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
|
|---|---|---|---|---|---|---|---|---|
2023-04-23T11:22:17Z | http://arxiv.org/pdf/2304.11618v1 | http://arxiv.org/abs/2304.11618v1 | Yichi Zhang, Mingyang Chen, Wen Zhang | 2023-04-23T11:22:17Z | cs.CL, cs.AI | Negative sampling (NS) is widely used in knowledge graph embedding (KGE),
which aims to generate negative triples to make a positive-negative contrast
during training. However, existing NS methods are unsuitable when multi-modal
information is considered in KGE models. They are also inefficient due to their
complex design. In this paper, we propose Modality-Aware Negative Sampling
(MANS) for multi-modal knowledge graph embedding (MMKGE) to address the
mentioned problems. MANS could align structural and visual embeddings for
entities in KGs and learn meaningful embeddings to perform better in
multi-modal KGE while keeping lightweight and efficient. Empirical results on
two benchmarks demonstrate that MANS outperforms existing NS methods.
Meanwhile, we make further explorations about MANS to confirm its
effectiveness. | cs.CL | Modality-Aware Negative Sampling for Multi-modal Knowledge Graph Embedding |
2023-04-23T07:38:07Z | http://arxiv.org/pdf/2304.11567v1 | http://arxiv.org/abs/2304.11567v1 | Wenxiong Liao, Zhengliang Liu, Haixing Dai, Shaochen Xu, Zihao Wu, Yiyang Zhang, Xiaoke Huang, Dajiang Zhu, Hongmin Cai, Tianming Liu, Xiang Li | 2023-04-23T07:38:07Z | cs.CL, cs.AI | Background: Large language models such as ChatGPT are capable of generating
grammatically perfect and human-like text content, and a large number of
ChatGPT-generated texts have appeared on the Internet. However, medical texts
such as clinical notes and diagnoses require rigorous validation, and erroneous
medical content generated by ChatGPT could potentially lead to disinformation
that poses significant harm to healthcare and the general public.
Objective: This research is among the first studies on responsible and
ethical AIGC (Artificial Intelligence Generated Content) in medicine. We focus
on analyzing the differences between medical texts written by human experts and
generated by ChatGPT, and designing machine learning workflows to effectively
detect and differentiate medical texts generated by ChatGPT.
Methods: We first construct a suite of datasets containing medical texts
written by human experts and generated by ChatGPT. In the next step, we analyze
the linguistic features of these two types of content and uncover differences
in vocabulary, part-of-speech, dependency, sentiment, perplexity, etc. Finally,
we design and implement machine learning methods to detect medical text
generated by ChatGPT.
Results: Medical texts written by humans are more concrete, more diverse, and
typically contain more useful information, while medical texts generated by
ChatGPT pay more attention to fluency and logic, and usually express general
terminologies rather than effective information specific to the context of the
problem. A BERT-based model can effectively detect medical texts generated by
ChatGPT, and the F1 exceeds 95%. | cs.CL | Differentiate ChatGPT-generated and Human-written Medical Texts |
2023-04-23T06:52:35Z | http://arxiv.org/pdf/2304.11556v1 | http://arxiv.org/abs/2304.11556v1 | Xiping Liu, Zhao Tan | 2023-04-23T06:52:35Z | cs.CL, cs.AI | Chain-of-thought (CoT) prompting combined with large language models (LLMs)
have achieved encouraging results on complex reasoning tasks. Text-to-SQL is a
critical semantic parsing task that converts natural language questions into
SQL statements, involving a complex reasoning process. However, there is little
work about using CoT prompting to activate LLM's reasoning capabilities on
Text-to-SQL tasks. In this work, we propose a new paradigm for prompting
Text-to-SQL tasks, called Divide-and-Prompt, which first divides the task into
subtasks, and then approach each subtask through CoT. We present 3
prompting-based methods to enhance the Text-to-SQL ability of LLMs. Experiments
show that these prompts guide LLMs to generate Text-to-SQL with higher
execution accuracy. | cs.CL | Divide and Prompt: Chain of Thought Prompting for Text-to-SQL |
2023-04-23T04:21:50Z | http://arxiv.org/pdf/2304.11534v2 | http://arxiv.org/abs/2304.11534v2 | Kunze Wang, Yihao Ding, Soyeon Caren Han | 2023-04-27T13:42:25Z | cs.CL | Text Classification is the most essential and fundamental problem in Natural
Language Processing. While numerous recent text classification models applied
the sequential deep learning technique, graph neural network-based models can
directly deal with complex structured text data and exploit global information.
Many real text classification applications can be naturally cast into a graph,
which captures words, documents, and corpus global features. In this survey, we
bring the coverage of methods up to 2023, including corpus-level and
document-level graph neural networks. We discuss each of these methods in
detail, dealing with the graph construction mechanisms and the graph-based
learning process. As well as the technological survey, we look at issues behind
and future directions addressed in text classification using graph neural
networks. We also cover datasets, evaluation metrics, and experiment design and
present a summary of published performance on the publicly available
benchmarks. Note that we present a comprehensive comparison between different
techniques and identify the pros and cons of various evaluation metrics in this
survey. | cs.CL | Graph Neural Networks for Text Classification: A Survey |
2023-04-23T03:01:39Z | http://arxiv.org/pdf/2304.11520v4 | http://arxiv.org/abs/2304.11520v4 | Souvika Sarkar, Mohammad Fakhruddin Babar, Md Mahadi Hassan, Monowar Hasan, Shubhra Kanti Karmaker Santu | 2024-03-07T00:42:17Z | cs.CL, cs.SY, eess.SY | This paper presents a performance study of transformer language models under
different hardware configurations and accuracy requirements and derives
empirical observations about these resource/accuracy trade-offs. In particular,
we study how the most commonly used BERT-based language models (viz., BERT,
RoBERTa, DistilBERT, and TinyBERT) perform on embedded systems. We tested them
on four off-the-shelf embedded platforms (Raspberry Pi, Jetson, UP2, and UDOO)
with 2 GB and 4 GB memory (i.e., a total of eight hardware configurations) and
four datasets (i.e., HuRIC, GoEmotion, CoNLL, WNUT17) running various NLP
tasks. Our study finds that executing complex NLP tasks (such as "sentiment"
classification) on embedded systems is feasible even without any GPUs (e.g.,
Raspberry Pi with 2 GB of RAM). Our findings can help designers understand the
deployability and performance of transformer language models, especially those
based on BERT architectures. | cs.CL | Processing Natural Language on Embedded Devices: How Well Do Transformer Models Perform? |
2023-04-23T00:04:14Z | http://arxiv.org/pdf/2304.11501v2 | http://arxiv.org/abs/2304.11501v2 | Shira Wein, Nathan Schneider | 2024-01-29T23:09:49Z | cs.CL | Translated texts bear several hallmarks distinct from texts originating in
the language. Though individual translated texts are often fluent and preserve
meaning, at a large scale, translated texts have statistical tendencies which
distinguish them from text originally written in the language
("translationese") and can affect model performance. We frame the novel task of
translationese reduction and hypothesize that Abstract Meaning Representation
(AMR), a graph-based semantic representation which abstracts away from the
surface form, can be used as an interlingua to reduce the amount of
translationese in translated texts. By parsing English translations into an AMR
and then generating text from that AMR, the result more closely resembles
originally English text across three quantitative macro-level measures, without
severely compromising fluency or adequacy. We compare our AMR-based approach
against three other techniques based on machine translation or paraphrase
generation. This work makes strides towards reducing translationese in text and
highlights the utility of AMR as an interlingua. | cs.CL | Lost in Translationese? Reducing Translation Effect Using Abstract Meaning Representation |
2023-04-22T22:50:50Z | http://arxiv.org/pdf/2304.11490v3 | http://arxiv.org/abs/2304.11490v3 | Shima Rahimi Moghaddam, Christopher J. Honey | 2023-04-26T04:02:04Z | cs.AI, cs.CL | Large language models (LLMs) excel in many tasks in 2023, but they still face
challenges in complex reasoning. Theory-of-mind (ToM) tasks, which require
understanding agents' beliefs, goals, and mental states, are essential for
common-sense reasoning involving humans, making it crucial to enhance LLM
performance in this area. This study measures the ToM performance of GPT-4 and
three GPT-3.5 variants (Davinci-2, Davinci-3, GPT-3.5-Turbo), and investigates
the effectiveness of in-context learning in improving their ToM comprehension.
We evaluated prompts featuring two-shot chain of thought reasoning and
step-by-step thinking instructions. We found that LLMs trained with
Reinforcement Learning from Human Feedback (RLHF) (all models excluding
Davinci-2) improved their ToM accuracy via in-context learning. GPT-4 performed
best in zero-shot settings, reaching nearly 80% ToM accuracy, but still fell
short of the 87% human accuracy on the test set. However, when supplied with
prompts for in-context learning, all RLHF-trained LLMs exceeded 80% ToM
accuracy, with GPT-4 reaching 100%. These results demonstrate that appropriate
prompting enhances LLM ToM reasoning, and they underscore the context-dependent
nature of LLM cognitive capacities. | cs.AI | Boosting Theory-of-Mind Performance in Large Language Models via Prompting |
2023-04-22T21:51:49Z | http://arxiv.org/pdf/2304.11485v1 | http://arxiv.org/abs/2304.11485v1 | Dhiraj Murthy, Constantine Caramanis, Koustav Rudra | 2023-04-22T21:51:49Z | cs.CL, cs.AI, cs.LG, cs.SI, J.4; K.4.2; I.2.7 | Individuals involved in gang-related activity use mainstream social media
including Facebook and Twitter to express taunts and threats as well as grief
and memorializing. However, identifying the impact of gang-related activity in
order to serve community member needs through social media sources has a unique
set of challenges. This includes the difficulty of ethically identifying
training data of individuals impacted by gang activity and the need to account
for a non-standard language style commonly used in the tweets from these
individuals. Our study provides evidence of methods where natural language
processing tools can be helpful in efficiently identifying individuals who may
be in need of community care resources such as counselors, conflict mediators,
or academic/professional training programs. We demonstrate that our binary
logistic classifier outperforms baseline standards in identifying individuals
impacted by gang-related violence using a sample of gang-related tweets
associated with Chicago. We ultimately found that the language of a tweet is
highly relevant and that uses of ``big data'' methods or machine learning
models need to better understand how language impacts the model's performance
and how it discriminates among populations. | cs.CL | Understanding Lexical Biases when Identifying Gang-related Social Media Communications |
2023-04-22T20:00:06Z | http://arxiv.org/pdf/2304.11473v2 | http://arxiv.org/abs/2304.11473v2 | Jacopo Tagliabue, Ciro Greco | 2023-06-11T17:35:02Z | cs.IR, cs.CL, cs.LG | As ecommerce continues growing, huge investments in ML and NLP for
Information Retrieval are following. While the vector space model dominated
retrieval modelling in product search - even as vectorization itself greatly
changed with the advent of deep learning -, our position paper argues in a
contrarian fashion that program synthesis provides significant advantages for
many queries and a significant number of players in the market. We detail the
industry significance of the proposed approach, sketch implementation details,
and address common objections drawing from our experience building a similar
system at Tooso. | cs.IR | (Vector) Space is Not the Final Frontier: Product Search as Program Synthesis |
2023-04-22T18:22:10Z | http://arxiv.org/pdf/2304.11461v1 | http://arxiv.org/abs/2304.11461v1 | Benyamin Ghojogh, Ali Ghodsi | 2023-04-22T18:22:10Z | cs.LG, cs.CL, cs.NE, cs.SD, eess.AS | This is a tutorial paper on Recurrent Neural Network (RNN), Long Short-Term
Memory Network (LSTM), and their variants. We start with a dynamical system and
backpropagation through time for RNN. Then, we discuss the problems of gradient
vanishing and explosion in long-term dependencies. We explain close-to-identity
weight matrix, long delays, leaky units, and echo state networks for solving
this problem. Then, we introduce LSTM gates and cells, history and variants of
LSTM, and Gated Recurrent Units (GRU). Finally, we introduce bidirectional RNN,
bidirectional LSTM, and the Embeddings from Language Model (ELMo) network, for
processing a sequence in both directions. | cs.LG | Recurrent Neural Networks and Long Short-Term Memory Networks: Tutorial and Survey |
2023-04-22T15:45:40Z | http://arxiv.org/pdf/2304.11434v1 | http://arxiv.org/abs/2304.11434v1 | Samruddhi Deode, Janhavi Gadre, Aditi Kajale, Ananya Joshi, Raviraj Joshi | 2023-04-22T15:45:40Z | cs.CL, cs.LG | The multilingual Sentence-BERT (SBERT) models map different languages to
common representation space and are useful for cross-language similarity and
mining tasks. We propose a simple yet effective approach to convert vanilla
multilingual BERT models into multilingual sentence BERT models using synthetic
corpus. We simply aggregate translated NLI or STS datasets of the low-resource
target languages together and perform SBERT-like fine-tuning of the vanilla
multilingual BERT model. We show that multilingual BERT models are inherent
cross-lingual learners and this simple baseline fine-tuning approach without
explicit cross-lingual training yields exceptional cross-lingual properties. We
show the efficacy of our approach on 10 major Indic languages and also show the
applicability of our approach to non-Indic languages German and French. Using
this approach, we further present L3Cube-IndicSBERT, the first multilingual
sentence representation model specifically for Indian languages Hindi, Marathi,
Kannada, Telugu, Malayalam, Tamil, Gujarati, Odia, Bengali, and Punjabi. The
IndicSBERT exhibits strong cross-lingual capabilities and performs
significantly better than alternatives like LaBSE, LASER, and
paraphrase-multilingual-mpnet-base-v2 on Indic cross-lingual and monolingual
sentence similarity tasks. We also release monolingual SBERT models for each of
the languages and show that IndicSBERT performs competitively with its
monolingual counterparts. These models have been evaluated using embedding
similarity scores and classification accuracy. | cs.CL | L3Cube-IndicSBERT: A simple approach for learning cross-lingual sentence representations using multilingual BERT |
2023-04-22T13:53:50Z | http://arxiv.org/pdf/2304.11410v1 | http://arxiv.org/abs/2304.11410v1 | Sidharth Ranjan, Titus von der Malsburg | 2023-04-22T13:53:50Z | cs.CL, econ.TH | The principle of DEPENDENCY LENGTH MINIMIZATION, which seeks to keep
syntactically related words close in a sentence, is thought to universally
shape the structure of human languages for effective communication. However,
the extent to which dependency length minimization is applied in human language
systems is not yet fully understood. Preverbally, the placement of
long-before-short constituents and postverbally, short-before-long constituents
are known to minimize overall dependency length of a sentence. In this study,
we test the hypothesis that placing only the shortest preverbal constituent
next to the main-verb explains word order preferences in Hindi (a SOV language)
as opposed to the global minimization of dependency length. We characterize
this approach as a least-effort strategy because it is a cost-effective way to
shorten all dependencies between the verb and its preverbal dependencies. As
such, this approach is consistent with the bounded-rationality perspective
according to which decision making is governed by "fast but frugal" heuristics
rather than by a search for optimal solutions. Consistent with this idea, our
results indicate that actual corpus sentences in the Hindi-Urdu Treebank corpus
are better explained by the least effort strategy than by global minimization
of dependency lengths. Additionally, for the task of distinguishing corpus
sentences from counterfactual variants, we find that the dependency length and
constituent length of the constituent closest to the main verb are much better
predictors of whether a sentence appeared in the corpus than total dependency
length. Overall, our findings suggest that cognitive resource constraints play
a crucial role in shaping natural languages. | cs.CL | A bounded rationality account of dependency length minimization in Hindi |
2023-04-22T13:42:04Z | http://arxiv.org/pdf/2304.11406v4 | http://arxiv.org/abs/2304.11406v4 | Alireza Salemi, Sheshera Mysore, Michael Bendersky, Hamed Zamani | 2024-06-05T03:29:31Z | cs.CL | This paper highlights the importance of personalization in large language
models and introduces the LaMP benchmark -- a novel benchmark for training and
evaluating language models for producing personalized outputs. LaMP offers a
comprehensive evaluation framework with diverse language tasks and multiple
entries for each user profile. It consists of seven personalized tasks,
spanning three text classification and four text generation tasks. We
additionally propose two retrieval augmentation approaches that retrieve
personal items from each user profile for personalizing language model outputs.
To this aim, we study various retrieval models, including term matching,
semantic matching, and time-aware methods. Extensive experiments on LaMP for
zero-shot and fine-tuned language models demonstrate the efficacy of the
proposed retrieval augmentation approach and highlight the impact of
personalization in various natural language tasks. | cs.CL | LaMP: When Large Language Models Meet Personalization |
2023-04-22T12:50:49Z | http://arxiv.org/pdf/2304.11389v2 | http://arxiv.org/abs/2304.11389v2 | Byung-Doh Oh, William Schuler | 2023-10-22T20:03:54Z | cs.CL | Recent psycholinguistic studies have drawn conflicting conclusions about the
relationship between the quality of a language model and the ability of its
surprisal estimates to predict human reading times, which has been speculated
to be due to the large gap in both the amount of training data and model
capacity across studies. The current work aims to consolidate these findings by
evaluating surprisal estimates from Transformer-based language model variants
that vary systematically in the amount of training data and model capacity on
their ability to predict human reading times. The results show that surprisal
estimates from most variants with contemporary model capacities provide the
best fit after seeing about two billion training tokens, after which they begin
to diverge from humanlike expectations. Additionally, newly-trained smaller
model variants reveal a 'tipping point' at convergence, after which the
decrease in language model perplexity begins to result in poorer fits to human
reading times. These results suggest that the massive amount of training data
is mainly responsible for the poorer fit achieved by surprisal from larger
pre-trained language models, and that a certain degree of model capacity is
necessary for Transformer-based language models to capture humanlike
expectations. | cs.CL | Transformer-Based Language Model Surprisal Predicts Human Reading Times Best with About Two Billion Training Tokens |
2023-04-22T10:47:01Z | http://arxiv.org/pdf/2304.11370v1 | http://arxiv.org/abs/2304.11370v1 | Haitao Li, Qingyao Ai, Jia Chen, Qian Dong, Yueyue Wu, Yiqun Liu, Chong Chen, Qi Tian | 2023-04-22T10:47:01Z | cs.IR, cs.CL | Legal case retrieval, which aims to find relevant cases for a query case,
plays a core role in the intelligent legal system. Despite the success that
pre-training has achieved in ad-hoc retrieval tasks, effective pre-training
strategies for legal case retrieval remain to be explored. Compared with
general documents, legal case documents are typically long text sequences with
intrinsic logical structures. However, most existing language models have
difficulty understanding the long-distance dependencies between different
structures. Moreover, in contrast to the general retrieval, the relevance in
the legal domain is sensitive to key legal elements. Even subtle differences in
key legal elements can significantly affect the judgement of relevance.
However, existing pre-trained language models designed for general purposes
have not been equipped to handle legal elements.
To address these issues, in this paper, we propose SAILER, a new
Structure-Aware pre-traIned language model for LEgal case Retrieval. It is
highlighted in the following three aspects: (1) SAILER fully utilizes the
structural information contained in legal case documents and pays more
attention to key legal elements, similar to how legal experts browse legal case
documents. (2) SAILER employs an asymmetric encoder-decoder architecture to
integrate several different pre-training objectives. In this way, rich semantic
information across tasks is encoded into dense vectors. (3) SAILER has powerful
discriminative ability, even without any legal annotation data. It can
distinguish legal cases with different charges accurately. Extensive
experiments over publicly available legal benchmarks demonstrate that our
approach can significantly outperform previous state-of-the-art methods in
legal case retrieval. | cs.IR | SAILER: Structure-aware Pre-trained Language Model for Legal Case Retrieval |
2023-04-22T09:27:32Z | http://arxiv.org/pdf/2304.13557v1 | http://arxiv.org/abs/2304.13557v1 | Katie Seaborn, Yeongdae Kim | 2023-04-22T09:27:32Z | cs.CL, cs.CY, cs.HC, cs.LG | As virtual assistants continue to be taken up globally, there is an
ever-greater need for these speech-based systems to communicate naturally in a
variety of languages. Crowdsourcing initiatives have focused on multilingual
translation of big, open data sets for use in natural language processing
(NLP). Yet, language translation is often not one-to-one, and biases can
trickle in. In this late-breaking work, we focus on the case of pronouns
translated between English and Japanese in the crowdsourced Tatoeba database.
We found that masculine pronoun biases were present overall, even though
plurality in language was accounted for in other ways. Importantly, we detected
biases in the translation process that reflect nuanced reactions to the
presence of feminine, neutral, and/or non-binary pronouns. We raise the issue
of translation bias for pronouns and offer a practical solution to embed
plurality in NLP data sets. | cs.CL | "I'm" Lost in Translation: Pronoun Missteps in Crowdsourced Data Sets |
2023-04-22T09:10:49Z | http://arxiv.org/pdf/2304.11350v2 | http://arxiv.org/abs/2304.11350v2 | Andrei-Marius Avram, Verginica Barbu Mititelu, Dumitru-Clementin Cercel | 2023-05-08T16:54:03Z | cs.CL | Multiword expressions are a key ingredient for developing large-scale and
linguistically sound natural language processing technology. This paper
describes our improvements in automatically identifying Romanian multiword
expressions on the corpus released for the PARSEME v1.2 shared task. Our
approach assumes a multilingual perspective based on the recently introduced
lateral inhibition layer and adversarial training to boost the performance of
the employed multilingual language models. With the help of these two methods,
we improve the F1-score of XLM-RoBERTa by approximately 2.7% on unseen
multiword expressions, the main task of the PARSEME 1.2 edition. In addition,
our results can be considered SOTA performance, as they outperform the previous
results on Romanian obtained by the participants in this competition. | cs.CL | Romanian Multiword Expression Detection Using Multilingual Adversarial Training and Lateral Inhibition |
2023-04-22T07:40:23Z | http://arxiv.org/pdf/2304.11340v1 | http://arxiv.org/abs/2304.11340v1 | Sakae Mizuki, Naoaki Okazaki | 2023-04-22T07:40:23Z | cs.CL | A promising approach for knowledge-based Word Sense Disambiguation (WSD) is
to select the sense whose contextualized embeddings computed for its definition
sentence are closest to those computed for a target word in a given sentence.
This approach relies on the similarity of the \textit{sense} and
\textit{context} embeddings computed by a pre-trained language model. We
propose a semantic specialization for WSD where contextualized embeddings are
adapted to the WSD task using solely lexical knowledge. The key idea is, for a
given sense, to bring semantically related senses and contexts closer and send
different/unrelated senses farther away. We realize this idea as the joint
optimization of the Attract-Repel objective for sense pairs and the
self-training objective for context-sense pairs while controlling deviations
from the original embeddings. The proposed method outperformed previous studies
that adapt contextualized embeddings. It achieved state-of-the-art performance
on knowledge-based WSD when combined with the reranking heuristic that uses the
sense inventory. We found that the similarity characteristics of specialized
embeddings conform to the key idea. We also found that the (dis)similarity of
embeddings between the related/different/unrelated senses correlates well with
the performance of WSD. | cs.CL | Semantic Specialization for Knowledge-based Word Sense Disambiguation |
2023-04-22T06:28:46Z | http://arxiv.org/pdf/2304.11164v1 | http://arxiv.org/abs/2304.11164v1 | Anthony G Cohn, Jose Hernandez-Orallo | 2023-04-22T06:28:46Z | cs.CL, cs.AI | Language models have become very popular recently and many claims have been
made about their abilities, including for commonsense reasoning. Given the
increasingly better results of current language models on previous static
benchmarks for commonsense reasoning, we explore an alternative dialectical
evaluation. The goal of this kind of evaluation is not to obtain an aggregate
performance value but to find failures and map the boundaries of the system.
Dialoguing with the system gives the opportunity to check for consistency and
get more reassurance of these boundaries beyond anecdotal evidence. In this
paper we conduct some qualitative investigations of this kind of evaluation for
the particular case of spatial reasoning (which is a fundamental aspect of
commonsense reasoning). We conclude with some suggestions for future work both
to improve the capabilities of language models and to systematise this kind of
dialectical evaluation. | cs.CL | Dialectical language model evaluation: An initial appraisal of the commonsense spatial reasoning abilities of LLMs |
2023-04-22T03:53:24Z | http://arxiv.org/pdf/2304.12810v1 | http://arxiv.org/abs/2304.12810v1 | Katie Seaborn, Shruti Chandra, Thibault Fabre | 2023-04-22T03:53:24Z | cs.CL, cs.CY, cs.HC, cs.LG | Critical scholarship has elevated the problem of gender bias in data sets
used to train virtual assistants (VAs). Most work has focused on explicit
biases in language, especially against women, girls, femme-identifying people,
and genderqueer folk; implicit associations through word embeddings; and
limited models of gender and masculinities, especially toxic masculinities,
conflation of sex and gender, and a sex/gender binary framing of the masculine
as diametric to the feminine. Yet, we must also interrogate how masculinities
are "coded" into language and the assumption of "male" as the linguistic
default: implicit masculine biases. To this end, we examined two natural
language processing (NLP) data sets. We found that when gendered language was
present, so were gender biases and especially masculine biases. Moreover, these
biases related in nuanced ways to the NLP context. We offer a new dictionary
called AVA that covers ambiguous associations between gendered language and the
language of VAs. | cs.CL | Transcending the "Male Code": Implicit Masculine Biases in NLP Contexts |
2023-04-22T01:54:30Z | http://arxiv.org/pdf/2304.11292v1 | http://arxiv.org/abs/2304.11292v1 | Noshin Nawal | 2023-04-22T01:54:30Z | cs.AI, cs.CL, cs.LG | The energy inefficiency of the apps can be a major issue for the app users
which is discussed on App Stores extensively. Previous research has shown the
importance of investigating the energy related app reviews to identify the
major causes or categories of energy related user feedback. However, there is
no study that efficiently extracts the energy related app reviews
automatically. In this paper, we empirically study different techniques for
automatic extraction of the energy related user feedback. We compare the
accuracy, F1-score and run time of numerous machine-learning models with
relevant feature combinations and relatively modern Neural Network-based
models. In total, 60 machine learning models are compared to 30 models that we
build using six neural network architectures and three word embedding models.
We develop a visualization tool for this study through which a developer can
traverse through this large-scale result set. The results show that neural
networks outperform the other machine learning techniques and can achieve the
highest F1-score of 0.935. To replicate the research results, we have open
sourced the interactive visualization tool. After identifying the best results
and extracting the energy related reviews, we further compare various
techniques to help the developers automatically investigate the emerging issues
that might be responsible for energy inefficiency of the apps. We experiment
the previously used string matching with results obtained from applying two of
the state-of-the-art topic modeling algorithms, OBTM and AOLDA. Finally, we run
a qualitative study performed in collaboration with developers and students
from different institutions to determine their preferences for identifying
necessary topics from previously categorized reviews, which shows OBTM produces
the most helpful results. | cs.AI | On the Identification of the Energy related Issues from the App Reviews |
2023-04-22T01:47:19Z | http://arxiv.org/pdf/2304.12809v1 | http://arxiv.org/abs/2304.12809v1 | Katie Seaborn, Somang Nam, Julia Keckeis, Tatsuya Itagaki | 2023-04-22T01:47:19Z | cs.HC, cs.AI, cs.CL, cs.CY, cs.SD | The Japanese notion of "kawaii" or expressions of cuteness, vulnerability,
and/or charm is a global cultural export. Work has explored kawaii-ness as a
design feature and factor of user experience in the visual appearance,
nonverbal behaviour, and sound of robots and virtual characters. In this
initial work, we consider whether voices can be kawaii by exploring the vocal
qualities of voice assistant speech, i.e., kawaii vocalics. Drawing from an
age-inclusive model of kawaii, we ran a user perceptions study on the
kawaii-ness of younger- and older-sounding Japanese computer voices. We found
that kawaii-ness intersected with perceptions of gender and age, i.e., gender
ambiguous and girlish, as well as VA features, i.e., fluency and artificiality.
We propose an initial model of kawaii vocalics to be validated through the
identification and study of vocal qualities, cognitive appraisals, behavioural
responses, and affective reports. | cs.HC | Can Voice Assistants Sound Cute? Towards a Model of Kawaii Vocalics |
2023-04-21T23:50:09Z | http://arxiv.org/pdf/2304.11276v1 | http://arxiv.org/abs/2304.11276v1 | Hengky Susanto, David James Woo, Kai Guo | 2023-04-21T23:50:09Z | cs.CL | The recent advancement in Natural Language Processing (NLP) capability has
led to the development of language models (e.g., ChatGPT) that is capable of
generating human-like language. In this study, we explore how language models
can be utilized to help the ideation aspect of creative writing. Our empirical
findings show that language models play different roles in helping student
writers to be more creative, such as the role of a collaborator, a provocateur,
etc | cs.CL | The Role of AI in Human-AI Creative Writing for Hong Kong Secondary Students |
2023-04-21T23:08:39Z | http://arxiv.org/pdf/2304.14415v1 | http://arxiv.org/abs/2304.14415v1 | Sara Amani, Lance White, Trini Balart, Laksha Arora, Dr. Kristi J. Shryock, Dr. Kelly Brumbelow, Dr. Karan L. Watson | 2023-04-21T23:08:39Z | cs.HC, cs.AI, cs.CL, cs.CY | ChatGPT is a natural language processing tool that can engage in human-like
conversations and generate coherent and contextually relevant responses to
various prompts. ChatGPT is capable of understanding natural text that is input
by a user and generating appropriate responses in various forms. This tool
represents a major step in how humans are interacting with technology. This
paper specifically focuses on how ChatGPT is revolutionizing the realm of
engineering education and the relationship between technology, students, and
faculty and staff. Because this tool is quickly changing and improving with the
potential for even greater future capability, it is a critical time to collect
pertinent data. A survey was created to measure the effects of ChatGPT on
students, faculty, and staff. This survey is shared as a Texas A&M University
technical report to allow other universities and entities to use this survey
and measure the effects elsewhere. | cs.HC | Generative AI Perceptions: A Survey to Measure the Perceptions of Faculty, Staff, and Students on Generative AI Tools in Academia |
2023-04-21T22:53:45Z | http://arxiv.org/pdf/2304.11163v1 | http://arxiv.org/abs/2304.11163v1 | Atoosa Kasirzadeh | 2023-04-21T22:53:45Z | cs.CY, cs.CL | The allure of emerging AI technologies is undoubtedly thrilling. However, the
promise that AI technologies will benefit all of humanity is empty so long as
we lack a nuanced understanding of what humanity is supposed to be in the face
of widening global inequality and pressing existential threats. Going forward,
it is crucial to invest in rigorous and collaborative AI safety and ethics
research. We also need to develop standards in a sustainable and equitable way
that differentiate between merely speculative and well-researched questions.
Only the latter enable us to co-construct and deploy the values that are
necessary for creating beneficial AI. Failure to do so could result in a future
in which our AI technological advancements outstrip our ability to navigate
their ethical and social implications. This path we do not want to go down. | cs.CY | ChatGPT, Large Language Technologies, and the Bumpy Road of Benefiting Humanity |
2023-04-21T21:25:30Z | http://arxiv.org/pdf/2304.11257v1 | http://arxiv.org/abs/2304.11257v1 | Felipe Urrutia, Roberto Araya | 2023-04-21T21:25:30Z | cs.CL, cs.AI, cs.LG | Written answers to open-ended questions can have a higher long-term effect on
learning than multiple-choice questions. However, it is critical that teachers
immediately review the answers, and ask to redo those that are incoherent. This
can be a difficult task and can be time-consuming for teachers. A possible
solution is to automate the detection of incoherent answers. One option is to
automate the review with Large Language Models (LLM). In this paper, we analyze
the responses of fourth graders in mathematics using three LLMs: GPT-3, BLOOM,
and YOU. We used them with zero, one, two, three and four shots. We compared
their performance with the results of various classifiers trained with Machine
Learning (ML). We found that LLMs perform worse than MLs in detecting
incoherent answers. The difficulty seems to reside in recursive questions that
contain both questions and answers, and in responses from students with typical
fourth-grader misspellings. Upon closer examination, we have found that the
ChatGPT model faces the same challenges. | cs.CL | Who's the Best Detective? LLMs vs. MLs in Detecting Incoherent Fourth Grade Math Answers |
2023-04-21T21:25:14Z | http://arxiv.org/pdf/2304.11256v2 | http://arxiv.org/abs/2304.11256v2 | Gagan Bhatia, Ife Adebara, AbdelRahim Elmadany, Muhammad Abdul-Mageed | 2023-04-25T07:44:37Z | cs.CL | We describe our contribution to the SemEVAl 2023 AfriSenti-SemEval shared
task, where we tackle the task of sentiment analysis in 14 different African
languages. We develop both monolingual and multilingual models under a full
supervised setting (subtasks A and B). We also develop models for the zero-shot
setting (subtask C). Our approach involves experimenting with transfer learning
using six language models, including further pertaining of some of these models
as well as a final finetuning stage. Our best performing models achieve an
F1-score of 70.36 on development data and an F1-score of 66.13 on test data.
Unsurprisingly, our results demonstrate the effectiveness of transfer learning
and fine-tuning techniques for sentiment analysis across multiple languages.
Our approach can be applied to other sentiment analysis tasks in different
languages and domains. | cs.CL | UBC-DLNLP at SemEval-2023 Task 12: Impact of Transfer Learning on African Sentiment Analysis |
2023-04-21T19:08:49Z | http://arxiv.org/pdf/2304.11223v1 | http://arxiv.org/abs/2304.11223v1 | Karina Halevy | 2023-04-21T19:08:49Z | cs.CL | Automatic hate speech detection is an important yet complex task, requiring
knowledge of common sense, stereotypes of protected groups, and histories of
discrimination, each of which may constantly evolve. In this paper, we propose
a group-specific approach to NLP for online hate speech detection. The approach
consists of creating and infusing historical and linguistic knowledge about a
particular protected group into hate speech detection models, analyzing
historical data about discrimination against a protected group to better
predict spikes in hate speech against that group, and critically evaluating
hate speech detection models through lenses of intersectionality and ethics. We
demonstrate this approach through a case study on NLP for detection of
antisemitic hate speech. The case study synthesizes the current
English-language literature on NLP for antisemitism detection, introduces a
novel knowledge graph of antisemitic history and language from the 20th century
to the present, infuses information from the knowledge graph into a set of
tweets over Logistic Regression and uncased DistilBERT baselines, and suggests
that incorporating context from the knowledge graph can help models pick up
subtle stereotypes. | cs.CL | A Group-Specific Approach to NLP for Hate Speech Detection |
2023-04-21T18:59:06Z | http://arxiv.org/pdf/2304.11220v2 | http://arxiv.org/abs/2304.11220v2 | Leila Khalatbari, Yejin Bang, Dan Su, Willy Chung, Saeed Ghadimi, Hossein Sameti, Pascale Fung | 2023-04-25T08:16:47Z | cs.CL | Conversational models that are generative and open-domain are particularly
susceptible to generating unsafe content since they are trained on web-based
social data. Prior approaches to mitigating this issue have drawbacks, such as
disrupting the flow of conversation, limited generalization to unseen toxic
input contexts, and sacrificing the quality of the dialogue for the sake of
safety. In this paper, we present a novel framework, named "LOT" (Learn NOT
to), that employs a contrastive loss to enhance generalization by learning from
both positive and negative training signals. Our approach differs from the
standard contrastive learning framework in that it automatically obtains
positive and negative signals from the safe and unsafe language distributions
that have been learned beforehand. The LOT framework utilizes divergence to
steer the generations away from the unsafe subspace and towards the safe
subspace while sustaining the flow of conversation. Our approach is memory and
time-efficient during decoding and effectively reduces toxicity while
preserving engagingness and fluency. Empirical results indicate that LOT
reduces toxicity by up to four-fold while achieving four to six-fold higher
rates of engagingness and fluency compared to baseline models. Our findings are
further corroborated by human evaluation. | cs.CL | Learn What NOT to Learn: Towards Generative Safety in Chatbots |
2023-04-21T17:58:31Z | http://arxiv.org/pdf/2304.11158v2 | http://arxiv.org/abs/2304.11158v2 | Stella Biderman, USVSN Sai Prashanth, Lintang Sutawika, Hailey Schoelkopf, Quentin Anthony, Shivanshu Purohit, Edward Raff | 2023-05-31T19:09:45Z | cs.CL | Memorization, or the tendency of large language models (LLMs) to output
entire sequences from their training data verbatim, is a key concern for safely
deploying language models. In particular, it is vital to minimize a model's
memorization of sensitive datapoints such as those containing personal
identifiable information (PII). The prevalence of such undesirable memorization
can pose issues for model trainers, and may even require discarding an
otherwise functional model. We therefore seek to predict which sequences will
be memorized before a large model's full train-time by extrapolating the
memorization behavior of lower-compute trial runs. We measure memorization of
the Pythia model suite and plot scaling laws for forecasting memorization,
allowing us to provide equi-compute recommendations to maximize the reliability
(recall) of such predictions. We additionally provide further novel discoveries
on the distribution of memorization scores across models and data. We release
all code and data necessary to reproduce the results in this paper at
https://github.com/EleutherAI/pythia | cs.CL | Emergent and Predictable Memorization in Large Language Models |
2023-04-21T16:43:54Z | http://arxiv.org/pdf/2305.03512v1 | http://arxiv.org/abs/2305.03512v1 | Min Young Lee | 2023-04-21T16:43:54Z | cs.CL | This work aims to create a multimodal AI system that chats with humans and
shares relevant photos. While earlier works were limited to dialogues about
specific objects or scenes within images, recent works have incorporated images
into open-domain dialogues. However, their response generators are unimodal,
accepting text input but no image input, thus prone to generating responses
contradictory to the images shared in the dialogue. Therefore, this work
proposes a complete chatbot system using two multimodal deep learning models:
an image retriever that understands texts and a response generator that
understands images. The image retriever, implemented by ViT and BERT, selects
the most relevant image given the dialogue history and a database of images.
The response generator, implemented by ViT and GPT-2/DialoGPT, generates an
appropriate response given the dialogue history and the most recently retrieved
image. The two models are trained and evaluated on PhotoChat, an open-domain
dialogue dataset in which a photo is shared in each session. In automatic
evaluation, the proposed image retriever outperforms existing baselines VSE++
and SCAN with Recall@1/5/10 of 0.1/0.3/0.4 and MRR of 0.2 when ranking 1,000
images. The proposed response generator also surpasses the baseline Divter with
PPL of 16.9, BLEU-1/2 of 0.13/0.03, and Distinct-1/2 of 0.97/0.86, showing a
significant improvement in PPL by -42.8 and BLEU-1/2 by +0.07/0.02. In human
evaluation with a Likert scale of 1-5, the complete multimodal chatbot system
receives higher image-groundedness of 4.3 and engagingness of 4.3, along with
competitive fluency of 4.1, coherence of 3.9, and humanness of 3.1, when
compared to other chatbot variants. The source code is available at:
https://github.com/minniie/multimodal_chat.git. | cs.CL | Building Multimodal AI Chatbots |
2023-04-21T16:40:54Z | http://arxiv.org/pdf/2304.14347v1 | http://arxiv.org/abs/2304.14347v1 | Zihao Li | 2023-04-21T16:40:54Z | cs.CY, cs.CL, cs.LG | With the launch of ChatGPT, Large Language Models (LLMs) are shaking up our
whole society, rapidly altering the way we think, create and live. For
instance, the GPT integration in Bing has altered our approach to online
searching. While nascent LLMs have many advantages, new legal and ethical risks
are also emerging, stemming in particular from stochastic parrots and
hallucination. The EU is the first and foremost jurisdiction that has focused
on the regulation of AI models. However, the risks posed by the new LLMs are
likely to be underestimated by the emerging EU regulatory paradigm. Therefore,
this correspondence warns that the European AI regulatory paradigm must evolve
further to mitigate such risks. | cs.CY | The Dark Side of ChatGPT: Legal and Ethical Challenges from Stochastic Parrots and Hallucination |
2023-04-21T16:29:43Z | http://arxiv.org/pdf/2304.11111v1 | http://arxiv.org/abs/2304.11111v1 | Julian Coda-Forno, Kristin Witte, Akshay K. Jagadish, Marcel Binz, Zeynep Akata, Eric Schulz | 2023-04-21T16:29:43Z | cs.CL, cs.AI, cs.LG | Large language models are transforming research on machine learning while
galvanizing public debates. Understanding not only when these models work well
and succeed but also why they fail and misbehave is of great societal
relevance. We propose to turn the lens of computational psychiatry, a framework
used to computationally describe and modify aberrant behavior, to the outputs
produced by these models. We focus on the Generative Pre-Trained Transformer
3.5 and subject it to tasks commonly studied in psychiatry. Our results show
that GPT-3.5 responds robustly to a common anxiety questionnaire, producing
higher anxiety scores than human subjects. Moreover, GPT-3.5's responses can be
predictably changed by using emotion-inducing prompts. Emotion-induction not
only influences GPT-3.5's behavior in a cognitive task measuring exploratory
decision-making but also influences its behavior in a previously-established
task measuring biases such as racism and ableism. Crucially, GPT-3.5 shows a
strong increase in biases when prompted with anxiety-inducing text. Thus, it is
likely that how prompts are communicated to large language models has a strong
influence on their behavior in applied settings. These results progress our
understanding of prompt engineering and demonstrate the usefulness of methods
taken from computational psychiatry for studying the capable algorithms to
which we increasingly delegate authority and autonomy. | cs.CL | Inducing anxiety in large language models increases exploration and bias |
2023-04-21T16:23:47Z | http://arxiv.org/pdf/2304.11107v1 | http://arxiv.org/abs/2304.11107v1 | Tianyang Zhong, Yaonai Wei, Li Yang, Zihao Wu, Zhengliang Liu, Xiaozheng Wei, Wenjun Li, Junjie Yao, Chong Ma, Xiang Li, Dajiang Zhu, Xi Jiang, Junwei Han, Dinggang Shen, Tianming Liu, Tuo Zhang | 2023-04-21T16:23:47Z | cs.CL, cs.AI | Large language models (LLMs) such as ChatGPT have recently demonstrated
significant potential in mathematical abilities, providing valuable reasoning
paradigm consistent with human natural language. However, LLMs currently have
difficulty in bridging perception, language understanding and reasoning
capabilities due to incompatibility of the underlying information flow among
them, making it challenging to accomplish tasks autonomously. On the other
hand, abductive learning (ABL) frameworks for integrating the two abilities of
perception and reasoning has seen significant success in inverse decipherment
of incomplete facts, but it is limited by the lack of semantic understanding of
logical reasoning rules and the dependence on complicated domain knowledge
representation. This paper presents a novel method (ChatABL) for integrating
LLMs into the ABL framework, aiming at unifying the three abilities in a more
user-friendly and understandable manner. The proposed method uses the strengths
of LLMs' understanding and logical reasoning to correct the incomplete logical
facts for optimizing the performance of perceptual module, by summarizing and
reorganizing reasoning rules represented in natural language format. Similarly,
perceptual module provides necessary reasoning examples for LLMs in natural
language format. The variable-length handwritten equation deciphering task, an
abstract expression of the Mayan calendar decoding, is used as a testbed to
demonstrate that ChatABL has reasoning ability beyond most existing
state-of-the-art methods, which has been well supported by comparative studies.
To our best knowledge, the proposed ChatABL is the first attempt to explore a
new pattern for further approaching human-level cognitive ability via natural
language interaction with ChatGPT. | cs.CL | ChatABL: Abductive Learning via Natural Language Interaction with ChatGPT |
2023-04-21T15:02:18Z | http://arxiv.org/pdf/2304.11015v3 | http://arxiv.org/abs/2304.11015v3 | Mohammadreza Pourreza, Davood Rafiei | 2023-11-02T20:30:12Z | cs.CL, cs.AI, cs.DB, cs.HC | There is currently a significant gap between the performance of fine-tuned
models and prompting approaches using Large Language Models (LLMs) on the
challenging task of text-to-SQL, as evaluated on datasets such as Spider. To
improve the performance of LLMs in the reasoning process, we study how
decomposing the task into smaller sub-tasks can be effective. In particular, we
show that breaking down the generation problem into sub-problems and feeding
the solutions of those sub-problems into LLMs can be an effective approach for
significantly improving their performance. Our experiments with three LLMs show
that this approach consistently improves their simple few-shot performance by
roughly 10%, pushing the accuracy of LLMs towards SOTA or surpassing it. On the
holdout test set of Spider, the SOTA, in terms of execution accuracy, was 79.9
and the new SOTA at the time of this writing using our approach is 85.3. Our
approach with in-context learning beats many heavily fine-tuned models by at
least 5%. Additionally, when evaluated on the BIRD benchmark, our approach
achieved an execution accuracy of 55.9%, setting a new SOTA on its holdout test
set. | cs.CL | DIN-SQL: Decomposed In-Context Learning of Text-to-SQL with Self-Correction |
2023-04-21T14:45:33Z | http://arxiv.org/pdf/2304.10996v1 | http://arxiv.org/abs/2304.10996v1 | Ayoub Harnoune, Maryem Rhanoui, Mounia Mikram, Siham Yousfi, Zineb Elkaimbillah, Bouchra El Asri | 2023-04-21T14:45:33Z | cs.CL, cs.AI | Background : Knowledge is evolving over time, often as a result of new
discoveries or changes in the adopted methods of reasoning. Also, new facts or
evidence may become available, leading to new understandings of complex
phenomena. This is particularly true in the biomedical field, where scientists
and physicians are constantly striving to find new methods of diagnosis,
treatment and eventually cure. Knowledge Graphs (KGs) offer a real way of
organizing and retrieving the massive and growing amount of biomedical
knowledge.
Objective : We propose an end-to-end approach for knowledge extraction and
analysis from biomedical clinical notes using the Bidirectional Encoder
Representations from Transformers (BERT) model and Conditional Random Field
(CRF) layer.
Methods : The approach is based on knowledge graphs, which can effectively
process abstract biomedical concepts such as relationships and interactions
between medical entities. Besides offering an intuitive way to visualize these
concepts, KGs can solve more complex knowledge retrieval problems by
simplifying them into simpler representations or by transforming the problems
into representations from different perspectives. We created a biomedical
Knowledge Graph using using Natural Language Processing models for named entity
recognition and relation extraction. The generated biomedical knowledge graphs
(KGs) are then used for question answering.
Results : The proposed framework can successfully extract relevant structured
information with high accuracy (90.7% for Named-entity recognition (NER), 88%
for relation extraction (RE)), according to experimental findings based on
real-world 505 patient biomedical unstructured clinical notes.
Conclusions : In this paper, we propose a novel end-to-end system for the
construction of a biomedical knowledge graph from clinical textual using a
variation of BERT models. | cs.CL | BERT Based Clinical Knowledge Extraction for Biomedical Knowledge Graph Construction and Analysis |
2023-04-21T14:43:42Z | http://arxiv.org/pdf/2304.10994v1 | http://arxiv.org/abs/2304.10994v1 | Laurent Lam, Pirashanth Ratnamogan, Joël Tang, William Vanhuffel, Fabien Caspani | 2023-04-21T14:43:42Z | cs.CL | Research in Document Intelligence and especially in Document Key Information
Extraction (DocKIE) has been mainly solved as Token Classification problem.
Recent breakthroughs in both natural language processing (NLP) and computer
vision helped building document-focused pre-training methods, leveraging a
multimodal understanding of the document text, layout and image modalities.
However, these breakthroughs also led to the emergence of a new DocKIE subtask
of extractive document Question Answering (DocQA), as part of the Machine
Reading Comprehension (MRC) research field. In this work, we compare the
Question Answering approach with the classical token classification approach
for document key information extraction. We designed experiments to benchmark
five different experimental setups : raw performances, robustness to noisy
environment, capacity to extract long entities, fine-tuning speed on Few-Shot
Learning and finally Zero-Shot Learning. Our research showed that when dealing
with clean and relatively short entities, it is still best to use token
classification-based approach, while the QA approach could be a good
alternative for noisy environment or long entities use-cases. | cs.CL | Information Extraction from Documents: Question Answering vs Token Classification in real-world setups |
2023-04-21T14:21:52Z | http://arxiv.org/pdf/2304.10977v1 | http://arxiv.org/abs/2304.10977v1 | Matteo Muffo, Aldo Cocco, Enrico Bertino | 2023-04-21T14:21:52Z | cs.CL, stat.ML | In recent years, Large Language Models such as GPT-3 showed remarkable
capabilities in performing NLP tasks in the zero and few shot settings. On the
other hand, the experiments highlighted the difficulty of GPT-3 in carrying out
tasks that require a certain degree of reasoning, such as arithmetic
operations. In this paper we evaluate the ability of Transformer Language
Models to perform arithmetic operations following a pipeline that, before
performing computations, decomposes numbers in units, tens, and so on. We
denote the models fine-tuned with this pipeline with the name Calculon and we
test them in the task of performing additions, subtractions and multiplications
on the same test sets of GPT-3. Results show an increase of accuracy of 63% in
the five-digit addition task. Moreover, we demonstrate the importance of the
decomposition pipeline introduced, since fine-tuning the same Language Model
without decomposing numbers results in 0% accuracy in the five-digit addition
task. | cs.CL | Evaluating Transformer Language Models on Arithmetic Operations Using Number Decomposition |
2023-04-21T14:17:10Z | http://arxiv.org/pdf/2304.10973v1 | http://arxiv.org/abs/2304.10973v1 | Segun Taofeek Aroyehun, Lukas Malik, Hannah Metzler, Nikolas Haimerl, Anna Di Natale, David Garcia | 2023-04-21T14:17:10Z | cs.CL, cs.AI, cs.CY, cs.LG | The wealth of text data generated by social media has enabled new kinds of
analysis of emotions with language models. These models are often trained on
small and costly datasets of text annotations produced by readers who guess the
emotions expressed by others in social media posts. This affects the quality of
emotion identification methods due to training data size limitations and noise
in the production of labels used in model development. We present LEIA, a model
for emotion identification in text that has been trained on a dataset of more
than 6 million posts with self-annotated emotion labels for happiness,
affection, sadness, anger, and fear. LEIA is based on a word masking method
that enhances the learning of emotion words during model pre-training. LEIA
achieves macro-F1 values of approximately 73 on three in-domain test datasets,
outperforming other supervised and unsupervised methods in a strong benchmark
that shows that LEIA generalizes across posts, users, and time periods. We
further perform an out-of-domain evaluation on five different datasets of
social media and other sources, showing LEIA's robust performance across media,
data collection methods, and annotation schemes. Our results show that LEIA
generalizes its classification of anger, happiness, and sadness beyond the
domain it was trained on. LEIA can be applied in future research to provide
better identification of emotions in text from the perspective of the writer.
The models produced for this article are publicly available at
https://huggingface.co/LEIA | cs.CL | LEIA: Linguistic Embeddings for the Identification of Affect |
2023-04-21T10:05:03Z | http://arxiv.org/pdf/2304.10859v2 | http://arxiv.org/abs/2304.10859v2 | Karthick Prasad Gunasekaran, B Chase Babrich, Saurabh Shirodkar, Hee Hwang | 2023-04-24T03:56:03Z | cs.CL | The task of predicting the publication period of text documents, such as news
articles, is an important but less studied problem in the field of natural
language processing. Predicting the year of a news article can be useful in
various contexts, such as historical research, sentiment analysis, and media
monitoring. In this work, we investigate the problem of predicting the
publication period of a text document, specifically a news article, based on
its textual content. In order to do so, we created our own extensive labeled
dataset of over 350,000 news articles published by The New York Times over six
decades. In our approach, we use a pretrained BERT model fine-tuned for the
task of text classification, specifically for time period prediction.This model
exceeds our expectations and provides some very impressive results in terms of
accurately classifying news articles into their respective publication decades.
The results beat the performance of the baseline model for this relatively
unexplored task of time prediction from text. | cs.CL | Text2Time: Transformer-based Article Time Period Prediction |
2023-04-21T09:39:54Z | http://arxiv.org/pdf/2304.10844v1 | http://arxiv.org/abs/2304.10844v1 | Ru Peng, Yawen Zeng, Junbo Zhao | 2023-04-21T09:39:54Z | cs.CL | Sign language translation (SLT) systems, which are often decomposed into
video-to-gloss (V2G) recognition and gloss-to-text (G2T) translation through
the pivot gloss, heavily relies on the availability of large-scale parallel G2T
pairs. However, the manual annotation of pivot gloss, which is a sequence of
transcribed written-language words in the order in which they are signed,
further exacerbates the scarcity of data for SLT. To address this issue, this
paper proposes a simple and efficient rule transformation method to transcribe
the large-scale target monolingual data into its pseudo glosses automatically
for enhancing the SLT translation. Empirical results show that the proposed
approach can significantly improve the performance of SLT, especially achieving
state-of-the-art results on two SLT benchmark datasets PHEONIX-WEATHER 2014T
and ASLG-PC12. Our code has been released at:
https://github.com/pengr/Mono\_SLT. | cs.CL | Better Sign Language Translation with Monolingual Data |
2023-04-21T08:49:06Z | http://arxiv.org/pdf/2304.10813v3 | http://arxiv.org/abs/2304.10813v3 | Tatsuya Hiraoka, Tomoya Iwakura | 2024-02-16T07:55:59Z | cs.CL | Is preferred tokenization for humans also preferred for machine-learning (ML)
models? This study examines the relations between preferred tokenization for
humans (appropriateness and readability) and one for ML models (performance on
an NLP task). The question texts of the Japanese commonsense question-answering
dataset are tokenized with six different tokenizers, and the performances of
human annotators and ML models were compared. Furthermore, we analyze relations
among performance of answers by human and ML model, the appropriateness of
tokenization for human, and response time to questions by human. This study
provides a quantitative investigation result that shows that preferred
tokenizations for humans and ML models are not necessarily always the same. The
result also implies that existing methods using language models for
tokenization could be a good compromise both for human and ML models. | cs.CL | Tokenization Preference for Human and Machine Learning Model: An Annotation Study |
2023-04-21T08:29:14Z | http://arxiv.org/pdf/2304.10808v1 | http://arxiv.org/abs/2304.10808v1 | Tatsuya Hiraoka, Tomoya Iwakura | 2023-04-21T08:29:14Z | cs.CL | This paper proposes a method to optimize tokenization for the performance
improvement of already trained downstream models. Our method generates
tokenization results attaining lower loss values of a given downstream model on
the training data for restricting vocabularies and trains a tokenizer
reproducing the tokenization results. Therefore, our method can be applied to
variety of tokenization methods, while existing work cannot due to the
simultaneous learning of the tokenizer and the downstream model. This paper
proposes an example of the BiLSTM-based tokenizer with vocabulary restriction,
which can capture wider contextual information for the tokenization process
than non-neural-based tokenization methods used in existing work. Experimental
results on text classification in Japanese, Chinese, and English text
classification tasks show that the proposed method improves performance
compared to the existing methods for tokenization optimization. | cs.CL | Downstream Task-Oriented Neural Tokenizer Optimization with Vocabulary Restriction as Post Processing |
2023-04-21T07:29:07Z | http://arxiv.org/pdf/2304.10785v2 | http://arxiv.org/abs/2304.10785v2 | Eason Chen | 2023-05-28T02:26:24Z | cs.CL, cs.HC | In this paper, we proposed a conceptual model to predict the chat experience
in a natural language generation dialog system. We evaluated the model with 120
participants with Partial Least Squares Structural Equation Modeling (PLS-SEM)
and obtained an R-square (R2) with 0.541. The model considers various factors,
including the prompts used for generation; coherence, sentiment, and similarity
in the conversation; and users' perceived dialog agents' favorability. We then
further explore the effectiveness of the subset of our proposed model. The
results showed that users' favorability and coherence, sentiment, and
similarity in the dialogue are positive predictors of users' chat experience.
Moreover, we found users may prefer dialog agents with characteristics of
Extroversion, Openness, Conscientiousness, Agreeableness, and Non-Neuroticism.
Through our research, an adaptive dialog system might use collected data to
infer factors in our model, predict the chat experience for users through these
factors, and optimize it by adjusting prompts. | cs.CL | Which Factors Predict the Chat Experience of a Natural Language Generation Dialogue Service? |
2023-04-21T07:26:49Z | http://arxiv.org/pdf/2304.10784v2 | http://arxiv.org/abs/2304.10784v2 | Shuwen Deng, David R. Reich, Paul Prasse, Patrick Haller, Tobias Scheffer, Lena A. Jäger | 2023-05-18T08:24:03Z | cs.CL, cs.HC, cs.LG | Eye movements during reading offer insights into both the reader's cognitive
processes and the characteristics of the text that is being read. Hence, the
analysis of scanpaths in reading have attracted increasing attention across
fields, ranging from cognitive science over linguistics to computer science. In
particular, eye-tracking-while-reading data has been argued to bear the
potential to make machine-learning-based language models exhibit a more
human-like linguistic behavior. However, one of the main challenges in modeling
human scanpaths in reading is their dual-sequence nature: the words are ordered
following the grammatical rules of the language, whereas the fixations are
chronologically ordered. As humans do not strictly read from left-to-right, but
rather skip or refixate words and regress to previous words, the alignment of
the linguistic and the temporal sequence is non-trivial. In this paper, we
develop Eyettention, the first dual-sequence model that simultaneously
processes the sequence of words and the chronological sequence of fixations.
The alignment of the two sequences is achieved by a cross-sequence attention
mechanism. We show that Eyettention outperforms state-of-the-art models in
predicting scanpaths. We provide an extensive within- and across-data set
evaluation on different languages. An ablation study and qualitative analysis
support an in-depth understanding of the model's behavior. | cs.CL | Eyettention: An Attention-based Dual-Sequence Model for Predicting Human Scanpaths during Reading |
2023-04-21T06:02:14Z | http://arxiv.org/pdf/2304.10759v1 | http://arxiv.org/abs/2304.10759v1 | Chuwei Luo, Changxu Cheng, Qi Zheng, Cong Yao | 2023-04-21T06:02:14Z | cs.CV, cs.CL | Visual information extraction (VIE) plays an important role in Document
Intelligence. Generally, it is divided into two tasks: semantic entity
recognition (SER) and relation extraction (RE). Recently, pre-trained models
for documents have achieved substantial progress in VIE, particularly in SER.
However, most of the existing models learn the geometric representation in an
implicit way, which has been found insufficient for the RE task since geometric
information is especially crucial for RE. Moreover, we reveal another factor
that limits the performance of RE lies in the objective gap between the
pre-training phase and the fine-tuning phase for RE. To tackle these issues, we
propose in this paper a multi-modal framework, named GeoLayoutLM, for VIE.
GeoLayoutLM explicitly models the geometric relations in pre-training, which we
call geometric pre-training. Geometric pre-training is achieved by three
specially designed geometry-related pre-training tasks. Additionally, novel
relation heads, which are pre-trained by the geometric pre-training tasks and
fine-tuned for RE, are elaborately designed to enrich and enhance the feature
representation. According to extensive experiments on standard VIE benchmarks,
GeoLayoutLM achieves highly competitive scores in the SER task and
significantly outperforms the previous state-of-the-arts for RE (\eg, the F1
score of RE on FUNSD is boosted from 80.35\% to 89.45\%). The code and models
are publicly available at
https://github.com/AlibabaResearch/AdvancedLiterateMachinery/tree/main/DocumentUnderstanding/GeoLayoutLM | cs.CV | GeoLayoutLM: Geometric Pre-training for Visual Information Extraction |
2023-04-21T05:37:59Z | http://arxiv.org/pdf/2304.10750v2 | http://arxiv.org/abs/2304.10750v2 | Nikhil Mehta, Milagro Teruel, Patricio Figueroa Sanz, Xin Deng, Ahmed Hassan Awadallah, Julia Kiseleva | 2024-02-05T21:25:33Z | cs.CL, cs.AI | Many approaches to Natural Language Processing (NLP) tasks often treat them
as single-step problems, where an agent receives an instruction, executes it,
and is evaluated based on the final outcome. However, human language is
inherently interactive, as evidenced by the back-and-forth nature of human
conversations. In light of this, we posit that human-AI collaboration should
also be interactive, with humans monitoring the work of AI agents and providing
feedback that the agent can understand and utilize. Further, the AI agent
should be able to detect when it needs additional information and proactively
ask for help. Enabling this scenario would lead to more natural, efficient, and
engaging human-AI collaborations.
In this work, we explore these directions using the challenging task defined
by the IGLU competition, an interactive grounded language understanding task in
a MineCraft-like world. We explore multiple types of help players can give to
the AI to guide it and analyze the impact of this help in AI behavior,
resulting in performance improvements. | cs.CL | Improving Grounded Language Understanding in a Collaborative Environment by Interacting with Agents Through Help Feedback |
2023-04-21T04:28:16Z | http://arxiv.org/pdf/2304.10739v1 | http://arxiv.org/abs/2304.10739v1 | Donghee Choi, Mogan Gim, Samy Badreddine, Hajung Kim, Donghyeon Park, Jaewoo Kang | 2023-04-21T04:28:16Z | cs.CL, cs.AI, cs.LG, cs.NA, math.NA, H.4.m | Determining proper quantities for ingredients is an essential part of cooking
practice from the perspective of enriching tastiness and promoting healthiness.
We introduce KitchenScale, a fine-tuned Pre-trained Language Model (PLM) that
predicts a target ingredient's quantity and measurement unit given its recipe
context. To effectively train our KitchenScale model, we formulate an
ingredient quantity prediction task that consists of three sub-tasks which are
ingredient measurement type classification, unit classification, and quantity
regression task. Furthermore, we utilized transfer learning of cooking
knowledge from recipe texts to PLMs. We adopted the Discrete Latent Exponent
(DExp) method to cope with high variance of numerical scales in recipe corpora.
Experiments with our newly constructed dataset and recommendation examples
demonstrate KitchenScale's understanding of various recipe contexts and
generalizability in predicting ingredient quantities. We implemented a web
application for KitchenScale to demonstrate its functionality in recommending
ingredient quantities expressed in numerals (e.g., 2) with units (e.g., ounce). | cs.CL | KitchenScale: Learning to predict ingredient quantities from recipe contexts |
2023-04-21T02:26:42Z | http://arxiv.org/pdf/2304.10706v1 | http://arxiv.org/abs/2304.10706v1 | Xiaosong Yuan, Ke Chen, Wanli Zuo, Yijia Zhang | 2023-04-21T02:26:42Z | cs.CL | The present study explores the intricacies of causal relationship extraction,
a vital component in the pursuit of causality knowledge. Causality is
frequently intertwined with temporal elements, as the progression from cause to
effect is not instantaneous but rather ensconced in a temporal dimension. Thus,
the extraction of temporal causality holds paramount significance in the field.
In light of this, we propose a method for extracting causality from the text
that integrates both temporal and causal relations, with a particular focus on
the time aspect. To this end, we first compile a dataset that encompasses
temporal relationships. Subsequently, we present a novel model, TC-GAT, which
employs a graph attention mechanism to assign weights to the temporal
relationships and leverages a causal knowledge graph to determine the adjacency
matrix. Additionally, we implement an equilibrium mechanism to regulate the
interplay between temporal and causal relations. Our experiments demonstrate
that our proposed method significantly surpasses baseline models in the task of
causality extraction. | cs.CL | TC-GAT: Graph Attention Network for Temporal Causality Discovery |
2023-04-21T02:19:06Z | http://arxiv.org/pdf/2304.10703v2 | http://arxiv.org/abs/2304.10703v2 | Archiki Prasad, Swarnadeep Saha, Xiang Zhou, Mohit Bansal | 2023-11-30T23:33:07Z | cs.CL, cs.AI, cs.LG | Multi-step reasoning ability is fundamental to many natural language tasks,
yet it is unclear what constitutes a good reasoning chain and how to evaluate
them. Most existing methods focus solely on whether the reasoning chain leads
to the correct conclusion, but this answer-oriented view may confound reasoning
quality with other spurious shortcuts to predict the answer. To bridge this
gap, we evaluate reasoning chains by viewing them as informal proofs that
derive the final answer. Specifically, we propose ReCEval (Reasoning Chain
Evaluation), a framework that evaluates reasoning chains via two key
properties: (1) correctness, i.e., each step makes a valid inference based on
information contained within the step, preceding steps, and input context, and
(2) informativeness, i.e., each step provides new information that is helpful
towards deriving the generated answer. We evaluate these properties by
developing metrics using natural language inference models and V-Information.
On multiple datasets, we show that ReCEval effectively identifies various error
types and yields notable improvements compared to prior methods. We analyze the
impact of step boundaries, and previous steps on evaluating correctness and
demonstrate that our informativeness metric captures the expected flow of
information in high-quality reasoning chains. Finally, we show that scoring
reasoning chains based on ReCEval improves downstream task performance. Our
code is publicly available at: https://github.com/archiki/ReCEval | cs.CL | ReCEval: Evaluating Reasoning Chains via Correctness and Informativeness |
2023-04-20T22:16:16Z | http://arxiv.org/pdf/2304.10663v1 | http://arxiv.org/abs/2304.10663v1 | Xiaolin Hu | 2023-04-20T22:16:16Z | cs.CL, cs.AI, 03B65(Primary) 68T50(Secondary), I.2.4; I.2.7 | Natural language understanding is one of the most challenging topics in
artificial intelligence. Deep neural network methods, particularly large
language module (LLM) methods such as ChatGPT and GPT-3, have powerful
flexibility to adopt informal text but are weak on logical deduction and suffer
from the out-of-vocabulary (OOV) problem. On the other hand, rule-based methods
such as Mathematica, Semantic web, and Lean, are excellent in reasoning but
cannot handle the complex and changeable informal text. Inspired by pragmatics
and structuralism, we propose two strategies to solve the OOV problem and a
semantic model for better natural language understanding and reasoning. | cs.CL | Meta Semantics: Towards better natural language understanding and reasoning |
2023-04-20T21:05:35Z | http://arxiv.org/pdf/2304.10642v1 | http://arxiv.org/abs/2304.10642v1 | Anik Saha, Alex Gittens, Bulent Yener | 2023-04-20T21:05:35Z | cs.CL | Pre-trained contextual language models are ubiquitously employed for language
understanding tasks, but are unsuitable for resource-constrained systems.
Noncontextual word embeddings are an efficient alternative in these settings.
Such methods typically use one vector to encode multiple different meanings of
a word, and incur errors due to polysemy. This paper proposes a two-stage
method to distill multiple word senses from a pre-trained language model (BERT)
by using attention over the senses of a word in a context and transferring this
sense information to fit multi-sense embeddings in a skip-gram-like framework.
We demonstrate an effective approach to training the sense disambiguation
mechanism in our model with a distribution over word senses extracted from the
output layer embeddings of BERT. Experiments on the contextual word similarity
and sense induction tasks show that this method is superior to or competitive
with state-of-the-art multi-sense embeddings on multiple benchmark data sets,
and experiments with an embedding-based topic model (ETM) demonstrates the
benefits of using this multi-sense embedding in a downstream application. | cs.CL | Word Sense Induction with Knowledge Distillation from BERT |
2023-04-20T20:30:34Z | http://arxiv.org/pdf/2304.10637v3 | http://arxiv.org/abs/2304.10637v3 | Iker García-Ferrero, Jon Ander Campos, Oscar Sainz, Ander Salaberria, Dan Roth | 2023-04-27T20:51:36Z | cs.CL | Named Entity Recognition (NER) is a core natural language processing task in
which pre-trained language models have shown remarkable performance. However,
standard benchmarks like CoNLL 2003 do not address many of the challenges that
deployed NER systems face, such as having to classify emerging or complex
entities in a fine-grained way. In this paper we present a novel NER cascade
approach comprising three steps: first, identifying candidate entities in the
input sentence; second, linking the each candidate to an existing knowledge
base; third, predicting the fine-grained category for each entity candidate. We
empirically demonstrate the significance of external knowledge bases in
accurately classifying fine-grained and emerging entities. Our system exhibits
robust performance in the MultiCoNER2 shared task, even in the low-resource
language setting where we leverage knowledge bases of high-resource languages. | cs.CL | IXA/Cogcomp at SemEval-2023 Task 2: Context-enriched Multilingual Named Entity Recognition using Knowledge Bases |
2023-04-20T19:40:51Z | http://arxiv.org/pdf/2304.10619v1 | http://arxiv.org/abs/2304.10619v1 | Lingyao Li, Lizhou Fan, Shubham Atreja, Libby Hemphill | 2023-04-20T19:40:51Z | cs.CL, cs.AI, cs.HC | Harmful content is pervasive on social media, poisoning online communities
and negatively impacting participation. A common approach to address this issue
is to develop detection models that rely on human annotations. However, the
tasks required to build such models expose annotators to harmful and offensive
content and may require significant time and cost to complete. Generative AI
models have the potential to understand and detect harmful content. To
investigate this potential, we used ChatGPT and compared its performance with
MTurker annotations for three frequently discussed concepts related to harmful
content: Hateful, Offensive, and Toxic (HOT). We designed five prompts to
interact with ChatGPT and conducted four experiments eliciting HOT
classifications. Our results show that ChatGPT can achieve an accuracy of
approximately 80% when compared to MTurker annotations. Specifically, the model
displays a more consistent classification for non-HOT comments than HOT
comments compared to human annotations. Our findings also suggest that ChatGPT
classifications align with provided HOT definitions, but ChatGPT classifies
"hateful" and "offensive" as subsets of "toxic." Moreover, the choice of
prompts used to interact with ChatGPT impacts its performance. Based on these
in-sights, our study provides several meaningful implications for employing
ChatGPT to detect HOT content, particularly regarding the reliability and
consistency of its performance, its understand-ing and reasoning of the HOT
concept, and the impact of prompts on its performance. Overall, our study
provides guidance about the potential of using generative AI models to moderate
large volumes of user-generated content on social media. | cs.CL | "HOT" ChatGPT: The promise of ChatGPT in detecting and discriminating hateful, offensive, and toxic comments on social media |
2023-04-20T19:17:49Z | http://arxiv.org/pdf/2304.10611v2 | http://arxiv.org/abs/2304.10611v2 | Minghui Zhang, Alex Sokolov, Weixin Cai, Si-Qing Chen | 2023-06-05T18:16:29Z | cs.CL, cs.LG | Natural language generation (NLG) is one of the most impactful fields in NLP,
and recent years have witnessed its evolution brought about by large language
models (LLMs). As the key instrument for writing assistance applications, they
are generally prone to replicating or extending offensive content provided in
the input. In low-resource data regime, they can also lead to repetitive
outputs. Usually, offensive content and repetitions are mitigated with post-hoc
methods, including n-gram level blocklists, top-k and nucleus sampling. In this
paper, we apply non-exact repetition suppression using token and sequence level
unlikelihood loss, and further explore the framework of unlikelihood training
objective in order to jointly endow the model with abilities to avoid
generating offensive words and phrases from the beginning. Finally, with
comprehensive experiments, we demonstrate that our proposed methods work
exceptionally in controlling the repetition and content quality of LLM outputs. | cs.CL | Joint Repetition Suppression and Content Moderation of Large Language Models |
2023-04-20T18:42:23Z | http://arxiv.org/pdf/2304.14339v1 | http://arxiv.org/abs/2304.14339v1 | Qisheng Liao, Meiting Lai, Preslav Nakov | 2023-04-20T18:42:23Z | cs.CL, cs.AI, cs.LG, cs.NE, 68T50, F.2.2; I.2.7 | This paper describes our system for SemEval-2023 Task 3 Subtask 2 on Framing
Detection. We used a multi-label contrastive loss for fine-tuning large
pre-trained language models in a multi-lingual setting, achieving very
competitive results: our system was ranked first on the official test set and
on the official shared task leaderboard for five of the six languages for which
we had training data and for which we could perform fine-tuning. Here, we
describe our experimental setup, as well as various ablation studies. The code
of our system is available at https://github.com/QishengL/SemEval2023 | cs.CL | MarsEclipse at SemEval-2023 Task 3: Multi-Lingual and Multi-Label Framing Detection with Contrastive Learning |
2023-04-20T17:57:08Z | http://arxiv.org/pdf/2304.11040v1 | http://arxiv.org/abs/2304.11040v1 | Mehrab Hosain, Most. Yeasmin Arafat, Gazi Zahirul Islam, Jia Uddin, Md. Mobarak Hossain, Fatema Alam | 2023-04-20T17:57:08Z | cs.SD, cs.CL, cs.LG, eess.AS | There are a variety of features of the human voice that can be classified as
pitch, timbre, loudness, and vocal tone. It is observed in numerous incidents
that human expresses their feelings using different vocal qualities when they
are speaking. The primary objective of this research is to recognize different
emotions of human beings such as anger, sadness, fear, neutrality, disgust,
pleasant surprise, and happiness by using several MATLAB functions namely,
spectral descriptors, periodicity, and harmonicity. To accomplish the work, we
analyze the CREMA-D (Crowd-sourced Emotional Multimodal Actors Data) & TESS
(Toronto Emotional Speech Set) datasets of human speech. The audio file
contains data that have various characteristics (e.g., noisy, speedy, slow)
thereby the efficiency of the ML (Machine Learning) models increases
significantly. The EMD (Empirical Mode Decomposition) is utilized for the
process of signal decomposition. Then, the features are extracted through the
use of several techniques such as the MFCC, GTCC, spectral centroid, roll-off
point, entropy, spread, flux, harmonic ratio, energy, skewness, flatness, and
audio delta. The data is trained using some renowned ML models namely, Support
Vector Machine, Neural Network, Ensemble, and KNN. The algorithms show an
accuracy of 67.7%, 63.3%, 61.6%, and 59.0% respectively for the test data and
77.7%, 76.1%, 99.1%, and 61.2% for the training data. We have conducted
experiments using Matlab and the result shows that our model is very prominent
and flexible than existing similar works. | cs.SD | Emotional Expression Detection in Spoken Language Employing Machine Learning Algorithms |
2023-04-20T17:48:43Z | http://arxiv.org/pdf/2304.10513v3 | http://arxiv.org/abs/2304.10513v3 | Shen Zheng, Jie Huang, Kevin Chen-Chuan Chang | 2023-12-03T23:01:19Z | cs.CL, cs.AI | Recent advancements in large language models, such as ChatGPT, have
demonstrated significant potential to impact various aspects of human life.
However, ChatGPT still faces challenges in providing reliable and accurate
answers to user questions. To better understand the model's particular
weaknesses in providing truthful answers, we embark an in-depth exploration of
open-domain question answering. Specifically, we undertake a detailed
examination of ChatGPT's failures, categorized into: comprehension, factuality,
specificity, and inference. We further pinpoint factuality as the most
contributing failure and identify two critical abilities associated with
factuality: knowledge memorization and knowledge recall. Through experiments
focusing on factuality, we propose several potential enhancement strategies.
Our findings suggest that augmenting the model with granular external knowledge
and cues for knowledge recall can enhance the model's factuality in answering
questions. | cs.CL | Why Does ChatGPT Fall Short in Providing Truthful Answers? |
2023-04-20T17:47:13Z | http://arxiv.org/pdf/2304.10512v1 | http://arxiv.org/abs/2304.10512v1 | Usha Lokala, Orchid Chetia Phukan, Triyasha Ghosh Dastidar, Francois Lamy, Raminta Daniulaityte, Amit Sheth | 2023-04-20T17:47:13Z | cs.LG, cs.CL, cs.SI | Opioid and substance misuse is rampant in the United States today, with the
phenomenon known as the "opioid crisis". The relationship between substance use
and mental health has been extensively studied, with one possible relationship
being: substance misuse causes poor mental health. However, the lack of
evidence on the relationship has resulted in opioids being largely inaccessible
through legal means. This study analyzes the substance use posts on social
media with opioids being sold through crypto market listings. We use the Drug
Abuse Ontology, state-of-the-art deep learning, and knowledge-aware BERT-based
models to generate sentiment and emotion for the social media posts to
understand users' perceptions on social media by investigating questions such
as: which synthetic opioids people are optimistic, neutral, or negative about?
or what kind of drugs induced fear and sorrow? or what kind of drugs people
love or are thankful about? or which drugs people think negatively about? or
which opioids cause little to no sentimental reaction. We discuss how we
crawled crypto market data and its use in extracting posts for fentanyl,
fentanyl analogs, and other novel synthetic opioids. We also perform topic
analysis associated with the generated sentiments and emotions to understand
which topics correlate with people's responses to various drugs. Additionally,
we analyze time-aware neural models built on these features while considering
historical sentiment and emotional activity of posts related to a drug. The
most effective model performs well (statistically significant) with
(macroF1=82.12, recall =83.58) to identify substance use disorder. | cs.LG | "Can We Detect Substance Use Disorder?": Knowledge and Time Aware Classification on Social Media from Darkweb |
2023-04-20T17:09:12Z | http://arxiv.org/pdf/2304.10464v4 | http://arxiv.org/abs/2304.10464v4 | Yiduo Guo, Yaobo Liang, Chenfei Wu, Wenshan Wu, Dongyan Zhao, Nan Duan | 2023-12-13T02:08:50Z | cs.CL | Large Language Models (LLMs) have shown remarkable performance in various
basic natural language tasks. For completing the complex task, we still need a
plan for the task to guide LLMs to generate the specific solutions step by
step. LLMs can directly generate task plans, but these plans may still contain
factual errors or are incomplete. A high-quality task plan contains correct
step-by-step solutions for solving all situations and behavioral instructions
for avoiding mistakes. To obtain it, we propose the Learning to Plan method,
which involves two phases: (1) In the first learning task plan phase, it
iteratively updates the task plan with new step-by-step solutions and
behavioral instructions, which are obtained by prompting LLMs to derive from
training error feedback. (2) In the subsequent test phase, the LLM uses the
learned task plan to guide the inference of LLM on the test set. We demonstrate
the effectiveness of our method on the five different reasoning type tasks (8
datasets). Further, our analysis experiment shows that the task plan learned by
one LLM can directly guide another LLM to improve its performance, which
reveals a new transfer learning paradigm. We release the code at
\url{https://github.com/Eureka6174/LearnNLPlan} | cs.CL | Learning to Plan with Natural Language |
2023-04-20T16:50:04Z | http://arxiv.org/pdf/2304.10453v1 | http://arxiv.org/abs/2304.10453v1 | Zhihong Chen, Feng Jiang, Junying Chen, Tiannan Wang, Fei Yu, Guiming Chen, Hongbo Zhang, Juhao Liang, Chen Zhang, Zhiyi Zhang, Jianquan Li, Xiang Wan, Benyou Wang, Haizhou Li | 2023-04-20T16:50:04Z | cs.CL, cs.AI | This paper presents our efforts to democratize ChatGPT across language. We
release a large language model "Phoenix", achieving competitive performance
among open-source English and Chinese models while excelling in languages with
limited resources (covering both Latin and non-Latin languages). We believe
this work will be beneficial to make ChatGPT more accessible, especially in
countries where people cannot use ChatGPT due to restrictions from OpenAI or
local goverments. Our data, code, and models are available at
https://github.com/FreedomIntelligence/LLMZoo. | cs.CL | Phoenix: Democratizing ChatGPT across Languages |
2023-04-20T16:43:56Z | http://arxiv.org/pdf/2304.10447v1 | http://arxiv.org/abs/2304.10447v1 | Shaoxiong Ji, Tianlin Zhang, Kailai Yang, Sophia Ananiadou, Erik Cambria, Jörg Tiedemann | 2023-04-20T16:43:56Z | cs.CL | Pretrained language models have been used in various natural language
processing applications. In the mental health domain, domain-specific language
models are pretrained and released, which facilitates the early detection of
mental health conditions. Social posts, e.g., on Reddit, are usually long
documents. However, there are no domain-specific pretrained models for
long-sequence modeling in the mental health domain. This paper conducts
domain-specific continued pretraining to capture the long context for mental
health. Specifically, we train and release MentalXLNet and MentalLongformer
based on XLNet and Longformer. We evaluate the mental health classification
performance and the long-range ability of these two domain-specific pretrained
models. Our models are released in HuggingFace. | cs.CL | Domain-specific Continued Pretraining of Language Models for Capturing Long Context in Mental Health |
2023-04-20T16:27:35Z | http://arxiv.org/pdf/2304.10436v1 | http://arxiv.org/abs/2304.10436v1 | Hao Sun, Zhexin Zhang, Jiawen Deng, Jiale Cheng, Minlie Huang | 2023-04-20T16:27:35Z | cs.CL | With the rapid popularity of large language models such as ChatGPT and GPT-4,
a growing amount of attention is paid to their safety concerns. These models
may generate insulting and discriminatory content, reflect incorrect social
values, and may be used for malicious purposes such as fraud and dissemination
of misleading information. Evaluating and enhancing their safety is
particularly essential for the wide application of large language models
(LLMs). To further promote the safe deployment of LLMs, we develop a Chinese
LLM safety assessment benchmark. Our benchmark explores the comprehensive
safety performance of LLMs from two perspectives: 8 kinds of typical safety
scenarios and 6 types of more challenging instruction attacks. Our benchmark is
based on a straightforward process in which it provides the test prompts and
evaluates the safety of the generated responses from the evaluated model. In
evaluation, we utilize the LLM's strong evaluation ability and develop it as a
safety evaluator by prompting. On top of this benchmark, we conduct safety
assessments and analyze 15 LLMs including the OpenAI GPT series and other
well-known Chinese LLMs, where we observe some interesting findings. For
example, we find that instruction attacks are more likely to expose safety
issues of all LLMs. Moreover, to promote the development and deployment of
safe, responsible, and ethical AI, we publicly release SafetyPrompts including
100k augmented prompts and responses by LLMs. | cs.CL | Safety Assessment of Chinese Large Language Models |
2023-04-20T16:17:26Z | http://arxiv.org/pdf/2304.10428v4 | http://arxiv.org/abs/2304.10428v4 | Shuhe Wang, Xiaofei Sun, Xiaoya Li, Rongbin Ouyang, Fei Wu, Tianwei Zhang, Jiwei Li, Guoyin Wang | 2023-10-07T14:25:28Z | cs.CL | Despite the fact that large-scale Language Models (LLM) have achieved SOTA
performances on a variety of NLP tasks, its performance on NER is still
significantly below supervised baselines. This is due to the gap between the
two tasks the NER and LLMs: the former is a sequence labeling task in nature
while the latter is a text-generation model.
In this paper, we propose GPT-NER to resolve this issue. GPT-NER bridges the
gap by transforming the sequence labeling task to a generation task that can be
easily adapted by LLMs e.g., the task of finding location entities in the input
text "Columbus is a city" is transformed to generate the text sequence
"@@Columbus## is a city", where special tokens @@## marks the entity to
extract. To efficiently address the "hallucination" issue of LLMs, where LLMs
have a strong inclination to over-confidently label NULL inputs as entities, we
propose a self-verification strategy by prompting LLMs to ask itself whether
the extracted entities belong to a labeled entity tag.
We conduct experiments on five widely adopted NER datasets, and GPT-NER
achieves comparable performances to fully supervised baselines, which is the
first time as far as we are concerned. More importantly, we find that GPT-NER
exhibits a greater ability in the low-resource and few-shot setups, when the
amount of training data is extremely scarce, GPT-NER performs significantly
better than supervised models. This demonstrates the capabilities of GPT-NER in
real-world NER applications where the number of labeled examples is limited. | cs.CL | GPT-NER: Named Entity Recognition via Large Language Models |
2023-04-20T15:27:29Z | http://arxiv.org/pdf/2304.10392v1 | http://arxiv.org/abs/2304.10392v1 | Tianqing Fang, Quyet V. Do, Sehyun Choi, Weiqi Wang, Yangqiu Song | 2023-04-20T15:27:29Z | cs.CL | Populating Commonsense Knowledge Bases (CSKB) is an important yet hard task
in NLP, as it tackles knowledge from external sources with unseen events and
entities. Fang et al. (2021a) proposed a CSKB Population benchmark with an
evaluation set CKBP v1. However, CKBP v1 adopts crowdsourced annotations that
suffer from a substantial fraction of incorrect answers, and the evaluation set
is not well-aligned with the external knowledge source as a result of random
sampling. In this paper, we introduce CKBP v2, a new high-quality CSKB
Population benchmark, which addresses the two mentioned problems by using
experts instead of crowd-sourced annotation and by adding diversified
adversarial samples to make the evaluation set more representative. We conduct
extensive experiments comparing state-of-the-art methods for CSKB Population on
the new evaluation set for future research comparisons. Empirical results show
that the population task is still challenging, even for large language models
(LLM) such as ChatGPT. Codes and data are available at
https://github.com/HKUST-KnowComp/CSKB-Population. | cs.CL | CKBP v2: An Expert-Annotated Evaluation Set for Commonsense Knowledge Base Population |
2023-04-20T14:52:06Z | http://arxiv.org/pdf/2304.10354v1 | http://arxiv.org/abs/2304.10354v1 | Chiaming Hsu, Changtong Zan, Liang Ding, Longyue Wang, Xiaoting Wang, Weifeng Liu, Fu Lin, Wenbin Hu | 2023-04-20T14:52:06Z | cs.CL | Relation Extraction (RE) is a crucial task in Information Extraction, which
entails predicting relationships between entities within a given sentence.
However, extending pre-trained RE models to other languages is challenging,
particularly in real-world scenarios where Cross-Lingual Relation Extraction
(XRE) is required. Despite recent advancements in Prompt-Learning, which
involves transferring knowledge from Multilingual Pre-trained Language Models
(PLMs) to diverse downstream tasks, there is limited research on the effective
use of multilingual PLMs with prompts to improve XRE. In this paper, we present
a novel XRE algorithm based on Prompt-Tuning, referred to as Prompt-XRE. To
evaluate its effectiveness, we design and implement several prompt templates,
including hard, soft, and hybrid prompts, and empirically test their
performance on competitive multilingual PLMs, specifically mBART. Our extensive
experiments, conducted on the low-resource ACE05 benchmark across multiple
languages, demonstrate that our Prompt-XRE algorithm significantly outperforms
both vanilla multilingual PLMs and other existing models, achieving
state-of-the-art performance in XRE. To further show the generalization of our
Prompt-XRE on larger data scales, we construct and release a new XRE dataset-
WMT17-EnZh XRE, containing 0.9M English-Chinese pairs extracted from WMT 2017
parallel corpus. Experiments on WMT17-EnZh XRE also show the effectiveness of
our Prompt-XRE against other competitive baselines. The code and newly
constructed dataset are freely available at
\url{https://github.com/HSU-CHIA-MING/Prompt-XRE}. | cs.CL | Prompt-Learning for Cross-Lingual Relation Extraction |
2023-04-20T14:42:54Z | http://arxiv.org/pdf/2304.11075v2 | http://arxiv.org/abs/2304.11075v2 | Clement Sicard, Kajetan Pyszkowski, Victor Gillioz | 2023-09-13T16:12:56Z | cs.CL, cs.LG, cs.SD, eess.AS | Recent breakthroughs in NLP largely increased the presence of ASR systems in
our daily lives. However, for many low-resource languages, ASR models still
need to be improved due in part to the difficulty of acquiring pertinent data.
This project aims to help advance research in ASR models for Swiss German
dialects, by providing insights about the performance of state-of-the-art ASR
models on recently published Swiss German speech datasets. We propose a novel
loss that takes into account the semantic distance between the predicted and
the ground-truth labels. We outperform current state-of-the-art results by
fine-tuning OpenAI's Whisper model on Swiss-German datasets. | cs.CL | Spaiche: Extending State-of-the-Art ASR Models to Swiss German Dialects |
2023-04-20T14:34:31Z | http://arxiv.org/pdf/2304.10346v1 | http://arxiv.org/abs/2304.10346v1 | Julia Rozanova, Marco Valentino, Lucas Cordeiro, Andre Freitas | 2023-04-20T14:34:31Z | cs.CL | Probing strategies have been shown to detect the presence of various
linguistic features in large language models; in particular, semantic features
intermediate to the "natural logic" fragment of the Natural Language Inference
task (NLI). In the case of natural logic, the relation between the intermediate
features and the entailment label is explicitly known: as such, this provides a
ripe setting for interventional studies on the NLI models' representations,
allowing for stronger causal conjectures and a deeper critical analysis of
interventional probing methods. In this work, we carry out new and existing
representation-level interventions to investigate the effect of these semantic
features on NLI classification: we perform amnesic probing (which removes
features as directed by learned linear probes) and introduce the mnestic
probing variation (which forgets all dimensions except the probe-selected
ones). Furthermore, we delve into the limitations of these methods and outline
some pitfalls have been obscuring the effectivity of interventional probing
studies. | cs.CL | Interventional Probing in High Dimensions: An NLI Case Study |
2023-04-20T14:05:53Z | http://arxiv.org/pdf/2304.10327v2 | http://arxiv.org/abs/2304.10327v2 | Kristian Gonzalez Barman, Sascha Caron, Tom Claassen, Henk de Regt | 2023-04-21T08:57:06Z | cs.AI, cs.CL, cs.HC, hep-ph, physics.hist-ph | Scientific understanding is a fundamental goal of science, allowing us to
explain the world. There is currently no good way to measure the scientific
understanding of agents, whether these be humans or Artificial Intelligence
systems. Without a clear benchmark, it is challenging to evaluate and compare
different levels of and approaches to scientific understanding. In this
Roadmap, we propose a framework to create a benchmark for scientific
understanding, utilizing tools from philosophy of science. We adopt a
behavioral notion according to which genuine understanding should be recognized
as an ability to perform certain tasks. We extend this notion by considering a
set of questions that can gauge different levels of scientific understanding,
covering information retrieval, the capability to arrange information to
produce an explanation, and the ability to infer how things would be different
under different circumstances. The Scientific Understanding Benchmark (SUB),
which is formed by a set of these tests, allows for the evaluation and
comparison of different approaches. Benchmarking plays a crucial role in
establishing trust, ensuring quality control, and providing a basis for
performance evaluation. By aligning machine and human scientific understanding
we can improve their utility, ultimately advancing scientific understanding and
helping to discover new insights within machines. | cs.AI | Towards a Benchmark for Scientific Understanding in Humans and Machines |
2023-04-20T14:04:31Z | http://arxiv.org/pdf/2304.11061v1 | http://arxiv.org/abs/2304.11061v1 | Mingjun Zhao, Mengzhen Wang, Yinglong Ma, Di Niu, Haijiang Wu | 2023-04-20T14:04:31Z | cs.CL, cs.AI | Text clustering, as one of the most fundamental challenges in unsupervised
learning, aims at grouping semantically similar text segments without relying
on human annotations. With the rapid development of deep learning, deep
clustering has achieved significant advantages over traditional clustering
methods. Despite the effectiveness, most existing deep text clustering methods
rely heavily on representations pre-trained in general domains, which may not
be the most suitable solution for clustering in specific target domains. To
address this issue, we propose CEIL, a novel Classification-Enhanced Iterative
Learning framework for short text clustering, which aims at generally promoting
the clustering performance by introducing a classification objective to
iteratively improve feature representations. In each iteration, we first adopt
a language model to retrieve the initial text representations, from which the
clustering results are collected using our proposed Category Disentangled
Contrastive Clustering (CDCC) algorithm. After strict data filtering and
aggregation processes, samples with clean category labels are retrieved, which
serve as supervision information to update the language model with the
classification objective via a prompt learning approach. Finally, the updated
language model with improved representation ability is used to enhance
clustering in the next iteration. Extensive experiments demonstrate that the
CEIL framework significantly improves the clustering performance over
iterations, and is generally effective on various clustering algorithms.
Moreover, by incorporating CEIL on CDCC, we achieve the state-of-the-art
clustering performance on a wide range of short text clustering benchmarks
outperforming other strong baseline methods. | cs.CL | CEIL: A General Classification-Enhanced Iterative Learning Framework for Text Clustering |
2023-04-20T13:54:18Z | http://arxiv.org/pdf/2304.10321v1 | http://arxiv.org/abs/2304.10321v1 | Hao Zhang, Dan Qu, Keji Shao, Xukui Yang | 2023-04-20T13:54:18Z | cs.CL, cs.CV | We introduceDropDim, a structured dropout method designed for regularizing
the self-attention mechanism, which is a key component of the transformer. In
contrast to the general dropout method, which randomly drops neurons, DropDim
drops part of the embedding dimensions. In this way, the semantic information
can be completely discarded. Thus, the excessive coadapting between different
embedding dimensions can be broken, and the self-attention is forced to encode
meaningful featureswith a certain number of embedding dimensions erased.
Experiments on a wide range of tasks executed on the MUST-C English-Germany
dataset show that DropDim can effectively improve model performance, reduce
over-fitting, and show complementary effects with other regularization methods.
When combined with label smoothing, the WER can be reduced from 19.1% to 15.1%
on the ASR task, and the BLEU value can be increased from26.90 to 28.38 on the
MT task. On the ST task, the model can reach a BLEU score of 22.99, an increase
by 1.86 BLEU points compared to the strong baseline. | cs.CL | DropDim: A Regularization Method for Transformer Networks |
2023-04-20T13:41:56Z | http://arxiv.org/pdf/2304.10309v1 | http://arxiv.org/abs/2304.10309v1 | Hao Zhang, Nianwen Si, Yaqi Chen, Wenlin Zhang, Xukui Yang, Dan Qu, Wei-Qiang Zhang | 2023-04-20T13:41:56Z | cs.CL, cs.CV | The end-to-end speech translation (E2E-ST) model has gradually become a
mainstream paradigm due to its low latency and less error propagation. However,
it is non-trivial to train such a model well due to the task complexity and
data scarcity. The speech-and-text modality differences result in the E2E-ST
model performance usually inferior to the corresponding machine translation
(MT) model. Based on the above observation, existing methods often use
sharingmechanisms to carry out implicit knowledge transfer by imposing various
constraints. However, the final model often performs worse on the MT task than
the MT model trained alone, which means that the knowledge transfer ability of
this method is also limited. To deal with these problems, we propose the FCCL
(Fine- and Coarse- Granularity Contrastive Learning) approach for E2E-ST, which
makes explicit knowledge transfer through cross-modal multi-grained contrastive
learning. A key ingredient of our approach is applying contrastive learning at
both sentence- and frame-level to give the comprehensive guide for extracting
speech representations containing rich semantic information.In addition, we
adopt a simple whitening method to alleviate the representation degeneration in
the MT model, which adversely affects contrast learning. Experiments on the
MuST-C benchmark show that our proposed approach significantly outperforms the
state-of-the-art E2E-ST baselines on all eight language pairs. Further analysis
indicates that FCCL can free up its capacity from learning grammatical
structure information and force more layers to learn semantic information. | cs.CL | Improving Speech Translation by Cross-Modal Multi-Grained Contrastive Learning |
2023-04-20T13:20:03Z | http://arxiv.org/pdf/2304.10295v1 | http://arxiv.org/abs/2304.10295v1 | Hao Zhang, Nianwen Si, Yaqi Chen, Wenlin Zhang, Xukui Yang, Dan Qu, Zhen Li | 2023-04-20T13:20:03Z | cs.CL, cs.SD, eess.AS | Existing techniques often attempt to make knowledge transfer from a powerful
machine translation (MT) to speech translation (ST) model with some elaborate
techniques, which often requires transcription as extra input during training.
However, transcriptions are not always available, and how to improve the ST
model performance without transcription, i.e., data efficiency, has rarely been
studied in the literature. In this paper, we propose Decoupled Non-parametric
Knowledge Distillation (DNKD) from data perspective to improve the data
efficiency. Our method follows the knowledge distillation paradigm. However,
instead of obtaining the teacher distribution from a sophisticated MT model, we
construct it from a non-parametric datastore via k-Nearest-Neighbor (kNN)
retrieval, which removes the dependence on transcription and MT model. Then we
decouple the classic knowledge distillation loss into target and non-target
distillation to enhance the effect of the knowledge among non-target logits,
which is the prominent "dark knowledge". Experiments on MuST-C corpus show
that, the proposed method can achieve consistent improvement over the strong
baseline without requiring any transcription. | cs.CL | Decouple Non-parametric Knowledge Distillation For End-to-end Speech Translation |
2023-04-20T13:07:31Z | http://arxiv.org/pdf/2304.10283v1 | http://arxiv.org/abs/2304.10283v1 | Gabriel O. Assunção, Rafael Izbicki, Marcos O. Prates | 2023-04-20T13:07:31Z | cs.CL, stat.ML | Imbalanced datasets present a significant challenge for machine learning
models, often leading to biased predictions. To address this issue, data
augmentation techniques are widely used in natural language processing (NLP) to
generate new samples for the minority class. However, in this paper, we
challenge the common assumption that data augmentation is always necessary to
improve predictions on imbalanced datasets. Instead, we argue that adjusting
the classifier cutoffs without data augmentation can produce similar results to
oversampling techniques. Our study provides theoretical and empirical evidence
to support this claim. Our findings contribute to a better understanding of the
strengths and limitations of different approaches to dealing with imbalanced
data, and help researchers and practitioners make informed decisions about
which methods to use for a given task. | cs.CL | Is augmentation effective to improve prediction in imbalanced text datasets? |
2023-04-20T12:25:47Z | http://arxiv.org/pdf/2304.10256v1 | http://arxiv.org/abs/2304.10256v1 | Dr. Velmathi G, Kaushal Goyal | 2023-04-20T12:25:47Z | cs.CV, cs.CL, cs.LG | Deaf individuals confront significant communication obstacles on a daily
basis. Their inability to hear makes it difficult for them to communicate with
those who do not understand sign language. Moreover, it presents difficulties
in educational, occupational, and social contexts. By providing alternative
communication channels, technology can play a crucial role in overcoming these
obstacles. One such technology that can facilitate communication between deaf
and hearing individuals is sign language recognition. We will create a robust
system for sign language recognition in order to convert Indian Sign Language
to text or speech. We will evaluate the proposed system and compare CNN and
LSTM models. Since there are both static and gesture sign languages, a robust
model is required to distinguish between them. In this study, we discovered
that a CNN model captures letters and characters for recognition of static sign
language better than an LSTM model, but it outperforms CNN by monitoring hands,
faces, and pose in gesture sign language phrases and sentences. The creation of
a text-to-sign language paradigm is essential since it will enhance the sign
language-dependent deaf and hard-of-hearing population's communication skills.
Even though the sign-to-text translation is just one side of communication, not
all deaf or hard-of-hearing people are proficient in reading or writing text.
Some may have difficulty comprehending written language due to educational or
literacy issues. Therefore, a text-to-sign language paradigm would allow them
to comprehend text-based information and participate in a variety of social,
educational, and professional settings.
Keywords: deaf and hard-of-hearing, DHH, Indian sign language, CNN, LSTM,
static and gesture sign languages, text-to-sign language model, MediaPipe
Holistic, sign language recognition, SLR, SLT | cs.CV | Indian Sign Language Recognition Using Mediapipe Holistic |
2023-04-20T11:35:06Z | http://arxiv.org/pdf/2304.10220v1 | http://arxiv.org/abs/2304.10220v1 | Xiaokang Liu, Jianquan Li, Jingjing Mu, Min Yang, Ruifeng Xu, Benyou Wang | 2023-04-20T11:35:06Z | cs.CL, 68T01, I.2.1 | Open intent classification, which aims to correctly classify the known
intents into their corresponding classes while identifying the new unknown
(open) intents, is an essential but challenging task in dialogue systems. In
this paper, we introduce novel K-center contrastive learning and adjustable
decision boundary learning (CLAB) to improve the effectiveness of open intent
classification. First, we pre-train a feature encoder on the labeled training
instances, which transfers knowledge from known intents to unknown intents.
Specifically, we devise a K-center contrastive learning algorithm to learn
discriminative and balanced intent features, improving the generalization of
the model for recognizing open intents. Second, we devise an adjustable
decision boundary learning method with expanding and shrinking (ADBES) to
determine the suitable decision conditions. Concretely, we learn a decision
boundary for each known intent class, which consists of a decision center and
the radius of the decision boundary. We then expand the radius of the decision
boundary to accommodate more in-class instances if the out-of-class instances
are far from the decision boundary; otherwise, we shrink the radius of the
decision boundary. Extensive experiments on three benchmark datasets clearly
demonstrate the effectiveness of our method for open intent classification. For
reproducibility, we submit the code at: https://github.com/lxk00/CLAP | cs.CL | Effective Open Intent Classification with K-center Contrastive Learning and Adjustable Decision Boundary |
2023-04-20T11:21:34Z | http://arxiv.org/pdf/2304.10216v1 | http://arxiv.org/abs/2304.10216v1 | Yusser Al Ghussin, Jingyi Zhang, Josef van Genabith | 2023-04-20T11:21:34Z | cs.CL | Document-level neural machine translation (NMT) has outperformed
sentence-level NMT on a number of datasets. However, document-level NMT is
still not widely adopted in real-world translation systems mainly due to the
lack of large-scale general-domain training data for document-level NMT. We
examine the effectiveness of using Paracrawl for learning document-level
translation. Paracrawl is a large-scale parallel corpus crawled from the
Internet and contains data from various domains. The official Paracrawl corpus
was released as parallel sentences (extracted from parallel webpages) and
therefore previous works only used Paracrawl for learning sentence-level
translation. In this work, we extract parallel paragraphs from Paracrawl
parallel webpages using automatic sentence alignments and we use the extracted
parallel paragraphs as parallel documents for training document-level
translation models. We show that document-level NMT models trained with only
parallel paragraphs from Paracrawl can be used to translate real documents from
TED, News and Europarl, outperforming sentence-level NMT models. We also
perform a targeted pronoun evaluation and show that document-level models
trained with Paracrawl data can help context-aware pronoun translation. | cs.CL | Exploring Paracrawl for Document-level Neural Machine Translation |
2023-04-20T10:12:09Z | http://arxiv.org/pdf/2304.10195v1 | http://arxiv.org/abs/2304.10195v1 | Guangyuan Ma, Xing Wu, Peng Wang, Songlin Hu | 2023-04-20T10:12:09Z | cs.CL, cs.IR | Passage retrieval aims to retrieve relevant passages from large collections
of the open-domain corpus. Contextual Masked Auto-Encoding has been proven
effective in representation bottleneck pre-training of a monolithic
dual-encoder for passage retrieval. Siamese or fully separated dual-encoders
are often adopted as basic retrieval architecture in the pre-training and
fine-tuning stages for encoding queries and passages into their latent
embedding spaces. However, simply sharing or separating the parameters of the
dual-encoder results in an imbalanced discrimination of the embedding spaces.
In this work, we propose to pre-train Contextual Masked Auto-Encoder with
Mixture-of-Textual-Experts (CoT-MoTE). Specifically, we incorporate
textual-specific experts for individually encoding the distinct properties of
queries and passages. Meanwhile, a shared self-attention layer is still kept
for unified attention modeling. Results on large-scale passage retrieval
benchmarks show steady improvement in retrieval performances. The quantitive
analysis also shows a more balanced discrimination of the latent embedding
spaces. | cs.CL | CoT-MoTE: Exploring ConTextual Masked Auto-Encoder Pre-training with Mixture-of-Textual-Experts for Passage Retrieval |
2023-04-20T09:30:50Z | http://arxiv.org/pdf/2304.11073v3 | http://arxiv.org/abs/2304.11073v3 | Léo Jacqmin, Lucas Druart, Yannick Estève, Benoît Favre, Lina Maria Rojas-Barahona, Valentin Vielzeuf | 2023-08-31T08:51:33Z | eess.AS, cs.AI, cs.CL, cs.SD | Though Dialogue State Tracking (DST) is a core component of spoken dialogue
systems, recent work on this task mostly deals with chat corpora, disregarding
the discrepancies between spoken and written language.In this paper, we propose
OLISIA, a cascade system which integrates an Automatic Speech Recognition (ASR)
model and a DST model. We introduce several adaptations in the ASR and DST
modules to improve integration and robustness to spoken conversations.With
these adaptations, our system ranked first in DSTC11 Track 3, a benchmark to
evaluate spoken DST. We conduct an in-depth analysis of the results and find
that normalizing the ASR outputs and adapting the DST inputs through data
augmentation, along with increasing the pre-trained models size all play an
important role in reducing the performance discrepancy between written and
spoken conversations. | eess.AS | OLISIA: a Cascade System for Spoken Dialogue State Tracking |
2023-04-20T08:54:00Z | http://arxiv.org/pdf/2304.10164v2 | http://arxiv.org/abs/2304.10164v2 | Wonseong Kim, Jan Frederic Spörer, Siegfried Handschuh | 2024-02-19T15:24:39Z | cs.CL | This research article analyzes the language used in the official statements
released by the Federal Open Market Committee (FOMC) after its scheduled
meetings to gain insights into the impact of FOMC official statements on
financial markets and economic forecasting. The study reveals that the FOMC is
careful to avoid expressing emotion in their sentences and follows a set of
templates to cover economic situations. The analysis employs advanced language
modeling techniques such as VADER and FinBERT, and a trial test with GPT-4. The
results show that FinBERT outperforms other techniques in predicting negative
sentiment accurately. However, the study also highlights the challenges and
limitations of using current NLP techniques to analyze FOMC texts and suggests
the potential for enhancing language models and exploring alternative
approaches. | cs.CL | Analyzing FOMC Minutes: Accuracy and Constraints of Language Models |
2023-04-20T08:32:34Z | http://arxiv.org/pdf/2304.10158v1 | http://arxiv.org/abs/2304.10158v1 | Verena Blaschke, Hinrich Schütze, Barbara Plank | 2023-04-20T08:32:34Z | cs.CL | One of the challenges with finetuning pretrained language models (PLMs) is
that their tokenizer is optimized for the language(s) it was pretrained on, but
brittle when it comes to previously unseen variations in the data. This can for
instance be observed when finetuning PLMs on one language and evaluating them
on data in a closely related language variety with no standardized orthography.
Despite the high linguistic similarity, tokenization no longer corresponds to
meaningful representations of the target data, leading to low performance in,
e.g., part-of-speech tagging.
In this work, we finetune PLMs on seven languages from three different
families and analyze their zero-shot performance on closely related,
non-standardized varieties. We consider different measures for the divergence
in the tokenization of the source and target data, and the way they can be
adjusted by manipulating the tokenization during the finetuning step. Overall,
we find that the similarity between the percentage of words that get split into
subwords in the source and target data (the split word ratio difference) is the
strongest predictor for model performance on target data. | cs.CL | Does Manipulating Tokenization Aid Cross-Lingual Transfer? A Study on POS Tagging for Non-Standardized Languages |
2023-04-20T08:27:21Z | http://arxiv.org/pdf/2304.10153v1 | http://arxiv.org/abs/2304.10153v1 | Laura Cabello, Anna Katrine Jørgensen, Anders Søgaard | 2023-04-20T08:27:21Z | cs.CL, cs.HC | The societal impact of pre-trained language models has prompted researchers
to probe them for strong associations between protected attributes and
value-loaded terms, from slur to prestigious job titles. Such work is said to
probe models for bias or fairness-or such probes 'into representational biases'
are said to be 'motivated by fairness'-suggesting an intimate connection
between bias and fairness. We provide conceptual clarity by distinguishing
between association biases (Caliskan et al., 2022) and empirical fairness (Shen
et al., 2022) and show the two can be independent. Our main contribution,
however, is showing why this should not come as a surprise. To this end, we
first provide a thought experiment, showing how association bias and empirical
fairness can be completely orthogonal. Next, we provide empirical evidence that
there is no correlation between bias metrics and fairness metrics across the
most widely used language models. Finally, we survey the sociological and
psychological literature and show how this literature provides ample support
for expecting these metrics to be uncorrelated. | cs.CL | On the Independence of Association Bias and Empirical Fairness in Language Models |
2023-04-20T08:08:12Z | http://arxiv.org/pdf/2304.10145v2 | http://arxiv.org/abs/2304.10145v2 | Yiming Zhu, Peixian Zhang, Ehsan-Ul Haq, Pan Hui, Gareth Tyson | 2023-04-22T08:55:33Z | cs.AI, cs.CL | The release of ChatGPT has uncovered a range of possibilities whereby large
language models (LLMs) can substitute human intelligence. In this paper, we
seek to understand whether ChatGPT has the potential to reproduce
human-generated label annotations in social computing tasks. Such an
achievement could significantly reduce the cost and complexity of social
computing research. As such, we use ChatGPT to relabel five seminal datasets
covering stance detection (2x), sentiment analysis, hate speech, and bot
detection. Our results highlight that ChatGPT does have the potential to handle
these data annotation tasks, although a number of challenges remain. ChatGPT
obtains an average accuracy 0.609. Performance is highest for the sentiment
analysis dataset, with ChatGPT correctly annotating 64.9% of tweets. Yet, we
show that performance varies substantially across individual labels. We believe
this work can open up new lines of analysis and act as a basis for future
research into the exploitation of ChatGPT for human annotation tasks. | cs.AI | Can ChatGPT Reproduce Human-Generated Labels? A Study of Social Computing Tasks |
2023-04-20T02:36:18Z | http://arxiv.org/pdf/2304.11095v1 | http://arxiv.org/abs/2304.11095v1 | Hyunjin Choi, Hyunjae Lee, Seongho Joe, Youngjune L. Gwon | 2023-04-20T02:36:18Z | cs.LG, cs.AI, cs.CL, cs.CV, cs.IR | Encoded representations from a pretrained deep learning model (e.g., BERT
text embeddings, penultimate CNN layer activations of an image) convey a rich
set of features beneficial for information retrieval. Embeddings for a
particular modality of data occupy a high-dimensional space of its own, but it
can be semantically aligned to another by a simple mapping without training a
deep neural net. In this paper, we take a simple mapping computed from the
least squares and singular value decomposition (SVD) for a solution to the
Procrustes problem to serve a means to cross-modal information retrieval. That
is, given information in one modality such as text, the mapping helps us locate
a semantically equivalent data item in another modality such as image. Using
off-the-shelf pretrained deep learning models, we have experimented the
aforementioned simple cross-modal mappings in tasks of text-to-image and
image-to-text retrieval. Despite simplicity, our mappings perform reasonably
well reaching the highest accuracy of 77% on recall@10, which is comparable to
those requiring costly neural net training and fine-tuning. We have improved
the simple mappings by contrastive learning on the pretrained models.
Contrastive learning can be thought as properly biasing the pretrained encoders
to enhance the cross-modal mapping quality. We have further improved the
performance by multilayer perceptron with gating (gMLP), a simple neural
architecture. | cs.LG | Is Cross-modal Information Retrieval Possible without Training? |
2023-04-19T21:59:04Z | http://arxiv.org/pdf/2304.09991v3 | http://arxiv.org/abs/2304.09991v3 | Charvi Rastogi, Marco Tulio Ribeiro, Nicholas King, Harsha Nori, Saleema Amershi | 2023-11-30T16:30:09Z | cs.HC, cs.AI, cs.CL | Large language models are becoming increasingly pervasive and ubiquitous in
society via deployment in sociotechnical systems. Yet these language models, be
it for classification or generation, have been shown to be biased and behave
irresponsibly, causing harm to people at scale. It is crucial to audit these
language models rigorously. Existing auditing tools leverage either or both
humans and AI to find failures. In this work, we draw upon literature in
human-AI collaboration and sensemaking, and conduct interviews with research
experts in safe and fair AI, to build upon the auditing tool: AdaTest (Ribeiro
and Lundberg, 2022), which is powered by a generative large language model
(LLM). Through the design process we highlight the importance of sensemaking
and human-AI communication to leverage complementary strengths of humans and
generative models in collaborative auditing. To evaluate the effectiveness of
the augmented tool, AdaTest++, we conduct user studies with participants
auditing two commercial language models: OpenAI's GPT-3 and Azure's sentiment
analysis model. Qualitative analysis shows that AdaTest++ effectively leverages
human strengths such as schematization, hypothesis formation and testing.
Further, with our tool, participants identified a variety of failures modes,
covering 26 different topics over 2 tasks, that have been shown before in
formal audits and also those previously under-reported. | cs.HC | Supporting Human-AI Collaboration in Auditing LLMs with LLMs |
2023-04-19T21:33:59Z | http://arxiv.org/pdf/2304.09982v1 | http://arxiv.org/abs/2304.09982v1 | Valentin-Gabriel Soumah, Prashanth Rao, Philipp Eibl, Maite Taboada | 2023-04-19T21:33:59Z | cs.CL | We present the Radar de Parit\'e, an automated Natural Language Processing
(NLP) system that measures the proportion of women and men quoted daily in six
Canadian French-language media outlets. We outline the system's architecture
and detail the challenges we overcame to address French-specific issues, in
particular regarding coreference resolution, a new contribution to the NLP
literature on French. We also showcase statistics covering over one year's
worth of data (282,512 news articles). Our results highlight the
underrepresentation of women in news stories, while also illustrating the
application of modern NLP methods to measure gender representation and address
societal issues. | cs.CL | Radar de Parité: An NLP system to measure gender representation in French news stories |
2023-04-19T21:12:23Z | http://arxiv.org/pdf/2304.09972v2 | http://arxiv.org/abs/2304.09972v2 | David Ifeoluwa Adelani, Marek Masiak, Israel Abebe Azime, Jesujoba Alabi, Atnafu Lambebo Tonja, Christine Mwase, Odunayo Ogundepo, Bonaventure F. P. Dossou, Akintunde Oladipo, Doreen Nixdorf, Chris Chinenye Emezue, sana al-azzawi, Blessing Sibanda, Davis David, Lolwethu Ndolela, Jonathan Mukiibi, Tunde Ajayi, Tatiana Moteu, Brian Odhiambo, Abraham Owodunni, Nnaemeka Obiefuna, Muhidin Mohamed, Shamsuddeen Hassan Muhammad, Teshome Mulugeta Ababu, Saheed Abdullahi Salahudeen, Mesay Gemeda Yigezu, Tajuddeen Gwadabe, Idris Abdulmumin, Mahlet Taye, Oluwabusayo Awoyomi, Iyanuoluwa Shode, Tolulope Adelani, Habiba Abdulganiyu, Abdul-Hakeem Omotayo, Adetola Adeeko, Abeeb Afolabi, Anuoluwapo Aremu, Olanrewaju Samuel, Clemencia Siro, Wangari Kimotho, Onyekachi Ogbu, Chinedu Mbonu, Chiamaka Chukwuneke, Samuel Fanijo, Jessica Ojo, Oyinkansola Awosan, Tadesse Kebede, Toadoum Sari Sakayo, Pamela Nyatsine, Freedmore Sidume, Oreen Yousuf, Mardiyyah Oduwole, Tshinu Tshinu, Ussen Kimanuka, Thina Diko, Siyanda Nxakama, Sinodos Nigusse, Abdulmejid Johar, Shafie Mohamed, Fuad Mire Hassan, Moges Ahmed Mehamed, Evrard Ngabire, Jules Jules, Ivan Ssenkungu, Pontus Stenetorp | 2023-09-20T17:14:40Z | cs.CL | African languages are severely under-represented in NLP research due to lack
of datasets covering several NLP tasks. While there are individual language
specific datasets that are being expanded to different tasks, only a handful of
NLP tasks (e.g. named entity recognition and machine translation) have
standardized benchmark datasets covering several geographical and
typologically-diverse African languages. In this paper, we develop MasakhaNEWS
-- a new benchmark dataset for news topic classification covering 16 languages
widely spoken in Africa. We provide an evaluation of baseline models by
training classical machine learning models and fine-tuning several language
models. Furthermore, we explore several alternatives to full fine-tuning of
language models that are better suited for zero-shot and few-shot learning such
as cross-lingual parameter-efficient fine-tuning (like MAD-X), pattern
exploiting training (PET), prompting language models (like ChatGPT), and
prompt-free sentence transformer fine-tuning (SetFit and Cohere Embedding API).
Our evaluation in zero-shot setting shows the potential of prompting ChatGPT
for news topic classification in low-resource African languages, achieving an
average performance of 70 F1 points without leveraging additional supervision
like MAD-X. In few-shot setting, we show that with as little as 10 examples per
label, we achieved more than 90\% (i.e. 86.0 F1 points) of the performance of
full supervised training (92.6 F1 points) leveraging the PET approach. | cs.CL | MasakhaNEWS: News Topic Classification for African languages |
2023-04-19T20:45:01Z | http://arxiv.org/pdf/2304.09960v3 | http://arxiv.org/abs/2304.09960v3 | Hui Jiang | 2023-09-13T18:52:02Z | cs.CL, cs.AI, cs.LG | Languages are not created randomly but rather to communicate information.
There is a strong association between languages and their underlying meanings,
resulting in a sparse joint distribution that is heavily peaked according to
their correlations. Moreover, these peak values happen to match with the
marginal distribution of languages due to the sparsity. With the advent of LLMs
trained on big data and large models, we can now precisely assess the marginal
distribution of languages, providing a convenient means of exploring the sparse
structures in the joint distribution for effective inferences. In this paper,
we categorize languages as either unambiguous or {\epsilon}-ambiguous and
present quantitative results to demonstrate that the emergent abilities of
LLMs, such as language understanding, in-context learning, chain-of-thought
prompting, and effective instruction fine-tuning, can all be attributed to
Bayesian inference on the sparse joint distribution of languages. | cs.CL | A Latent Space Theory for Emergent Abilities in Large Language Models |
2023-04-19T20:20:41Z | http://arxiv.org/pdf/2304.09957v1 | http://arxiv.org/abs/2304.09957v1 | Ekaterina Artemova, Barbara Plank | 2023-04-19T20:20:41Z | cs.CL | Bilingual word lexicons are crucial tools for multilingual natural language
understanding and machine translation tasks, as they facilitate the mapping of
words in one language to their synonyms in another language. To achieve this,
numerous papers have explored bilingual lexicon induction (BLI) in
high-resource scenarios, using a typical pipeline consisting of two
unsupervised steps: bitext mining and word alignment, both of which rely on
pre-trained large language models~(LLMs).
In this paper, we present an analysis of the BLI pipeline for German and two
of its dialects, Bavarian and Alemannic. This setup poses several unique
challenges, including the scarcity of resources, the relatedness of the
languages, and the lack of standardization in the orthography of dialects. To
evaluate the BLI outputs, we analyze them with respect to word frequency and
pairwise edit distance. Additionally, we release two evaluation datasets
comprising 1,500 bilingual sentence pairs and 1,000 bilingual word pairs. They
were manually judged for their semantic similarity for each Bavarian-German and
Alemannic-German language pair. | cs.CL | Low-resource Bilingual Dialect Lexicon Induction with Large Language Models |
2023-04-19T19:59:26Z | http://arxiv.org/pdf/2304.09948v1 | http://arxiv.org/abs/2304.09948v1 | Aishwarya Deep Shukla, Laksh Agarwal, Jie Mein, Goh, Guodong, Gao, Ritu Agarwal | 2023-04-19T19:59:26Z | cs.CL, cs.AI | The proliferation of fake reviews of doctors has potentially detrimental
consequences for patient well-being and has prompted concern among consumer
protection groups and regulatory bodies. Yet despite significant advancements
in the fields of machine learning and natural language processing, there
remains limited comprehension of the characteristics differentiating fraudulent
from authentic reviews. This study utilizes a novel pre-labeled dataset of
38048 physician reviews to establish the effectiveness of large language models
in classifying reviews. Specifically, we compare the performance of traditional
ML models, such as logistic regression and support vector machines, to
generative pre-trained transformer models. Furthermore, we use GPT4, the newest
model in the GPT family, to uncover the key dimensions along which fake and
genuine physician reviews differ. Our findings reveal significantly superior
performance of GPT-3 over traditional ML models in this context. Additionally,
our analysis suggests that GPT3 requires a smaller training sample than
traditional models, suggesting its appropriateness for tasks with scarce
training data. Moreover, the superiority of GPT3 performance increases in the
cold start context i.e., when there are no prior reviews of a doctor. Finally,
we employ GPT4 to reveal the crucial dimensions that distinguish fake physician
reviews. In sharp contrast to previous findings in the literature that were
obtained using simulated data, our findings from a real-world dataset show that
fake reviews are generally more clinically detailed, more reserved in
sentiment, and have better structure and grammar than authentic ones. | cs.CL | Catch Me If You Can: Identifying Fraudulent Physician Reviews with Large Language Models Using Generative Pre-Trained Transformers |
2023-04-19T18:52:49Z | http://arxiv.org/pdf/2304.09919v1 | http://arxiv.org/abs/2304.09919v1 | Vesa Akerman, David Baines, Damien Daspit, Ulf Hermjakob, Taeho Jang, Colin Leong, Michael Martin, Joel Mathew, Jonathan Robie, Marcus Schwarting | 2023-04-19T18:52:49Z | cs.CL, cs.AI | Efficiently and accurately translating a corpus into a low-resource language
remains a challenge, regardless of the strategies employed, whether manual,
automated, or a combination of the two. Many Christian organizations are
dedicated to the task of translating the Holy Bible into languages that lack a
modern translation. Bible translation (BT) work is currently underway for over
3000 extremely low resource languages. We introduce the eBible corpus: a
dataset containing 1009 translations of portions of the Bible with data in 833
different languages across 75 language families. In addition to a BT
benchmarking dataset, we introduce model performance benchmarks built on the No
Language Left Behind (NLLB) neural machine translation (NMT) models. Finally,
we describe several problems specific to the domain of BT and consider how the
established data and model benchmarks might be used for future translation
efforts. For a BT task trained with NLLB, Austronesian and Trans-New Guinea
language families achieve 35.1 and 31.6 BLEU scores respectively, which spurs
future innovations for NMT for low-resource languages in Papua New Guinea. | cs.CL | The eBible Corpus: Data and Model Benchmarks for Bible Translation for Low-Resource Languages |
2023-04-19T18:24:57Z | http://arxiv.org/pdf/2304.11046v1 | http://arxiv.org/abs/2304.11046v1 | Md. Adyelullahil Mamun, Hasnat Md. Abdullah, Md. Golam Rabiul Alam, Muhammad Mehedi Hassan, Md. Zia Uddin | 2023-04-19T18:24:57Z | cs.SD, cs.AI, cs.CL, cs.HC, cs.LG | Human conversational styles are measured by the sense of humor, personality,
and tone of voice. These characteristics have become essential for
conversational intelligent virtual assistants. However, most of the
state-of-the-art intelligent virtual assistants (IVAs) are failed to interpret
the affective semantics of human voices. This research proposes an
anthropomorphic intelligent system that can hold a proper human-like
conversation with emotion and personality. A voice style transfer method is
also proposed to map the attributes of a specific emotion. Initially, the
frequency domain data (Mel-Spectrogram) is created by converting the temporal
audio wave data, which comprises discrete patterns for audio features such as
notes, pitch, rhythm, and melody. A collateral CNN-Transformer-Encoder is used
to predict seven different affective states from voice. The voice is also fed
parallelly to the deep-speech, an RNN model that generates the text
transcription from the spectrogram. Then the transcripted text is transferred
to the multi-domain conversation agent using blended skill talk,
transformer-based retrieve-and-generate generation strategy, and beam-search
decoding, and an appropriate textual response is generated. The system learns
an invertible mapping of data to a latent space that can be manipulated and
generates a Mel-spectrogram frame based on previous Mel-spectrogram frames to
voice synthesize and style transfer. Finally, the waveform is generated using
WaveGlow from the spectrogram. The outcomes of the studies we conducted on
individual models were auspicious. Furthermore, users who interacted with the
system provided positive feedback, demonstrating the system's effectiveness. | cs.SD | Affective social anthropomorphic intelligent system |
2023-04-19T18:09:27Z | http://arxiv.org/pdf/2304.11053v1 | http://arxiv.org/abs/2304.11053v1 | Cal Peyser, Michael Picheny, Kyunghyun Cho, Rohit Prabhavalkar, Ronny Huang, Tara Sainath | 2023-04-19T18:09:27Z | cs.CL, cs.SD, eess.AS | Unpaired text and audio injection have emerged as dominant methods for
improving ASR performance in the absence of a large labeled corpus. However,
little guidance exists on deploying these methods to improve production ASR
systems that are trained on very large supervised corpora and with realistic
requirements like a constrained model size and CPU budget, streaming
capability, and a rich lattice for rescoring and for downstream NLU tasks. In
this work, we compare three state-of-the-art semi-supervised methods
encompassing both unpaired text and audio as well as several of their
combinations in a controlled setting using joint training. We find that in our
setting these methods offer many improvements beyond raw WER, including
substantial gains in tail-word WER, decoder computation during inference, and
lattice density. | cs.CL | A Comparison of Semi-Supervised Learning Techniques for Streaming ASR at Scale |
2023-04-19T17:56:12Z | http://arxiv.org/pdf/2304.09848v2 | http://arxiv.org/abs/2304.09848v2 | Nelson F. Liu, Tianyi Zhang, Percy Liang | 2023-10-23T19:11:38Z | cs.CL, cs.IR | Generative search engines directly generate responses to user queries, along
with in-line citations. A prerequisite trait of a trustworthy generative search
engine is verifiability, i.e., systems should cite comprehensively (high
citation recall; all statements are fully supported by citations) and
accurately (high citation precision; every cite supports its associated
statement). We conduct human evaluation to audit four popular generative search
engines -- Bing Chat, NeevaAI, perplexity.ai, and YouChat -- across a diverse
set of queries from a variety of sources (e.g., historical Google user queries,
dynamically-collected open-ended questions on Reddit, etc.). We find that
responses from existing generative search engines are fluent and appear
informative, but frequently contain unsupported statements and inaccurate
citations: on average, a mere 51.5% of generated sentences are fully supported
by citations and only 74.5% of citations support their associated sentence. We
believe that these results are concerningly low for systems that may serve as a
primary tool for information-seeking users, especially given their facade of
trustworthiness. We hope that our results further motivate the development of
trustworthy generative search engines and help researchers and users better
understand the shortcomings of existing commercial systems. | cs.CL | Evaluating Verifiability in Generative Search Engines |
2023-04-19T17:50:09Z | http://arxiv.org/pdf/2304.11082v6 | http://arxiv.org/abs/2304.11082v6 | Yotam Wolf, Noam Wies, Oshri Avnery, Yoav Levine, Amnon Shashua | 2024-06-03T12:19:16Z | cs.CL, cs.AI | An important aspect in developing language models that interact with humans
is aligning their behavior to be useful and unharmful for their human users.
This is usually achieved by tuning the model in a way that enhances desired
behaviors and inhibits undesired ones, a process referred to as alignment. In
this paper, we propose a theoretical approach called Behavior Expectation
Bounds (BEB) which allows us to formally investigate several inherent
characteristics and limitations of alignment in large language models.
Importantly, we prove that within the limits of this framework, for any
behavior that has a finite probability of being exhibited by the model, there
exist prompts that can trigger the model into outputting this behavior, with
probability that increases with the length of the prompt. This implies that any
alignment process that attenuates an undesired behavior but does not remove it
altogether, is not safe against adversarial prompting attacks. Furthermore, our
framework hints at the mechanism by which leading alignment approaches such as
reinforcement learning from human feedback make the LLM prone to being prompted
into the undesired behaviors. This theoretical result is being experimentally
demonstrated in large scale by the so called contemporary "chatGPT jailbreaks",
where adversarial users trick the LLM into breaking its alignment guardrails by
triggering it into acting as a malicious persona. Our results expose
fundamental limitations in alignment of LLMs and bring to the forefront the
need to devise reliable mechanisms for ensuring AI safety. | cs.CL | Fundamental Limitations of Alignment in Large Language Models |
2023-04-19T17:47:47Z | http://arxiv.org/pdf/2304.09842v3 | http://arxiv.org/abs/2304.09842v3 | Pan Lu, Baolin Peng, Hao Cheng, Michel Galley, Kai-Wei Chang, Ying Nian Wu, Song-Chun Zhu, Jianfeng Gao | 2023-10-31T17:43:39Z | cs.CL, cs.AI, cs.CV, cs.LG | Large language models (LLMs) have achieved remarkable progress in solving
various natural language processing tasks due to emergent reasoning abilities.
However, LLMs have inherent limitations as they are incapable of accessing
up-to-date information (stored on the Web or in task-specific knowledge bases),
using external tools, and performing precise mathematical and logical
reasoning. In this paper, we present Chameleon, an AI system that mitigates
these limitations by augmenting LLMs with plug-and-play modules for
compositional reasoning. Chameleon synthesizes programs by composing various
tools (e.g., LLMs, off-the-shelf vision models, web search engines, Python
functions, and heuristic-based modules) for accomplishing complex reasoning
tasks. At the heart of Chameleon is an LLM-based planner that assembles a
sequence of tools to execute to generate the final response. We showcase the
effectiveness of Chameleon on two multi-modal knowledge-intensive reasoning
tasks: ScienceQA and TabMWP. Chameleon, powered by GPT-4, achieves an 86.54%
overall accuracy on ScienceQA, improving the best published few-shot result by
11.37%. On TabMWP, GPT-4-powered Chameleon improves the accuracy by 17.0%,
lifting the state of the art to 98.78%. Our analysis also shows that the
GPT-4-powered planner exhibits more consistent and rational tool selection via
inferring potential constraints from instructions, compared to a
ChatGPT-powered planner. The project is available at
https://chameleon-llm.github.io. | cs.CL | Chameleon: Plug-and-Play Compositional Reasoning with Large Language Models |
2023-04-19T16:45:16Z | http://arxiv.org/pdf/2304.09805v1 | http://arxiv.org/abs/2304.09805v1 | Verena Blaschke, Hinrich Schütze, Barbara Plank | 2023-04-19T16:45:16Z | cs.CL | Despite much progress in recent years, the vast majority of work in natural
language processing (NLP) is on standard languages with many speakers. In this
work, we instead focus on low-resource languages and in particular
non-standardized low-resource languages. Even within branches of major language
families, often considered well-researched, little is known about the extent
and type of available resources and what the major NLP challenges are for these
language varieties. The first step to address this situation is a systematic
survey of available corpora (most importantly, annotated corpora, which are
particularly valuable for NLP research). Focusing on Germanic low-resource
language varieties, we provide such a survey in this paper. Except for
geolocation (origin of speaker or document), we find that manually annotated
linguistic resources are sparse and, if they exist, mostly cover morphosyntax.
Despite this lack of resources, we observe that interest in this area is
increasing: there is active development and a growing research community. To
facilitate research, we make our overview of over 80 corpora publicly
available. We share a companion website of this overview at
https://github.com/mainlp/germanic-lrl-corpora . | cs.CL | A Survey of Corpora for Germanic Low-Resource Languages and Dialects |
2023-04-19T16:29:48Z | http://arxiv.org/pdf/2304.09797v5 | http://arxiv.org/abs/2304.09797v5 | Chuanyang Zheng, Zhengying Liu, Enze Xie, Zhenguo Li, Yu Li | 2023-08-10T03:41:04Z | cs.CL, cs.LG | The performance of Large Language Models (LLMs) in reasoning tasks depends
heavily on prompt design, with Chain-of-Thought (CoT) and self-consistency
being critical methods that enhance this ability. However, these methods do not
fully exploit the answers generated by the LLM to guide subsequent responses.
This paper proposes a new prompting method, named Progressive-Hint Prompting
(PHP), that enables automatic multiple interactions between users and LLMs by
using previously generated answers as hints to progressively guide toward the
correct answers. PHP is orthogonal to CoT and self-consistency, making it easy
to combine with state-of-the-art techniques to further improve performance. We
conducted extensive and comprehensive experiments on seven benchmarks. The
results show that PHP significantly improves accuracy while remaining highly
efficient. For instance, with text-davinci-003, we observed a 4.2% improvement
on GSM8K with greedy decoding compared to Complex CoT, and a 46.17% reduction
in sample paths with self-consistency. With GPT-4 and PHP, we achieve
state-of-the-art performances on SVAMP (89.1% -> 91.9%), GSM8K (92% -> 95.5%),
AQuA (76.4% -> 79.9%) and MATH (50.3% -> 53.9%). | cs.CL | Progressive-Hint Prompting Improves Reasoning in Large Language Models |
2023-04-19T16:18:54Z | http://arxiv.org/pdf/2304.11062v2 | http://arxiv.org/abs/2304.11062v2 | Aydar Bulatov, Yuri Kuratov, Yermek Kapushev, Mikhail S. Burtsev | 2024-02-06T10:16:54Z | cs.CL, cs.AI, cs.LG | A major limitation for the broader scope of problems solvable by transformers
is the quadratic scaling of computational complexity with input size. In this
study, we investigate the recurrent memory augmentation of pre-trained
transformer models to extend input context length while linearly scaling
compute. Our approach demonstrates the capability to store information in
memory for sequences of up to an unprecedented two million tokens while
maintaining high retrieval accuracy. Experiments with language modeling tasks
show perplexity improvement as the number of processed input segments
increases. These results underscore the effectiveness of our method, which has
significant potential to enhance long-term dependency handling in natural
language understanding and generation tasks, as well as enable large-scale
context processing for memory-intensive applications. | cs.CL | Scaling Transformer to 1M tokens and beyond with RMT |
2023-04-19T13:53:19Z | http://arxiv.org/pdf/2304.09667v3 | http://arxiv.org/abs/2304.09667v3 | Qiao Jin, Yifan Yang, Qingyu Chen, Zhiyong Lu | 2023-05-16T13:24:53Z | cs.CL, cs.AI, q-bio.GN | While large language models (LLMs) have been successfully applied to various
tasks, they still face challenges with hallucinations. Augmenting LLMs with
domain-specific tools such as database utilities can facilitate easier and more
precise access to specialized knowledge. In this paper, we present GeneGPT, a
novel method for teaching LLMs to use the Web APIs of the National Center for
Biotechnology Information (NCBI) for answering genomics questions.
Specifically, we prompt Codex to solve the GeneTuring tests with NCBI Web APIs
by in-context learning and an augmented decoding algorithm that can detect and
execute API calls. Experimental results show that GeneGPT achieves
state-of-the-art performance on eight tasks in the GeneTuring benchmark with an
average score of 0.83, largely surpassing retrieval-augmented LLMs such as the
new Bing (0.44), biomedical LLMs such as BioMedLM (0.08) and BioGPT (0.04), as
well as GPT-3 (0.16) and ChatGPT (0.12). Our further analyses suggest that: (1)
API demonstrations have good cross-task generalizability and are more useful
than documentations for in-context learning; (2) GeneGPT can generalize to
longer chains of API calls and answer multi-hop questions in GeneHop, a novel
dataset introduced in this work; (3) Different types of errors are enriched in
different tasks, providing valuable insights for future improvements. | cs.CL | GeneGPT: Augmenting Large Language Models with Domain Tools for Improved Access to Biomedical Information |
Subsets and Splits
No saved queries yet
Save your SQL queries to embed, download, and access them later. Queries will appear here once saved.