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/2307.11785v1
Adversarial Conversational Shaping for Intelligent Agents
Piotr Tarasiewicz, Sultan Kenjeyev, Ilana Sebag, Shehab Alshehabi
cs.CL, cs.LG
The recent emergence of deep learning methods has enabled the research community to achieve state-of-the art results in several domains including natural language processing. However, the current robocall system remains unstable and inaccurate: text generator and chat-bots can be tedious and misunderstand human-like di...
http://arxiv.org/pdf/2307.11785v1
2023-07-20T12:44:47
2023-07-20T12:44:47
{ "title": [ -0.013473511, 0.0003118515, -0.016647339, -0.038238525, 0.0052261353, 0.00705719, -0.019317627, -0.016799927, -0.016586304, 0.030197144, -0.00919342, 0.013427734, -0.014007568, -0.017959595, -0.03564453, -0.015853882, -0.020004272, 0...
http://arxiv.org/abs/2307.10826v1
Yelp Reviews and Food Types: A Comparative Analysis of Ratings, Sentiments, and Topics
Wenyu Liao, Yiqing Shi, Yujia Hu, Wei Quan
cs.CL
This study examines the relationship between Yelp reviews and food types, investigating how ratings, sentiments, and topics vary across different types of food. Specifically, we analyze how ratings and sentiments of reviews vary across food types, cluster food types based on ratings and sentiments, infer review topics ...
http://arxiv.org/pdf/2307.10826v1
2023-07-20T12:41:35
2023-07-20T12:41:35
{ "title": [ 0.012062073, 0.021377563, -0.029052734, 0.0024585724, 0.021911621, -0.041290283, 0.016403198, -0.054779053, -0.032409668, 0.03137207, -0.02758789, 0.026443481, -0.013404846, -0.023712158, 0.0019817352, -0.0045928955, 0.009422302, 0.0...
http://arxiv.org/abs/2307.10814v1
Cross-Corpus Multilingual Speech Emotion Recognition: Amharic vs. Other Languages
Ephrem Afele Retta, Richard Sutcliffe, Jabar Mahmood, Michael Abebe Berwo, Eiad Almekhlafi, Sajjad Ahmed Khan, Shehzad Ashraf Chaudhry, Mustafa Mhamed, Jun Feng
cs.CL, cs.NE, cs.SD, eess.AS
In a conventional Speech emotion recognition (SER) task, a classifier for a given language is trained on a pre-existing dataset for that same language. However, where training data for a language does not exist, data from other languages can be used instead. We experiment with cross-lingual and multilingual SER, workin...
http://arxiv.org/pdf/2307.10814v1
2023-07-20T12:24:23
2023-07-20T12:24:23
{ "title": [ -0.012237549, -0.0018377304, -0.027679443, -0.04019165, -0.016494751, -0.008598328, -0.06463623, 0.008255005, 0.00051546097, 0.046875, -0.017288208, -0.0075683594, 0.052459717, -0.033447266, -0.025253296, 0.011367798, -0.010437012, 0...
http://arxiv.org/abs/2307.10802v1
Meta-Transformer: A Unified Framework for Multimodal Learning
Yiyuan Zhang, Kaixiong Gong, Kaipeng Zhang, Hongsheng Li, Yu Qiao, Wanli Ouyang, Xiangyu Yue
cs.CV, cs.AI, cs.CL, cs.LG, cs.MM
Multimodal learning aims to build models that can process and relate information from multiple modalities. Despite years of development in this field, it still remains challenging to design a unified network for processing various modalities ($\textit{e.g.}$ natural language, 2D images, 3D point clouds, audio, video, t...
http://arxiv.org/pdf/2307.10802v1
2023-07-20T12:10:29
2023-07-20T12:10:29
{ "title": [ -0.019744873, 0.005580902, -0.047210693, -0.045074463, 0.010261536, -0.010871887, -0.040039062, -0.03213501, -0.0036392212, 0.03463745, -0.0060195923, 0.009819031, -0.0075950623, -0.021514893, -0.028701782, -0.0010566711, -0.0059509277, ...
http://arxiv.org/abs/2307.10799v1
Layer-wise Representation Fusion for Compositional Generalization
Yafang Zheng, Lei Lin, Zhaohong Lai, Binling Wang, Shan Liu, Biao Fu, Wenhao Rao, Peigen Ye, Yidong Chen, Xiaodong Shi
cs.CL
Despite successes across a broad range of applications, sequence-to-sequence models' construct of solutions are argued to be less compositional than human-like generalization. There is mounting evidence that one of the reasons hindering compositional generalization is representations of the encoder and decoder uppermos...
http://arxiv.org/pdf/2307.10799v1
2023-07-20T12:01:40
2023-07-20T12:01:40
{ "title": [ 0.004219055, 0.005428314, -0.029251099, -0.08294678, 0.017913818, 0.0072517395, -0.057769775, 0.0029850006, 0.016342163, 0.03942871, -0.015380859, 0.0071144104, -0.009162903, -0.01499939, -0.039031982, 0.0005173683, -0.01625061, 0.06...
http://arxiv.org/abs/2307.10778v1
Extreme Multi-Label Skill Extraction Training using Large Language Models
Jens-Joris Decorte, Severine Verlinden, Jeroen Van Hautte, Johannes Deleu, Chris Develder, Thomas Demeester
cs.CL
Online job ads serve as a valuable source of information for skill requirements, playing a crucial role in labor market analysis and e-recruitment processes. Since such ads are typically formatted in free text, natural language processing (NLP) technologies are required to automatically process them. We specifically fo...
http://arxiv.org/pdf/2307.10778v1
2023-07-20T11:29:15
2023-07-20T11:29:15
{ "title": [ -0.017120361, 0.004852295, -0.036987305, -0.035491943, -0.0038204193, -0.012702942, -0.016296387, -0.010253906, -0.010726929, 0.015197754, 0.011024475, 0.0038890839, -0.0047416687, -0.039794922, -0.011016846, 0.012336731, -0.02355957, ...
http://arxiv.org/abs/2307.10757v1
Vesper: A Compact and Effective Pretrained Model for Speech Emotion Recognition
Weidong Chen, Xiaofen Xing, Peihao Chen, Xiangmin Xu
cs.SD, cs.CL, eess.AS
This paper presents a paradigm that adapts general large-scale pretrained models (PTMs) to speech emotion recognition task. Although PTMs shed new light on artificial general intelligence, they are constructed with general tasks in mind, and thus, their efficacy for specific tasks can be further improved. Additionally,...
http://arxiv.org/pdf/2307.10757v1
2023-07-20T10:42:16
2023-07-20T10:42:16
{ "title": [ 0.0022277832, -0.005176544, -0.04208374, -0.06994629, 0.009262085, -0.010124207, -0.038513184, 0.010116577, 0.024246216, 0.034729004, -0.007583618, -0.016860962, 0.023376465, -0.028030396, -0.057922363, 0.0082473755, -0.006046295, 0....
http://arxiv.org/abs/2307.10751v1
Exploring Perspectives on the Impact of Artificial Intelligence on the Creativity of Knowledge Work: Beyond Mechanised Plagiarism and Stochastic Parrots
Advait Sarkar
cs.HC, cs.AI, cs.CL
Artificial Intelligence (AI), and in particular generative models, are transformative tools for knowledge work. They problematise notions of creativity, originality, plagiarism, the attribution of credit, and copyright ownership. Critics of generative models emphasise the reliance on large amounts of training data, and...
http://arxiv.org/pdf/2307.10751v1
2023-07-20T10:26:57
2023-07-20T10:26:57
{ "title": [ -0.04373169, -0.013832092, -0.056488037, 0.007648468, -0.008506775, -0.038360596, 0.019943237, -0.007896423, -0.006629944, 0.028869629, -0.02557373, 0.002002716, -0.0061149597, -0.010391235, -0.050750732, -0.022827148, -0.012489319, ...
http://arxiv.org/abs/2307.10719v1
LLM Censorship: A Machine Learning Challenge or a Computer Security Problem?
David Glukhov, Ilia Shumailov, Yarin Gal, Nicolas Papernot, Vardan Papyan
cs.AI, cs.CL, cs.CR, cs.LG
Large language models (LLMs) have exhibited impressive capabilities in comprehending complex instructions. However, their blind adherence to provided instructions has led to concerns regarding risks of malicious use. Existing defence mechanisms, such as model fine-tuning or output censorship using LLMs, have proven to ...
http://arxiv.org/pdf/2307.10719v1
2023-07-20T09:25:02
2023-07-20T09:25:02
{ "title": [ -0.03692627, 0.015182495, -0.026428223, 0.024398804, -0.011543274, -0.02015686, 0.00013136864, -0.0368042, -0.02809143, 0.013221741, 0.0008459091, -0.004764557, -0.014663696, -0.05697632, -0.015342712, -0.018478394, -0.0368042, 0.056...
http://arxiv.org/abs/2307.10700v2
Topics, Authors, and Networks in Large Language Model Research: Trends from a Survey of 17K arXiv Papers
Rajiv Movva, Sidhika Balachandar, Kenny Peng, Gabriel Agostini, Nikhil Garg, Emma Pierson
cs.DL, cs.CL, cs.CY
Large language model (LLM) research is dramatically impacting society, making it essential to understand the topics and values it prioritizes, the authors and institutions driving it, and its networks of collaboration. Due to the recent growth of the field, many of these fundamental attributes lack systematic descripti...
http://arxiv.org/pdf/2307.10700v2
2023-10-23T17:04:35
2023-07-20T08:45:00
{ "title": [ -0.05227661, -0.00094366074, -0.0635376, -0.043151855, -0.013450623, -0.007865906, -0.062347412, -0.00687027, -0.012863159, 0.017715454, 0.0060653687, 0.029632568, -0.0158844, -0.018951416, -0.02420044, 0.022705078, -0.01687622, 0.02...
http://arxiv.org/abs/2307.10666v1
A Dataset and Strong Baselines for Classification of Czech News Texts
Hynek Kydlíček, Jindřich Libovický
cs.CL
Pre-trained models for Czech Natural Language Processing are often evaluated on purely linguistic tasks (POS tagging, parsing, NER) and relatively simple classification tasks such as sentiment classification or article classification from a single news source. As an alternative, we present CZEch~NEws~Classification~dat...
http://arxiv.org/pdf/2307.10666v1
2023-07-20T07:47:08
2023-07-20T07:47:08
{ "title": [ -0.014831543, 0.0036411285, -0.04776001, -0.040527344, 0.012931824, -0.010108948, -0.025665283, -0.0034637451, 0.00014746189, 0.051361084, -0.046203613, 0.017990112, 0.017196655, 0.003835678, -0.054260254, 0.002090454, -0.008926392, ...
http://arxiv.org/abs/2307.10652v5
Exploring the Landscape of Natural Language Processing Research
Tim Schopf, Karim Arabi, Florian Matthes
cs.CL, I.2.7
As an efficient approach to understand, generate, and process natural language texts, research in natural language processing (NLP) has exhibited a rapid spread and wide adoption in recent years. Given the increasing research work in this area, several NLP-related approaches have been surveyed in the research community...
http://arxiv.org/pdf/2307.10652v5
2023-09-24T20:18:07
2023-07-20T07:33:30
{ "title": [ -0.025787354, -0.0073661804, -0.04296875, -0.041534424, 0.008178711, -0.012748718, -0.017181396, -0.00031280518, -0.011123657, 0.025390625, -0.034820557, 0.004436493, -0.014732361, -0.0060577393, -0.048217773, -0.006126404, -0.016189575,...
http://arxiv.org/abs/2307.10635v1
SciBench: Evaluating College-Level Scientific Problem-Solving Abilities of Large Language Models
Xiaoxuan Wang, Ziniu Hu, Pan Lu, Yanqiao Zhu, Jieyu Zhang, Satyen Subramaniam, Arjun R. Loomba, Shichang Zhang, Yizhou Sun, Wei Wang
cs.CL, cs.AI, cs.LG
Recent advances in large language models (LLMs) have demonstrated notable progress on many mathematical benchmarks. However, most of these benchmarks only feature problems grounded in junior and senior high school subjects, contain only multiple-choice questions, and are confined to a limited scope of elementary arithm...
http://arxiv.org/pdf/2307.10635v1
2023-07-20T07:01:57
2023-07-20T07:01:57
{ "title": [ -0.03527832, -0.019378662, -0.027236938, -0.043395996, -0.0014753342, -0.0012159348, -0.037994385, -0.011940002, 0.010269165, 0.015510559, -0.0067596436, -0.008453369, -0.008918762, -0.029342651, -0.046173096, -0.0027713776, -0.024505615...
http://arxiv.org/abs/2307.10634v1
Generative Language Models on Nucleotide Sequences of Human Genes
Musa Nuri Ihtiyar, Arzucan Ozgur
q-bio.GN, cs.CL, cs.LG
Language models, primarily transformer-based ones, obtained colossal success in NLP. To be more precise, studies like BERT in NLU and works such as GPT-3 for NLG are very crucial. DNA sequences are very close to natural language in terms of structure, so if the DNA-related bioinformatics domain is concerned, discrimina...
http://arxiv.org/pdf/2307.10634v1
2023-07-20T06:59:02
2023-07-20T06:59:02
{ "title": [ -0.035064697, -0.017456055, -0.049560547, -0.07989502, -0.0024318695, 0.013168335, -0.07409668, 0.01889038, 0.018844604, 0.03491211, -0.005329132, -0.0069770813, 0.012954712, -0.027175903, -0.043945312, 0.019805908, -0.009315491, 0.0...
http://arxiv.org/abs/2307.10633v1
Multi-Method Self-Training: Improving Code Generation With Text, And Vice Versa
Shriyash K. Upadhyay, Etan J. Ginsberg
cs.CL, cs.LG
Large Language Models have many methods for solving the same problem. This introduces novel strengths (different methods may work well for different problems) and weaknesses (it may be difficult for users to know which method to use). In this paper, we introduce Multi-Method Self-Training (MMST), where one method is tr...
http://arxiv.org/pdf/2307.10633v1
2023-07-20T06:58:55
2023-07-20T06:58:55
{ "title": [ -0.022399902, -0.010391235, -0.03527832, -0.053955078, -0.004020691, -0.007621765, -0.04663086, 0.007865906, -0.012374878, 0.030761719, -0.009086609, 0.03353882, -0.026535034, -0.018371582, -0.015960693, 0.0073051453, -0.0032348633, ...
http://arxiv.org/abs/2307.10587v1
A Deep Dive into the Disparity of Word Error Rates Across Thousands of NPTEL MOOC Videos
Anand Kumar Rai, Siddharth D Jaiswal, Animesh Mukherjee
cs.CL, cs.HC
Automatic speech recognition (ASR) systems are designed to transcribe spoken language into written text and find utility in a variety of applications including voice assistants and transcription services. However, it has been observed that state-of-the-art ASR systems which deliver impressive benchmark results, struggl...
http://arxiv.org/pdf/2307.10587v1
2023-07-20T05:03:00
2023-07-20T05:03:00
{ "title": [ 0.017089844, 0.00844574, -0.034484863, 0.0009279251, -0.016784668, -0.025985718, 0.019592285, -0.019088745, -0.008071899, 0.010719299, -0.088256836, 0.03265381, 0.0056533813, -0.004055023, 0.029022217, -0.061401367, -0.01197052, 0.02...
http://arxiv.org/abs/2307.10558v1
Instruction-following Evaluation through Verbalizer Manipulation
Shiyang Li, Jun Yan, Hai Wang, Zheng Tang, Xiang Ren, Vijay Srinivasan, Hongxia Jin
cs.CL
While instruction-tuned models have shown remarkable success in various natural language processing tasks, accurately evaluating their ability to follow instructions remains challenging. Existing benchmarks primarily focus on common instructions that align well with what the model learned during training. However, prof...
http://arxiv.org/pdf/2307.10558v1
2023-07-20T03:54:24
2023-07-20T03:54:24
{ "title": [ 0.00774765, 0.027572632, -0.01878357, -0.05770874, -0.00017476082, -0.015716553, -0.027832031, -0.04498291, -0.01939392, 0.034576416, -0.021774292, 0.0066108704, -0.023788452, -0.023269653, -0.026321411, -0.023880005, -0.023452759, 0...
http://arxiv.org/abs/2307.10549v1
Dynamic Large Language Models on Blockchains
Yuanhao Gong
cs.CV, cs.AI, cs.CL, econ.GN, q-fin.EC
Training and deploying the large language models requires a large mount of computational resource because the language models contain billions of parameters and the text has thousands of tokens. Another problem is that the large language models are static. They are fixed after the training process. To tackle these issu...
http://arxiv.org/pdf/2307.10549v1
2023-07-20T03:26:57
2023-07-20T03:26:57
{ "title": [ -0.043121338, -0.0067329407, -0.06604004, -0.08721924, 0.012268066, 0.011177063, -0.055633545, 0.03768921, -0.0104599, 0.040100098, -0.016525269, 0.00085639954, 0.00017392635, 0.004245758, -0.04699707, 0.00027680397, -0.0060691833, 0...
http://arxiv.org/abs/2307.11779v1
The Extractive-Abstractive Axis: Measuring Content "Borrowing" in Generative Language Models
Nedelina Teneva
cs.CL, cs.AI, cs.CY, cs.LG
Generative language models produce highly abstractive outputs by design, in contrast to extractive responses in search engines. Given this characteristic of LLMs and the resulting implications for content Licensing & Attribution, we propose the the so-called Extractive-Abstractive axis for benchmarking generative model...
http://arxiv.org/pdf/2307.11779v1
2023-07-20T02:12:00
2023-07-20T02:12:00
{ "title": [ -0.028640747, -0.012138367, -0.029464722, -0.03100586, -0.014595032, -0.0077285767, -0.033813477, -0.02142334, 0.008934021, 0.027130127, -0.021408081, 0.030197144, -0.00642395, -0.020843506, -0.035980225, 0.012428284, -0.02394104, 0....
http://arxiv.org/abs/2307.10522v1
Gender-tuning: Empowering Fine-tuning for Debiasing Pre-trained Language Models
Somayeh Ghanbarzadeh, Yan Huang, Hamid Palangi, Radames Cruz Moreno, Hamed Khanpour
cs.CL
Recent studies have revealed that the widely-used Pre-trained Language Models (PLMs) propagate societal biases from the large unmoderated pre-training corpora. Existing solutions require debiasing training processes and datasets for debiasing, which are resource-intensive and costly. Furthermore, these methods hurt the...
http://arxiv.org/pdf/2307.10522v1
2023-07-20T01:48:51
2023-07-20T01:48:51
{ "title": [ -0.035247803, -0.0060272217, -0.046783447, -0.07849121, -0.006099701, -0.0033740997, -0.06768799, 0.0032634735, -0.0069389343, 0.035125732, -0.010925293, -0.0044670105, 0.024429321, -0.014785767, -0.03817749, 0.0024871826, -0.023254395, ...
http://arxiv.org/abs/2307.10514v1
Building Socio-culturally Inclusive Stereotype Resources with Community Engagement
Sunipa Dev, Jaya Goyal, Dinesh Tewari, Shachi Dave, Vinodkumar Prabhakaran
cs.CL, cs.AI, cs.HC
With rapid development and deployment of generative language models in global settings, there is an urgent need to also scale our measurements of harm, not just in the number and types of harms covered, but also how well they account for local cultural contexts, including marginalized identities and the social biases e...
http://arxiv.org/pdf/2307.10514v1
2023-07-20T01:26:34
2023-07-20T01:26:34
{ "title": [ 0.042541504, 0.017593384, -0.010650635, -0.06939697, 0.010383606, -0.012123108, -0.018615723, -0.017120361, -0.023834229, 0.050872803, -0.031799316, 0.008613586, 0.010482788, -0.021224976, -0.010124207, -0.006893158, 0.0033435822, 0....
http://arxiv.org/abs/2307.10512v1
IvyGPT: InteractiVe Chinese pathwaY language model in medical domain
Rongsheng Wang, Yaofei Duan, ChanTong Lam, Jiexi Chen, Jiangsheng Xu, Haoming Chen, Xiaohong Liu, Patrick Cheong-Iao Pang, Tao Tan
cs.CL, cs.AI
General large language models (LLMs) such as ChatGPT have shown remarkable success. However, such LLMs have not been widely adopted for medical purposes, due to poor accuracy and inability to provide medical advice. We propose IvyGPT, an LLM based on LLaMA that is trained and fine-tuned with high-quality medical questi...
http://arxiv.org/pdf/2307.10512v1
2023-07-20T01:11:14
2023-07-20T01:11:14
{ "title": [ -0.03729248, -0.025009155, -0.036590576, -0.046051025, -0.012245178, 0.0051994324, -0.043304443, -0.0034370422, -0.00749588, 0.024795532, 0.018249512, -0.010925293, -0.0061912537, -0.0541687, -0.03164673, 0.008842468, -0.0065727234, ...
http://arxiv.org/abs/2307.11778v1
Transsion TSUP's speech recognition system for ASRU 2023 MADASR Challenge
Xiaoxiao Li, Gaosheng Zhang, An Zhu, Weiyong Li, Shuming Fang, Xiaoyue Yang, Jianchao Zhu
cs.CL, cs.SD, eess.AS
This paper presents a speech recognition system developed by the Transsion Speech Understanding Processing Team (TSUP) for the ASRU 2023 MADASR Challenge. The system focuses on adapting ASR models for low-resource Indian languages and covers all four tracks of the challenge. For tracks 1 and 2, the acoustic model utili...
http://arxiv.org/pdf/2307.11778v1
2023-07-20T00:55:01
2023-07-20T00:55:01
{ "title": [ 0.002752304, -0.004146576, -0.038391113, -0.048614502, -0.010498047, -0.0040245056, -0.032958984, 0.014831543, -0.022384644, 0.0680542, -0.010627747, -0.0036468506, 0.04333496, -0.04031372, 0.001247406, 0.0035953522, 0.028167725, 0.0...
http://arxiv.org/abs/2307.10511v2
General Debiasing for Multimodal Sentiment Analysis
Teng Sun, Juntong Ni, Wenjie Wang, Liqiang Jing, Yinwei Wei, Liqiang Nie
cs.CL
Existing work on Multimodal Sentiment Analysis (MSA) utilizes multimodal information for prediction yet unavoidably suffers from fitting the spurious correlations between multimodal features and sentiment labels. For example, if most videos with a blue background have positive labels in a dataset, the model will rely o...
http://arxiv.org/pdf/2307.10511v2
2023-08-07T09:08:23
2023-07-20T00:36:41
{ "title": [ -0.009643555, 0.023254395, -0.039001465, -0.06173706, 0.024414062, -0.003129959, -0.044677734, -0.034240723, 0.015411377, 0.04220581, -0.020828247, -0.01121521, 0.010848999, -0.017684937, -0.05682373, 0.00091934204, -0.018493652, 0.0...
http://arxiv.org/abs/2307.10490v4
Abusing Images and Sounds for Indirect Instruction Injection in Multi-Modal LLMs
Eugene Bagdasaryan, Tsung-Yin Hsieh, Ben Nassi, Vitaly Shmatikov
cs.CR, cs.AI, cs.CL, cs.LG
We demonstrate how images and sounds can be used for indirect prompt and instruction injection in multi-modal LLMs. An attacker generates an adversarial perturbation corresponding to the prompt and blends it into an image or audio recording. When the user asks the (unmodified, benign) model about the perturbed image or...
http://arxiv.org/pdf/2307.10490v4
2023-10-03T17:03:10
2023-07-19T23:03:20
{ "title": [ 0.0041999817, -0.0057411194, -0.0009160042, -0.005935669, -0.03152466, -0.008552551, -0.024917603, -0.04397583, -0.004676819, 0.020812988, -0.059173584, 0.034820557, 0.007205963, -0.023468018, 0.007949829, -0.022094727, -0.008956909, ...
http://arxiv.org/abs/2307.10488v1
SPRINT: A Unified Toolkit for Evaluating and Demystifying Zero-shot Neural Sparse Retrieval
Nandan Thakur, Kexin Wang, Iryna Gurevych, Jimmy Lin
cs.IR, cs.CL, cs.LG
Traditionally, sparse retrieval systems relied on lexical representations to retrieve documents, such as BM25, dominated information retrieval tasks. With the onset of pre-trained transformer models such as BERT, neural sparse retrieval has led to a new paradigm within retrieval. Despite the success, there has been lim...
http://arxiv.org/pdf/2307.10488v1
2023-07-19T22:48:02
2023-07-19T22:48:02
{ "title": [ -0.014091492, -0.007499695, -0.045715332, -0.040496826, 0.0050201416, -0.0137786865, -0.012680054, 0.0044555664, 0.0309906, 0.036956787, -0.0027446747, 0.0022220612, -0.008026123, -0.011978149, -0.059295654, 0.021255493, 0.014053345, ...
http://arxiv.org/abs/2307.10485v2
FinGPT: Democratizing Internet-scale Data for Financial Large Language Models
Xiao-Yang Liu, Guoxuan Wang, Hongyang Yang, Daochen Zha
cs.CL, cs.LG, q-fin.GN
Large language models (LLMs) have demonstrated remarkable proficiency in understanding and generating human-like texts, which may potentially revolutionize the finance industry. However, existing LLMs often fall short in the financial field, which is mainly attributed to the disparities between general text data and fi...
http://arxiv.org/pdf/2307.10485v2
2023-11-14T16:34:00
2023-07-19T22:43:57
{ "title": [ -0.03378296, -0.015136719, -0.072631836, -0.06286621, 0.005256653, -0.01007843, -0.051757812, 0.010169983, 0.009857178, 0.025375366, -0.015426636, -0.008682251, 0.008323669, -0.013717651, -0.043121338, 0.006729126, -0.014808655, 0.03...
http://arxiv.org/abs/2307.10476v1
What can we learn from Data Leakage and Unlearning for Law?
Jaydeep Borkar
cs.CR, cs.CL
Large Language Models (LLMs) have a privacy concern because they memorize training data (including personally identifiable information (PII) like emails and phone numbers) and leak it during inference. A company can train an LLM on its domain-customized data which can potentially also include their users' PII. In order...
http://arxiv.org/pdf/2307.10476v1
2023-07-19T22:14:58
2023-07-19T22:14:58
{ "title": [ -0.008430481, -0.005680084, -0.021118164, -0.0045433044, -0.0209198, -0.016921997, -0.0049552917, -0.0002539158, -0.0052223206, 0.030410767, -0.037017822, 0.024398804, -0.032562256, -0.015205383, -0.022140503, -0.011238098, 0.013557434, ...
http://arxiv.org/abs/2307.10475v2
Findings of Factify 2: Multimodal Fake News Detection
S Suryavardan, Shreyash Mishra, Megha Chakraborty, Parth Patwa, Anku Rani, Aman Chadha, Aishwarya Reganti, Amitava Das, Amit Sheth, Manoj Chinnakotla, Asif Ekbal, Srijan Kumar
cs.CL, cs.CV
With social media usage growing exponentially in the past few years, fake news has also become extremely prevalent. The detrimental impact of fake news emphasizes the need for research focused on automating the detection of false information and verifying its accuracy. In this work, we present the outcome of the Factif...
http://arxiv.org/pdf/2307.10475v2
2023-09-12T18:51:05
2023-07-19T22:14:49
{ "title": [ -0.015686035, 0.014526367, -0.04547119, -0.037628174, 0.012252808, -0.0079193115, -0.023864746, -0.032562256, -0.04321289, 0.02949524, 0.0012702942, 0.021713257, -0.00466156, -0.02508545, -0.004283905, -0.0062561035, -0.03427124, 0.0...
http://arxiv.org/abs/2307.10472v1
Can Instruction Fine-Tuned Language Models Identify Social Bias through Prompting?
Omkar Dige, Jacob-Junqi Tian, David Emerson, Faiza Khan Khattak
cs.CL, cs.AI, cs.CY, cs.LG
As the breadth and depth of language model applications continue to expand rapidly, it is increasingly important to build efficient frameworks for measuring and mitigating the learned or inherited social biases of these models. In this paper, we present our work on evaluating instruction fine-tuned language models' abi...
http://arxiv.org/pdf/2307.10472v1
2023-07-19T22:03:40
2023-07-19T22:03:40
{ "title": [ 0.017730713, -0.0023231506, 0.0075645447, -0.016662598, -0.03933716, -0.011711121, -0.033691406, -0.034423828, -0.024749756, 0.011459351, -0.037872314, -0.00030612946, 0.0011167526, -0.017211914, 0.010551453, -0.036499023, -0.019256592, ...
http://arxiv.org/abs/2307.10457v3
Improving the Reusability of Pre-trained Language Models in Real-world Applications
Somayeh Ghanbarzadeh, Hamid Palangi, Yan Huang, Radames Cruz Moreno, Hamed Khanpour
cs.CL
The reusability of state-of-the-art Pre-trained Language Models (PLMs) is often limited by their generalization problem, where their performance drastically decreases when evaluated on examples that differ from the training dataset, known as Out-of-Distribution (OOD)/unseen examples. This limitation arises from PLMs' r...
http://arxiv.org/pdf/2307.10457v3
2023-08-08T04:18:34
2023-07-19T21:00:16
{ "title": [ -0.008796692, -0.016235352, -0.034484863, -0.08465576, 0.007980347, 0.0023403168, -0.06994629, 0.0017175674, 0.004802704, 0.044647217, -0.0040359497, -0.008972168, 0.011299133, -0.01423645, -0.047210693, 0.020355225, -0.004562378, 0....
http://arxiv.org/abs/2307.10443v2
Integrating a Heterogeneous Graph with Entity-aware Self-attention using Relative Position Labels for Reading Comprehension Model
Shima Foolad, Kourosh Kiani
cs.CL, cs.LG
Despite the significant progress made by transformer models in machine reading comprehension tasks, they still fall short in handling complex reasoning tasks due to the absence of explicit knowledge in the input sequence. To address this limitation, many recent works have proposed injecting external knowledge into the ...
http://arxiv.org/pdf/2307.10443v2
2023-07-22T08:36:18
2023-07-19T20:17:37
{ "title": [ -0.0014476776, 0.016342163, -0.03353882, -0.02973938, -0.005908966, -0.014984131, -0.026611328, -0.020629883, 0.0070610046, 0.052246094, -0.014259338, -0.00058460236, 0.031951904, -0.01020813, -0.041381836, 0.0037784576, -0.012619019, ...
http://arxiv.org/abs/2307.10442v1
Thrust: Adaptively Propels Large Language Models with External Knowledge
Xinran Zhao, Hongming Zhang, Xiaoman Pan, Wenlin Yao, Dong Yu, Jianshu Chen
cs.CL
Although large-scale pre-trained language models (PTLMs) are shown to encode rich knowledge in their model parameters, the inherent knowledge in PTLMs can be opaque or static, making external knowledge necessary. However, the existing information retrieval techniques could be costly and may even introduce noisy and som...
http://arxiv.org/pdf/2307.10442v1
2023-07-19T20:16:46
2023-07-19T20:16:46
{ "title": [ -0.019241333, -0.021408081, -0.060699463, -0.049591064, -0.03704834, -0.020996094, -0.04434204, 0.020950317, -0.034973145, 0.008575439, -0.031036377, -0.002866745, 0.010093689, -0.02923584, 0.021057129, -0.016204834, 0.014022827, 0.0...
http://arxiv.org/abs/2307.10432v2
PharmacyGPT: The AI Pharmacist
Zhengliang Liu, Zihao Wu, Mengxuan Hu, Bokai Zhao, Lin Zhao, Tianyi Zhang, Haixing Dai, Xianyan Chen, Ye Shen, Sheng Li, Brian Murray, Tianming Liu, Andrea Sikora
cs.CL
In this study, we introduce PharmacyGPT, a novel framework to assess the capabilities of large language models (LLMs) such as ChatGPT and GPT-4 in emulating the role of clinical pharmacists. Our methodology encompasses the utilization of LLMs to generate comprehensible patient clusters, formulate medication plans, and ...
http://arxiv.org/pdf/2307.10432v2
2023-07-21T02:22:14
2023-07-19T19:40:34
{ "title": [ -0.008735657, -0.0178833, -0.042785645, -0.0061569214, -0.049224854, -0.020339966, -0.034240723, -0.009056091, -0.024429321, 0.0075302124, -0.048950195, 0.018341064, -0.01651001, -0.038330078, 0.024124146, -0.018981934, -0.008270264, ...
http://arxiv.org/abs/2307.10172v2
DialogStudio: Towards Richest and Most Diverse Unified Dataset Collection for Conversational AI
Jianguo Zhang, Kun Qian, Zhiwei Liu, Shelby Heinecke, Rui Meng, Ye Liu, Zhou Yu, Huan Wang, Silvio Savarese, Caiming Xiong
cs.CL, cs.AI
Despite advancements in conversational AI, language models encounter challenges to handle diverse conversational tasks, and existing dialogue dataset collections often lack diversity and comprehensiveness. To tackle these issues, we introduce DialogStudio: the largest and most diverse collection of dialogue datasets, u...
http://arxiv.org/pdf/2307.10172v2
2023-07-20T17:59:35
2023-07-19T17:57:53
{ "title": [ -0.016906738, -0.010047913, -0.059570312, -0.062927246, 0.013130188, -0.014755249, -0.044006348, 0.0032444, 0.016708374, 0.0395813, -0.022277832, 0.014831543, 0.00053310394, -0.0046157837, -0.056365967, 0.0104522705, -0.01399231, 0.0...
http://arxiv.org/abs/2307.10169v1
Challenges and Applications of Large Language Models
Jean Kaddour, Joshua Harris, Maximilian Mozes, Herbie Bradley, Roberta Raileanu, Robert McHardy
cs.CL, cs.AI, cs.LG
Large Language Models (LLMs) went from non-existent to ubiquitous in the machine learning discourse within a few years. Due to the fast pace of the field, it is difficult to identify the remaining challenges and already fruitful application areas. In this paper, we aim to establish a systematic set of open problems and...
http://arxiv.org/pdf/2307.10169v1
2023-07-19T17:55:13
2023-07-19T17:55:13
{ "title": [ -0.02947998, -0.012451172, -0.039245605, -0.076538086, 0.005519867, 0.01071167, -0.06781006, 0.0077285767, 0.010307312, 0.033050537, -0.0073928833, -0.015220642, 0.006580353, -0.024536133, -0.042175293, 0.024658203, -0.009742737, 0.0...
http://arxiv.org/abs/2307.10168v2
LLMs as Workers in Human-Computational Algorithms? Replicating Crowdsourcing Pipelines with LLMs
Tongshuang Wu, Haiyi Zhu, Maya Albayrak, Alexis Axon, Amanda Bertsch, Wenxing Deng, Ziqi Ding, Bill Guo, Sireesh Gururaja, Tzu-Sheng Kuo, Jenny T. Liang, Ryan Liu, Ihita Mandal, Jeremiah Milbauer, Xiaolin Ni, Namrata Padmanabhan, Subhashini Ramkumar, Alexis Sudjianto, Jordan Taylor, Ying-Jui Tseng, Patricia Vaidos, Zhi...
cs.CL, cs.HC
LLMs have shown promise in replicating human-like behavior in crowdsourcing tasks that were previously thought to be exclusive to human abilities. However, current efforts focus mainly on simple atomic tasks. We explore whether LLMs can replicate more complex crowdsourcing pipelines. We find that modern LLMs can simula...
http://arxiv.org/pdf/2307.10168v2
2023-07-20T02:29:25
2023-07-19T17:54:43
{ "title": [ -0.04208374, -0.005908966, -0.020431519, -0.01689148, -0.044555664, -0.0063056946, -0.04449463, -0.00983429, -0.017501831, 0.014518738, -0.0496521, 0.037231445, -0.02041626, -0.0072402954, -0.015808105, -0.023269653, -0.008865356, 0....
http://arxiv.org/abs/2307.10156v1
Exploring Transformer Extrapolation
Zhen Qin, Yiran Zhong, Hui Deng
cs.CL
Length extrapolation has attracted considerable attention recently since it allows transformers to be tested on longer sequences than those used in training. Previous research has shown that this property can be attained by using carefully designed Relative Positional Encodings (RPEs). While these methods perform well ...
http://arxiv.org/pdf/2307.10156v1
2023-07-19T17:37:03
2023-07-19T17:37:03
{ "title": [ 0.04144287, 0.022521973, 0.009178162, -0.00071144104, -0.027252197, -0.025100708, -0.004802704, -0.047668457, -0.035827637, 0.04055786, -0.03503418, -0.0013809204, 0.012229919, -0.03149414, 0.018005371, -0.011100769, 0.012588501, 0.0...
http://arxiv.org/abs/2307.10098v1
Gradient Sparsification For Masked Fine-Tuning of Transformers
James O' Neill, Sourav Dutta
cs.CL, cs.LG
Fine-tuning pretrained self-supervised language models is widely adopted for transfer learning to downstream tasks. Fine-tuning can be achieved by freezing gradients of the pretrained network and only updating gradients of a newly added classification layer, or by performing gradient updates on all parameters. Gradual ...
http://arxiv.org/pdf/2307.10098v1
2023-07-19T16:13:13
2023-07-19T16:13:13
{ "title": [ 0.002281189, 0.009384155, -0.032562256, -0.061401367, 0.020233154, 0.01209259, -0.037261963, -0.01991272, -0.015312195, 0.05166626, -0.016525269, 0.031707764, 0.023361206, -0.02368164, -0.010734558, -0.0019950867, 0.0042648315, 0.042...
http://arxiv.org/abs/2307.10088v2
Android in the Wild: A Large-Scale Dataset for Android Device Control
Christopher Rawles, Alice Li, Daniel Rodriguez, Oriana Riva, Timothy Lillicrap
cs.LG, cs.CL, cs.HC
There is a growing interest in device-control systems that can interpret human natural language instructions and execute them on a digital device by directly controlling its user interface. We present a dataset for device-control research, Android in the Wild (AITW), which is orders of magnitude larger than current dat...
http://arxiv.org/pdf/2307.10088v2
2023-10-27T14:24:31
2023-07-19T15:57:24
{ "title": [ -0.034210205, -0.010353088, -0.040771484, -0.038085938, 0.013557434, -0.0129776, -0.032562256, -0.010215759, -0.039520264, 0.024475098, -0.0132369995, 0.034301758, -0.026550293, -0.011787415, -0.0135269165, -0.00007122755, -0.004928589, ...
http://arxiv.org/abs/2307.10025v2
An Empirical Study on Fertility Proposals Using Multi-Grained Topic Analysis Methods
Yulin Zhou
cs.HC, cs.AI, cs.CL, cs.SI
Fertility issues are closely related to population security, in 60 years China's population for the first time in a negative growth trend, the change of fertility policy is of great concern to the community. 2023 "two sessions" proposal "suggests that the country in the form of legislation, the birth of the registratio...
http://arxiv.org/pdf/2307.10025v2
2023-07-22T15:07:57
2023-07-19T15:09:50
{ "title": [ 0.016189575, 0.015335083, -0.039123535, -0.037078857, 0.030044556, -0.024642944, -0.0044059753, -0.033966064, -0.0008854866, 0.0256958, -0.007572174, 0.015914917, -0.008003235, -0.015945435, -0.045440674, -0.0013723373, -0.013916016, ...
http://arxiv.org/abs/2307.10016v1
When Dialects Collide: How Socioeconomic Mixing Affects Language Use
Thomas Louf, José J. Ramasco, David Sánchez, Márton Karsai
physics.soc-ph, cs.CL, cs.SI
The socioeconomic background of people and how they use standard forms of language are not independent, as demonstrated in various sociolinguistic studies. However, the extent to which these correlations may be influenced by the mixing of people from different socioeconomic classes remains relatively unexplored from a ...
http://arxiv.org/pdf/2307.10016v1
2023-07-19T14:55:50
2023-07-19T14:55:50
{ "title": [ -0.01777649, 0.01675415, -0.048858643, -0.030853271, 0.015640259, -0.026443481, -0.0077781677, -0.0055007935, -0.044708252, 0.035461426, 0.0036067963, 0.01374054, -0.015281677, -0.030410767, -0.025482178, -0.0050086975, -0.0076446533, ...
http://arxiv.org/abs/2307.09998v3
Generating Mathematical Derivations with Large Language Models
Jordan Meadows, Marco Valentino, Andre Freitas
cs.CL, math.HO
The derivation of mathematical results in specialised fields, using Large Language Models (LLMs), is an emerging research direction that can help identify models' limitations, and potentially support mathematical discovery. In this paper, we leverage a symbolic engine to generate derivations of equations at scale, and ...
http://arxiv.org/pdf/2307.09998v3
2023-08-08T12:23:49
2023-07-19T14:13:02
{ "title": [ -0.042144775, -0.0058250427, -0.04498291, -0.071899414, 0.005203247, 0.012283325, -0.076416016, 0.0020065308, 0.0072364807, 0.03161621, 0.00982666, -0.004425049, -0.0014762878, -0.015701294, -0.039398193, 0.030532837, -0.0039711, 0.0...
http://arxiv.org/abs/2307.09959v1
GUIDO: A Hybrid Approach to Guideline Discovery & Ordering from Natural Language Texts
Nils Freyer, Dustin Thewes, Matthias Meinecke
cs.CL
Extracting workflow nets from textual descriptions can be used to simplify guidelines or formalize textual descriptions of formal processes like business processes and algorithms. The task of manually extracting processes, however, requires domain expertise and effort. While automatic process model extraction is desira...
http://arxiv.org/pdf/2307.09959v1
2023-07-19T13:01:03
2023-07-19T13:01:03
{ "title": [ -0.05114746, -0.01927185, -0.047058105, -0.0769043, -0.008262634, 0.027236938, -0.064086914, 0.0077438354, -0.017547607, 0.02444458, -0.026351929, 0.03665161, -0.00919342, -0.01360321, -0.03265381, -0.005771637, -0.022277832, 0.04483...
http://arxiv.org/abs/2307.09923v1
Large Language Models can accomplish Business Process Management Tasks
Michael Grohs, Luka Abb, Nourhan Elsayed, Jana-Rebecca Rehse
cs.CL
Business Process Management (BPM) aims to improve organizational activities and their outcomes by managing the underlying processes. To achieve this, it is often necessary to consider information from various sources, including unstructured textual documents. Therefore, researchers have developed several BPM-specific s...
http://arxiv.org/pdf/2307.09923v1
2023-07-19T11:54:46
2023-07-19T11:54:46
{ "title": [ 0.01763916, -0.021957397, -0.042236328, -0.103271484, -0.035095215, -0.007850647, -0.087402344, 0.041900635, -0.015266418, 0.045318604, -0.0029067993, -0.036102295, 0.046051025, -0.053222656, 0.025115967, -0.012710571, 0.0074310303, ...
http://arxiv.org/abs/2307.09885v1
Test-takers have a say: understanding the implications of the use of AI in language tests
Dawen Zhang, Thong Hoang, Shidong Pan, Yongquan Hu, Zhenchang Xing, Mark Staples, Xiwei Xu, Qinghua Lu, Aaron Quigley
cs.CY, cs.AI, cs.CL, cs.HC
Language tests measure a person's ability to use a language in terms of listening, speaking, reading, or writing. Such tests play an integral role in academic, professional, and immigration domains, with entities such as educational institutions, professional accreditation bodies, and governments using them to assess c...
http://arxiv.org/pdf/2307.09885v1
2023-07-19T10:28:59
2023-07-19T10:28:59
{ "title": [ -0.005706787, -0.0043029785, -0.022384644, -0.012954712, -0.013931274, -0.024749756, 0.026794434, -0.008293152, -0.027328491, 0.017028809, -0.038604736, 0.010787964, -0.007827759, -0.018341064, -0.014526367, -0.030258179, -0.0057373047, ...
http://arxiv.org/abs/2307.09813v1
DAPrompt: Deterministic Assumption Prompt Learning for Event Causality Identification
Wei Xiang, Chuanhong Zhan, Bang Wang
cs.CL
Event Causality Identification (ECI) aims at determining whether there is a causal relation between two event mentions. Conventional prompt learning designs a prompt template to first predict an answer word and then maps it to the final decision. Unlike conventional prompts, we argue that predicting an answer word may ...
http://arxiv.org/pdf/2307.09813v1
2023-07-19T08:02:20
2023-07-19T08:02:20
{ "title": [ 0.02293396, 0.007751465, -0.04244995, -0.095214844, 0.029449463, -0.019226074, -0.06390381, -0.020751953, 0.003326416, -0.015716553, -0.009864807, 0.008300781, 0.0015029907, -0.050689697, 0.017578125, -0.0037879944, -0.008491516, 0.0...
http://arxiv.org/abs/2307.10323v1
IncDSI: Incrementally Updatable Document Retrieval
Varsha Kishore, Chao Wan, Justin Lovelace, Yoav Artzi, Kilian Q. Weinberger
cs.IR, cs.CL, cs.LG
Differentiable Search Index is a recently proposed paradigm for document retrieval, that encodes information about a corpus of documents within the parameters of a neural network and directly maps queries to corresponding documents. These models have achieved state-of-the-art performances for document retrieval across ...
http://arxiv.org/pdf/2307.10323v1
2023-07-19T07:20:30
2023-07-19T07:20:30
{ "title": [ -0.014282227, -0.015579224, -0.048919678, -0.07727051, 0.015823364, -0.0063667297, -0.009895325, 0.008666992, 0.042419434, 0.045074463, -0.04046631, 0.0076065063, 0.01322937, 0.0009784698, -0.060913086, 0.014404297, 0.0046539307, 0.0...
http://arxiv.org/abs/2307.09793v1
On the Origin of LLMs: An Evolutionary Tree and Graph for 15,821 Large Language Models
Sarah Gao, Andrew Kean Gao
cs.DL, cs.CL, I.2.1; H.5.0
Since late 2022, Large Language Models (LLMs) have become very prominent with LLMs like ChatGPT and Bard receiving millions of users. Hundreds of new LLMs are announced each week, many of which are deposited to Hugging Face, a repository of machine learning models and datasets. To date, nearly 16,000 Text Generation mo...
http://arxiv.org/pdf/2307.09793v1
2023-07-19T07:17:43
2023-07-19T07:17:43
{ "title": [ -0.057617188, -0.019378662, -0.058013916, -0.035217285, -0.0317688, -0.00705719, -0.07940674, 0.0079956055, -0.018981934, 0.017745972, -0.022705078, 0.009674072, -0.0066604614, -0.03616333, -0.012580872, 0.0033950806, 0.00004839897, ...
http://arxiv.org/abs/2307.09782v2
ZeroQuant-FP: A Leap Forward in LLMs Post-Training W4A8 Quantization Using Floating-Point Formats
Xiaoxia Wu, Zhewei Yao, Yuxiong He
cs.LG, cs.AI, cs.CL
In the complex domain of large language models (LLMs), striking a balance between computational efficiency and maintaining model quality is a formidable challenge. Navigating the inherent limitations of uniform quantization, particularly when dealing with outliers, and motivated by the launch of NVIDIA's H100 hardware,...
http://arxiv.org/pdf/2307.09782v2
2023-07-20T18:47:20
2023-07-19T06:58:03
{ "title": [ -0.028503418, 0.0005812645, -0.06506348, -0.02293396, -0.02255249, -0.045410156, -0.032043457, 0.019119263, 0.004009247, 0.040527344, -0.006877899, 0.012199402, 0.042999268, -0.041778564, 0.006954193, 0.0206604, 0.0016202927, 0.02401...
http://arxiv.org/abs/2307.09744v1
Enhancing conversational quality in language learning chatbots: An evaluation of GPT4 for ASR error correction
Long Mai, Julie Carson-Berndsen
cs.CL, cs.AI
The integration of natural language processing (NLP) technologies into educational applications has shown promising results, particularly in the language learning domain. Recently, many spoken open-domain chatbots have been used as speaking partners, helping language learners improve their language skills. However, one...
http://arxiv.org/pdf/2307.09744v1
2023-07-19T04:25:21
2023-07-19T04:25:21
{ "title": [ 0.00491333, 0.00819397, -0.0013837814, -0.023345947, -0.015914917, 0.003112793, -0.021621704, -0.022018433, -0.010925293, 0.045074463, -0.012184143, 0.0013284683, 0.011207581, -0.014442444, -0.053375244, -0.00554657, -0.022827148, 0....
http://arxiv.org/abs/2307.10314v1
Mood Classification of Bangla Songs Based on Lyrics
Maliha Mahajebin, Mohammad Rifat Ahmmad Rashid, Nafees Mansoor
cs.IR, cs.CL, cs.LG, cs.SD, eess.AS
Music can evoke various emotions, and with the advancement of technology, it has become more accessible to people. Bangla music, which portrays different human emotions, lacks sufficient research. The authors of this article aim to analyze Bangla songs and classify their moods based on the lyrics. To achieve this, this...
http://arxiv.org/pdf/2307.10314v1
2023-07-19T03:31:41
2023-07-19T03:31:41
{ "title": [ 0.011199951, -0.001156807, -0.034240723, -0.029922485, -0.011428833, -0.022979736, -0.0033779144, -0.0007405281, 0.0128479, 0.024276733, -0.039031982, -0.020141602, 0.0028209686, -0.032470703, -0.04385376, -0.027832031, -0.0032348633, ...
http://arxiv.org/abs/2307.09706v1
RaTE: a Reproducible automatic Taxonomy Evaluation by Filling the Gap
Tianjian Gao, Phillipe Langlais
cs.CL, cs.AI, cs.LG
Taxonomies are an essential knowledge representation, yet most studies on automatic taxonomy construction (ATC) resort to manual evaluation to score proposed algorithms. We argue that automatic taxonomy evaluation (ATE) is just as important as taxonomy construction. We propose RaTE, an automatic label-free taxonomy sco...
http://arxiv.org/pdf/2307.09706v1
2023-07-19T01:37:31
2023-07-19T01:37:31
{ "title": [ -0.016677856, -0.0011367798, -0.05154419, -0.044281006, 0.020217896, -0.01537323, -0.012519836, -0.028945923, 0.010520935, -0.003107071, -0.045562744, 0.010665894, -0.032043457, -0.014480591, -0.027770996, -0.006500244, -0.016220093, ...
http://arxiv.org/abs/2307.09705v1
CValues: Measuring the Values of Chinese Large Language Models from Safety to Responsibility
Guohai Xu, Jiayi Liu, Ming Yan, Haotian Xu, Jinghui Si, Zhuoran Zhou, Peng Yi, Xing Gao, Jitao Sang, Rong Zhang, Ji Zhang, Chao Peng, Fei Huang, Jingren Zhou
cs.CL
With the rapid evolution of large language models (LLMs), there is a growing concern that they may pose risks or have negative social impacts. Therefore, evaluation of human values alignment is becoming increasingly important. Previous work mainly focuses on assessing the performance of LLMs on certain knowledge and re...
http://arxiv.org/pdf/2307.09705v1
2023-07-19T01:22:40
2023-07-19T01:22:40
{ "title": [ -0.029251099, 0.0047569275, -0.040924072, -0.05633545, 0.007713318, -0.0060310364, -0.048797607, -0.003314972, -0.017791748, 0.021408081, -0.022277832, 0.0070381165, -0.0054626465, -0.04208374, -0.007129669, -0.016357422, -0.012565613, ...
http://arxiv.org/abs/2307.09702v4
Efficient Guided Generation for Large Language Models
Brandon T. Willard, Rémi Louf
cs.CL, cs.LG
In this article we show how the problem of neural text generation can be constructively reformulated in terms of transitions between the states of a finite-state machine. This framework leads to an efficient approach to guiding text generation with regular expressions and context-free grammars by allowing the construct...
http://arxiv.org/pdf/2307.09702v4
2023-08-19T21:27:51
2023-07-19T01:14:49
{ "title": [ -0.046447754, -0.02684021, -0.04611206, -0.099609375, -0.0010623932, 0.03515625, -0.099853516, 0.04269409, 0.013534546, 0.0256958, -0.014625549, 0.0034675598, 0.015029907, -0.019104004, -0.026000977, 0.021881104, -0.014030457, 0.0273...
http://arxiv.org/abs/2307.09701v1
Efficiency Pentathlon: A Standardized Arena for Efficiency Evaluation
Hao Peng, Qingqing Cao, Jesse Dodge, Matthew E. Peters, Jared Fernandez, Tom Sherborne, Kyle Lo, Sam Skjonsberg, Emma Strubell, Darrell Plessas, Iz Beltagy, Evan Pete Walsh, Noah A. Smith, Hannaneh Hajishirzi
cs.CL
Rising computational demands of modern natural language processing (NLP) systems have increased the barrier to entry for cutting-edge research while posing serious environmental concerns. Yet, progress on model efficiency has been impeded by practical challenges in model evaluation and comparison. For example, hardware...
http://arxiv.org/pdf/2307.09701v1
2023-07-19T01:05:33
2023-07-19T01:05:33
{ "title": [ -0.041931152, 0.011367798, -0.030181885, 0.0038528442, -0.0034236908, -0.0140686035, 0.008041382, -0.053131104, 0.013641357, 0.017852783, -0.0017328262, 0.0020980835, -0.02180481, -0.035125732, -0.036621094, 0.011878967, -0.023391724, ...
http://arxiv.org/abs/2307.10303v1
Analyzing sports commentary in order to automatically recognize events and extract insights
Yanis Miraoui
cs.CL, cs.LG, stat.ML
In this paper, we carefully investigate how we can use multiple different Natural Language Processing techniques and methods in order to automatically recognize the main actions in sports events. We aim to extract insights by analyzing live sport commentaries from different sources and by classifying these major action...
http://arxiv.org/pdf/2307.10303v1
2023-07-18T18:51:06
2023-07-18T18:51:06
{ "title": [ 0.018478394, -0.008338928, -0.028320312, -0.04574585, -0.018005371, -0.027236938, -0.0064353943, 0.008483887, -0.009666443, 0.060943604, -0.04449463, 0.009933472, 0.04837036, -0.016021729, -0.018707275, -0.013290405, 0.03466797, -0.0...
http://arxiv.org/abs/2307.09532v1
Can Model Fusing Help Transformers in Long Document Classification? An Empirical Study
Damith Premasiri, Tharindu Ranasinghe, Ruslan Mitkov
cs.CL
Text classification is an area of research which has been studied over the years in Natural Language Processing (NLP). Adapting NLP to multiple domains has introduced many new challenges for text classification and one of them is long document classification. While state-of-the-art transformer models provide excellent ...
http://arxiv.org/pdf/2307.09532v1
2023-07-18T18:21:26
2023-07-18T18:21:26
{ "title": [ 0.0022468567, 0.023483276, -0.026916504, -0.045135498, -0.004787445, 0.00022351742, -0.045928955, -0.03111267, -0.012504578, 0.024383545, -0.040496826, 0.0010643005, 0.0056648254, -0.014640808, -0.027526855, -0.036834717, -0.020431519, ...
http://arxiv.org/abs/2307.09476v1
Overthinking the Truth: Understanding how Language Models Process False Demonstrations
Danny Halawi, Jean-Stanislas Denain, Jacob Steinhardt
cs.LG, cs.AI, cs.CL
Modern language models can imitate complex patterns through few-shot learning, enabling them to complete challenging tasks without fine-tuning. However, imitation can also lead models to reproduce inaccuracies or harmful content if present in the context. We study harmful imitation through the lens of a model's interna...
http://arxiv.org/pdf/2307.09476v1
2023-07-18T17:56:50
2023-07-18T17:56:50
{ "title": [ 0.0004911423, -0.0087890625, -0.01235199, -0.032806396, -0.007537842, -0.012786865, -0.02319336, -0.03111267, -0.028244019, 0.018447876, -0.0058898926, -0.0022468567, 0.0146484375, -0.03466797, -0.021896362, 0.00051546097, -0.02255249, ...
http://arxiv.org/abs/2307.09474v1
ChatSpot: Bootstrapping Multimodal LLMs via Precise Referring Instruction Tuning
Liang Zhao, En Yu, Zheng Ge, Jinrong Yang, Haoran Wei, Hongyu Zhou, Jianjian Sun, Yuang Peng, Runpei Dong, Chunrui Han, Xiangyu Zhang
cs.CL, cs.CV
Human-AI interactivity is a critical aspect that reflects the usability of multimodal large language models (MLLMs). However, existing end-to-end MLLMs only allow users to interact with them through language instructions, leading to the limitation of the interactive accuracy and efficiency. In this study, we present pr...
http://arxiv.org/pdf/2307.09474v1
2023-07-18T17:56:06
2023-07-18T17:56:06
{ "title": [ -0.09094238, -0.020812988, -0.046203613, -0.03640747, -0.049621582, 0.015838623, -0.09576416, 0.011817932, -0.016418457, 0.036071777, -0.02798462, 0.032409668, 0.008308411, -0.014984131, 0.0039138794, 0.0039405823, -0.016693115, 0.02...
http://arxiv.org/abs/2307.09456v2
A comparative analysis of SRGAN models
Fatemeh Rezapoor Nikroo, Ajinkya Deshmukh, Anantha Sharma, Adrian Tam, Kaarthik Kumar, Cleo Norris, Aditya Dangi
cs.CV, cs.CL, eess.IV
In this study, we evaluate the performance of multiple state-of-the-art SRGAN (Super Resolution Generative Adversarial Network) models, ESRGAN, Real-ESRGAN and EDSR, on a benchmark dataset of real-world images which undergo degradation using a pipeline. Our results show that some models seem to significantly increase t...
http://arxiv.org/pdf/2307.09456v2
2023-07-19T14:27:57
2023-07-18T17:35:45
{ "title": [ 0.0124435425, 0.010391235, -0.023727417, -0.04006958, 0.0056533813, -0.0158844, -0.04550171, -0.021697998, -0.0317688, 0.033172607, 0.052368164, -0.00088119507, -0.009536743, -0.062561035, -0.007713318, 0.02017212, -0.0011863708, 0.0...
http://arxiv.org/abs/2307.09455v2
Pseudo Outlier Exposure for Out-of-Distribution Detection using Pretrained Transformers
Jaeyoung Kim, Kyuheon Jung, Dongbin Na, Sion Jang, Eunbin Park, Sungchul Choi
cs.CL, 68T50
For real-world language applications, detecting an out-of-distribution (OOD) sample is helpful to alert users or reject such unreliable samples. However, modern over-parameterized language models often produce overconfident predictions for both in-distribution (ID) and OOD samples. In particular, language models suffer...
http://arxiv.org/pdf/2307.09455v2
2023-07-19T04:13:11
2023-07-18T17:29:23
{ "title": [ 0.043518066, 0.015625, 0.011077881, -0.034301758, 0.013435364, 0.0061454773, 0.024993896, -0.04550171, 0.0013341904, 0.02619934, -0.019851685, 0.008255005, -0.018493652, -0.041229248, -0.039123535, -0.019714355, -0.0034446716, 0.0404...
http://arxiv.org/abs/2307.09416v2
Let's ViCE! Mimicking Human Cognitive Behavior in Image Generation Evaluation
Federico Betti, Jacopo Staiano, Lorenzo Baraldi, Lorenzo Baraldi, Rita Cucchiara, Nicu Sebe
cs.CV, cs.CL
Research in Image Generation has recently made significant progress, particularly boosted by the introduction of Vision-Language models which are able to produce high-quality visual content based on textual inputs. Despite ongoing advancements in terms of generation quality and realism, no methodical frameworks have be...
http://arxiv.org/pdf/2307.09416v2
2023-07-19T08:27:50
2023-07-18T16:33:30
{ "title": [ -0.0068855286, 0.007972717, -0.0075912476, -0.012031555, 0.0055236816, -0.015213013, 0.0060310364, -0.03933716, 0.012031555, 0.02305603, -0.021453857, 0.00970459, -0.014945984, -0.019714355, -0.042938232, -0.004699707, -0.0018615723, ...
http://arxiv.org/abs/2307.09384v2
Zero-shot Query Reformulation for Conversational Search
Dayu Yang, Yue Zhang, Hui Fang
cs.IR, cs.CL
As the popularity of voice assistants continues to surge, conversational search has gained increased attention in Information Retrieval. However, data sparsity issues in conversational search significantly hinder the progress of supervised conversational search methods. Consequently, researchers are focusing more on ze...
http://arxiv.org/pdf/2307.09384v2
2023-10-23T17:24:02
2023-07-18T16:05:25
{ "title": [ 0.0033931732, -0.01133728, -0.024932861, -0.058380127, -0.00178051, 0.0028858185, -0.035339355, 0.0076560974, 0.04510498, 0.04876709, -0.014152527, -0.011726379, 0.007293701, -0.0074882507, -0.07543945, 0.011436462, -0.005104065, -0....
http://arxiv.org/abs/2307.09312v2
Multi-Modal Discussion Transformer: Integrating Text, Images and Graph Transformers to Detect Hate Speech on Social Media
Liam Hebert, Gaurav Sahu, Yuxuan Guo, Nanda Kishore Sreenivas, Lukasz Golab, Robin Cohen
cs.CL, cs.LG, cs.MM, cs.SI
We present the Multi-Modal Discussion Transformer (mDT), a novel multi-modal graph-based transformer model for detecting hate speech in online social networks, such as Reddit discussions. In contrast to traditional comment-only methods, our approach to labelling a comment as hate speech involves a holistic analysis of ...
http://arxiv.org/pdf/2307.09312v2
2023-08-31T15:32:01
2023-07-18T14:57:12
{ "title": [ 0.054016113, 0.02999878, -0.015731812, -0.027679443, 0.0126571655, -0.0287323, -0.0070610046, -0.04119873, -0.013832092, 0.03652954, -0.040924072, 0.031707764, 0.04824829, -0.026046753, 0.0010251999, -0.016036987, -0.0029449463, 0.07...
http://arxiv.org/abs/2307.10292v1
The Language Labyrinth: Constructive Critique on the Terminology Used in the AI Discourse
Rainer Rehak
cs.CY, cs.AI, cs.CL, cs.NE, K.4.0; K.4.1; K.4.2; I.2.0; I.2.1; K.4.3
In the interdisciplinary field of artificial intelligence (AI) the problem of clear terminology is especially momentous. This paper claims, that AI debates are still characterised by a lack of critical distance to metaphors like 'training', 'learning' or 'deciding'. As consequence, reflections regarding responsibility ...
http://arxiv.org/pdf/2307.10292v1
2023-07-18T14:32:21
2023-07-18T14:32:21
{ "title": [ -0.036987305, -0.013328552, -0.019699097, 0.013298035, -0.021514893, -0.016540527, 0.014442444, -0.03552246, -0.026382446, 0.027740479, -0.010520935, 0.003774643, 0.0032196045, -0.027816772, -0.033721924, -0.013381958, -0.0067596436, ...
http://arxiv.org/abs/2307.09288v2
Llama 2: Open Foundation and Fine-Tuned Chat Models
Hugo Touvron, Louis Martin, Kevin Stone, Peter Albert, Amjad Almahairi, Yasmine Babaei, Nikolay Bashlykov, Soumya Batra, Prajjwal Bhargava, Shruti Bhosale, Dan Bikel, Lukas Blecher, Cristian Canton Ferrer, Moya Chen, Guillem Cucurull, David Esiobu, Jude Fernandes, Jeremy Fu, Wenyin Fu, Brian Fuller, Cynthia Gao, Vedanu...
cs.CL, cs.AI
In this work, we develop and release Llama 2, a collection of pretrained and fine-tuned large language models (LLMs) ranging in scale from 7 billion to 70 billion parameters. Our fine-tuned LLMs, called Llama 2-Chat, are optimized for dialogue use cases. Our models outperform open-source chat models on most benchmarks ...
http://arxiv.org/pdf/2307.09288v2
2023-07-19T17:08:59
2023-07-18T14:31:57
{ "title": [ -0.02128601, -0.0050468445, -0.030960083, -0.030975342, -0.03274536, 0.009094238, -0.060913086, -0.008140564, -0.052825928, 0.023483276, 0.019561768, 0.03314209, -0.013519287, -0.058166504, 0.043548584, -0.009109497, -0.008071899, 0....
http://arxiv.org/abs/2307.10291v2
Mutual Reinforcement Effects in Japanese Sentence Classification and Named Entity Recognition Tasks
Chengguang Gan, Qinghao Zhang, Tatsunori Mori
cs.CL
Information extraction(IE) is a crucial subfield within natural language processing. However, for the traditionally segmented approach to sentence classification and Named Entity Recognition, the intricate interactions between these individual subtasks remain largely uninvestigated. In this study, we propose an integra...
http://arxiv.org/pdf/2307.10291v2
2023-07-21T02:34:58
2023-07-18T14:30:36
{ "title": [ -0.007881165, 0.019683838, -0.008651733, -0.015197754, 0.0036621094, -0.017120361, 0.0017089844, -0.019515991, -0.016693115, 0.008453369, -0.038269043, 0.01751709, -0.010635376, -0.008529663, -0.0418396, -0.039031982, -0.029449463, 0...
http://arxiv.org/abs/2307.09274v1
Improving Text Semantic Similarity Modeling through a 3D Siamese Network
Jianxiang Zang, Hui Liu
cs.CL, cs.AI
Siamese networks have gained popularity as a method for modeling text semantic similarity. Traditional methods rely on pooling operation to compress the semantic representations from Transformer blocks in encoding, resulting in two-dimensional semantic vectors and the loss of hierarchical semantic information from Tran...
http://arxiv.org/pdf/2307.09274v1
2023-07-18T14:11:58
2023-07-18T14:11:58
{ "title": [ -0.029785156, 0.014404297, -0.056640625, -0.043945312, 0.002708435, -0.00048279762, -0.04019165, -0.016235352, -0.009475708, 0.062469482, -0.004135132, 0.024276733, 0.0056915283, 0.0036697388, -0.06323242, 0.005138397, -0.0013980865, ...
http://arxiv.org/abs/2307.09270v1
Linearized Relative Positional Encoding
Zhen Qin, Weixuan Sun, Kaiyue Lu, Hui Deng, Dongxu Li, Xiaodong Han, Yuchao Dai, Lingpeng Kong, Yiran Zhong
cs.CL
Relative positional encoding is widely used in vanilla and linear transformers to represent positional information. However, existing encoding methods of a vanilla transformer are not always directly applicable to a linear transformer, because the latter requires a decomposition of the query and key representations int...
http://arxiv.org/pdf/2307.09270v1
2023-07-18T13:56:43
2023-07-18T13:56:43
{ "title": [ 0.0029697418, -0.00039577484, -0.021728516, -0.032196045, -0.03515625, -0.010269165, -0.049804688, -0.044006348, 0.0017347336, 0.07397461, -0.0077552795, 0.0016517639, 0.032440186, -0.03744507, -0.0036735535, 0.037139893, 0.025253296, ...
http://arxiv.org/abs/2307.09255v1
Text vectorization via transformer-based language models and n-gram perplexities
Mihailo Škorić
cs.CL, 68T50
As the probability (and thus perplexity) of a text is calculated based on the product of the probabilities of individual tokens, it may happen that one unlikely token significantly reduces the probability (i.e., increase the perplexity) of some otherwise highly probable input, while potentially representing a simple ty...
http://arxiv.org/pdf/2307.09255v1
2023-07-18T13:38:39
2023-07-18T13:38:39
{ "title": [ -0.028366089, -0.00844574, 0.011619568, -0.019195557, -0.028717041, 0.012756348, -0.055786133, -0.042266846, 0.0031013489, 0.032348633, -0.029251099, 0.013244629, 0.023223877, -0.009002686, -0.040283203, 0.008148193, -0.007980347, 0....
http://arxiv.org/abs/2307.09254v1
PAC Neural Prediction Set Learning to Quantify the Uncertainty of Generative Language Models
Sangdon Park, Taesoo Kim
cs.LG, cs.CL, stat.ML
Uncertainty learning and quantification of models are crucial tasks to enhance the trustworthiness of the models. Importantly, the recent surge of generative language models (GLMs) emphasizes the need for reliable uncertainty quantification due to the concerns on generating hallucinated facts. In this paper, we propose...
http://arxiv.org/pdf/2307.09254v1
2023-07-18T13:36:24
2023-07-18T13:36:24
{ "title": [ -0.05001831, 0.0027770996, -0.06329346, -0.04815674, -0.016494751, -0.01411438, -0.07702637, 0.011795044, -0.0028152466, 0.026641846, -0.02041626, 0.023727417, -0.0053710938, -0.036071777, -0.02961731, 0.009994507, -0.0008764267, 0.0...
http://arxiv.org/abs/2307.09249v1
UniTabE: Pretraining a Unified Tabular Encoder for Heterogeneous Tabular Data
Yazheng Yang, Yuqi Wang, Guang Liu, Ledell Wu, Qi Liu
cs.LG, cs.AI, cs.CL
Recent advancements in Natural Language Processing (NLP) have witnessed the groundbreaking impact of pretrained models, yielding impressive outcomes across various tasks. This study seeks to extend the power of pretraining methodologies to tabular data, a domain traditionally overlooked, yet inherently challenging due ...
http://arxiv.org/pdf/2307.09249v1
2023-07-18T13:28:31
2023-07-18T13:28:31
{ "title": [ -0.014312744, 0.008216858, -0.061157227, -0.07672119, 0.013885498, -0.0069503784, -0.06173706, -0.013671875, 0.015182495, 0.062927246, -0.020904541, 0.0149383545, 0.031433105, -0.0012998581, -0.03845215, 0.023086548, 0.0027217865, 0....
http://arxiv.org/abs/2307.09209v1
Automated Ableism: An Exploration of Explicit Disability Biases in Sentiment and Toxicity Analysis Models
Pranav Narayanan Venkit, Mukund Srinath, Shomir Wilson
cs.CL, cs.AI, cs.CY, cs.LG
We analyze sentiment analysis and toxicity detection models to detect the presence of explicit bias against people with disability (PWD). We employ the bias identification framework of Perturbation Sensitivity Analysis to examine conversations related to PWD on social media platforms, specifically Twitter and Reddit, i...
http://arxiv.org/pdf/2307.09209v1
2023-07-18T12:45:54
2023-07-18T12:45:54
{ "title": [ 0.017211914, 0.002916336, -0.019622803, -0.061676025, 0.018966675, -0.009117126, 0.00039100647, -0.0055007935, 0.0151901245, 0.039764404, -0.039520264, 0.025390625, 0.0020427704, 0.0012350082, -0.06591797, -0.0044441223, -0.033996582, ...
http://arxiv.org/abs/2307.09162v3
Unveiling Gender Bias in Terms of Profession Across LLMs: Analyzing and Addressing Sociological Implications
Vishesh Thakur
cs.CL
Gender bias in artificial intelligence (AI) and natural language processing has garnered significant attention due to its potential impact on societal perceptions and biases. This research paper aims to analyze gender bias in Large Language Models (LLMs) with a focus on multiple comparisons between GPT-2 and GPT-3.5, s...
http://arxiv.org/pdf/2307.09162v3
2023-08-31T20:02:47
2023-07-18T11:38:45
{ "title": [ -0.05609131, 0.00932312, -0.068359375, -0.0054626465, -0.016036987, -0.03250122, -0.029769897, -0.0009531975, -0.018707275, 0.011764526, -0.058929443, 0.01398468, -0.00491333, -0.014335632, -0.0027999878, -0.02319336, -0.0309906, 0.0...
http://arxiv.org/abs/2307.09084v1
Attention over pre-trained Sentence Embeddings for Long Document Classification
Amine Abdaoui, Sourav Dutta
cs.CL
Despite being the current de-facto models in most NLP tasks, transformers are often limited to short sequences due to their quadratic attention complexity on the number of tokens. Several attempts to address this issue were studied, either by reducing the cost of the self-attention computation or by modeling smaller se...
http://arxiv.org/pdf/2307.09084v1
2023-07-18T09:06:35
2023-07-18T09:06:35
{ "title": [ 0.002702713, 0.011306763, -0.021560669, -0.053100586, 0.009208679, -0.015182495, -0.020507812, -0.00737381, 0.0149002075, 0.019744873, -0.020492554, 0.0023784637, 0.0012960434, -0.0065078735, -0.027404785, -0.006275177, -0.020446777, ...
http://arxiv.org/abs/2307.09059v1
Unleashing the Imagination of Text: A Novel Framework for Text-to-image Person Retrieval via Exploring the Power of Words
Delong Liu, Haiwen Li
cs.CL, cs.AI, cs.CV
The goal of Text-to-image person retrieval is to retrieve person images from a large gallery that match the given textual descriptions. The main challenge of this task lies in the significant differences in information representation between the visual and textual modalities. The textual modality conveys abstract and p...
http://arxiv.org/pdf/2307.09059v1
2023-07-18T08:23:46
2023-07-18T08:23:46
{ "title": [ 0.015586853, -0.0006866455, -0.009750366, -0.026412964, 0.0062408447, -0.010597229, 0.0060424805, -0.03024292, 0.032989502, 0.05593872, -0.038970947, 0.021575928, 0.03652954, 0.0059013367, -0.051330566, 0.0076828003, -0.0011997223, 0...
http://arxiv.org/abs/2307.09021v1
Towards a Neural Era in Dialogue Management for Collaboration: A Literature Survey
Amogh Mannekote
cs.CL
Dialogue-based human-AI collaboration can revolutionize collaborative problem-solving, creative exploration, and social support. To realize this goal, the development of automated agents proficient in skills such as negotiating, following instructions, establishing common ground, and progressing shared tasks is essenti...
http://arxiv.org/pdf/2307.09021v1
2023-07-18T07:20:43
2023-07-18T07:20:43
{ "title": [ -0.031311035, -0.0057792664, -0.052734375, -0.034942627, -0.0003311634, -0.0068511963, -0.028152466, 0.017349243, -0.0050086975, 0.046539307, 0.010719299, -0.0029964447, -0.0066490173, -0.024459839, -0.06021118, 0.0035705566, -0.0362854,...
http://arxiv.org/abs/2307.09014v1
Exploring acceptance of autonomous vehicle policies using KeyBERT and SNA: Targeting engineering students
Jinwoo Ha, Dongsoo Kim
cs.SI, cs.AI, cs.CL, cs.RO
This study aims to explore user acceptance of Autonomous Vehicle (AV) policies with improved text-mining methods. Recently, South Korean policymakers have viewed Autonomous Driving Car (ADC) and Autonomous Driving Robot (ADR) as next-generation means of transportation that will reduce the cost of transporting passenger...
http://arxiv.org/pdf/2307.09014v1
2023-07-18T07:03:29
2023-07-18T07:03:29
{ "title": [ -0.02229309, 0.024902344, -0.027679443, 0.03338623, 0.016326904, -0.028747559, 0.06732178, -0.04333496, -0.08099365, 0.008934021, -0.01285553, 0.027709961, -0.061035156, -0.027542114, 0.004638672, -0.034820557, -0.01979065, -0.006225...
http://arxiv.org/abs/2307.09009v3
How is ChatGPT's behavior changing over time?
Lingjiao Chen, Matei Zaharia, James Zou
cs.CL, cs.AI, cs.LG
GPT-3.5 and GPT-4 are the two most widely used large language model (LLM) services. However, when and how these models are updated over time is opaque. Here, we evaluate the March 2023 and June 2023 versions of GPT-3.5 and GPT-4 on several diverse tasks: 1) math problems, 2) sensitive/dangerous questions, 3) opinion su...
http://arxiv.org/pdf/2307.09009v3
2023-10-31T16:13:44
2023-07-18T06:56:08
{ "title": [ 0.017120361, -0.010757446, -0.007499695, 0.00920105, -0.0149383545, -0.012573242, 0.004131317, -0.003211975, 0.009963989, 0.034210205, -0.025802612, 0.031677246, -0.015037537, -0.0007863045, -0.00058841705, 0.013511658, 0.012794495, ...
http://arxiv.org/abs/2307.09007v1
On the (In)Effectiveness of Large Language Models for Chinese Text Correction
Yinghui Li, Haojing Huang, Shirong Ma, Yong Jiang, Yangning Li, Feng Zhou, Hai-Tao Zheng, Qingyu Zhou
cs.CL
Recently, the development and progress of Large Language Models (LLMs) have amazed the entire Artificial Intelligence community. As an outstanding representative of LLMs and the foundation model that set off this wave of research on LLMs, ChatGPT has attracted more and more researchers to study its capabilities and per...
http://arxiv.org/pdf/2307.09007v1
2023-07-18T06:48:52
2023-07-18T06:48:52
{ "title": [ -0.027191162, 0.0043029785, -0.01499939, -0.022827148, -0.006500244, -0.00048351288, -0.036254883, -0.023086548, 0.005016327, 0.04336548, -0.027557373, -0.008544922, 0.042938232, -0.017349243, -0.054992676, 0.007835388, -0.029251099, ...
http://arxiv.org/abs/2307.10274v2
Zero-shot Domain-sensitive Speech Recognition with Prompt-conditioning Fine-tuning
Feng-Ting Liao, Yung-Chieh Chan, Yi-Chang Chen, Chan-Jan Hsu, Da-shan Shiu
eess.AS, cs.CL, cs.LG
In this work, we propose a method to create domain-sensitive speech recognition models that utilize textual domain information by conditioning its generation on a given text prompt. This is accomplished by fine-tuning a pre-trained, end-to-end model (Whisper) to learn from demonstrations with prompt examples. We show t...
http://arxiv.org/pdf/2307.10274v2
2023-10-06T03:41:30
2023-07-18T06:45:43
{ "title": [ 0.009223938, 0.0044403076, -0.050872803, -0.07867432, 0.013771057, -0.017166138, -0.030395508, 0.026123047, -0.00995636, 0.04348755, -0.015686035, -0.0019836426, -0.009490967, -0.012237549, -0.06536865, -0.0075149536, -0.0051345825, ...
http://arxiv.org/abs/2307.11772v3
AutoAlign: Fully Automatic and Effective Knowledge Graph Alignment enabled by Large Language Models
Rui Zhang, Yixin Su, Bayu Distiawan Trisedya, Xiaoyan Zhao, Min Yang, Hong Cheng, Jianzhong Qi
cs.IR, cs.CL, cs.LG
The task of entity alignment between knowledge graphs (KGs) aims to identify every pair of entities from two different KGs that represent the same entity. Many machine learning-based methods have been proposed for this task. However, to our best knowledge, existing methods all require manually crafted seed alignments, ...
http://arxiv.org/pdf/2307.11772v3
2023-11-13T10:56:22
2023-07-18T04:43:24
{ "title": [ -0.045410156, -0.015388489, -0.072753906, -0.07763672, -0.0059547424, 0.009887695, -0.07556152, 0.011650085, 0.00081300735, 0.04425049, -0.0036754608, 0.0074653625, 0.0234375, -0.017181396, -0.03286743, 0.014335632, 0.0006132126, 0.0...
http://arxiv.org/abs/2307.08945v2
Mitigating Label Bias via Decoupled Confident Learning
Yunyi Li, Maria De-Arteaga, Maytal Saar-Tsechansky
cs.LG, cs.CL
Growing concerns regarding algorithmic fairness have led to a surge in methodologies to mitigate algorithmic bias. However, such methodologies largely assume that observed labels in training data are correct. This is problematic because bias in labels is pervasive across important domains, including healthcare, hiring,...
http://arxiv.org/pdf/2307.08945v2
2023-09-29T16:06:05
2023-07-18T03:28:03
{ "title": [ -0.010108948, 0.02142334, -0.045074463, -0.046295166, 0.019851685, -0.0044555664, -0.028015137, -0.046203613, -0.0032691956, 0.009666443, -0.027801514, 0.0025978088, -0.015159607, -0.04147339, -0.015144348, -0.008338928, -0.022232056, ...
http://arxiv.org/abs/2307.08941v2
NTK-approximating MLP Fusion for Efficient Language Model Fine-tuning
Tianxin Wei, Zeming Guo, Yifan Chen, Jingrui He
cs.LG, cs.CL
Fine-tuning a pre-trained language model (PLM) emerges as the predominant strategy in many natural language processing applications. However, even fine-tuning the PLMs and doing inference are expensive, especially on edge devices with low computing power. Some general approaches (e.g. quantization and distillation) hav...
http://arxiv.org/pdf/2307.08941v2
2023-08-05T01:10:06
2023-07-18T03:12:51
{ "title": [ -0.041625977, -0.014175415, -0.030944824, -0.04812622, -0.03262329, 0.012374878, -0.068725586, 0.0034332275, -0.0052375793, 0.026412964, -0.029449463, 0.013885498, 0.03967285, -0.019332886, -0.007247925, 0.01965332, -0.03237915, 0.04...
http://arxiv.org/abs/2307.08931v1
Teach model to answer questions after comprehending the document
Ruiqing Sun, Ping Jian
cs.CL
Multi-choice Machine Reading Comprehension (MRC) is a challenging extension of Natural Language Processing (NLP) that requires the ability to comprehend the semantics and logical relationships between entities in a given text. The MRC task has traditionally been viewed as a process of answering questions based on the g...
http://arxiv.org/pdf/2307.08931v1
2023-07-18T02:38:02
2023-07-18T02:38:02
{ "title": [ 0.01928711, -0.00069379807, 0.013473511, -0.012191772, -0.04449463, 0.0037326813, -0.03326416, -0.037506104, 0.002658844, 0.027999878, -0.06011963, -0.019485474, 0.0069351196, -0.008712769, -0.004852295, -0.046173096, 0.026794434, 0....
http://arxiv.org/abs/2307.08925v1
Federated Large Language Model: A Position Paper
Chaochao Chen, Xiaohua Feng, Jun Zhou, Jianwei Yin, Xiaolin Zheng
cs.LG, cs.AI, cs.CL
Large scale language models (LLM) have received significant attention and found diverse applications across various domains, but their development encounters challenges in real-world scenarios. These challenges arise due to the scarcity of public domain data availability and the need to maintain privacy with respect to...
http://arxiv.org/pdf/2307.08925v1
2023-07-18T02:09:14
2023-07-18T02:09:14
{ "title": [ -0.0262146, -0.0064315796, -0.06222534, -0.093811035, 0.025512695, 0.012199402, -0.05429077, 0.030303955, -0.0036811829, 0.05142212, -0.018508911, -0.008285522, 0.015838623, -0.0048599243, -0.049591064, 0.008918762, -0.018859863, 0.0...
http://arxiv.org/abs/2307.08922v1
Large Language Models Perform Diagnostic Reasoning
Cheng-Kuang Wu, Wei-Lin Chen, Hsin-Hsi Chen
cs.CL
We explore the extension of chain-of-thought (CoT) prompting to medical reasoning for the task of automatic diagnosis. Motivated by doctors' underlying reasoning process, we present Diagnostic-Reasoning CoT (DR-CoT). Empirical results demonstrate that by simply prompting large language models trained only on general te...
http://arxiv.org/pdf/2307.08922v1
2023-07-18T01:43:00
2023-07-18T01:43:00
{ "title": [ 0.027664185, -0.017501831, -0.029769897, -0.12445068, -0.05441284, 0.0052719116, -0.11224365, 0.03225708, -0.025375366, 0.022949219, -0.03604126, -0.033416748, 0.04510498, -0.053344727, 0.053833008, -0.011749268, 0.020584106, 0.02952...
http://arxiv.org/abs/2307.11771v2
An Integrated NPL Approach to Sentiment Analysis in Satisfaction Surveys
Edson B. Pinto-Luque
cs.CL
The research project aims to apply an integrated approach to natural language processing NLP to satisfaction surveys. It will focus on understanding and extracting relevant information from survey responses, analyzing feelings, and identifying recurring word patterns. NLP techniques will be used to determine emotional ...
http://arxiv.org/pdf/2307.11771v2
2023-08-02T00:33:22
2023-07-18T00:23:35
{ "title": [ 0.014282227, 0.008071899, -0.017593384, -0.05657959, 0.024780273, -0.0047569275, -0.0037899017, -0.011932373, 0.03970337, 0.017196655, -0.03894043, -0.018936157, -0.026351929, -0.014930725, -0.06970215, -0.017990112, -0.022750854, 0....
http://arxiv.org/abs/2307.08859v1
Curriculum Learning for Graph Neural Networks: A Multiview Competence-based Approach
Nidhi Vakil, Hadi Amiri
cs.LG, cs.AI, cs.CL
A curriculum is a planned sequence of learning materials and an effective one can make learning efficient and effective for both humans and machines. Recent studies developed effective data-driven curriculum learning approaches for training graph neural networks in language applications. However, existing curriculum le...
http://arxiv.org/pdf/2307.08859v1
2023-07-17T21:33:35
2023-07-17T21:33:35
{ "title": [ -0.013633728, -0.0075263977, -0.04348755, -0.026168823, -0.002626419, -0.031463623, -0.011306763, 0.0037441254, -0.00856781, 0.004234314, 0.009765625, 0.007434845, -0.052093506, -0.04421997, 0.0022792816, -0.006996155, -0.0051841736, ...
http://arxiv.org/abs/2307.08813v2
Comparative Performance Evaluation of Large Language Models for Extracting Molecular Interactions and Pathway Knowledge
Gilchan Park, Byung-Jun Yoon, Xihaier Luo, Vanessa López-Marrero, Shinjae Yoo, Shantenu Jha
cs.CL, cs.LG
Understanding protein interactions and pathway knowledge is crucial for unraveling the complexities of living systems and investigating the underlying mechanisms of biological functions and complex diseases. While existing databases provide curated biological data from literature and other sources, they are often incom...
http://arxiv.org/pdf/2307.08813v2
2023-10-18T13:52:33
2023-07-17T20:01:11
{ "title": [ -0.04574585, -0.028518677, -0.0625, -0.07080078, -0.01889038, 0.0033283234, -0.07751465, 0.035583496, 0.02709961, 0.017364502, -0.014457703, -0.0007548332, -0.002292633, -0.026489258, -0.04019165, 0.016540527, 0.0012521744, 0.0168304...
http://arxiv.org/abs/2307.08767v1
A mixed policy to improve performance of language models on math problems
Gang Chen
cs.CL, cs.AI, cs.LG, 68T10, I.2.6
When to solve math problems, most language models take a sampling strategy to predict next word according conditional probabilities. In the math reasoning step, it may generate wrong answer. Considering math problems are deterministic, we propose a mixed policy exploration approach to solve math problems with reinforce...
http://arxiv.org/pdf/2307.08767v1
2023-07-17T18:27:49
2023-07-17T18:27:49
{ "title": [ -0.016311646, -0.008407593, -0.017044067, -0.028793335, -0.012451172, -0.022888184, -0.011711121, 0.0057029724, -0.020126343, 0.026626587, -0.018615723, 0.0087890625, -0.011772156, -0.023742676, -0.019561768, -0.0055389404, -0.007865906,...
http://arxiv.org/abs/2307.08701v4
AlpaGasus: Training A Better Alpaca with Fewer Data
Lichang Chen, Shiyang Li, Jun Yan, Hai Wang, Kalpa Gunaratna, Vikas Yadav, Zheng Tang, Vijay Srinivasan, Tianyi Zhou, Heng Huang, Hongxia Jin
cs.CL
Large language models~(LLMs) strengthen instruction-following capability through instruction-finetuning (IFT) on supervised instruction/response data. However, widely used IFT datasets (e.g., Alpaca's 52k data) surprisingly contain many low-quality instances with incorrect or irrelevant responses, which are misleading ...
http://arxiv.org/pdf/2307.08701v4
2023-11-04T21:39:59
2023-07-17T17:59:40
{ "title": [ -0.039520264, -0.014816284, -0.05114746, 0.013656616, -0.034118652, -0.030670166, -0.009689331, -0.01802063, -0.01448822, 0.024276733, -0.04550171, 0.009918213, -0.014465332, -0.03189087, 0.0005016327, -0.032684326, 0.0021839142, 0.0...
http://arxiv.org/abs/2307.08689v1
COLLIE: Systematic Construction of Constrained Text Generation Tasks
Shunyu Yao, Howard Chen, Austin W. Hanjie, Runzhe Yang, Karthik Narasimhan
cs.CL, cs.AI, cs.LG
Text generation under constraints have seen increasing interests in natural language processing, especially with the rapidly improving capabilities of large language models. However, existing benchmarks for constrained generation usually focus on fixed constraint types (e.g.,generate a sentence containing certain words...
http://arxiv.org/pdf/2307.08689v1
2023-07-17T17:48:51
2023-07-17T17:48:51
{ "title": [ -0.06768799, -0.022903442, -0.052368164, -0.034362793, -0.018371582, 0.02494812, -0.06341553, -0.0011501312, 0.006893158, 0.024963379, -0.005241394, 0.018966675, 0.010231018, -0.017120361, -0.029510498, 0.012916565, -0.028701782, 0.0...
http://arxiv.org/abs/2307.08678v1
Do Models Explain Themselves? Counterfactual Simulatability of Natural Language Explanations
Yanda Chen, Ruiqi Zhong, Narutatsu Ri, Chen Zhao, He He, Jacob Steinhardt, Zhou Yu, Kathleen McKeown
cs.CL, cs.AI, cs.LG
Large language models (LLMs) are trained to imitate humans to explain human decisions. However, do LLMs explain themselves? Can they help humans build mental models of how LLMs process different inputs? To answer these questions, we propose to evaluate $\textbf{counterfactual simulatability}$ of natural language explan...
http://arxiv.org/pdf/2307.08678v1
2023-07-17T17:41:47
2023-07-17T17:41:47
{ "title": [ -0.01965332, -0.013923645, -0.0024471283, -0.061401367, -0.02319336, 0.017074585, -0.055358887, -0.034332275, -0.002746582, 0.009925842, -0.018234253, -0.0055122375, -0.036254883, -0.014877319, -0.051483154, -0.009788513, -0.007648468, ...
http://arxiv.org/abs/2307.08655v1
Multilingual Speech-to-Speech Translation into Multiple Target Languages
Hongyu Gong, Ning Dong, Sravya Popuri, Vedanuj Goswami, Ann Lee, Juan Pino
cs.CL, cs.SD, eess.AS
Speech-to-speech translation (S2ST) enables spoken communication between people talking in different languages. Despite a few studies on multilingual S2ST, their focus is the multilinguality on the source side, i.e., the translation from multiple source languages to one target language. We present the first work on mul...
http://arxiv.org/pdf/2307.08655v1
2023-07-17T17:12:44
2023-07-17T17:12:44
{ "title": [ -0.01675415, -0.013153076, -0.031051636, -0.06561279, -0.005508423, -0.0038986206, -0.05392456, 0.012306213, -0.026184082, 0.05935669, 0.0054855347, 0.010482788, 0.025253296, -0.035614014, -0.020370483, 0.017105103, 0.008377075, 0.01...
http://arxiv.org/abs/2307.08621v4
Retentive Network: A Successor to Transformer for Large Language Models
Yutao Sun, Li Dong, Shaohan Huang, Shuming Ma, Yuqing Xia, Jilong Xue, Jianyong Wang, Furu Wei
cs.CL, cs.LG
In this work, we propose Retentive Network (RetNet) as a foundation architecture for large language models, simultaneously achieving training parallelism, low-cost inference, and good performance. We theoretically derive the connection between recurrence and attention. Then we propose the retention mechanism for sequen...
http://arxiv.org/pdf/2307.08621v4
2023-08-09T08:53:08
2023-07-17T16:40:01
{ "title": [ -0.023544312, -0.010009766, -0.03189087, -0.059326172, 0.00088119507, 0.0057792664, -0.05834961, -0.0110321045, 0.02243042, 0.02255249, -0.015701294, -0.005580902, 0.028717041, -0.023101807, -0.030426025, 0.024093628, -0.01335907, 0....
http://arxiv.org/abs/2307.08597v1
Multimodal Diffusion Segmentation Model for Object Segmentation from Manipulation Instructions
Yui Iioka, Yu Yoshida, Yuiga Wada, Shumpei Hatanaka, Komei Sugiura
cs.CV, cs.CL, cs.RO
In this study, we aim to develop a model that comprehends a natural language instruction (e.g., "Go to the living room and get the nearest pillow to the radio art on the wall") and generates a segmentation mask for the target everyday object. The task is challenging because it requires (1) the understanding of the refe...
http://arxiv.org/pdf/2307.08597v1
2023-07-17T16:07:07
2023-07-17T16:07:07
{ "title": [ -0.021438599, 0.017669678, -0.050567627, -0.013793945, 0.005897522, -0.021057129, -0.008758545, -0.016403198, -0.0011730194, 0.014160156, -0.022109985, 0.015548706, 0.003145218, -0.029281616, -0.0047416687, 0.0032157898, 0.0024471283, ...
http://arxiv.org/abs/2307.08586v1
Syntax-Aware Complex-Valued Neural Machine Translation
Yang Liu, Yuexian Hou
cs.CL
Syntax has been proven to be remarkably effective in neural machine translation (NMT). Previous models obtained syntax information from syntactic parsing tools and integrated it into NMT models to improve translation performance. In this work, we propose a method to incorporate syntax information into a complex-valued ...
http://arxiv.org/pdf/2307.08586v1
2023-07-17T15:58:05
2023-07-17T15:58:05
{ "title": [ -0.011375427, -0.016494751, 0.019210815, -0.050720215, -0.02331543, 0.025283813, -0.0446167, 0.006340027, 0.006877899, 0.058166504, -0.026855469, 0.00016951561, 0.029937744, -0.011154175, -0.04736328, 0.016571045, -0.0055007935, 0.00...
http://arxiv.org/abs/2307.08580v1
The Resume Paradox: Greater Language Differences, Smaller Pay Gaps
Joshua R. Minot, Marc Maier, Bradford Demarest, Nicholas Cheney, Christopher M. Danforth, Peter Sheridan Dodds, Morgan R. Frank
physics.soc-ph, cs.CL
Over the past decade, the gender pay gap has remained steady with women earning 84 cents for every dollar earned by men on average. Many studies explain this gap through demand-side bias in the labor market represented through employers' job postings. However, few studies analyze potential bias from the worker supply-s...
http://arxiv.org/pdf/2307.08580v1
2023-07-17T15:49:35
2023-07-17T15:49:35
{ "title": [ -0.018753052, 0.0044288635, -0.05215454, -0.022506714, 0.00617218, -0.031219482, -0.004550934, -0.023590088, -0.03277588, 0.013946533, -0.02482605, -0.011520386, -0.0022087097, -0.036865234, -0.027175903, -0.031311035, -0.011039734, ...