancing long-form generations using
approach of merging multiple samples. | cs.CL | Atomic Self-Consistency for Better Long Form Generations |
2024-05-21T18:02:07Z | http://arxiv.org/pdf/2405.13127v1 | http://arxiv.org/abs/2405.13127v1 | Sara Sarto, Marcella Cornia, Lorenzo Baraldi, Alessandro Nicolosi, Rita Cucchiara | 2024-05-21T18:02:07Z | cs.CV, cs.AI, cs.CL, cs.MM | The objective of image captioning models is to bridge the gap between the
visual and linguistic modalities by generating natural language descriptions
that accurately reflect the content of input images. In recent years,
researchers have leveraged deep learning-based models and made advances in the
extraction of visual features and the design of multimodal connections to
tackle this task. This work presents a novel approach towards developing image
captioning models that utilize an external kNN memory to improve the generation
process. Specifically, we propose two model variants that incorporate a
knowledge retriever component that is based on visual similarities, a
differentiable encoder to represent input images, and a kNN-augmented language
model to predict tokens based on contextual cues and text retrieved from the
external memory. We experimentally validate our approach on COCO and nocaps
datasets and demonstrate that incorporating an explicit external memory can
significantly enhance the quality of captions, especially with a larger
retrieval corpus. This work provides valuable insights into retrieval-augmented
captioning models and opens up new avenues for improving image captioning at a
larger scale. | cs.CV | Towards Retrieval-Augmented Architectures for Image Captioning |
2024-05-21T17:59:29Z | http://arxiv.org/pdf/2405.12981v1 | http://arxiv.org/abs/2405.12981v1 | William Brandon, Mayank Mishra, Aniruddha Nrusimha, Rameswar Panda, Jonathan Ragan Kelly | 2024-05-21T17:59:29Z | cs.LG, cs.CL | Key-value (KV) caching plays an essential role in accelerating decoding for
transformer-based autoregressive large language models (LLMs). However, the
amount of memory required to store the KV cache can become prohibitive at long
sequence lengths and large batch sizes. Since the invention of the transformer,
two of the most effective interventions discovered for reducing the size of the
KV cache have been Multi-Query Attention (MQA) and its generalization,
Grouped-Query Attention (GQA). MQA and GQA both modify the design of the
attention block so that multiple query heads can share a single key/value head,
reducing the number of distinct key/value heads by a large factor while only
minimally degrading accuracy. In this paper, we show that it is possible to
take Multi-Query Attention a step further by also sharing key and value heads
between adjacent layers, yielding a new attention design we call Cross-Layer
Attention (CLA). With CLA, we find that it is possible to reduce the size of
the KV cache by another 2x while maintaining nearly the same accuracy as
unmodified MQA. In experiments training 1B- and 3B-parameter models from
scratch, we demonstrate that CLA provides a Pareto improvement over the
memory/accuracy tradeoffs which are possible with traditional MQA, enabling
inference with longer sequence lengths and larger batch sizes than would
otherwise be possible | cs.LG | Reducing Transformer Key-Value Cache Size with Cross-Layer Attention |
2024-05-21T17:12:19Z | http://arxiv.org/pdf/2405.12939v1 | http://arxiv.org/abs/2405.12939v1 | Zhangyue Yin, Qiushi Sun, Qipeng Guo, Zhiyuan Zeng, Xiaonan Li, Tianxiang Sun, Cheng Chang, Qinyuan Cheng, Ding Wang, Xiaofeng Mou, Xipeng Qiu, XuanJing Huang | 2024-05-21T17:12:19Z | cs.CL | Recent advancements in Chain-of-Thought prompting have facilitated
significant breakthroughs for Large Language Models (LLMs) in complex reasoning
tasks. Current research enhances the reasoning performance of LLMs by sampling
multiple reasoning chains and ensembling based on the answer frequency.
However, this approach fails in scenarios where the correct answers are in the
minority. We identify this as a primary factor constraining the reasoning
capabilities of LLMs, a limitation that cannot be resolved solely based on the
predicted answers. To address this shortcoming, we introduce a hierarchical
reasoning aggregation framework AoR (Aggregation of Reasoning), which selects
answers based on the evaluation of reasoning chains. Additionally, AoR
incorporates dynamic sampling, adjusting the number of reasoning chains in
accordance with the complexity of the task. Experimental results on a series of
complex reasoning tasks show that AoR outperforms prominent ensemble methods.
Further analysis reveals that AoR not only adapts various LLMs but also
achieves a superior performance ceiling when compared to current methods. | cs.CL | Aggregation of Reasoning: A Hierarchical Framework for Enhancing Answer Selection in Large Language Models |
2024-05-21T17:04:44Z | http://arxiv.org/pdf/2405.12933v2 | http://arxiv.org/abs/2405.12933v2 | Bilgehan Sel, Priya Shanmugasundaram, Mohammad Kachuee, Kun Zhou, Ruoxi Jia, Ming Jin | 2024-06-02T18:48:56Z | cs.CL, cs.AI, cs.LG | Large Language Models (LLMs) have shown remarkable capabilities in tasks such
as summarization, arithmetic reasoning, and question answering. However, they
encounter significant challenges in the domain of moral reasoning and ethical
decision-making, especially in complex scenarios with multiple stakeholders.
This paper introduces the Skin-in-the-Game (SKIG) framework, aimed at enhancing
moral reasoning in LLMs by exploring decisions' consequences from multiple
stakeholder perspectives. Central to SKIG's mechanism is simulating
accountability for actions, which, alongside empathy exercises and risk
assessment, is pivotal to its effectiveness. We validate SKIG's performance
across various moral reasoning benchmarks with proprietary and opensource LLMs,
and investigate its crucial components through extensive ablation analyses. | cs.CL | Skin-in-the-Game: Decision Making via Multi-Stakeholder Alignment in LLMs |
2024-05-21T16:56:36Z | http://arxiv.org/pdf/2405.12929v1 | http://arxiv.org/abs/2405.12929v1 | Anjali Yadav, Tanya Garg, Matej Klemen, Matej Ulcar, Basant Agarwal, Marko Robnik Sikonja | 2024-05-21T16:56:36Z | cs.CL | Code-mixed discourse combines multiple languages in a single text. It is
commonly used in informal discourse in countries with several official
languages, but also in many other countries in combination with English or
neighboring languages. As recently large language models have dominated most
natural language processing tasks, we investigated their performance in
code-mixed settings for relevant tasks. We first created four new bilingual
pre-trained masked language models for English-Hindi and English-Slovene
languages, specifically aimed to support informal language. Then we performed
an evaluation of monolingual, bilingual, few-lingual, and massively
multilingual models on several languages, using two tasks that frequently
contain code-mixed text, in particular, sentiment analysis and offensive
language detection in social media texts. The results show that the most
successful classifiers are fine-tuned bilingual models and multilingual models,
specialized for social media texts, followed by non-specialized massively
multilingual and monolingual models, while huge generative models are not
competitive. For our affective problems, the models mostly perform slightly
better on code-mixed data compared to non-code-mixed data. | cs.CL | Code-mixed Sentiment and Hate-speech Prediction |
2024-05-21T16:38:13Z | http://arxiv.org/pdf/2405.12915v1 | http://arxiv.org/abs/2405.12915v1 | Xingyuan Pan, Luyang Huang, Liyan Kang, Zhicheng Liu, Yu Lu, Shanbo Cheng | 2024-05-21T16:38:13Z | cs.CL | Large Language Models (LLMs) have demonstrated remarkable abilities in
general scenarios. Instruction finetuning empowers them to align with humans in
various tasks. Nevertheless, the Diversity and Quality of the instruction data
remain two main challenges for instruction finetuning. With regard to this, in
this paper, we propose a novel gradient-based method to automatically select
high-quality and diverse instruction finetuning data for machine translation.
Our key innovation centers around analyzing how individual training examples
influence the model during training. Specifically, we select training examples
that exert beneficial influences on the model as high-quality ones by means of
Influence Function plus a small high-quality seed dataset. Moreover, to enhance
the diversity of the training data we maximize the variety of influences they
have on the model by clustering on their gradients and resampling. Extensive
experiments on WMT22 and FLORES translation tasks demonstrate the superiority
of our methods, and in-depth analysis further validates their effectiveness and
generalization. | cs.CL | G-DIG: Towards Gradient-based DIverse and hiGh-quality Instruction Data Selection for Machine Translation |
2024-05-21T16:30:25Z | http://arxiv.org/pdf/2405.12910v1 | http://arxiv.org/abs/2405.12910v1 | Holli Sargeant, Ahmed Izzidien, Felix Steffek | 2024-05-21T16:30:25Z | cs.CL, cs.AI, cs.CY, cs.LG | This paper addresses a critical gap in legal analytics by developing and
applying a novel taxonomy for topic modelling summary judgment cases in the
United Kingdom. Using a curated dataset of summary judgment cases, we use the
Large Language Model Claude 3 Opus to explore functional topics and trends. We
find that Claude 3 Opus correctly classified the topic with an accuracy of
87.10%. The analysis reveals distinct patterns in the application of summary
judgments across various legal domains. As case law in the United Kingdom is
not originally labelled with keywords or a topic filtering option, the findings
not only refine our understanding of the thematic underpinnings of summary
judgments but also illustrate the potential of combining traditional and
AI-driven approaches in legal classification. Therefore, this paper provides a
new and general taxonomy for UK law. The implications of this work serve as a
foundation for further research and policy discussions in the field of judicial
administration and computational legal research methodologies. | cs.CL | Topic Modelling Case Law Using a Large Language Model and a New Taxonomy for UK Law: AI Insights into Summary Judgment |
2024-05-21T16:14:55Z | http://arxiv.org/pdf/2405.12900v1 | http://arxiv.org/abs/2405.12900v1 | San Kim, Gary Geunbae Lee | 2024-05-21T16:14:55Z | cs.CL, cs.AI, I.2.7 | Recent advancements in open-domain dialogue systems have been propelled by
the emergence of high-quality large language models (LLMs) and various
effective training methodologies. Nevertheless, the presence of toxicity within
these models presents a significant challenge that can potentially diminish the
user experience. In this study, we introduce an innovative training algorithm,
an improvement upon direct preference optimization (DPO), called adversarial
DPO (ADPO). The ADPO algorithm is designed to train models to assign higher
probability distributions to preferred responses and lower distributions to
unsafe responses, which are self-generated using the toxic control token. We
demonstrate that ADPO enhances the model's resilience against harmful
conversations while minimizing performance degradation. Furthermore, we
illustrate that ADPO offers a more stable training procedure compared to the
traditional DPO. To the best of our knowledge, this is the first adaptation of
the DPO algorithm that directly incorporates harmful data into the generative
model, thereby reducing the need to artificially create safe dialogue data. | cs.CL | Adversarial DPO: Harnessing Harmful Data for Reducing Toxicity with Minimal Impact on Coherence and Evasiveness in Dialogue Agents |
2024-05-21T15:55:09Z | http://arxiv.org/pdf/2405.12884v1 | http://arxiv.org/abs/2405.12884v1 | Abdurahmman Alzahrani, Eyad Babkier, Faisal Yanbaawi, Firas Yanbaawi, Hassan Alhuzali | 2024-05-21T15:55:09Z | cs.CL, cs.AI | In the current era of digital communication and widespread use of social
media, it is crucial to develop an understanding of persuasive techniques
employed in written text. This knowledge is essential for effectively
discerning accurate information and making informed decisions. To address this
need, this paper presents a comprehensive empirical study focused on
identifying persuasive techniques in Arabic social media content. To achieve
this objective, we utilize Pre-trained Language Models (PLMs) and leverage the
ArAlEval dataset, which encompasses two tasks: binary classification to
determine the presence or absence of persuasion techniques, and multi-label
classification to identify the specific types of techniques employed in the
text. Our study explores three different learning approaches by harnessing the
power of PLMs: feature extraction, fine-tuning, and prompt engineering
techniques. Through extensive experimentation, we find that the fine-tuning
approach yields the highest results on the aforementioned dataset, achieving an
f1-micro score of 0.865 and an f1-weighted score of 0.861. Furthermore, our
analysis sheds light on an interesting finding. While the performance of the
GPT model is relatively lower compared to the other approaches, we have
observed that by employing few-shot learning techniques, we can enhance its
results by up to 20\%. This offers promising directions for future research and
exploration in this topic\footnote{Upon Acceptance, the source code will be
released on GitHub.}. | cs.CL | Investigating Persuasion Techniques in Arabic: An Empirical Study Leveraging Large Language Models |
2024-05-21T15:44:31Z | http://arxiv.org/pdf/2405.12875v1 | http://arxiv.org/abs/2405.12875v1 | Xiaofei Yu, Yitong Li, Jie Ma | 2024-05-21T15:44:31Z | cs.CV, cs.CL | Remote sensing image change captioning (RSICC) aims at generating human-like
language to describe the semantic changes between bi-temporal remote sensing
image pairs. It provides valuable insights into environmental dynamics and land
management. Unlike conventional change captioning task, RSICC involves not only
retrieving relevant information across different modalities and generating
fluent captions, but also mitigating the impact of pixel-level differences on
terrain change localization. The pixel problem due to long time span decreases
the accuracy of generated caption. Inspired by the remarkable generative power
of diffusion model, we propose a probabilistic diffusion model for RSICC to
solve the aforementioned problems. In training process, we construct a noise
predictor conditioned on cross modal features to learn the distribution from
the real caption distribution to the standard Gaussian distribution under the
Markov chain. Meanwhile, a cross-mode fusion and a stacking self-attention
module are designed for noise predictor in the reverse process. In testing
phase, the well-trained noise predictor helps to estimate the mean value of the
distribution and generate change captions step by step. Extensive experiments
on the LEVIR-CC dataset demonstrate the effectiveness of our Diffusion-RSCC and
its individual components. The quantitative results showcase superior
performance over existing methods across both traditional and newly augmented
metrics. The code and materials will be available online at
https://github.com/Fay-Y/Diffusion-RSCC. | cs.CV | Diffusion-RSCC: Diffusion Probabilistic Model for Change Captioning in Remote Sensing Images |
2024-05-21T15:13:12Z | http://arxiv.org/pdf/2405.12856v2 | http://arxiv.org/abs/2405.12856v2 | James Requeima, John Bronskill, Dami Choi, Richard E. Turner, David Duvenaud | 2024-05-25T22:07:48Z | stat.ML, cs.CL, cs.LG | Machine learning practitioners often face significant challenges in formally
integrating their prior knowledge and beliefs into predictive models, limiting
the potential for nuanced and context-aware analyses. Moreover, the expertise
needed to integrate this prior knowledge into probabilistic modeling typically
limits the application of these models to specialists. Our goal is to build a
regression model that can process numerical data and make probabilistic
predictions at arbitrary locations, guided by natural language text which
describes a user's prior knowledge. Large Language Models (LLMs) provide a
useful starting point for designing such a tool since they 1) provide an
interface where users can incorporate expert insights in natural language and
2) provide an opportunity for leveraging latent problem-relevant knowledge
encoded in LLMs that users may not have themselves. We start by exploring
strategies for eliciting explicit, coherent numerical predictive distributions
from LLMs. We examine these joint predictive distributions, which we call LLM
Processes, over arbitrarily-many quantities in settings such as forecasting,
multi-dimensional regression, black-box optimization, and image modeling. We
investigate the practical details of prompting to elicit coherent predictive
distributions, and demonstrate their effectiveness at regression. Finally, we
demonstrate the ability to usefully incorporate text into numerical
predictions, improving predictive performance and giving quantitative structure
that reflects qualitative descriptions. This lets us begin to explore the rich,
grounded hypothesis space that LLMs implicitly encode. | stat.ML | LLM Processes: Numerical Predictive Distributions Conditioned on Natural Language |
2024-05-21T14:24:01Z | http://arxiv.org/pdf/2405.12819v1 | http://arxiv.org/abs/2405.12819v1 | Libo Qin, Qiguang Chen, Xiachong Feng, Yang Wu, Yongheng Zhang, Yinghui Li, Min Li, Wanxiang Che, Philip S. Yu | 2024-05-21T14:24:01Z | cs.CL, cs.AI | While large language models (LLMs) like ChatGPT have shown impressive
capabilities in Natural Language Processing (NLP) tasks, a systematic
investigation of their potential in this field remains largely unexplored. This
study aims to address this gap by exploring the following questions: (1) How
are LLMs currently applied to NLP tasks in the literature? (2) Have traditional
NLP tasks already been solved with LLMs? (3) What is the future of the LLMs for
NLP? To answer these questions, we take the first step to provide a
comprehensive overview of LLMs in NLP. Specifically, we first introduce a
unified taxonomy including (1) parameter-frozen application and (2)
parameter-tuning application to offer a unified perspective for understanding
the current progress of LLMs in NLP. Furthermore, we summarize the new
frontiers and the associated challenges, aiming to inspire further
groundbreaking advancements. We hope this work offers valuable insights into
the {potential and limitations} of LLMs in NLP, while also serving as a
practical guide for building effective LLMs in NLP. | cs.CL | Large Language Models Meet NLP: A Survey |
2024-05-21T13:52:33Z | http://arxiv.org/pdf/2405.13095v1 | http://arxiv.org/abs/2405.13095v1 | Himanshu Maheshwari, Sambaran Bandyopadhyay, Aparna Garimella, Anandhavelu Natarajan | 2024-05-21T13:52:33Z | cs.CL, cs.AI | Automatically generating a presentation from the text of a long document is a
challenging and useful problem. In contrast to a flat summary, a presentation
needs to have a better and non-linear narrative, i.e., the content of a slide
can come from different and non-contiguous parts of the given document.
However, it is difficult to incorporate such non-linear mapping of content to
slides and ensure that the content is faithful to the document. LLMs are prone
to hallucination and their performance degrades with the length of the input
document. Towards this, we propose a novel graph based solution where we learn
a graph from the input document and use a combination of graph neural network
and LLM to generate a presentation with attribution of content for each slide.
We conduct thorough experiments to show the merit of our approach compared to
directly using LLMs for this task. | cs.CL | Presentations are not always linear! GNN meets LLM for Document-to-Presentation Transformation with Attribution |
2024-05-21T13:51:48Z | http://arxiv.org/pdf/2405.12801v1 | http://arxiv.org/abs/2405.12801v1 | Jonghyun Song, Cheyon Jin, Wenlong Zhao, Jay-Yoon Lee | 2024-05-21T13:51:48Z | cs.CL, cs.IR, cs.LG | A common retrieve-and-rerank paradigm involves retrieving a broad set of
relevant candidates using a scalable bi-encoder, followed by expensive but more
accurate cross-encoders to a limited candidate set. However, this small subset
often leads to error propagation from the bi-encoders, thereby restricting the
performance of the overall pipeline. To address these issues, we propose the
Comparing Multiple Candidates (CMC) framework, which compares a query and
multiple candidate embeddings jointly through shallow self-attention layers.
While providing contextualized representations, CMC is scalable enough to
handle multiple comparisons simultaneously, where comparing 2K candidates takes
only twice as long as comparing 100. Practitioners can use CMC as a lightweight
and effective reranker to improve top-1 accuracy. Moreover, when integrated
with another retriever, CMC reranking can function as a virtually enhanced
retriever. This configuration adds only negligible latency compared to using a
single retriever (virtual), while significantly improving recall at K
(enhanced).} Through experiments, we demonstrate that CMC, as a virtually
enhanced retriever, significantly improves Recall@k (+6.7, +3.5%-p for R@16,
R@64) compared to the initial retrieval stage on the ZeSHEL dataset. Meanwhile,
we conduct experiments for direct reranking on entity, passage, and dialogue
ranking. The results indicate that CMC is not only faster (11x) than
cross-encoders but also often more effective, with improved prediction
performance in Wikipedia entity linking (+0.7%-p) and DSTC7 dialogue ranking
(+3.3%-p). The code and link to datasets are available at
https://github.com/yc-song/cmc | cs.CL | Comparing Neighbors Together Makes it Easy: Jointly Comparing Multiple Candidates for Efficient and Effective Retrieval |
2024-05-21T13:38:15Z | http://arxiv.org/pdf/2405.12788v1 | http://arxiv.org/abs/2405.12788v1 | Yafu Li, Huajian Zhang, Jianhao Yan, Yongjing Yin, Yue Zhang | 2024-05-21T13:38:15Z | cs.CL | Recent advances have made non-autoregressive (NAT) translation comparable to
autoregressive methods (AT). However, their evaluation using BLEU has been
shown to weakly correlate with human annotations. Limited research compares
non-autoregressive translation and autoregressive translation comprehensively,
leaving uncertainty about the true proximity of NAT to AT. To address this gap,
we systematically evaluate four representative NAT methods across various
dimensions, including human evaluation. Our empirical results demonstrate that
despite narrowing the performance gap, state-of-the-art NAT still underperforms
AT under more reliable evaluation metrics. Furthermore, we discover that
explicitly modeling dependencies is crucial for generating natural language and
generalizing to out-of-distribution sequences. | cs.CL | What Have We Achieved on Non-autoregressive Translation? |
2024-05-21T13:24:07Z | http://arxiv.org/pdf/2405.12775v1 | http://arxiv.org/abs/2405.12775v1 | Hanlei Zhang, Hua Xu, Fei Long, Xin Wang, Kai Gao | 2024-05-21T13:24:07Z | cs.MM, cs.AI, cs.CL | Discovering the semantics of multimodal utterances is essential for
understanding human language and enhancing human-machine interactions. Existing
methods manifest limitations in leveraging nonverbal information for discerning
complex semantics in unsupervised scenarios. This paper introduces a novel
unsupervised multimodal clustering method (UMC), making a pioneering
contribution to this field. UMC introduces a unique approach to constructing
augmentation views for multimodal data, which are then used to perform
pre-training to establish well-initialized representations for subsequent
clustering. An innovative strategy is proposed to dynamically select
high-quality samples as guidance for representation learning, gauged by the
density of each sample's nearest neighbors. Besides, it is equipped to
automatically determine the optimal value for the top-$K$ parameter in each
cluster to refine sample selection. Finally, both high- and low-quality samples
are used to learn representations conducive to effective clustering. We build
baselines on benchmark multimodal intent and dialogue act datasets. UMC shows
remarkable improvements of 2-6\% scores in clustering metrics over
state-of-the-art methods, marking the first successful endeavor in this domain.
The complete code and data are available at https://github.com/thuiar/UMC. | cs.MM | Unsupervised Multimodal Clustering for Semantics Discovery in Multimodal Utterances |
2024-05-21T12:55:15Z | http://arxiv.org/pdf/2405.12744v1 | http://arxiv.org/abs/2405.12744v1 | Rochelle Choenni, Anne Lauscher, Ekaterina Shutova | 2024-05-21T12:55:15Z | cs.CL | Texts written in different languages reflect different culturally-dependent
beliefs of their writers. Thus, we expect multilingual LMs (MLMs), that are
jointly trained on a concatenation of text in multiple languages, to encode
different cultural values for each language. Yet, as the 'multilinguality' of
these LMs is driven by cross-lingual sharing, we also have reason to belief
that cultural values bleed over from one language into another. This limits the
use of MLMs in practice, as apart from being proficient in generating text in
multiple languages, creating language technology that can serve a community
also requires the output of LMs to be sensitive to their biases (Naous et al.,
2023). Yet, little is known about how cultural values emerge and evolve in MLMs
(Hershcovich et al., 2022a). We are the first to study how languages can exert
influence on the cultural values encoded for different test languages, by
studying how such values are revised during fine-tuning. Focusing on the
fine-tuning stage allows us to study the interplay between value shifts when
exposed to new linguistic experience from different data sources and languages.
Lastly, we use a training data attribution method to find patterns in the
fine-tuning examples, and the languages that they come from, that tend to
instigate value shifts. | cs.CL | The Echoes of Multilinguality: Tracing Cultural Value Shifts during LM Fine-tuning |
2024-05-21T12:16:20Z | http://arxiv.org/pdf/2405.12715v1 | http://arxiv.org/abs/2405.12715v1 | Hoang Ngo, Dat Quoc Nguyen | 2024-05-21T12:16:20Z | cs.IR, cs.CL | We present the first domain-adapted and fully-trained large language model,
RecGPT-7B, and its instruction-following variant, RecGPT-7B-Instruct, for
text-based recommendation. Experimental results on rating prediction and
sequential recommendation tasks show that our model, RecGPT-7B-Instruct,
outperforms previous strong baselines. We are releasing our RecGPT models as
well as their pre-training and fine-tuning datasets to facilitate future
research and downstream applications in text-based recommendation. Public
"huggingface" links to our RecGPT models and datasets are available at:
https://github.com/VinAIResearch/RecGPT | cs.IR | RecGPT: Generative Pre-training for Text-based Recommendation |
2024-05-21T12:04:55Z | http://arxiv.org/pdf/2405.12712v1 | http://arxiv.org/abs/2405.12712v1 | Ranim Khojah, Francisco Gomes de Oliveira Neto, Philipp Leitner | 2024-05-21T12:04:55Z | cs.SE, cs.AI, cs.CL, cs.HC | Software developers use natural language to interact not only with other
humans, but increasingly also with chatbots. These interactions have different
properties and flow differently based on what goal the developer wants to
achieve and who they interact with. In this paper, we aim to understand the
dynamics of conversations that occur during modern software development after
the integration of AI and chatbots, enabling a deeper recognition of the
advantages and disadvantages of including chatbot interactions in addition to
human conversations in collaborative work. We compile existing conversation
attributes with humans and NLU-based chatbots and adapt them to the context of
software development. Then, we extend the comparison to include LLM-powered
chatbots based on an observational study. We present similarities and
differences between human-to-human and human-to-bot conversations, also
distinguishing between NLU- and LLM-based chatbots. Furthermore, we discuss how
understanding the differences among the conversation styles guides the
developer on how to shape their expectations from a conversation and
consequently support the communication within a software team. We conclude that
the recent conversation styles that we observe with LLM-chatbots can not
replace conversations with humans due to certain attributes regarding social
aspects despite their ability to support productivity and decrease the
developers' mental load. | cs.SE | From Human-to-Human to Human-to-Bot Conversations in Software Engineering |
2024-05-21T11:52:14Z | http://arxiv.org/pdf/2405.12705v1 | http://arxiv.org/abs/2405.12705v1 | Omar Hamed, Souhail Bakkali, Marie-Francine Moens, Matthew Blaschko, Jordy Van Landeghem | 2024-05-21T11:52:14Z | cs.CV, cs.CL, cs.LG | This work addresses the need for a balanced approach between performance and
efficiency in scalable production environments for visually-rich document
understanding (VDU) tasks. Currently, there is a reliance on large document
foundation models that offer advanced capabilities but come with a heavy
computational burden. In this paper, we propose a multimodal early exit (EE)
model design that incorporates various training strategies, exit layer types
and placements. Our goal is to achieve a Pareto-optimal balance between
predictive performance and efficiency for multimodal document image
classification. Through a comprehensive set of experiments, we compare our
approach with traditional exit policies and showcase an improved
performance-efficiency trade-off. Our multimodal EE design preserves the
model's predictive capabilities, enhancing both speed and latency. This is
achieved through a reduction of over 20% in latency, while fully retaining the
baseline accuracy. This research represents the first exploration of multimodal
EE design within the VDU community, highlighting as well the effectiveness of
calibration in improving confidence scores for exiting at different layers.
Overall, our findings contribute to practical VDU applications by enhancing
both performance and efficiency. | cs.CV | Multimodal Adaptive Inference for Document Image Classification with Anytime Early Exiting |
2024-05-21T11:50:16Z | http://arxiv.org/pdf/2405.12701v1 | http://arxiv.org/abs/2405.12701v1 | Minbyul Jeong, Hyeon Hwang, Chanwoong Yoon, Taewhoo Lee, Jaewoo Kang | 2024-05-21T11:50:16Z | cs.CL, cs.AI | In the medical domain, numerous scenarios necessitate the long-form
generation ability of large language models (LLMs). Specifically, when
addressing patients' questions, it is essential that the model's response
conveys factual claims, highlighting the need for an automated method to
evaluate those claims. Thus, we introduce MedLFQA, a benchmark dataset
reconstructed using long-form question-answering datasets related to the
biomedical domain. We use MedLFQA to facilitate the automatic evaluations of
factuality. We also propose OLAPH, a simple and novel framework that enables
the improvement of factuality through automatic evaluations. The OLAPH
framework iteratively trains LLMs to mitigate hallucinations using sampling
predictions and preference optimization. In other words, we iteratively set the
highest-scoring response as a preferred response derived from sampling
predictions and train LLMs to align with the preferred response that improves
factuality. We highlight that, even on evaluation metrics not used during
training, LLMs trained with our OLAPH framework demonstrate significant
performance improvement in factuality. Our findings reveal that a 7B LLM
trained with our OLAPH framework can provide long answers comparable to the
medical experts' answers in terms of factuality. We believe that our work could
shed light on gauging the long-text generation ability of LLMs in the medical
domain. Our code and datasets are available at
https://github.com/dmis-lab/OLAPH}{https://github.com/dmis-lab/OLAPH. | cs.CL | OLAPH: Improving Factuality in Biomedical Long-form Question Answering |
2024-05-21T11:22:27Z | http://arxiv.org/pdf/2405.12689v2 | http://arxiv.org/abs/2405.12689v2 | Yafu Li, Zhilin Wang, Leyang Cui, Wei Bi, Shuming Shi, Yue Zhang | 2024-05-29T07:09:59Z | cs.CL, cs.AI | AI-generated text detection has attracted increasing attention as powerful
language models approach human-level generation. Limited work is devoted to
detecting (partially) AI-paraphrased texts. However, AI paraphrasing is
commonly employed in various application scenarios for text refinement and
diversity. To this end, we propose a novel detection framework, paraphrased
text span detection (PTD), aiming to identify paraphrased text spans within a
text. Different from text-level detection, PTD takes in the full text and
assigns each of the sentences with a score indicating the paraphrasing degree.
We construct a dedicated dataset, PASTED, for paraphrased text span detection.
Both in-distribution and out-of-distribution results demonstrate the
effectiveness of PTD models in identifying AI-paraphrased text spans.
Statistical and model analysis explains the crucial role of the surrounding
context of the paraphrased text spans. Extensive experiments show that PTD
models can generalize to versatile paraphrasing prompts and multiple
paraphrased text spans. We release our resources at
https://github.com/Linzwcs/PASTED. | cs.CL | Spotting AI's Touch: Identifying LLM-Paraphrased Spans in Text |
2024-05-21T10:34:47Z | http://arxiv.org/pdf/2405.12669v2 | http://arxiv.org/abs/2405.12669v2 | Huangjun Shen, Liangying Shao, Wenbo Li, Zhibin Lan, Zhanyu Liu, Jinsong Su | 2024-05-23T03:56:56Z | cs.CL | In recent years, multi-modal machine translation has attracted significant
interest in both academia and industry due to its superior performance. It
takes both textual and visual modalities as inputs, leveraging visual context
to tackle the ambiguities in source texts. In this paper, we begin by offering
an exhaustive overview of 99 prior works, comprehensively summarizing
representative studies from the perspectives of dominant models, datasets, and
evaluation metrics. Afterwards, we analyze the impact of various factors on
model performance and finally discuss the possible research directions for this
task in the future. Over time, multi-modal machine translation has developed
more types to meet diverse needs. Unlike previous surveys confined to the early
stage of multi-modal machine translation, our survey thoroughly concludes these
emerging types from different aspects, so as to provide researchers with a
better understanding of its current state. | cs.CL | A Survey on Multi-modal Machine Translation: Tasks, Methods and Challenges |
2024-05-21T10:10:56Z | http://arxiv.org/pdf/2405.12656v1 | http://arxiv.org/abs/2405.12656v1 | Yu-Hsiang Lin, Huang-Ting Shieh, Chih-Yu Liu, Kuang-Ting Lee, Hsiao-Cheng Chang, Jing-Lun Yang, Yu-Sheng Lin | 2024-05-21T10:10:56Z | cs.CL, cs.AI | Extrapolation in Large language models (LLMs) for open-ended inquiry
encounters two pivotal issues: (1) hallucination and (2) expensive training
costs. These issues present challenges for LLMs in specialized domains and
personalized data, requiring truthful responses and low fine-tuning costs.
Existing works attempt to tackle the problem by augmenting the input of a
smaller language model with information from a knowledge graph (KG). However,
they have two limitations: (1) failing to extract relevant information from a
large one-hop neighborhood in KG and (2) applying the same augmentation
strategy for KGs with different characteristics that may result in low
performance. Moreover, open-ended inquiry typically yields multiple responses,
further complicating extrapolation. We propose a new task, the extreme
multi-label KG link prediction task, to enable a model to perform extrapolation
with multiple responses using structured real-world knowledge. Our retriever
identifies relevant one-hop neighbors by considering entity, relation, and
textual data together. Our experiments demonstrate that (1) KGs with different
characteristics require different augmenting strategies, and (2) augmenting the
language model's input with textual data improves task performance
significantly. By incorporating the retrieval-augmented framework with KG, our
framework, with a small parameter size, is able to extrapolate based on a given
KG. The code can be obtained on GitHub:
https://github.com/exiled1143/Retrieval-Augmented-Language-Model-for-Multi-Label-Knowledge-Graph-Link-Prediction.git | cs.CL | Retrieval-Augmented Language Model for Extreme Multi-Label Knowledge Graph Link Prediction |
2024-05-21T09:33:31Z | http://arxiv.org/pdf/2405.12630v1 | http://arxiv.org/abs/2405.12630v1 | Nicolo Micheletti, Samuel Belkadi, Lifeng Han, Goran Nenadic | 2024-05-21T09:33:31Z | cs.CL, cs.AI | Large Language Models (LLMs) have revolutionised the field of Natural
Language Processing (NLP) and have achieved state-of-the-art performance in
practically every task in this field. However, the prevalent approach used in
text generation, Causal Language Modelling (CLM), which generates text
sequentially from left to right, inherently limits the freedom of the model,
which does not decide when and where each token is generated. In contrast,
Masked Language Modelling (MLM), primarily used for language understanding
tasks, can generate tokens anywhere in the text and any order. This paper
conducts an extensive comparison of MLM and CLM approaches for text generation
tasks. To do so, we pre-train several language models of comparable sizes on
three different datasets, namely 1) medical discharge summaries, 2) movie plot
synopses, and 3) authorship verification datasets. To assess the quality of the
generations, we first employ quantitative metrics and then perform a
qualitative human evaluation to analyse coherence and grammatical correctness.
In addition, we evaluate the usefulness of the generated texts by using them in
three different downstream tasks: 1) Entity Recognition, 2) Text
Classification, and 3) Authorship Verification. The results show that MLM
consistently outperforms CLM in text generation across all datasets, with
higher quantitative scores and better coherence in the generated text. The
study also finds \textit{no strong correlation} between the quality of the
generated text and the performance of the models in the downstream tasks. With
this study, we show that MLM for text generation has great potential for future
research and provides direction for future studies in this area. | cs.CL | Exploration of Masked and Causal Language Modelling for Text Generation |
2024-05-21T09:16:38Z | http://arxiv.org/pdf/2405.12619v1 | http://arxiv.org/abs/2405.12619v1 | Hassan Alhuzali, Ashwag Alasmari, Hamad Alsaleh | 2024-05-21T09:16:38Z | cs.CL | Mental health disorders significantly impact people globally, regardless of
background, education, or socioeconomic status. However, access to adequate
care remains a challenge, particularly for underserved communities with limited
resources. Text mining tools offer immense potential to support mental
healthcare by assisting professionals in diagnosing and treating patients. This
study addresses the scarcity of Arabic mental health resources for developing
such tools. We introduce MentalQA, a novel Arabic dataset featuring
conversational-style question-and-answer (QA) interactions. To ensure data
quality, we conducted a rigorous annotation process using a well-defined schema
with quality control measures. Data was collected from a question-answering
medical platform. The annotation schema for mental health questions and
corresponding answers draws upon existing classification schemes with some
modifications. Question types encompass six distinct categories: diagnosis,
treatment, anatomy \& physiology, epidemiology, healthy lifestyle, and provider
choice. Answer strategies include information provision, direct guidance, and
emotional support. Three experienced annotators collaboratively annotated the
data to ensure consistency. Our findings demonstrate high inter-annotator
agreement, with Fleiss' Kappa of $0.61$ for question types and $0.98$ for
answer strategies. In-depth analysis revealed insightful patterns, including
variations in question preferences across age groups and a strong correlation
between question types and answer strategies. MentalQA offers a valuable
foundation for developing Arabic text mining tools capable of supporting mental
health professionals and individuals seeking information. | cs.CL | MentalQA: An Annotated Arabic Corpus for Questions and Answers of Mental Healthcare |
2024-05-21T09:12:20Z | http://arxiv.org/pdf/2405.12617v1 | http://arxiv.org/abs/2405.12617v1 | Hang Chen, Xinyu Yang, Jiaying Zhu, Wenya Wang | 2024-05-21T09:12:20Z | cs.CL, cs.AI | Emergence, broadly conceptualized as the ``intelligent'' behaviors of LLMs,
has recently been studied and proved challenging to quantify due to the lack of
a measurable definition. Most commonly, it has been estimated statistically
through model performances across extensive datasets and tasks, which consumes
significant resources. In addition, such estimation is difficult to interpret
and may not accurately reflect the models' intrinsic emergence. In this work,
we propose a quantifiable solution for estimating emergence. Inspired by
emergentism in dynamics, we quantify the strength of emergence by comparing the
entropy reduction of the macroscopic (semantic) level with that of the
microscopic (token) level, both of which are derived from the representations
within the transformer block. Using a low-cost estimator, our quantification
method demonstrates consistent behaviors across a suite of LMs (GPT-2, GEMMA,
etc.) under both in-context learning (ICL) and natural sentences. Empirical
results show that (1) our method gives consistent measurements which align with
existing observations based on performance metrics, validating the
effectiveness of our emergence quantification; (2) our proposed metric uncovers
novel emergence patterns such as the correlations between the variance of our
metric and the number of ``shots'' in ICL, which further suggests a new way of
interpreting hallucinations in LLMs; (3) we offer a potential solution towards
estimating the emergence of larger and closed-resource LMs via smaller LMs like
GPT-2. Our codes are available at:
https://github.com/Zodiark-ch/Emergence-of-LLMs/. | cs.CL | Quantifying Emergence in Large Language Models |
2024-05-21T09:06:36Z | http://arxiv.org/pdf/2405.12612v1 | http://arxiv.org/abs/2405.12612v1 | Peter Devine | 2024-05-21T09:06:36Z | cs.CL, cs.AI, cs.LG | Open source large language models (LLMs) have shown great improvements in
recent times. However, many of these models are focused solely on popular
spoken languages. We present a high quality dataset of more than 70k
prompt-response pairs in 74 languages which consist of human generated prompts
and synthetic responses. We use this dataset to train a state-of-the-art open
source English LLM to chat multilingually. We evaluate our model on MT-Bench
chat benchmarks in 6 languages, finding that our multilingual model outperforms
previous state-of-the-art open source LLMs across each language. We further
find that training on more multilingual data is beneficial to the performance
in a chosen target language (Japanese) compared to simply training on only data
in that language. These results indicate the necessity of training on large
amounts of high quality multilingual data to make a more accessible LLM. | cs.CL | Tagengo: A Multilingual Chat Dataset |
2024-05-21T08:57:44Z | http://arxiv.org/pdf/2405.12604v2 | http://arxiv.org/abs/2405.12604v2 | Jiaxu Liu, Xiangyu Yin, Sihao Wu, Jianhong Wang, Meng Fang, Xinping Yi, Xiaowei Huang | 2024-06-17T18:52:45Z | cs.CL, cs.AI | With the proliferation of red-teaming strategies for Large Language Models
(LLMs), the deficiency in the literature about improving the safety and
robustness of LLM defense strategies is becoming increasingly pronounced. This
paper introduces the LLM-based \textbf{sentinel} model as a plug-and-play
prefix module designed to reconstruct the input prompt with just a few ($<30$)
additional tokens, effectively reducing toxicity in responses from target LLMs.
The sentinel model naturally overcomes the \textit{parameter inefficiency} and
\textit{limited model accessibility} for fine-tuning large target models. We
employ an interleaved training regimen using Proximal Policy Optimization (PPO)
to optimize both red team and sentinel models dynamically, incorporating a
value head-sharing mechanism inspired by the multi-agent centralized critic to
manage the complex interplay between agents. Our extensive experiments across
text-to-text and text-to-image demonstrate the effectiveness of our approach in
mitigating toxic outputs, even when dealing with larger models like
\texttt{Llama-2}, \texttt{GPT-3.5} and \texttt{Stable-Diffusion}, highlighting
the potential of our framework in enhancing safety and robustness in various
applications. | cs.CL | Tiny Refinements Elicit Resilience: Toward Efficient Prefix-Model Against LLM Red-Teaming |
2024-05-21T08:35:10Z | http://arxiv.org/pdf/2405.12591v1 | http://arxiv.org/abs/2405.12591v1 | Peiyu Liu, Ze-Feng Gao, Wayne Xin Zhao, Yipeng Ma, Tao Wang, Ji-Rong Wen | 2024-05-21T08:35:10Z | cs.CL | Key-value~(KV) caching is an important technique to accelerate the inference
of large language models~(LLMs), but incurs significant memory overhead. To
compress the size of KV cache, existing methods often compromise precision or
require extra data for calibration, limiting their practicality in LLM
deployment. In this paper, we introduce \textbf{DecoQuant}, a novel data-free
low-bit quantization technique based on tensor decomposition methods, to
effectively compress KV cache. Our core idea is to adjust the outlier
distribution of the original matrix by performing tensor decomposition, so that
the quantization difficulties are migrated from the matrix to decomposed local
tensors. Specially, we find that outliers mainly concentrate on small local
tensors, while large tensors tend to have a narrower value range. Based on this
finding, we propose to apply low-bit quantization to the large tensor, while
maintaining high-precision representation for the small tensor. Furthermore, we
utilize the proposed quantization method to compress the KV cache of LLMs to
accelerate the inference and develop an efficient dequantization kernel
tailored specifically for DecoQuant. Through extensive experiments, DecoQuant
demonstrates remarkable efficiency gains, showcasing up to a $\sim$75\%
reduction in memory footprint while maintaining comparable generation quality. | cs.CL | Unlocking Data-free Low-bit Quantization with Matrix Decomposition for KV Cache Compression |
2024-05-21T08:23:54Z | http://arxiv.org/pdf/2405.12579v2 | http://arxiv.org/abs/2405.12579v2 | Guangyao Lu, Yulin Liu | 2024-05-23T08:02:37Z | cs.CL | Fact-checking based on commercial LLMs has become mainstream. Although these
methods offer high explainability, it falls short in accuracy compared to
traditional fine-tuning approaches, and data security is also a significant
concern. In this paper, we propose a self-instruction based fine-tuning
approach for fact-checking that balances accuracy and explainability. Our
method consists of Data Augmentation and Improved DPO fine-tuning. The former
starts by instructing the model to generate both positive and negative
explanations based on claim-evidence pairs and labels, then sampling the
dataset according to our customized difficulty standards. The latter employs
our proposed improved DPO to fine-tune the model using the generated samples.
We fine-tune the smallest-scale LLaMA-7B model and evaluate it on the
challenging fact-checking datasets FEVEROUS and HOVER, utilizing four
fine-tuning methods and three few-shot learning methods for comparison. The
experiments demonstrate that our approach not only retains accuracy comparable
to, or even surpassing, traditional fine-tuning methods, but also generates
fluent explanation text. Moreover, it also exhibit high generalization
performance. Our method is the first to leverage self-supervised learning for
fact-checking and innovatively combines contrastive learning and improved DPO
in fine-tuning LLMs, as shown in the experiments. | cs.CL | Mining the Explainability and Generalization: Fact Verification Based on Self-Instruction |
2024-05-21T08:22:14Z | http://arxiv.org/pdf/2405.13085v1 | http://arxiv.org/abs/2405.13085v1 | Yichi Zhang, Binbin Hu, Zhuo Chen, Lingbing Guo, Ziqi Liu, Zhiqiang Zhang, Lei Liang, Huajun Chen, Wen Zhang | 2024-05-21T08:22:14Z | cs.CL, cs.AI | Knowledge graphs (KGs) provide reliable external knowledge for a wide variety
of AI tasks in the form of structured triples. Knowledge graph pre-training
(KGP) aims to pre-train neural networks on large-scale KGs and provide unified
interfaces to enhance different downstream tasks, which is a key direction for
KG management, maintenance, and applications. Existing works often focus on
purely research questions in open domains, or they are not open source due to
data security and privacy in real scenarios. Meanwhile, existing studies have
not explored the training efficiency and transferability of KGP models in
depth. To address these problems, We propose a framework MuDoK to achieve
multi-domain collaborative pre-training and efficient prefix prompt tuning to
serve diverse downstream tasks like recommendation and text understanding. Our
design is a plug-and-play prompt learning approach that can be flexibly adapted
to different downstream task backbones. In response to the lack of open-source
benchmarks, we constructed a new multi-domain KGP benchmark called KPI with two
large-scale KGs and six different sub-domain tasks to evaluate our method and
open-sourced it for subsequent research. We evaluated our approach based on
constructed KPI benchmarks using diverse backbone models in heterogeneous
downstream tasks. The experimental results show that our framework brings
significant performance gains, along with its generality, efficiency, and
transferability. | cs.CL | Multi-domain Knowledge Graph Collaborative Pre-training and Prompt Tuning for Diverse Downstream Tasks |
2024-05-21T08:06:13Z | http://arxiv.org/pdf/2405.12564v1 | http://arxiv.org/abs/2405.12564v1 | Zhiyuan Liu, An Zhang, Hao Fei, Enzhi Zhang, Xiang Wang, Kenji Kawaguchi, Tat-Seng Chua | 2024-05-21T08:06:13Z | q-bio.QM, cs.CL, cs.MM | Language Models (LMs) excel in understanding textual descriptions of
proteins, as evident in biomedical question-answering tasks. However, their
capability falters with raw protein data, such as amino acid sequences, due to
a deficit in pretraining on such data. Conversely, Protein Language Models
(PLMs) can understand and convert protein data into high-quality
representations, but struggle to process texts. To address their limitations,
we introduce ProtT3, a framework for Protein-to-Text Generation for Text-based
Protein Understanding. ProtT3 empowers an LM to understand protein sequences of
amino acids by incorporating a PLM as its protein understanding module,
enabling effective protein-to-text generation. This collaboration between PLM
and LM is facilitated by a cross-modal projector (i.e., Q-Former) that bridges
the modality gap between the PLM's representation space and the LM's input
space. Unlike previous studies focusing on protein property prediction and
protein-text retrieval, we delve into the largely unexplored field of
protein-to-text generation. To facilitate comprehensive benchmarks and promote
future research, we establish quantitative evaluations for protein-text
modeling tasks, including protein captioning, protein question-answering, and
protein-text retrieval. Our experiments show that ProtT3 substantially
surpasses current baselines, with ablation studies further highlighting the
efficacy of its core components. Our code is available at
https://github.com/acharkq/ProtT3. | q-bio.QM | ProtT3: Protein-to-Text Generation for Text-based Protein Understanding |
2024-05-21T07:35:21Z | http://arxiv.org/pdf/2405.13084v1 | http://arxiv.org/abs/2405.13084v1 | Yucheng Cai, Si Chen, Yi Huang, Junlan Feng, Zhijian Ou | 2024-05-21T07:35:21Z | cs.CL, cs.AI | The 2nd FutureDial Challenge: Dialog Systems with Retrieval Augmented
Generation (FutureDial-RAG), Co-located with SLT 2024 | cs.CL | The 2nd FutureDial Challenge: Dialog Systems with Retrieval Augmented Generation (FutureDial-RAG) |
2024-05-21T06:46:37Z | http://arxiv.org/pdf/2405.12532v2 | http://arxiv.org/abs/2405.12532v2 | Dongjie Yang, XiaoDong Han, Yan Gao, Yao Hu, Shilin Zhang, Hai Zhao | 2024-06-05T09:01:24Z | cs.CL | Large Language Models (LLMs) have shown remarkable comprehension abilities
but face challenges in GPU memory usage during inference, hindering their
scalability for real-time applications like chatbots. To accelerate inference,
we store computed keys and values (KV cache) in the GPU memory. Existing
methods study the KV cache compression to reduce memory by pruning the
pre-computed KV cache. However, they neglect the inter-layer dependency between
layers and huge memory consumption in pre-computation. To explore these
deficiencies, we find that the number of crucial keys and values that influence
future generations decreases layer by layer and we can extract them by the
consistency in attention weights. Based on the findings, we propose
PyramidInfer, a method that compresses the KV cache by layer-wise retaining
crucial context. PyramidInfer saves significant memory by computing fewer keys
and values without sacrificing performance. Experimental results show
PyramidInfer improves 2.2x throughput compared to Accelerate with over 54% GPU
memory reduction in KV cache. | cs.CL | PyramidInfer: Pyramid KV Cache Compression for High-throughput LLM Inference |
Subsets and Splits
No saved queries yet
Save your SQL queries to embed, download, and access them later. Queries will appear here once saved.