id
stringlengths
33
34
title
stringlengths
6
218
authors
stringlengths
5
6.98k
categories
stringlengths
5
188
summary
stringlengths
34
3.26k
link_pdf
stringlengths
33
34
updated
timestamp[s]
published
timestamp[s]
embeddings
dict
http://arxiv.org/abs/2302.05737v1
A Reparameterized Discrete Diffusion Model for Text Generation
Lin Zheng, Jianbo Yuan, Lei Yu, Lingpeng Kong
cs.CL, cs.LG
This work studies discrete diffusion probabilistic models with applications to natural language generation. We derive an alternative yet equivalent formulation of the sampling from discrete diffusion processes and leverage this insight to develop a family of reparameterized discrete diffusion models. The derived generi...
http://arxiv.org/pdf/2302.05737v1
2023-02-11T16:26:57
2023-02-11T16:26:57
{ "title": [ -0.0368042, -0.003566742, -0.049560547, -0.03668213, -0.0010919571, 0.0029525757, -0.04550171, -0.013671875, 0.013038635, 0.04095459, -0.016571045, 0.019714355, -0.008926392, -0.007385254, -0.055999756, 0.018722534, -0.03466797, 0.04...
http://arxiv.org/abs/2302.05735v2
Divergence-Based Domain Transferability for Zero-Shot Classification
Alexander Pugantsov, Richard McCreadie
cs.CL
Transferring learned patterns from pretrained neural language models has been shown to significantly improve effectiveness across a variety of language-based tasks, meanwhile further tuning on intermediate tasks has been demonstrated to provide additional performance benefits, provided the intermediate task is sufficie...
http://arxiv.org/pdf/2302.05735v2
2023-02-28T11:26:32
2023-02-11T16:04:38
{ "title": [ 0.024307251, 0.016418457, -0.035247803, -0.043823242, 0.011817932, -0.021316528, 0.011703491, -0.0032539368, 0.01737976, 0.011161804, -0.007396698, 0.0029087067, -0.01889038, -0.022094727, -0.04107666, -0.0030899048, -0.004798889, 0....
http://arxiv.org/abs/2302.05729v2
A Brief Report on LawGPT 1.0: A Virtual Legal Assistant Based on GPT-3
Ha-Thanh Nguyen
cs.CL, cs.AI
LawGPT 1.0 is a virtual legal assistant built on the state-of-the-art language model GPT-3, fine-tuned for the legal domain. The system is designed to provide legal assistance to users in a conversational manner, helping them with tasks such as answering legal questions, generating legal documents, and providing legal ...
http://arxiv.org/pdf/2302.05729v2
2023-02-14T06:26:42
2023-02-11T15:50:20
{ "title": [ -0.028564453, -0.011375427, -0.070007324, -0.024902344, -0.0072250366, -0.032226562, -0.031280518, 0.013061523, 0.01234436, 0.05340576, -0.052246094, 0.0025939941, 0.059143066, -0.0036792755, -0.034240723, 0.0004386902, -0.02949524, ...
http://arxiv.org/abs/2302.05721v1
Synthesizing Human Gaze Feedback for Improved NLP Performance
Varun Khurana, Yaman Kumar Singla, Nora Hollenstein, Rajesh Kumar, Balaji Krishnamurthy
cs.HC, cs.CL
Integrating human feedback in models can improve the performance of natural language processing (NLP) models. Feedback can be either explicit (e.g. ranking used in training language models) or implicit (e.g. using human cognitive signals in the form of eyetracking). Prior eye tracking and NLP research reveal that cogni...
http://arxiv.org/pdf/2302.05721v1
2023-02-11T15:34:23
2023-02-11T15:34:23
{ "title": [ -0.015197754, -0.013435364, -0.0065193176, -0.021118164, -0.027053833, -0.0074157715, -0.014411926, -0.019836426, 0.04260254, 0.032348633, -0.06463623, -0.016342163, 0.015716553, 0.020858765, -0.077819824, -0.011436462, 0.0047073364, ...
http://arxiv.org/abs/2302.05717v1
Learning by Applying: A General Framework for Mathematical Reasoning via Enhancing Explicit Knowledge Learning
Jiayu Liu, Zhenya Huang, Chengxiang Zhai, Qi Liu
cs.AI, cs.CL, cs.LG
Mathematical reasoning is one of the crucial abilities of general artificial intelligence, which requires machines to master mathematical logic and knowledge from solving problems. However, existing approaches are not transparent (thus not interpretable) in terms of what knowledge has been learned and applied in the re...
http://arxiv.org/pdf/2302.05717v1
2023-02-11T15:15:41
2023-02-11T15:15:41
{ "title": [ -0.0011816025, 0.000998497, -0.05090332, -0.06451416, 0.025924683, -0.022201538, -0.0077705383, -0.0050621033, 0.009048462, 0.0053367615, -0.023452759, -0.006893158, -0.015426636, -0.031143188, -0.0206604, -0.009552002, 0.010269165, ...
http://arxiv.org/abs/2302.05711v1
Fair Enough: Standardizing Evaluation and Model Selection for Fairness Research in NLP
Xudong Han, Timothy Baldwin, Trevor Cohn
cs.CL, cs.CY, cs.LG
Modern NLP systems exhibit a range of biases, which a growing literature on model debiasing attempts to correct. However current progress is hampered by a plurality of definitions of bias, means of quantification, and oftentimes vague relation between debiasing algorithms and theoretical measures of bias. This paper se...
http://arxiv.org/pdf/2302.05711v1
2023-02-11T14:54:00
2023-02-11T14:54:00
{ "title": [ -0.024261475, -0.013015747, -0.04788208, -0.041748047, 0.009590149, -0.00844574, -0.028930664, -0.006290436, 0.013893127, 0.0115356445, -0.053833008, 0.01335144, 0.01499939, -0.0107040405, -0.04321289, 0.0019035339, -0.050842285, 0.0...
http://arxiv.org/abs/2302.05706v1
MTTM: Metamorphic Testing for Textual Content Moderation Software
Wenxuan Wang, Jen-tse Huang, Weibin Wu, Jianping Zhang, Yizhan Huang, Shuqing Li, Pinjia He, Michael Lyu
cs.CL, cs.SE
The exponential growth of social media platforms such as Twitter and Facebook has revolutionized textual communication and textual content publication in human society. However, they have been increasingly exploited to propagate toxic content, such as hate speech, malicious advertisement, and pornography, which can lea...
http://arxiv.org/pdf/2302.05706v1
2023-02-11T14:44:39
2023-02-11T14:44:39
{ "title": [ -0.008300781, 0.0013141632, -0.018463135, -0.05895996, 0.0000019669533, 0.018569946, -0.05267334, -0.0050735474, -0.0016727448, 0.051971436, 0.0066604614, 0.0053253174, 0.043121338, -0.028961182, -0.03387451, 0.014724731, -0.034088135, ...
http://arxiv.org/abs/2302.05703v1
HateProof: Are Hateful Meme Detection Systems really Robust?
Piush Aggarwal, Pranit Chawla, Mithun Das, Punyajoy Saha, Binny Mathew, Torsten Zesch, Animesh Mukherjee
cs.CL, cs.AI
Exploiting social media to spread hate has tremendously increased over the years. Lately, multi-modal hateful content such as memes has drawn relatively more traction than uni-modal content. Moreover, the availability of implicit content payloads makes them fairly challenging to be detected by existing hateful meme det...
http://arxiv.org/pdf/2302.05703v1
2023-02-11T14:36:11
2023-02-11T14:36:11
{ "title": [ 0.045837402, 0.012512207, -0.03805542, -0.020614624, 0.02508545, -0.038848877, 0.02659607, -0.03149414, -0.0038928986, 0.03729248, -0.01486969, 0.02432251, 0.019210815, -0.020141602, -0.042938232, 0.003704071, -0.010887146, 0.0500793...
http://arxiv.org/abs/2302.05698v3
Compositional Exemplars for In-context Learning
Jiacheng Ye, Zhiyong Wu, Jiangtao Feng, Tao Yu, Lingpeng Kong
cs.CL, cs.AI, cs.LG
Large pretrained language models (LMs) have shown impressive In-Context Learning (ICL) ability, where the model learns to do an unseen task via a prompt consisting of input-output examples as the demonstration, without any parameter updates. The performance of ICL is highly dominated by the quality of the selected in-c...
http://arxiv.org/pdf/2302.05698v3
2023-06-20T08:37:42
2023-02-11T14:02:08
{ "title": [ 0.025894165, -0.00060749054, -0.023666382, -0.07635498, 0.01121521, -0.004421234, -0.03564453, -0.009262085, 0.0069618225, 0.030578613, -0.032684326, 0.0046310425, -0.020767212, -0.019760132, -0.03463745, -0.018707275, -0.016357422, ...
http://arxiv.org/abs/2302.05674v1
Counter-GAP: Counterfactual Bias Evaluation through Gendered Ambiguous Pronouns
Zhongbin Xie, Vid Kocijan, Thomas Lukasiewicz, Oana-Maria Camburu
cs.CL, cs.AI
Bias-measuring datasets play a critical role in detecting biased behavior of language models and in evaluating progress of bias mitigation methods. In this work, we focus on evaluating gender bias through coreference resolution, where previous datasets are either hand-crafted or fail to reliably measure an explicitly d...
http://arxiv.org/pdf/2302.05674v1
2023-02-11T12:11:03
2023-02-11T12:11:03
{ "title": [ 0.00045275688, 0.013435364, -0.043762207, -0.018478394, 0.0019130707, -0.03186035, -0.005645752, -0.040100098, 0.027389526, 0.036895752, -0.02923584, -0.022064209, 0.021835327, -0.01398468, -0.06640625, -0.01663208, -0.047332764, 0.0...
http://arxiv.org/abs/2302.07733v1
Explaining text classifiers through progressive neighborhood approximation with realistic samples
Yi Cai, Arthur Zimek, Eirini Ntoutsi, Gerhard Wunder
cs.CL, cs.LG
The importance of neighborhood construction in local explanation methods has been already highlighted in the literature. And several attempts have been made to improve neighborhood quality for high-dimensional data, for example, texts, by adopting generative models. Although the generators produce more realistic sample...
http://arxiv.org/pdf/2302.07733v1
2023-02-11T11:42:39
2023-02-11T11:42:39
{ "title": [ -0.01687622, 0.018981934, 0.010688782, -0.02809143, 0.005657196, 0.0031528473, 0.005077362, -0.03488159, -0.011260986, -0.0062026978, -0.033813477, 0.008583069, -0.03753662, -0.00843811, -0.025466919, -0.03579712, -0.0066184998, 0.05...
http://arxiv.org/abs/2302.05658v2
DocILE Benchmark for Document Information Localization and Extraction
Štěpán Šimsa, Milan Šulc, Michal Uřičář, Yash Patel, Ahmed Hamdi, Matěj Kocián, Matyáš Skalický, Jiří Matas, Antoine Doucet, Mickaël Coustaty, Dimosthenis Karatzas
cs.CL, cs.AI, cs.LG
This paper introduces the DocILE benchmark with the largest dataset of business documents for the tasks of Key Information Localization and Extraction and Line Item Recognition. It contains 6.7k annotated business documents, 100k synthetically generated documents, and nearly~1M unlabeled documents for unsupervised pre-...
http://arxiv.org/pdf/2302.05658v2
2023-05-03T16:24:58
2023-02-11T11:32:10
{ "title": [ -0.045043945, 0.010864258, -0.02986145, -0.03967285, 0.0018978119, 0.0059928894, -0.032684326, -0.02708435, 0.020339966, 0.04043579, -0.022201538, 0.006706238, 0.011627197, -0.011260986, -0.039855957, 0.017471313, 0.0062332153, 0.012...
http://arxiv.org/abs/2302.05657v1
Dialectograms: Machine Learning Differences between Discursive Communities
Thyge Enggaard, August Lohse, Morten Axel Pedersen, Sune Lehmann
cs.CL
Word embeddings provide an unsupervised way to understand differences in word usage between discursive communities. A number of recent papers have focused on identifying words that are used differently by two or more communities. But word embeddings are complex, high-dimensional spaces and a focus on identifying differ...
http://arxiv.org/pdf/2302.05657v1
2023-02-11T11:32:08
2023-02-11T11:32:08
{ "title": [ 0.006778717, 0.008850098, -0.03491211, -0.028167725, 0.0033073425, -0.036712646, 0.013259888, -0.020111084, -0.00970459, 0.04989624, -0.037139893, 0.02810669, -0.000600338, -0.008613586, -0.028900146, 0.0052108765, -0.0020999908, 0.0...
http://arxiv.org/abs/2302.10276v2
See Your Heart: Psychological states Interpretation through Visual Creations
Likun Yang, Xiaokun Feng, Xiaotang Chen, Shiyu Zhang, Kaiqi Huang
cs.CV, cs.AI, cs.CL
In psychoanalysis, generating interpretations to one's psychological state through visual creations is facing significant demands. The two main tasks of existing studies in the field of computer vision, sentiment/emotion classification and affective captioning, can hardly satisfy the requirement of psychological interp...
http://arxiv.org/pdf/2302.10276v2
2023-03-16T10:01:05
2023-02-11T07:30:21
{ "title": [ 0.020462036, -0.010154724, 0.011474609, -0.021453857, 0.0009899139, -0.020935059, 0.003698349, -0.05166626, 0.0043182373, 0.013511658, -0.031677246, -0.00919342, -0.0032730103, -0.037384033, -0.03277588, -0.015281677, 0.026779175, 0....
http://arxiv.org/abs/2302.05619v1
Evaluating the Robustness of Discrete Prompts
Yoichi Ishibashi, Danushka Bollegala, Katsuhito Sudoh, Satoshi Nakamura
cs.CL, cs.AI
Discrete prompts have been used for fine-tuning Pre-trained Language Models for diverse NLP tasks. In particular, automatic methods that generate discrete prompts from a small set of training instances have reported superior performance. However, a closer look at the learnt prompts reveals that they contain noisy and c...
http://arxiv.org/pdf/2302.05619v1
2023-02-11T07:01:53
2023-02-11T07:01:53
{ "title": [ 0.02468872, 0.011352539, -0.02784729, -0.037353516, 0.00017654896, -0.024047852, -0.034362793, -0.04067993, -0.005584717, 0.02217102, -0.031280518, -0.009681702, -0.018981934, -0.024887085, -0.048553467, -0.021347046, -0.007484436, 0...
http://arxiv.org/abs/2302.05611v1
Metaphor Detection with Effective Context Denoising
Shun Wang, Yucheng Li, Chenghua Lin, Loïc Barrault, Frank Guerin
cs.CL
We propose a novel RoBERTa-based model, RoPPT, which introduces a target-oriented parse tree structure in metaphor detection. Compared to existing models, RoPPT focuses on semantically relevant information and achieves the state-of-the-art on several main metaphor datasets. We also compare our approach against several ...
http://arxiv.org/pdf/2302.05611v1
2023-02-11T05:53:51
2023-02-11T05:53:51
{ "title": [ -0.000073075294, 0.020523071, -0.020523071, -0.0056533813, 0.029724121, -0.018188477, 0.025772095, -0.046173096, 0.032226562, 0.045959473, -0.011489868, -0.008705139, 0.04550171, -0.011169434, -0.047424316, 0.020065308, -0.0107040405, ...
http://arxiv.org/abs/2302.05610v1
Emotion Detection From Social Media Posts
Md Mahbubur Rahman, Shaila Shova
cs.LG, cs.AI, cs.CL
Over the last few years, social media has evolved into a medium for expressing personal views, emotions, and even business and political proposals, recommendations, and advertisements. We address the topic of identifying emotions from text data obtained from social media posts like Twitter in this research. We have dep...
http://arxiv.org/pdf/2302.05610v1
2023-02-11T05:52:33
2023-02-11T05:52:33
{ "title": [ 0.0010623932, 0.007873535, -0.028747559, -0.030090332, 0.017211914, -0.019760132, 0.0053215027, -0.020263672, 0.012268066, 0.02784729, -0.03894043, 0.0055236816, -0.005569458, -0.011886597, -0.04638672, 0.0015830994, 0.010276794, 0.0...
http://arxiv.org/abs/2302.05608v1
Differentiable Outlier Detection Enable Robust Deep Multimodal Analysis
Zhu Wang, Sourav Medya, Sathya N. Ravi
cs.CV, cs.AI, cs.CL, cs.LG
Often, deep network models are purely inductive during training and while performing inference on unseen data. Thus, when such models are used for predictions, it is well known that they often fail to capture the semantic information and implicit dependencies that exist among objects (or concepts) on a population level...
http://arxiv.org/pdf/2302.05608v1
2023-02-11T05:46:21
2023-02-11T05:46:21
{ "title": [ -0.007850647, 0.02760315, -0.05432129, -0.051818848, 0.0236969, -0.025634766, -0.019927979, -0.02394104, -0.012489319, 0.0067481995, -0.022277832, 0.03213501, -0.01574707, -0.028289795, -0.0052948, -0.005748749, -0.021713257, 0.04800...
http://arxiv.org/abs/2302.05597v1
MatKB: Semantic Search for Polycrystalline Materials Synthesis Procedures
Xianjun Yang, Stephen Wilson, Linda Petzold
cs.CL, cs.AI
In this paper, we present a novel approach to knowledge extraction and retrieval using Natural Language Processing (NLP) techniques for material science. Our goal is to automatically mine structured knowledge from millions of research articles in the field of polycrystalline materials and make it easily accessible to t...
http://arxiv.org/pdf/2302.05597v1
2023-02-11T04:18:07
2023-02-11T04:18:07
{ "title": [ -0.06695557, -0.02507019, -0.042175293, -0.042877197, -0.025054932, 0.016281128, -0.072631836, 0.009292603, -0.0040779114, 0.02734375, 0.006214142, 0.021865845, -0.002210617, -0.033172607, -0.008262634, 0.028060913, -0.01751709, 0.02...
http://arxiv.org/abs/2302.07736v2
Is ChatGPT better than Human Annotators? Potential and Limitations of ChatGPT in Explaining Implicit Hate Speech
Fan Huang, Haewoon Kwak, Jisun An
cs.CL, cs.HC
Recent studies have alarmed that many online hate speeches are implicit. With its subtle nature, the explainability of the detection of such hateful speech has been a challenging problem. In this work, we examine whether ChatGPT can be used for providing natural language explanations (NLEs) for implicit hateful speech ...
http://arxiv.org/pdf/2302.07736v2
2023-03-15T19:16:45
2023-02-11T03:13:54
{ "title": [ 0.011169434, 0.0040893555, -0.00630188, -0.02798462, -0.009422302, -0.013298035, 0.0063819885, -0.015144348, -0.022842407, 0.043548584, -0.02154541, 0.014877319, -0.003578186, -0.0041770935, -0.031143188, -0.008071899, -0.020645142, ...
http://arxiv.org/abs/2302.05582v1
ASDF: A Differential Testing Framework for Automatic Speech Recognition Systems
Daniel Hao Xian Yuen, Andrew Yong Chen Pang, Zhou Yang, Chun Yong Chong, Mei Kuan Lim, David Lo
eess.AS, cs.CL, cs.SD, cs.SE
Recent years have witnessed wider adoption of Automated Speech Recognition (ASR) techniques in various domains. Consequently, evaluating and enhancing the quality of ASR systems is of great importance. This paper proposes ASDF, an Automated Speech Recognition Differential Testing Framework for testing ASR systems. ASDF...
http://arxiv.org/pdf/2302.05582v1
2023-02-11T02:53:12
2023-02-11T02:53:12
{ "title": [ -0.023651123, -0.006614685, -0.036468506, -0.074279785, 0.010322571, 0.013084412, -0.051696777, 0.016967773, 0.003583908, 0.030685425, -0.053955078, 0.019195557, 0.008728027, -0.0001462698, -0.038848877, -0.0015497208, -0.0062561035, ...
http://arxiv.org/abs/2302.05578v2
Characterizing Attribution and Fluency Tradeoffs for Retrieval-Augmented Large Language Models
Renat Aksitov, Chung-Ching Chang, David Reitter, Siamak Shakeri, Yunhsuan Sung
cs.CL, cs.AI
Despite recent progress, it has been difficult to prevent semantic hallucinations in generative Large Language Models. One common solution to this is augmenting LLMs with a retrieval system and making sure that the generated output is attributable to the retrieved information. Given this new added constraint, it is pla...
http://arxiv.org/pdf/2302.05578v2
2023-02-14T23:53:56
2023-02-11T02:43:34
{ "title": [ 0.0055770874, -0.00058555603, -0.03375244, -0.036987305, -0.002609253, -0.015838623, -0.030151367, -0.015777588, -0.0037899017, 0.045288086, 0.001947403, 0.007396698, 0.010917664, -0.015617371, -0.039154053, 0.0014801025, -0.023208618, ...
http://arxiv.org/abs/2302.05574v1
NapSS: Paragraph-level Medical Text Simplification via Narrative Prompting and Sentence-matching Summarization
Junru Lu, Jiazheng Li, Byron C. Wallace, Yulan He, Gabriele Pergola
cs.CL
Accessing medical literature is difficult for laypeople as the content is written for specialists and contains medical jargon. Automated text simplification methods offer a potential means to address this issue. In this work, we propose a summarize-then-simplify two-stage strategy, which we call NapSS, identifying the ...
http://arxiv.org/pdf/2302.05574v1
2023-02-11T02:20:25
2023-02-11T02:20:25
{ "title": [ -0.024368286, -0.007171631, -0.031829834, -0.0848999, -0.0128479, 0.026794434, -0.0715332, -0.003479004, -0.016662598, 0.039154053, 0.0024642944, 0.04647827, 0.002538681, -0.0121154785, -0.030426025, 0.0045700073, 0.0020160675, -0.00...
http://arxiv.org/abs/2302.05508v1
FairPy: A Toolkit for Evaluation of Social Biases and their Mitigation in Large Language Models
Hrishikesh Viswanath, Tianyi Zhang
cs.CL, cs.AI, cs.LG
Studies have shown that large pretrained language models exhibit biases against social groups based on race, gender etc, which they inherit from the datasets they are trained on. Various researchers have proposed mathematical tools for quantifying and identifying these biases. There have been methods proposed to mitiga...
http://arxiv.org/pdf/2302.05508v1
2023-02-10T20:54:10
2023-02-10T20:54:10
{ "title": [ -0.018341064, -0.0105896, -0.04699707, -0.05126953, 0.014175415, -0.018218994, -0.036071777, -0.014625549, 0.01878357, 0.021392822, -0.020950317, 0.006023407, 0.01890564, -0.01889038, -0.034942627, 0.013031006, -0.03918457, 0.0681152...
http://arxiv.org/abs/2302.05507v1
Long-Context Language Decision Transformers and Exponential Tilt for Interactive Text Environments
Nicolas Gontier, Pau Rodriguez, Issam Laradji, David Vazquez, Christopher Pal
cs.CL, cs.AI, cs.LG
Text-based game environments are challenging because agents must deal with long sequences of text, execute compositional actions using text and learn from sparse rewards. We address these challenges by proposing Long-Context Language Decision Transformers (LLDTs), a framework that is based on long transformer language ...
http://arxiv.org/pdf/2302.05507v1
2023-02-10T20:50:58
2023-02-10T20:50:58
{ "title": [ 0.027389526, -0.010398865, 0.011116028, -0.06048584, -0.013214111, -0.0018968582, -0.027404785, -0.011260986, 0.024993896, 0.035736084, -0.06750488, 0.017166138, 0.023208618, 0.0041770935, -0.025253296, -0.005836487, -0.009208679, 0....
http://arxiv.org/abs/2302.07738v2
Alloprof: a new French question-answer education dataset and its use in an information retrieval case study
Antoine Lefebvre-Brossard, Stephane Gazaille, Michel C. Desmarais
cs.CL, cs.IR, cs.LG
Teachers and students are increasingly relying on online learning resources to supplement the ones provided in school. This increase in the breadth and depth of available resources is a great thing for students, but only provided they are able to find answers to their queries. Question-answering and information retriev...
http://arxiv.org/pdf/2302.07738v2
2023-04-14T13:20:07
2023-02-10T20:23:27
{ "title": [ -0.014602661, -0.03213501, -0.029922485, -0.018981934, -0.015808105, -0.022445679, 0.008018494, 0.016159058, 0.018859863, 0.030349731, -0.028930664, 0.0236969, -0.029754639, -0.00033688545, -0.050689697, -0.004890442, -0.0047302246, ...
http://arxiv.org/abs/2302.07731v3
Combat AI With AI: Counteract Machine-Generated Fake Restaurant Reviews on Social Media
Alessandro Gambetti, Qiwei Han
cs.CL, cs.AI, cs.IR, cs.LG
Recent advances in generative models such as GPT may be used to fabricate indistinguishable fake customer reviews at a much lower cost, thus posing challenges for social media platforms to detect these machine-generated fake reviews. We propose to leverage the high-quality elite restaurant reviews verified by Yelp to g...
http://arxiv.org/pdf/2302.07731v3
2023-04-21T16:40:55
2023-02-10T19:40:10
{ "title": [ 0.05392456, -0.00013720989, 0.015380859, 0.028686523, 0.0010328293, -0.0340271, 0.06915283, -0.048919678, -0.050354004, 0.026123047, -0.035095215, 0.045837402, -0.007572174, -0.018676758, 0.010757446, -0.03173828, -0.011886597, 0.048...
http://arxiv.org/abs/2302.05454v1
Distillation of encoder-decoder transformers for sequence labelling
Marco Farina, Duccio Pappadopulo, Anant Gupta, Leslie Huang, Ozan İrsoy, Thamar Solorio
cs.CL, cs.IR
Driven by encouraging results on a wide range of tasks, the field of NLP is experiencing an accelerated race to develop bigger language models. This race for bigger models has also underscored the need to continue the pursuit of practical distillation approaches that can leverage the knowledge acquired by these big mod...
http://arxiv.org/pdf/2302.05454v1
2023-02-10T19:00:00
2023-02-10T19:00:00
{ "title": [ 0.01159668, 0.030715942, -0.05319214, -0.069885254, 0.0072135925, -0.027114868, -0.047058105, -0.025238037, 0.022491455, 0.06567383, -0.03665161, -0.011489868, 0.04119873, -0.026382446, -0.02708435, 0.020751953, 0.012367249, 0.042480...
http://arxiv.org/abs/2302.05406v1
Adversarial Transformer Language Models for Contextual Commonsense Inference
Pedro Colon-Hernandez, Henry Lieberman, Yida Xin, Claire Yin, Cynthia Breazeal, Peter Chin
cs.CL
Contextualized or discourse aware commonsense inference is the task of generating coherent commonsense assertions (i.e., facts) from a given story, and a particular sentence from that story. Some problems with the task are: lack of controllability for topics of the inferred facts; lack of commonsense knowledge during t...
http://arxiv.org/pdf/2302.05406v1
2023-02-10T18:21:13
2023-02-10T18:21:13
{ "title": [ 0.010559082, -0.007888794, -0.02796936, -0.053619385, 0.0016355515, 0.0045051575, -0.032592773, -0.011421204, 0.0023975372, 0.031219482, -0.021240234, 0.00035238266, 0.0069122314, -0.015701294, -0.02973938, -0.0067329407, -0.0132751465, ...
http://arxiv.org/abs/2302.05392v1
Span-based Named Entity Recognition by Generating and Compressing Information
Nhung T. H. Nguyen, Makoto Miwa, Sophia Ananiadou
cs.CL, cs.LG
The information bottleneck (IB) principle has been proven effective in various NLP applications. The existing work, however, only used either generative or information compression models to improve the performance of the target task. In this paper, we propose to combine the two types of IB models into one system to enh...
http://arxiv.org/pdf/2302.05392v1
2023-02-10T17:40:51
2023-02-10T17:40:51
{ "title": [ -0.03945923, -0.005207062, -0.06604004, -0.053741455, 0.023147583, 0.006252289, -0.025512695, 0.019439697, -0.0045166016, 0.016082764, -0.021316528, -0.00093746185, -0.011878967, -0.017059326, -0.037902832, -0.016113281, 0.0055885315, ...
http://arxiv.org/abs/2302.05244v5
A Song of Ice and Fire: Analyzing Textual Autotelic Agents in ScienceWorld
Laetitia Teodorescu, Xingdi Yuan, Marc-Alexandre Côté, Pierre-Yves Oudeyer
cs.AI, cs.CL, cs.LG
Building open-ended agents that can autonomously discover a diversity of behaviours is one of the long-standing goals of artificial intelligence. This challenge can be studied in the framework of autotelic RL agents, i.e. agents that learn by selecting and pursuing their own goals, self-organizing a learning curriculum...
http://arxiv.org/pdf/2302.05244v5
2023-02-24T10:15:45
2023-02-10T13:49:50
{ "title": [ -0.0110321045, -0.004776001, 0.0031433105, -0.0006766319, -0.008682251, 0.002817154, 0.014091492, -0.0317688, -0.057800293, 0.046447754, -0.032470703, 0.050933838, 0.0045204163, 0.0050315857, -0.018051147, -0.010253906, -0.0049095154, ...
http://arxiv.org/abs/2302.05206v1
The Wisdom of Hindsight Makes Language Models Better Instruction Followers
Tianjun Zhang, Fangchen Liu, Justin Wong, Pieter Abbeel, Joseph E. Gonzalez
cs.CL, cs.AI
Reinforcement learning has seen wide success in finetuning large language models to better align with instructions via human feedback. The so-called algorithm, Reinforcement Learning with Human Feedback (RLHF) demonstrates impressive performance on the GPT series models. However, the underlying Reinforcement Learning (...
http://arxiv.org/pdf/2302.05206v1
2023-02-10T12:16:38
2023-02-10T12:16:38
{ "title": [ 0.013389587, -0.028625488, -0.016601562, -0.042175293, -0.06506348, -0.008895874, -0.058929443, -0.0031929016, -0.013977051, 0.03274536, -0.028839111, -0.008590698, 0.009880066, -0.046722412, 0.007949829, -0.01235199, -0.0045814514, ...
http://arxiv.org/abs/2302.05138v2
Plan-then-Seam: Towards Efficient Table-to-Text Generation
Liang Li, Ruiying Geng, Chengyang Fang, Bing Li, Can Ma, Binhua Li, Yongbin Li
cs.CL
Table-to-text generation aims at automatically generating text to help people conveniently obtain salient information in tables. Recent works explicitly decompose the generation process into content planning and surface generation stages, employing two autoregressive networks for them respectively. However, they are co...
http://arxiv.org/pdf/2302.05138v2
2023-02-28T07:20:06
2023-02-10T09:43:15
{ "title": [ -0.03564453, -0.003929138, -0.051818848, -0.05670166, 0.009986877, 0.006855011, -0.057922363, -0.015083313, -0.0043754578, 0.05847168, -0.025802612, 0.038970947, 0.015686035, 0.017028809, -0.050872803, -0.0011758804, 0.00178051, 0.01...
http://arxiv.org/abs/2302.05137v1
Realistic Conversational Question Answering with Answer Selection based on Calibrated Confidence and Uncertainty Measurement
Soyeong Jeong, Jinheon Baek, Sung Ju Hwang, Jong C. Park
cs.CL, cs.AI, cs.IR
Conversational Question Answering (ConvQA) models aim at answering a question with its relevant paragraph and previous question-answer pairs that occurred during conversation multiple times. To apply such models to a real-world scenario, some existing work uses predicted answers, instead of unavailable ground-truth ans...
http://arxiv.org/pdf/2302.05137v1
2023-02-10T09:42:07
2023-02-10T09:42:07
{ "title": [ -0.0236969, 0.010978699, -0.055236816, -0.041381836, -0.001619339, -0.011741638, -0.042419434, -0.010108948, 0.018676758, 0.027542114, -0.003446579, 0.00982666, -0.019348145, -0.021209717, -0.064697266, 0.007423401, -0.003862381, -0....
http://arxiv.org/abs/2302.05128v1
Translating Natural Language to Planning Goals with Large-Language Models
Yaqi Xie, Chen Yu, Tongyao Zhu, Jinbin Bai, Ze Gong, Harold Soh
cs.CL, cs.AI, cs.RO
Recent large language models (LLMs) have demonstrated remarkable performance on a variety of natural language processing (NLP) tasks, leading to intense excitement about their applicability across various domains. Unfortunately, recent work has also shown that LLMs are unable to perform accurate reasoning nor solve pla...
http://arxiv.org/pdf/2302.05128v1
2023-02-10T09:17:52
2023-02-10T09:17:52
{ "title": [ -0.036895752, -0.014724731, -0.02633667, -0.049957275, -0.0034503937, 0.008506775, -0.04650879, -0.014579773, -0.008468628, 0.013153076, -0.018234253, 0.0046310425, -0.023788452, -0.015853882, -0.011062622, -0.015289307, -0.0036354065, ...
http://arxiv.org/abs/2302.05120v1
Step by Step Loss Goes Very Far: Multi-Step Quantization for Adversarial Text Attacks
Piotr Gaiński, Klaudia Bałazy
cs.CL
We propose a novel gradient-based attack against transformer-based language models that searches for an adversarial example in a continuous space of token probabilities. Our algorithm mitigates the gap between adversarial loss for continuous and discrete text representations by performing multi-step quantization in a q...
http://arxiv.org/pdf/2302.05120v1
2023-02-10T08:50:51
2023-02-10T08:50:51
{ "title": [ -0.01525116, -0.006149292, -0.024658203, -0.039978027, 0.000009953976, -0.011924744, -0.015144348, -0.007911682, -0.039093018, 0.05355835, 0.02268982, 0.02784729, -0.0070533752, -0.026535034, -0.029815674, 0.008270264, 0.012069702, 0...
http://arxiv.org/abs/2302.05110v1
Cross-Corpora Spoken Language Identification with Domain Diversification and Generalization
Spandan Dey, Md Sahidullah, Goutam Saha
eess.AS, cs.CL, cs.LG, cs.SD
This work addresses the cross-corpora generalization issue for the low-resourced spoken language identification (LID) problem. We have conducted the experiments in the context of Indian LID and identified strikingly poor cross-corpora generalization due to corpora-dependent non-lingual biases. Our contribution to this ...
http://arxiv.org/pdf/2302.05110v1
2023-02-10T08:21:42
2023-02-10T08:21:42
{ "title": [ -0.009536743, 0.028503418, -0.058502197, -0.06774902, 0.029647827, -0.009918213, -0.05319214, -0.0029907227, -0.014808655, 0.027236938, -0.018051147, 0.0067825317, -0.0054397583, -0.018432617, -0.03994751, -0.023880005, -0.020370483, ...
http://arxiv.org/abs/2302.05096v1
Selective In-Context Data Augmentation for Intent Detection using Pointwise V-Information
Yen-Ting Lin, Alexandros Papangelis, Seokhwan Kim, Sungjin Lee, Devamanyu Hazarika, Mahdi Namazifar, Di Jin, Yang Liu, Dilek Hakkani-Tur
cs.CL, cs.AI
This work focuses on in-context data augmentation for intent detection. Having found that augmentation via in-context prompting of large pre-trained language models (PLMs) alone does not improve performance, we introduce a novel approach based on PLMs and pointwise V-information (PVI), a metric that can measure the use...
http://arxiv.org/pdf/2302.05096v1
2023-02-10T07:37:49
2023-02-10T07:37:49
{ "title": [ 0.028945923, 0.012123108, -0.062561035, -0.05239868, 0.03845215, -0.039520264, 0.016555786, -0.017700195, 0.0055274963, 0.015899658, -0.04977417, 0.010871887, 0.0042800903, -0.025756836, -0.01374054, -0.01361084, -0.004371643, 0.0448...
http://arxiv.org/abs/2302.05061v1
ControversialQA: Exploring Controversy in Question Answering
Zhen Wang, Peide Zhu, Jie Yang
cs.CL
Controversy is widespread online. Previous studies mainly define controversy based on vague assumptions of its relation to sentiment such as hate speech and offensive words. This paper introduces the first question-answering dataset that defines content controversy by user perception, i.e., votes from plenty of users. ...
http://arxiv.org/pdf/2302.05061v1
2023-02-10T05:39:29
2023-02-10T05:39:29
{ "title": [ -0.011199951, -0.008407593, -0.029205322, -0.048339844, 0.010803223, -0.008323669, -0.014450073, -0.011543274, 0.02607727, 0.013191223, -0.036834717, -0.0025138855, -0.012893677, -0.0058288574, -0.060150146, -0.01386261, -0.011650085, ...
http://arxiv.org/abs/2302.05040v2
PATCorrect: Non-autoregressive Phoneme-augmented Transformer for ASR Error Correction
Ziji Zhang, Zhehui Wang, Rajesh Kamma, Sharanya Eswaran, Narayanan Sadagopan
cs.CL, cs.LG, cs.SD, eess.AS
Speech-to-text errors made by automatic speech recognition (ASR) systems negatively impact downstream models. Error correction models as a post-processing text editing method have been recently developed for refining the ASR outputs. However, efficient models that meet the low latency requirements of industrial grade p...
http://arxiv.org/pdf/2302.05040v2
2023-06-21T17:44:58
2023-02-10T04:05:24
{ "title": [ -0.0152282715, -0.011222839, -0.05609131, -0.06878662, 0.0047340393, -0.0029697418, -0.03564453, 0.023239136, -0.004787445, 0.056671143, -0.018066406, 0.0124435425, 0.044891357, -0.0041007996, -0.05014038, 0.011169434, -0.01423645, 0...
http://arxiv.org/abs/2302.05039v1
Predicting Desirable Revisions of Evidence and Reasoning in Argumentative Writing
Tazin Afrin, Diane Litman
cs.CL, cs.AI
We develop models to classify desirable evidence and desirable reasoning revisions in student argumentative writing. We explore two ways to improve classifier performance - using the essay context of the revision, and using the feedback students received before the revision. We perform both intrinsic and extrinsic eval...
http://arxiv.org/pdf/2302.05039v1
2023-02-10T03:59:59
2023-02-10T03:59:59
{ "title": [ -0.016693115, -0.001206398, -0.03375244, -0.04498291, -0.013427734, -0.0049095154, -0.026107788, -0.007659912, -0.03665161, 0.004623413, -0.021255493, 0.009170532, -0.009552002, -0.02368164, -0.02230835, -0.023025513, -0.026321411, 0...
http://arxiv.org/abs/2302.06419v1
AV-data2vec: Self-supervised Learning of Audio-Visual Speech Representations with Contextualized Target Representations
Jiachen Lian, Alexei Baevski, Wei-Ning Hsu, Michael Auli
eess.AS, cs.AI, cs.CL
Self-supervision has shown great potential for audio-visual speech recognition by vastly reducing the amount of labeled data required to build good systems. However, existing methods are either not entirely end-to-end or do not train joint representations of both modalities. In this paper, we introduce AV-data2vec whic...
http://arxiv.org/pdf/2302.06419v1
2023-02-10T02:55:52
2023-02-10T02:55:52
{ "title": [ 0.0032253265, -0.00096559525, -0.04724121, -0.078125, 0.009887695, -0.007549286, -0.0345459, 0.016418457, -0.027816772, 0.04827881, 0.003030777, 0.017547607, 0.0062789917, -0.025466919, -0.028579712, 0.0030097961, 0.003232956, 0.0478...
http://arxiv.org/abs/2302.05016v2
Is Multimodal Vision Supervision Beneficial to Language?
Avinash Madasu, Vasudev Lal
cs.CV, cs.AI, cs.CL
Vision (image and video) - Language (VL) pre-training is the recent popular paradigm that achieved state-of-the-art results on multi-modal tasks like image-retrieval, video-retrieval, visual question answering etc. These models are trained in an unsupervised way and greatly benefit from the complementary modality super...
http://arxiv.org/pdf/2302.05016v2
2023-04-15T00:04:54
2023-02-10T02:22:44
{ "title": [ -0.0029964447, -0.009262085, -0.04296875, -0.035736084, -0.04046631, -0.02709961, -0.04748535, 0.0062789917, -0.0018634796, 0.018981934, -0.019119263, 0.0055351257, 0.018722534, -0.025512695, -0.011520386, -0.0030879974, -0.0076179504, ...
http://arxiv.org/abs/2302.06470v1
POSGen: Personalized Opening Sentence Generation for Online Insurance Sales
Yu Li, Yi Zhang, Weijia Wu, Zimu Zhou, Qiang Li
cs.CL, Online Insurance Recommendation, Transfer Learning, Data-to-text Generation
The insurance industry is shifting their sales mode from offline to online, in expectation to reach massive potential customers in the digitization era. Due to the complexity and the nature of insurance products, a cost-effective online sales solution is to exploit chatbot AI to raise customers' attention and pass thos...
http://arxiv.org/pdf/2302.06470v1
2023-02-10T01:40:03
2023-02-10T01:40:03
{ "title": [ -0.028045654, -0.012794495, -0.03466797, -0.04446411, -0.017364502, 0.013198853, -0.039001465, 0.00031352043, -0.027389526, 0.026321411, -0.016983032, 0.044677734, -0.0045204163, -0.013931274, -0.004016876, -0.0062789917, -0.019866943, ...
http://arxiv.org/abs/2302.05008v1
Language-Aware Multilingual Machine Translation with Self-Supervised Learning
Haoran Xu, Jean Maillard, Vedanuj Goswami
cs.CL
Multilingual machine translation (MMT) benefits from cross-lingual transfer but is a challenging multitask optimization problem. This is partly because there is no clear framework to systematically learn language-specific parameters. Self-supervised learning (SSL) approaches that leverage large quantities of monolingua...
http://arxiv.org/pdf/2302.05008v1
2023-02-10T01:34:24
2023-02-10T01:34:24
{ "title": [ -0.028533936, -0.0012950897, -0.0209198, -0.059143066, -0.0037250519, 0.014877319, -0.04876709, 0.016204834, -0.017364502, 0.044036865, 0.0043792725, 0.0014820099, 0.010231018, -0.025527954, -0.035858154, 0.01386261, 0.000916481, 0.0...
http://arxiv.org/abs/2302.04985v1
Event Temporal Relation Extraction with Bayesian Translational Model
Xingwei Tan, Gabriele Pergola, Yulan He
cs.CL, cs.LG
Existing models to extract temporal relations between events lack a principled method to incorporate external knowledge. In this study, we introduce Bayesian-Trans, a Bayesian learning-based method that models the temporal relation representations as latent variables and infers their values via Bayesian inference and t...
http://arxiv.org/pdf/2302.04985v1
2023-02-10T00:11:19
2023-02-10T00:11:19
{ "title": [ 0.009399414, 0.008460999, -0.05947876, -0.107055664, 0.023956299, -0.009338379, -0.06951904, -0.0032520294, 0.034942627, 0.02494812, 0.014060974, -0.016143799, -0.0023841858, -0.030288696, -0.057861328, 0.013893127, -0.009437561, 0.0...
http://arxiv.org/abs/2302.04981v1
AutoNMT: A Framework to Streamline the Research of Seq2Seq Models
Salvador Carrión, Francisco Casacuberta
cs.CL, cs.AI, cs.LG
We present AutoNMT, a framework to streamline the research of seq-to-seq models by automating the data pipeline (i.e., file management, data preprocessing, and exploratory analysis), automating experimentation in a toolkit-agnostic manner, which allows users to use either their own models or existing seq-to-seq toolkit...
http://arxiv.org/pdf/2302.04981v1
2023-02-09T23:42:30
2023-02-09T23:42:30
{ "title": [ -0.029129028, -0.012924194, -0.04638672, -0.049621582, -0.0014743805, -0.0067443848, -0.053894043, -0.02734375, -0.017166138, 0.02923584, 0.021362305, 0.0039787292, 0.004524231, -0.051361084, -0.026901245, 0.022964478, -0.012878418, ...
http://arxiv.org/abs/2302.04975v1
Leveraging supplementary text data to kick-start automatic speech recognition system development with limited transcriptions
Nay San, Martijn Bartelds, Blaine Billings, Ella de Falco, Hendi Feriza, Johan Safri, Wawan Sahrozi, Ben Foley, Bradley McDonnell, Dan Jurafsky
cs.CL
Recent research using pre-trained transformer models suggests that just 10 minutes of transcribed speech may be enough to fine-tune such a model for automatic speech recognition (ASR) -- at least if we can also leverage vast amounts of text data (803 million tokens). But is that much text data necessary? We study the u...
http://arxiv.org/pdf/2302.04975v1
2023-02-09T23:30:49
2023-02-09T23:30:49
{ "title": [ -0.052734375, -0.023834229, -0.05126953, -0.06210327, -0.034576416, 0.002960205, -0.07147217, 0.024291992, -0.030319214, 0.040039062, -0.0836792, 0.043273926, -0.018432617, 0.021392822, -0.03866577, -0.030822754, 0.011978149, 0.03842...
http://arxiv.org/abs/2304.09082v1
Unsupervised clustering of file dialects according to monotonic decompositions of mixtures
Michael Robinson, Tate Altman, Denley Lam, Letitia W. Li
cs.PL, cs.CL, cs.IR, 62P30 (Primary), 06A07 (Secondary), D.3.4
This paper proposes an unsupervised classification method that partitions a set of files into non-overlapping dialects based upon their behaviors, determined by messages produced by a collection of programs that consume them. The pattern of messages can be used as the signature of a particular kind of behavior, with th...
http://arxiv.org/pdf/2304.09082v1
2023-02-09T21:15:36
2023-02-09T21:15:36
{ "title": [ -0.0107421875, 0.022735596, -0.05987549, -0.07019043, 0.024047852, -0.0016756058, -0.062805176, 0.0036506653, 0.00003963709, 0.057159424, 0.0000011920929, 0.019317627, 0.026504517, -0.01285553, -0.022216797, 0.028442383, 0.008110046, ...
http://arxiv.org/abs/2302.04931v1
In-Context Learning with Many Demonstration Examples
Mukai Li, Shansan Gong, Jiangtao Feng, Yiheng Xu, Jun Zhang, Zhiyong Wu, Lingpeng Kong
cs.CL, cs.AI
Large pre-training language models (PLMs) have shown promising in-context learning abilities. However, due to the backbone transformer architecture, existing PLMs are bottlenecked by the memory and computational cost when scaling up to a large context size, leaving instruction tuning and in-context learning of many dem...
http://arxiv.org/pdf/2302.04931v1
2023-02-09T20:53:12
2023-02-09T20:53:12
{ "title": [ 0.07299805, 0.017990112, -0.022521973, -0.07067871, 0.020721436, -0.046661377, -0.0027751923, -0.01537323, -0.009559631, 0.010612488, -0.04953003, 0.008491516, -0.009246826, -0.03479004, 0.032928467, -0.021972656, 0.026107788, 0.0219...
http://arxiv.org/abs/2302.04914v1
Flexible, Model-Agnostic Method for Materials Data Extraction from Text Using General Purpose Language Models
Maciej P. Polak, Shrey Modi, Anna Latosinska, Jinming Zhang, Ching-Wen Wang, Shanonan Wang, Ayan Deep Hazra, Dane Morgan
cond-mat.mtrl-sci, cs.AI, cs.CL
Accurate and comprehensive material databases extracted from research papers are critical for materials science and engineering but require significant human effort to develop. In this paper we present a simple method of extracting materials data from full texts of research papers suitable for quickly developing modest...
http://arxiv.org/pdf/2302.04914v1
2023-02-09T19:56:37
2023-02-09T19:56:37
{ "title": [ -0.055999756, -0.001241684, -0.07287598, -0.055114746, -0.026138306, -0.024047852, -0.07910156, 0.0057754517, 0.011787415, 0.027236938, -0.004760742, -0.020355225, -0.002817154, -0.03942871, -0.03237915, 0.008743286, -0.013710022, 0....
http://arxiv.org/abs/2302.04907v1
Binarized Neural Machine Translation
Yichi Zhang, Ankush Garg, Yuan Cao, Łukasz Lew, Behrooz Ghorbani, Zhiru Zhang, Orhan Firat
cs.CL, cs.LG
The rapid scaling of language models is motivating research using low-bitwidth quantization. In this work, we propose a novel binarization technique for Transformers applied to machine translation (BMT), the first of its kind. We identify and address the problem of inflated dot-product variance when using one-bit weigh...
http://arxiv.org/pdf/2302.04907v1
2023-02-09T19:27:34
2023-02-09T19:27:34
{ "title": [ -0.024139404, -0.0059661865, -0.017822266, -0.057617188, -0.008224487, 0.008216858, -0.045410156, 0.007335663, -0.008308411, 0.059417725, -0.0066871643, -0.007545471, 0.027801514, -0.018203735, -0.038970947, 0.024642944, -0.007381439, ...
http://arxiv.org/abs/2302.04870v1
Offsite-Tuning: Transfer Learning without Full Model
Guangxuan Xiao, Ji Lin, Song Han
cs.CL, cs.CV, cs.LG
Transfer learning is important for foundation models to adapt to downstream tasks. However, many foundation models are proprietary, so users must share their data with model owners to fine-tune the models, which is costly and raise privacy concerns. Moreover, fine-tuning large foundation models is computation-intensive...
http://arxiv.org/pdf/2302.04870v1
2023-02-09T18:59:55
2023-02-09T18:59:55
{ "title": [ -0.026000977, -0.005680084, -0.030532837, -0.03201294, 0.004234314, -0.008346558, -0.021377563, -0.028213501, -0.00983429, 0.008377075, -0.014556885, 0.01335907, -0.02923584, -0.019195557, -0.019622803, -0.0012016296, -0.004589081, 0...
http://arxiv.org/abs/2302.04865v1
Learning by Asking for Embodied Visual Navigation and Task Completion
Ying Shen, Ismini Lourentzou
cs.CV, cs.AI, cs.CL, cs.LG
The research community has shown increasing interest in designing intelligent embodied agents that can assist humans in accomplishing tasks. Despite recent progress on related vision-language benchmarks, most prior work has focused on building agents that follow instructions rather than endowing agents the ability to a...
http://arxiv.org/pdf/2302.04865v1
2023-02-09T18:59:41
2023-02-09T18:59:41
{ "title": [ -0.004585266, -0.01146698, 0.000238657, -0.004508972, -0.0064582825, -0.020263672, 0.016174316, -0.03604126, -0.005970001, 0.018417358, -0.0513916, 0.0054092407, -0.021514893, 0.005088806, -0.045898438, -0.025802612, 0.011047363, 0.0...
http://arxiv.org/abs/2302.04863v3
Knowledge is a Region in Weight Space for Fine-tuned Language Models
Almog Gueta, Elad Venezian, Colin Raffel, Noam Slonim, Yoav Katz, Leshem Choshen
cs.LG, cs.AI, cs.CL
Research on neural networks has focused on understanding a single model trained on a single dataset. However, relatively little is known about the relationships between different models, particularly those trained or tested on different datasets. We address this by studying how the weight space and the underlying loss ...
http://arxiv.org/pdf/2302.04863v3
2023-10-12T18:42:34
2023-02-09T18:59:18
{ "title": [ -0.022369385, -0.05203247, -0.043273926, -0.06726074, -0.062316895, -0.010795593, -0.04525757, 0.06616211, -0.021728516, 0.036621094, -0.03363037, 0.0039100647, -0.006439209, -0.019577026, -0.025222778, -0.02557373, 0.021896362, 0.01...
http://arxiv.org/abs/2302.04858v2
Re-ViLM: Retrieval-Augmented Visual Language Model for Zero and Few-Shot Image Captioning
Zhuolin Yang, Wei Ping, Zihan Liu, Vijay Korthikanti, Weili Nie, De-An Huang, Linxi Fan, Zhiding Yu, Shiyi Lan, Bo Li, Ming-Yu Liu, Yuke Zhu, Mohammad Shoeybi, Bryan Catanzaro, Chaowei Xiao, Anima Anandkumar
cs.CV, cs.AI, cs.CL, cs.IR, cs.LG
Augmenting pretrained language models (LMs) with a vision encoder (e.g., Flamingo) has obtained the state-of-the-art results in image-to-text generation. However, these models store all the knowledge within their parameters, thus often requiring enormous model parameters to model the abundant visual concepts and very r...
http://arxiv.org/pdf/2302.04858v2
2023-10-22T04:18:00
2023-02-09T18:57:56
{ "title": [ 0.012786865, -0.017578125, -0.009757996, -0.03616333, -0.02507019, -0.012214661, -0.010627747, -0.007461548, 0.030975342, 0.059570312, -0.009338379, 0.010948181, 0.008483887, -0.001906395, -0.07098389, 0.012626648, -0.0019950867, 0.0...
http://arxiv.org/abs/2302.04834v1
FrameBERT: Conceptual Metaphor Detection with Frame Embedding Learning
Yucheng Li, Shun Wang, Chenghua Lin, Frank Guerin, Loïc Barrault
cs.CL
In this paper, we propose FrameBERT, a RoBERTa-based model that can explicitly learn and incorporate FrameNet Embeddings for concept-level metaphor detection. FrameBERT not only achieves better or comparable performance to the state-of-the-art, but also is more explainable and interpretable compared to existing models,...
http://arxiv.org/pdf/2302.04834v1
2023-02-09T18:41:04
2023-02-09T18:41:04
{ "title": [ 0.0036258698, -0.0038642883, -0.013084412, -0.014503479, 0.0046958923, -0.015899658, 0.001616478, -0.032104492, 0.014518738, 0.036956787, -0.02067566, 0.0027484894, 0.039764404, -0.013153076, -0.037078857, 0.01210022, -0.022094727, 0...
http://arxiv.org/abs/2302.04813v3
Explanation Selection Using Unlabeled Data for Chain-of-Thought Prompting
Xi Ye, Greg Durrett
cs.CL
Recent work has shown how to prompt large language models with explanations to obtain strong performance on textual reasoning tasks, i.e., the chain-of-thought paradigm. However, subtly different explanations can yield widely varying downstream task accuracy. Explanations that have not been "tuned" for a task, such as ...
http://arxiv.org/pdf/2302.04813v3
2023-10-18T14:42:15
2023-02-09T18:02:34
{ "title": [ -0.0025997162, -0.0004339218, -0.023788452, -0.03564453, -0.015510559, -0.007827759, -0.026763916, -0.022460938, -0.022735596, -0.02142334, -0.032409668, 0.024169922, -0.037597656, -0.029632568, 0.00484848, -0.029632568, -0.00015962124, ...
http://arxiv.org/abs/2302.04811v1
A Large-Scale Multilingual Study of Visual Constraints on Linguistic Selection of Descriptions
Uri Berger, Lea Frermann, Gabriel Stanovsky, Omri Abend
cs.CL
We present a large, multilingual study into how vision constrains linguistic choice, covering four languages and five linguistic properties, such as verb transitivity or use of numerals. We propose a novel method that leverages existing corpora of images with captions written by native speakers, and apply it to nine co...
http://arxiv.org/pdf/2302.04811v1
2023-02-09T17:57:58
2023-02-09T17:57:58
{ "title": [ -0.025039673, -0.011795044, -0.02696228, -0.014213562, -0.017364502, -0.015174866, -0.01878357, -0.019927979, -0.008087158, 0.029785156, -0.030715942, 0.025390625, 0.009048462, -0.0068244934, -0.020614624, 0.005039215, 0.0005545616, ...
http://arxiv.org/abs/2302.04790v1
Massively Multilingual Language Models for Cross Lingual Fact Extraction from Low Resource Indian Languages
Bhavyajeet Singh, Pavan Kandru, Anubhav Sharma, Vasudeva Varma
cs.CL, cs.AI, cs.IR
Massive knowledge graphs like Wikidata attempt to capture world knowledge about multiple entities. Recent approaches concentrate on automatically enriching these KGs from text. However a lot of information present in the form of natural text in low resource languages is often missed out. Cross Lingual Information Extra...
http://arxiv.org/pdf/2302.04790v1
2023-02-09T17:29:56
2023-02-09T17:29:56
{ "title": [ -0.037109375, -0.01586914, -0.049560547, -0.07543945, -0.011131287, 0.005672455, -0.072265625, 0.017822266, -0.0024433136, 0.0368042, -0.005004883, 0.012786865, 0.002506256, -0.010726929, -0.038360596, 0.016998291, -0.02520752, 0.044...
http://arxiv.org/abs/2302.04761v1
Toolformer: Language Models Can Teach Themselves to Use Tools
Timo Schick, Jane Dwivedi-Yu, Roberto Dessì, Roberta Raileanu, Maria Lomeli, Luke Zettlemoyer, Nicola Cancedda, Thomas Scialom
cs.CL
Language models (LMs) exhibit remarkable abilities to solve new tasks from just a few examples or textual instructions, especially at scale. They also, paradoxically, struggle with basic functionality, such as arithmetic or factual lookup, where much simpler and smaller models excel. In this paper, we show that LMs can...
http://arxiv.org/pdf/2302.04761v1
2023-02-09T16:49:57
2023-02-09T16:49:57
{ "title": [ -0.0021305084, -0.008346558, -0.053619385, -0.028930664, -0.03466797, -0.02078247, -0.042755127, -0.0041999817, -0.007904053, 0.04083252, -0.0035495758, -0.015655518, 0.024658203, -0.042907715, -0.018173218, -0.013809204, 0.008407593, ...
http://arxiv.org/abs/2302.04725v1
Lightweight Transformers for Clinical Natural Language Processing
Omid Rohanian, Mohammadmahdi Nouriborji, Hannah Jauncey, Samaneh Kouchaki, ISARIC Clinical Characterisation Group, Lei Clifton, Laura Merson, David A. Clifton
cs.CL, cs.AI, cs.LG, 68T50, I.2.7
Specialised pre-trained language models are becoming more frequent in NLP since they can potentially outperform models trained on generic texts. BioBERT and BioClinicalBERT are two examples of such models that have shown promise in medical NLP tasks. Many of these models are overparametrised and resource-intensive, but...
http://arxiv.org/pdf/2302.04725v1
2023-02-09T16:07:31
2023-02-09T16:07:31
{ "title": [ -0.013870239, -0.015083313, -0.027694702, -0.090270996, -0.0009279251, 0.027435303, -0.071777344, 0.008422852, -0.013130188, 0.041015625, -0.03074646, -0.0035419464, 0.012420654, -0.007850647, -0.044128418, 0.0013065338, -0.0059661865, ...
http://arxiv.org/abs/2302.04716v1
Improving the Generalizability of Collaborative Dialogue Analysis with Multi-Feature Embeddings
Ayesha Enayet, Gita Sukthankar
cs.CL
Conflict prediction in communication is integral to the design of virtual agents that support successful teamwork by providing timely assistance. The aim of our research is to analyze discourse to predict collaboration success. Unfortunately, resource scarcity is a problem that teamwork researchers commonly face since ...
http://arxiv.org/pdf/2302.04716v1
2023-02-09T15:58:32
2023-02-09T15:58:32
{ "title": [ -0.0017337799, 0.012374878, -0.046813965, -0.06994629, 0.018966675, -0.008712769, -0.05581665, 0.0067825317, 0.015167236, 0.013694763, 0.0052871704, 0.002231598, -0.023971558, -0.018295288, -0.050048828, -0.008865356, -0.025268555, 0...
http://arxiv.org/abs/2302.04700v1
Augmenting NLP data to counter Annotation Artifacts for NLI Tasks
Armaan Singh Bhullar
cs.CL, cs.LG, I.2.7
In this paper, we explore Annotation Artifacts - the phenomena wherein large pre-trained NLP models achieve high performance on benchmark datasets but do not actually "solve" the underlying task and instead rely on some dataset artifacts (same across train, validation, and test sets) to figure out the right answer. We ...
http://arxiv.org/pdf/2302.04700v1
2023-02-09T15:34:53
2023-02-09T15:34:53
{ "title": [ 0.0035800934, -0.013648987, -0.019729614, -0.022735596, -0.016983032, -0.015098572, -0.005306244, -0.02168274, 0.006954193, 0.025787354, -0.029724121, 0.00945282, -0.005882263, -0.0020160675, -0.054138184, -0.0013389587, -0.018066406, ...
http://arxiv.org/abs/2302.04643v1
A Novel Approach for Auto-Formulation of Optimization Problems
Yuting Ning, Jiayu Liu, Longhu Qin, Tong Xiao, Shangzi Xue, Zhenya Huang, Qi Liu, Enhong Chen, Jinze Wu
cs.CL
In the Natural Language for Optimization (NL4Opt) NeurIPS 2022 competition, competitors focus on improving the accessibility and usability of optimization solvers, with the aim of subtask 1: recognizing the semantic entities that correspond to the components of the optimization problem; subtask 2: generating formulatio...
http://arxiv.org/pdf/2302.04643v1
2023-02-09T13:57:06
2023-02-09T13:57:06
{ "title": [ -0.03427124, 0.003501892, -0.009895325, -0.042266846, 0.0043411255, 0.020019531, -0.048858643, -0.019302368, -0.0026626587, 0.027175903, 0.00856781, 0.017181396, -0.020339966, -0.017745972, -0.029769897, 0.01625061, 0.009971619, -0.0...
http://arxiv.org/abs/2302.04618v1
Robust Question Answering against Distribution Shifts with Test-Time Adaptation: An Empirical Study
Hai Ye, Yuyang Ding, Juntao Li, Hwee Tou Ng
cs.CL
A deployed question answering (QA) model can easily fail when the test data has a distribution shift compared to the training data. Robustness tuning (RT) methods have been widely studied to enhance model robustness against distribution shifts before model deployment. However, can we improve a model after deployment? T...
http://arxiv.org/pdf/2302.04618v1
2023-02-09T13:10:53
2023-02-09T13:10:53
{ "title": [ 0.006252289, -0.00077581406, -0.03543091, -0.042236328, 0.005634308, -0.020690918, -0.012496948, -0.0052833557, 0.018936157, 0.029769897, -0.027633667, 0.02154541, -0.016357422, 0.004966736, -0.062561035, -0.008399963, -0.0045280457, ...
http://arxiv.org/abs/2302.04580v1
Generating a Structured Summary of Numerous Academic Papers: Dataset and Method
Shuaiqi Liu, Jiannong Cao, Ruosong Yang, Zhiyuan Wen
cs.CL, cs.AI, I.2.7; I.7
Writing a survey paper on one research topic usually needs to cover the salient content from numerous related papers, which can be modeled as a multi-document summarization (MDS) task. Existing MDS datasets usually focus on producing the structureless summary covering a few input documents. Meanwhile, previous structur...
http://arxiv.org/pdf/2302.04580v1
2023-02-09T11:42:07
2023-02-09T11:42:07
{ "title": [ -0.004535675, 0.0032730103, -0.060272217, -0.06793213, 0.0001155138, -0.02168274, -0.04824829, 0.008766174, -0.05984497, 0.036621094, -0.020568848, 0.025390625, -0.012619019, -0.014953613, 0.004299164, -0.018417358, 0.022460938, 0.00...
http://arxiv.org/abs/2302.04562v1
NLP-based Decision Support System for Examination of Eligibility Criteria from Securities Prospectuses at the German Central Bank
Christian Hänig, Markus Schlösser, Serhii Hamotskyi, Gent Zambaku, Janek Blankenburg
cs.CL, cs.AI
As part of its digitization initiative, the German Central Bank (Deutsche Bundesbank) wants to examine the extent to which natural Language Processing (NLP) can be used to make independent decisions upon the eligibility criteria of securities prospectuses. Every month, the Directorate General Markets at the German Cent...
http://arxiv.org/pdf/2302.04562v1
2023-02-09T11:00:58
2023-02-09T11:00:58
{ "title": [ 0.00030207634, -0.026184082, -0.030227661, -0.044403076, -0.036071777, -0.009109497, -0.026245117, 0.0115737915, 0.036712646, 0.0016489029, -0.068603516, -0.026000977, 0.039154053, -0.024765015, -0.005203247, -0.030059814, -0.03274536, ...
http://arxiv.org/abs/2302.04555v1
Data Augmentation for Robust Character Detection in Fantasy Novels
Arthur Amalvy, Vincent Labatut, Richard Dufour
cs.CL
Named Entity Recognition (NER) is a low-level task often used as a foundation for solving higher level NLP problems. In the context of character detection in novels, NER false negatives can be an issue as they possibly imply missing certain characters or relationships completely. In this article, we demonstrate that ap...
http://arxiv.org/pdf/2302.04555v1
2023-02-09T10:47:03
2023-02-09T10:47:03
{ "title": [ -0.015716553, 0.0069007874, -0.05142212, -0.016342163, 0.017944336, -0.02331543, 0.011642456, -0.012771606, 0.01209259, 0.027114868, -0.032836914, 0.0137786865, 0.020645142, -0.017745972, -0.036376953, 0.0065231323, -0.013046265, 0.0...
http://arxiv.org/abs/2302.04542v1
Efficient Attention via Control Variates
Lin Zheng, Jianbo Yuan, Chong Wang, Lingpeng Kong
cs.LG, cs.CL, cs.CV
Random-feature-based attention (RFA) is an efficient approximation of softmax attention with linear runtime and space complexity. However, the approximation gap between RFA and conventional softmax attention is not well studied. Built upon previous progress of RFA, we characterize this gap through the lens of control v...
http://arxiv.org/pdf/2302.04542v1
2023-02-09T10:16:20
2023-02-09T10:16:20
{ "title": [ 0.010757446, 0.0021762848, -0.034423828, -0.055023193, 0.012023926, -0.020812988, -0.00031256676, -0.016571045, 0.016784668, 0.023269653, -0.043823242, -0.00491333, 0.0043640137, -0.0061073303, -0.034423828, 0.00579834, 0.0031585693, ...
http://arxiv.org/abs/2302.04511v2
A Large-Scale Analysis of Persian Tweets Regarding Covid-19 Vaccination
Taha ShabaniMirzaei, Houmaan Chamani, Amirhossein Abaskohi, Zhivar Sourati Hassan Zadeh, Behnam Bahrak
cs.CL, cs.SI, I.2.7
The Covid-19 pandemic had an enormous effect on our lives, especially on people's interactions. By introducing Covid-19 vaccines, both positive and negative opinions were raised over the subject of taking vaccines or not. In this paper, using data gathered from Twitter, including tweets and user profiles, we offer a co...
http://arxiv.org/pdf/2302.04511v2
2023-10-03T22:27:03
2023-02-09T09:08:19
{ "title": [ 0.039031982, -0.0057525635, -0.03817749, -0.06976318, 0.010307312, -0.024459839, -0.008583069, 0.018112183, -0.046691895, 0.03857422, -0.0016803741, 0.03604126, 0.018478394, -0.03427124, 0.019714355, -0.0015525818, 0.0039482117, 0.03...
http://arxiv.org/abs/2302.04465v1
Detecting Contextomized Quotes in News Headlines by Contrastive Learning
Seonyeong Song, Hyeonho Song, Kunwoo Park, Jiyoung Han, Meeyoung Cha
cs.CL, cs.CY
Quotes are critical for establishing credibility in news articles. A direct quote enclosed in quotation marks has a strong visual appeal and is a sign of a reliable citation. Unfortunately, this journalistic practice is not strictly followed, and a quote in the headline is often "contextomized." Such a quote uses words...
http://arxiv.org/pdf/2302.04465v1
2023-02-09T07:04:11
2023-02-09T07:04:11
{ "title": [ 0.021087646, 0.02558899, -0.0071144104, -0.02911377, 0.026153564, -0.012649536, 0.020339966, -0.03955078, -0.012390137, 0.04360962, -0.039978027, 0.02684021, -0.020233154, 0.009651184, -0.0625, -0.023803711, -0.026428223, 0.06390381,...
http://arxiv.org/abs/2302.04460v2
Bag of Tricks for Training Data Extraction from Language Models
Weichen Yu, Tianyu Pang, Qian Liu, Chao Du, Bingyi Kang, Yan Huang, Min Lin, Shuicheng Yan
cs.CL, cs.AI, cs.CR, cs.LG
With the advance of language models, privacy protection is receiving more attention. Training data extraction is therefore of great importance, as it can serve as a potential tool to assess privacy leakage. However, due to the difficulty of this task, most of the existing methods are proof-of-concept and still not effe...
http://arxiv.org/pdf/2302.04460v2
2023-06-01T10:14:55
2023-02-09T06:46:42
{ "title": [ -0.035339355, -0.005886078, -0.053619385, -0.0592041, -0.0024604797, -0.017349243, -0.050323486, -0.002193451, -0.008155823, 0.026168823, -0.022506714, 0.0044517517, -0.009880066, -0.022354126, -0.04336548, 0.0056991577, -0.009208679, ...
http://arxiv.org/abs/2302.04459v1
Global Constraints with Prompting for Zero-Shot Event Argument Classification
Zizheng Lin, Hongming Zhang, Yangqiu Song
cs.CL
Determining the role of event arguments is a crucial subtask of event extraction. Most previous supervised models leverage costly annotations, which is not practical for open-domain applications. In this work, we propose to use global constraints with prompting to effectively tackles event argument classification witho...
http://arxiv.org/pdf/2302.04459v1
2023-02-09T06:39:29
2023-02-09T06:39:29
{ "title": [ 0.03390503, -0.007888794, -0.02772522, -0.10534668, -0.0027122498, -0.003326416, -0.050354004, 0.01374054, 0.013923645, -0.0076789856, 0.016647339, -0.008117676, -0.021438599, -0.05999756, 0.014160156, -0.004047394, 0.011672974, 0.02...
http://arxiv.org/abs/2302.04456v2
ERNIE-Music: Text-to-Waveform Music Generation with Diffusion Models
Pengfei Zhu, Chao Pang, Yekun Chai, Lei Li, Shuohuan Wang, Yu Sun, Hao Tian, Hua Wu
cs.SD, cs.AI, cs.CL, cs.MM, eess.AS
In recent years, the burgeoning interest in diffusion models has led to significant advances in image and speech generation. Nevertheless, the direct synthesis of music waveforms from unrestricted textual prompts remains a relatively underexplored domain. In response to this lacuna, this paper introduces a pioneering c...
http://arxiv.org/pdf/2302.04456v2
2023-09-21T09:30:00
2023-02-09T06:27:09
{ "title": [ -0.033569336, -0.011787415, -0.054504395, -0.059509277, 0.010101318, -0.0032405853, -0.025604248, 0.0154418945, 0.012924194, 0.04953003, -0.02130127, 0.020614624, 0.030303955, -0.01676941, -0.04812622, 0.021057129, -0.019210815, 0.03...
http://arxiv.org/abs/2302.06407v1
Correcting Real-Word Spelling Errors: A New Hybrid Approach
Seyed MohammadSadegh Dashti, Amid Khatibi Bardsiri, Vahid Khatibi Bardsiri
cs.CL, cs.AI
Spelling correction is one of the main tasks in the field of Natural Language Processing. Contrary to common spelling errors, real-word errors cannot be detected by conventional spelling correction methods. The real-word correction model proposed by Mays, Damerau and Mercer showed a great performance in different evalu...
http://arxiv.org/pdf/2302.06407v1
2023-02-09T06:03:11
2023-02-09T06:03:11
{ "title": [ -0.026855469, -0.01335144, -0.009346008, -0.008895874, -0.014450073, -0.0042495728, -0.00073862076, -0.015563965, 0.0015153885, 0.046813965, -0.030166626, -0.007587433, 0.017120361, -0.0051116943, -0.056793213, -0.0019893646, -0.03329467...
http://arxiv.org/abs/2302.04449v3
Read and Reap the Rewards: Learning to Play Atari with the Help of Instruction Manuals
Yue Wu, Yewen Fan, Paul Pu Liang, Amos Azaria, Yuanzhi Li, Tom M. Mitchell
cs.LG, cs.AI, cs.CL
High sample complexity has long been a challenge for RL. On the other hand, humans learn to perform tasks not only from interaction or demonstrations, but also by reading unstructured text documents, e.g., instruction manuals. Instruction manuals and wiki pages are among the most abundant data that could inform agents ...
http://arxiv.org/pdf/2302.04449v3
2023-10-26T08:23:48
2023-02-09T05:47:03
{ "title": [ -0.018630981, -0.015655518, -0.024368286, 0.041046143, -0.02784729, -0.025894165, 0.00019288063, -0.035247803, -0.004299164, 0.009559631, -0.04953003, 0.016784668, -0.0047798157, -0.018173218, 0.006259918, -0.03857422, 0.009521484, -...
http://arxiv.org/abs/2302.04443v1
Enhancing E-Commerce Recommendation using Pre-Trained Language Model and Fine-Tuning
Nuofan Xu, Chenhui Hu
cs.LG, cs.CL, cs.IR
Pretrained Language Models (PLM) have been greatly successful on a board range of natural language processing (NLP) tasks. However, it has just started being applied to the domain of recommendation systems. Traditional recommendation algorithms failed to incorporate the rich textual information in e-commerce datasets, ...
http://arxiv.org/pdf/2302.04443v1
2023-02-09T05:20:52
2023-02-09T05:20:52
{ "title": [ 0.008201599, -0.007297516, -0.04269409, -0.064331055, 0.0067825317, -0.011581421, -0.0340271, 0.000995636, -0.015533447, 0.048187256, -0.011428833, 0.012512207, 0.011856079, -0.0076560974, -0.037200928, 0.0055007935, 0.002866745, 0.0...
http://arxiv.org/abs/2302.04434v1
Real-Time Visual Feedback to Guide Benchmark Creation: A Human-and-Metric-in-the-Loop Workflow
Anjana Arunkumar, Swaroop Mishra, Bhavdeep Sachdeva, Chitta Baral, Chris Bryan
cs.CL, cs.AI, cs.HC, cs.LG
Recent research has shown that language models exploit `artifacts' in benchmarks to solve tasks, rather than truly learning them, leading to inflated model performance. In pursuit of creating better benchmarks, we propose VAIDA, a novel benchmark creation paradigm for NLP, that focuses on guiding crowdworkers, an under...
http://arxiv.org/pdf/2302.04434v1
2023-02-09T04:43:10
2023-02-09T04:43:10
{ "title": [ -0.031021118, -0.013946533, -0.003047943, -0.0064926147, -0.02330017, 0.0007019043, -0.017059326, -0.04272461, 0.0052490234, 0.036834717, -0.020339966, 0.014450073, -0.02848816, -0.012619019, -0.0602417, -0.009292603, 0.034210205, -0...
http://arxiv.org/abs/2302.04424v1
A Transformer-based Response Evaluator for Open-Domain Spoken Conversation
Vrindavan Harrison, Rishi Rajasekaran, Marilyn Walker
cs.CL
Many open-domain dialogue systems rely on multiple response generators, any of which can contribute a response to the dialogue in a particular context. Thus the ability to compare potential responses and then select the best plays an important role in ensuring a dialogue system is coherent and engaging. Dialogue cohere...
http://arxiv.org/pdf/2302.04424v1
2023-02-09T03:38:07
2023-02-09T03:38:07
{ "title": [ 0.0062217712, 0.017745972, -0.02671814, -0.068847656, 0.01739502, -0.004131317, -0.03930664, -0.0019187927, -0.009918213, 0.026260376, -0.009513855, 0.0016937256, -0.008850098, -0.013366699, -0.0703125, -0.01977539, -0.0140686035, 0....
http://arxiv.org/abs/2302.04415v3
Few-Shot Table-to-Text Generation with Prompt Planning and Knowledge Memorization
Zhixin Guo, Minyxuan Yan, Jiexing Qi, Jianping Zhou, Ziwei He, Zhouhan Lin, Guanjie Zheng, Xinbing Wang
cs.CL, cs.AI
Pre-trained language models (PLM) have achieved remarkable advancement in table-to-text generation tasks. However, the lack of labeled domain-specific knowledge and the topology gap between tabular data and text make it difficult for PLMs to yield faithful text. Low-resource generation likewise faces unique challenges ...
http://arxiv.org/pdf/2302.04415v3
2023-08-17T01:51:48
2023-02-09T03:04:11
{ "title": [ -0.011451721, -0.013122559, -0.034454346, -0.049713135, -0.020339966, 0.0023078918, -0.037628174, 0.008636475, -0.010238647, 0.049682617, -0.030456543, 0.036956787, 0.0015668869, 0.0023422241, -0.038604736, -0.015563965, 0.016723633, ...
http://arxiv.org/abs/2302.04391v6
The Re-Label Method For Data-Centric Machine Learning
Tong Guo
cs.LG, cs.CL
In industry deep learning application, our manually labeled data has a certain number of noisy data. To solve this problem and achieve more than 90 score in dev dataset, we present a simple method to find the noisy data and re-label the noisy data by human, given the model predictions as references in human labeling. I...
http://arxiv.org/pdf/2302.04391v6
2023-11-02T03:46:34
2023-02-09T01:09:57
{ "title": [ -0.028869629, -0.006652832, -0.031143188, -0.039489746, -0.013191223, -0.003643036, -0.02822876, -0.01272583, -0.012184143, 0.04324341, -0.027404785, 0.01574707, 0.00011175871, -0.010940552, -0.020736694, 0.014892578, 0.002298355, 0....
http://arxiv.org/abs/2302.06451v1
Ordered Memory Baselines
Daniel Borisov, Matthew D'Iorio, Jeffrey Hyacinthe
cs.CL, cs.LG
Natural language semantics can be modeled using the phrase-structured model, which can be represented using a tree-type architecture. As a result, recent advances in natural language processing have been made utilising recursive neural networks using memory models that allow them to infer tree-type representations of t...
http://arxiv.org/pdf/2302.06451v1
2023-02-08T22:36:25
2023-02-08T22:36:25
{ "title": [ 0.039093018, -0.005844116, -0.03616333, -0.057861328, -0.025650024, -0.03744507, -0.025344849, -0.0033512115, 0.0012617111, 0.058532715, -0.027801514, -0.027816772, 0.028289795, -0.03656006, -0.018493652, 0.003534317, 0.01763916, 0.0...
http://arxiv.org/abs/2302.04359v1
Sentiment analysis and opinion mining on educational data: A survey
Thanveer Shaik, Xiaohui Tao, Christopher Dann, Haoran Xie, Yan Li, Linda Galligan
cs.CL
Sentiment analysis AKA opinion mining is one of the most widely used NLP applications to identify human intentions from their reviews. In the education sector, opinion mining is used to listen to student opinions and enhance their learning-teaching practices pedagogically. With advancements in sentiment annotation tech...
http://arxiv.org/pdf/2302.04359v1
2023-02-08T22:14:08
2023-02-08T22:14:08
{ "title": [ -0.007247925, -0.0048332214, -0.01739502, -0.009544373, 0.012168884, -0.005104065, 0.02609253, -0.015266418, 0.009674072, 0.031982422, -0.053222656, -0.000037670135, -0.0340271, 0.0060043335, -0.07128906, -0.028625488, -0.027191162, ...
http://arxiv.org/abs/2302.04343v1
CRL+: A Novel Semi-Supervised Deep Active Contrastive Representation Learning-Based Text Classification Model for Insurance Data
Amir Namavar Jahromi, Ebrahim Pourjafari, Hadis Karimipour, Amit Satpathy, Lovell Hodge
cs.CL, cs.AI
Financial sector and especially the insurance industry collect vast volumes of text on a daily basis and through multiple channels (their agents, customer care centers, emails, social networks, and web in general). The information collected includes policies, expert and health reports, claims and complaints, results of...
http://arxiv.org/pdf/2302.04343v1
2023-02-08T21:23:52
2023-02-08T21:23:52
{ "title": [ -0.012893677, 0.011489868, -0.03555298, -0.049804688, 0.014091492, -0.009605408, -0.038360596, -0.012756348, -0.010055542, 0.026306152, -0.035583496, 0.045532227, -0.00409317, -0.019851685, -0.0022468567, -0.007621765, -0.020980835, ...
http://arxiv.org/abs/2302.04269v1
Diagnosing and Rectifying Vision Models using Language
Yuhui Zhang, Jeff Z. HaoChen, Shih-Cheng Huang, Kuan-Chieh Wang, James Zou, Serena Yeung
cs.LG, cs.AI, cs.CL, cs.CV
Recent multi-modal contrastive learning models have demonstrated the ability to learn an embedding space suitable for building strong vision classifiers, by leveraging the rich information in large-scale image-caption datasets. Our work highlights a distinct advantage of this multi-modal embedding space: the ability to...
http://arxiv.org/pdf/2302.04269v1
2023-02-08T18:59:42
2023-02-08T18:59:42
{ "title": [ -0.005962372, -0.021575928, -0.018463135, -0.034210205, -0.037628174, -0.011955261, -0.052581787, -0.021484375, 0.02746582, 0.027801514, -0.043426514, -0.045776367, 0.015930176, -0.02394104, -0.045562744, -0.009681702, 0.037872314, 0...
http://arxiv.org/abs/2302.04197v1
Hierarchical Event Grounding
Jiefu Ou, Adithya Pratapa, Rishubh Gupta, Teruko Mitamura
cs.CL
Event grounding aims at linking mention references in text corpora to events from a knowledge base (KB). Previous work on this task focused primarily on linking to a single KB event, thereby overlooking the hierarchical aspects of events. Events in documents are typically described at various levels of spatio-temporal ...
http://arxiv.org/pdf/2302.04197v1
2023-02-08T17:09:41
2023-02-08T17:09:41
{ "title": [ 0.026382446, 0.031143188, -0.08111572, -0.1307373, 0.02861023, -0.014274597, -0.07495117, 0.0026817322, 0.011352539, 0.056274414, 0.0013904572, -0.006454468, 0.06274414, -0.059692383, -0.00020194054, 0.035980225, 0.005443573, 0.01869...
http://arxiv.org/abs/2302.04185v1
Efficient Joint Learning for Clinical Named Entity Recognition and Relation Extraction Using Fourier Networks: A Use Case in Adverse Drug Events
Anthony Yazdani, Dimitrios Proios, Hossein Rouhizadeh, Douglas Teodoro
cs.CL, cs.LG
Current approaches for clinical information extraction are inefficient in terms of computational costs and memory consumption, hindering their application to process large-scale electronic health records (EHRs). We propose an efficient end-to-end model, the Joint-NER-RE-Fourier (JNRF), to jointly learn the tasks of nam...
http://arxiv.org/pdf/2302.04185v1
2023-02-08T16:44:27
2023-02-08T16:44:27
{ "title": [ 0.010124207, 0.0020256042, -0.040649414, -0.095336914, 0.0236969, 0.003704071, -0.059265137, 0.009269714, -0.0082473755, 0.030838013, -0.0049743652, 0.018157959, -0.0025234222, -0.020095825, -0.036071777, 0.0014419556, -0.013908386, ...
http://arxiv.org/abs/2302.04166v2
GPTScore: Evaluate as You Desire
Jinlan Fu, See-Kiong Ng, Zhengbao Jiang, Pengfei Liu
cs.CL
Generative Artificial Intelligence (AI) has enabled the development of sophisticated models that are capable of producing high-caliber text, images, and other outputs through the utilization of large pre-trained models. Nevertheless, assessing the quality of the generation is an even more arduous task than the generati...
http://arxiv.org/pdf/2302.04166v2
2023-02-13T16:19:05
2023-02-08T16:17:29
{ "title": [ 0.03878784, 0.020935059, -0.0087890625, -0.013809204, -0.0013952255, -0.017852783, -0.00724411, -0.03741455, 0.0012426376, 0.011070251, -0.030136108, -0.04309082, -0.036865234, -0.042510986, -0.02734375, -0.06341553, -0.003501892, -0...
http://arxiv.org/abs/2302.04156v1
Prompting for Multimodal Hateful Meme Classification
Rui Cao, Roy Ka-Wei Lee, Wen-Haw Chong, Jing Jiang
cs.CL, cs.IR, cs.MM
Hateful meme classification is a challenging multimodal task that requires complex reasoning and contextual background knowledge. Ideally, we could leverage an explicit external knowledge base to supplement contextual and cultural information in hateful memes. However, there is no known explicit external knowledge base...
http://arxiv.org/pdf/2302.04156v1
2023-02-08T16:04:08
2023-02-08T16:04:08
{ "title": [ 0.04031372, 0.01637268, -0.0231781, -0.03930664, 0.011245728, -0.029785156, -0.0014638901, -0.02243042, 0.022567749, 0.022460938, -0.02255249, 0.022521973, 0.008766174, -0.019561768, -0.039855957, 0.008346558, -0.0069847107, 0.033905...
http://arxiv.org/abs/2302.04116v1
Training-free Lexical Backdoor Attacks on Language Models
Yujin Huang, Terry Yue Zhuo, Qiongkai Xu, Han Hu, Xingliang Yuan, Chunyang Chen
cs.CR, cs.AI, cs.CL
Large-scale language models have achieved tremendous success across various natural language processing (NLP) applications. Nevertheless, language models are vulnerable to backdoor attacks, which inject stealthy triggers into models for steering them to undesirable behaviors. Most existing backdoor attacks, such as dat...
http://arxiv.org/pdf/2302.04116v1
2023-02-08T15:18:51
2023-02-08T15:18:51
{ "title": [ -0.031280518, -0.011833191, -0.023880005, -0.060028076, -0.0057792664, 0.0050849915, -0.04748535, 0.0051460266, -0.025100708, 0.034057617, 0.004447937, 0.014259338, -0.0034389496, -0.015823364, -0.03475952, 0.0070610046, -0.015197754, ...
http://arxiv.org/abs/2302.04084v2
Reception Reader: Exploring Text Reuse in Early Modern British Publications
David Rosson, Eetu Mäkelä, Ville Vaara, Ananth Mahadevan, Yann Ryan, Mikko Tolonen
cs.DL, cs.CL, cs.HC, J.5
The Reception Reader is a web tool for studying text reuse in the Early English Books Online (EEBO-TCP) and Eighteenth Century Collections Online (ECCO) data. Users can: 1) explore a visual overview of the reception of a work, or its incoming connections, across time based on shared text segments, 2) interactively surv...
http://arxiv.org/pdf/2302.04084v2
2023-04-18T13:41:20
2023-02-08T14:37:35
{ "title": [ -0.014083862, 0.012062073, -0.0068359375, -0.015434265, 0.013763428, -0.0018262863, -0.021362305, -0.04083252, -0.0011968613, 0.05960083, -0.01020813, 0.02494812, 0.014198303, -0.0021381378, -0.06677246, 0.01109314, 0.0024414062, 0.0...
http://arxiv.org/abs/2302.04054v7
Towards Inferential Reproducibility of Machine Learning Research
Michael Hagmann, Philipp Meier, Stefan Riezler
cs.LG, cs.AI, cs.CL, stat.AP, stat.ML
Reliability of machine learning evaluation -- the consistency of observed evaluation scores across replicated model training runs -- is affected by several sources of nondeterminism which can be regarded as measurement noise. Current tendencies to remove noise in order to enforce reproducibility of research results neg...
http://arxiv.org/pdf/2302.04054v7
2023-10-09T15:05:36
2023-02-08T13:47:00
{ "title": [ -0.04055786, -0.008361816, -0.040161133, -0.025970459, 0.0038757324, -0.0110321045, -0.027374268, -0.029342651, -0.0045204163, 0.009521484, -0.022201538, 0.007369995, -0.030853271, -0.020233154, -0.0413208, -0.00315094, -0.017822266, ...
http://arxiv.org/abs/2302.04045v1
Revisiting Offline Compression: Going Beyond Factorization-based Methods for Transformer Language Models
Mohammadreza Banaei, Klaudia Bałazy, Artur Kasymov, Rémi Lebret, Jacek Tabor, Karl Aberer
cs.CL
Recent transformer language models achieve outstanding results in many natural language processing (NLP) tasks. However, their enormous size often makes them impractical on memory-constrained devices, requiring practitioners to compress them to smaller networks. In this paper, we explore offline compression methods, me...
http://arxiv.org/pdf/2302.04045v1
2023-02-08T13:36:06
2023-02-08T13:36:06
{ "title": [ -0.048583984, -0.007888794, -0.03149414, -0.06341553, 0.00032377243, 0.023254395, -0.079589844, 0.017364502, -0.00022232533, 0.024734497, 0.020965576, -0.027801514, 0.029296875, -0.041412354, -0.02078247, 0.0049705505, -0.00028920174, ...
http://arxiv.org/abs/2302.06466v1
ChatGPT versus Traditional Question Answering for Knowledge Graphs: Current Status and Future Directions Towards Knowledge Graph Chatbots
Reham Omar, Omij Mangukiya, Panos Kalnis, Essam Mansour
cs.CL, cs.AI, cs.IR
Conversational AI and Question-Answering systems (QASs) for knowledge graphs (KGs) are both emerging research areas: they empower users with natural language interfaces for extracting information easily and effectively. Conversational AI simulates conversations with humans; however, it is limited by the data captured i...
http://arxiv.org/pdf/2302.06466v1
2023-02-08T13:03:27
2023-02-08T13:03:27
{ "title": [ -0.034851074, -0.025817871, -0.03677368, -0.034088135, -0.022445679, 0.008705139, -0.03164673, 0.011398315, -0.0004477501, 0.03353882, -0.026504517, 0.0068855286, 0.0017280579, -0.001036644, -0.04510498, 0.0006713867, -0.009780884, 0...
http://arxiv.org/abs/2302.04023v3
A Multitask, Multilingual, Multimodal Evaluation of ChatGPT on Reasoning, Hallucination, and Interactivity
Yejin Bang, Samuel Cahyawijaya, Nayeon Lee, Wenliang Dai, Dan Su, Bryan Wilie, Holy Lovenia, Ziwei Ji, Tiezheng Yu, Willy Chung, Quyet V. Do, Yan Xu, Pascale Fung
cs.CL, cs.AI
This paper proposes a framework for quantitatively evaluating interactive LLMs such as ChatGPT using publicly available data sets. We carry out an extensive technical evaluation of ChatGPT using 23 data sets covering 8 different common NLP application tasks. We evaluate the multitask, multilingual and multi-modal aspec...
http://arxiv.org/pdf/2302.04023v3
2023-11-09T07:42:55
2023-02-08T12:35:34
{ "title": [ 0.0099487305, -0.0050697327, -0.02406311, -0.023483276, -0.0115737915, -0.007167816, -0.029388428, -0.011871338, -0.015098572, 0.027618408, 0.0070762634, 0.029037476, -0.013114929, -0.025787354, -0.0040283203, 0.009605408, -0.03186035, ...
http://arxiv.org/abs/2302.04012v2
CodeLMSec Benchmark: Systematically Evaluating and Finding Security Vulnerabilities in Black-Box Code Language Models
Hossein Hajipour, Keno Hassler, Thorsten Holz, Lea Schönherr, Mario Fritz
cs.CR, cs.AI, cs.CL, cs.LG, cs.SE
Large language models (LLMs) for automatic code generation have achieved breakthroughs in several programming tasks. Their advances in competition-level programming problems have made them an essential pillar of AI-assisted pair programming, and tools such as GitHub Copilot have emerged as part of the daily programming...
http://arxiv.org/pdf/2302.04012v2
2023-10-23T14:08:09
2023-02-08T11:54:07
{ "title": [ -0.03277588, -0.016296387, -0.03881836, -0.04373169, -0.022277832, -0.021591187, -0.04272461, 0.008506775, -0.021514893, 0.035217285, 0.019302368, 0.021957397, -0.013191223, -0.028244019, -0.029342651, 0.0076942444, -0.022994995, 0.0...
http://arxiv.org/abs/2302.04001v1
Leveraging Summary Guidance on Medical Report Summarization
Yunqi Zhu, Xuebing Yang, Yuanyuan Wu, Wensheng Zhang
cs.CL, cs.AI
This study presents three deidentified large medical text datasets, named DISCHARGE, ECHO and RADIOLOGY, which contain 50K, 16K and 378K pairs of report and summary that are derived from MIMIC-III, respectively. We implement convincing baselines of automated abstractive summarization on the proposed datasets with pre-t...
http://arxiv.org/pdf/2302.04001v1
2023-02-08T11:21:58
2023-02-08T11:21:58
{ "title": [ 0.0020923615, 0.00087594986, -0.03744507, -0.10205078, -0.015716553, 0.005004883, -0.06573486, 0.012298584, -0.025131226, 0.047943115, -0.020080566, 0.028839111, 0.0051994324, -0.0041542053, -0.03274536, 0.012687683, 0.01096344, 0.02...
http://arxiv.org/abs/2302.06476v2
Is ChatGPT a General-Purpose Natural Language Processing Task Solver?
Chengwei Qin, Aston Zhang, Zhuosheng Zhang, Jiaao Chen, Michihiro Yasunaga, Diyi Yang
cs.CL, cs.AI
Spurred by advancements in scale, large language models (LLMs) have demonstrated the ability to perform a variety of natural language processing (NLP) tasks zero-shot -- i.e., without adaptation on downstream data. Recently, the debut of ChatGPT has drawn a great deal of attention from the natural language processing (...
http://arxiv.org/pdf/2302.06476v2
2023-02-15T17:46:20
2023-02-08T09:44:51
{ "title": [ -0.0038394928, -0.013587952, -0.02281189, -0.013298035, -0.043945312, 0.005180359, -0.039367676, -0.0140686035, -0.01586914, 0.025970459, -0.00157547, 0.005645752, 0.0036468506, -0.022338867, 0.013305664, 0.0065956116, -0.014419556, ...
http://arxiv.org/abs/2302.03950v1
Improving (Dis)agreement Detection with Inductive Social Relation Information From Comment-Reply Interactions
Yun Luo, Zihan Liu, Stan Z. Li, Yue Zhang
cs.CL, cs.AI
(Dis)agreement detection aims to identify the authors' attitudes or positions (\textit{{agree, disagree, neutral}}) towards a specific text. It is limited for existing methods merely using textual information for identifying (dis)agreements, especially for cross-domain settings. Social relation information can play an ...
http://arxiv.org/pdf/2302.03950v1
2023-02-08T09:09:47
2023-02-08T09:09:47
{ "title": [ -0.0049057007, -0.0009598732, -0.052124023, -0.049591064, 0.021865845, -0.027542114, -0.019500732, -0.0049743652, 0.0105896, 0.008552551, -0.05987549, 0.009101868, -0.016342163, 0.007797241, -0.045806885, -0.032958984, -0.018737793, ...