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/2304.10808v1
Downstream Task-Oriented Neural Tokenizer Optimization with Vocabulary Restriction as Post Processing
Tatsuya Hiraoka, Tomoya Iwakura
cs.CL
This paper proposes a method to optimize tokenization for the performance improvement of already trained downstream models. Our method generates tokenization results attaining lower loss values of a given downstream model on the training data for restricting vocabularies and trains a tokenizer reproducing the tokenizat...
http://arxiv.org/pdf/2304.10808v1
2023-04-21T08:29:14
2023-04-21T08:29:14
{ "title": [ -0.026367188, 0.011772156, -0.051574707, -0.06585693, 0.012420654, 0.004383087, -0.054138184, 0.0014772415, -0.012199402, 0.05404663, -0.013092041, 0.02041626, -0.0016937256, -0.0062446594, -0.046905518, 0.008430481, -0.008651733, 0....
http://arxiv.org/abs/2304.10785v2
Which Factors Predict the Chat Experience of a Natural Language Generation Dialogue Service?
Eason Chen
cs.CL, cs.HC
In this paper, we proposed a conceptual model to predict the chat experience in a natural language generation dialog system. We evaluated the model with 120 participants with Partial Least Squares Structural Equation Modeling (PLS-SEM) and obtained an R-square (R2) with 0.541. The model considers various factors, inclu...
http://arxiv.org/pdf/2304.10785v2
2023-05-28T02:26:24
2023-04-21T07:29:07
{ "title": [ 0.014129639, -0.0259552, -0.02645874, -0.023605347, -0.023788452, -0.017028809, 0.006729126, 0.024093628, 0.018554688, 0.024337769, -0.017501831, -0.024780273, 0.009162903, -0.04345703, -0.020492554, -0.013084412, -0.003578186, 0.036...
http://arxiv.org/abs/2304.10784v2
Eyettention: An Attention-based Dual-Sequence Model for Predicting Human Scanpaths during Reading
Shuwen Deng, David R. Reich, Paul Prasse, Patrick Haller, Tobias Scheffer, Lena A. Jäger
cs.CL, cs.HC, cs.LG
Eye movements during reading offer insights into both the reader's cognitive processes and the characteristics of the text that is being read. Hence, the analysis of scanpaths in reading have attracted increasing attention across fields, ranging from cognitive science over linguistics to computer science. In particular...
http://arxiv.org/pdf/2304.10784v2
2023-05-18T08:24:03
2023-04-21T07:26:49
{ "title": [ -0.0035762787, -0.012756348, 0.012901306, -0.020690918, -0.037139893, -0.011734009, -0.014976501, -0.029556274, 0.033813477, 0.051452637, -0.043914795, 0.003540039, 0.033996582, 0.0029697418, -0.0713501, -0.012969971, -0.0008120537, ...
http://arxiv.org/abs/2304.10759v1
GeoLayoutLM: Geometric Pre-training for Visual Information Extraction
Chuwei Luo, Changxu Cheng, Qi Zheng, Cong Yao
cs.CV, cs.CL
Visual information extraction (VIE) plays an important role in Document Intelligence. Generally, it is divided into two tasks: semantic entity recognition (SER) and relation extraction (RE). Recently, pre-trained models for documents have achieved substantial progress in VIE, particularly in SER. However, most of the e...
http://arxiv.org/pdf/2304.10759v1
2023-04-21T06:02:14
2023-04-21T06:02:14
{ "title": [ -0.013038635, 0.002046585, -0.030410767, -0.054107666, 0.008049011, 0.002603531, -0.019638062, -0.02696228, -0.00041770935, 0.029785156, -0.010055542, 0.012878418, -0.020629883, -0.01247406, -0.050567627, -0.0009994507, 0.02130127, 0...
http://arxiv.org/abs/2304.10750v1
Improving Grounded Language Understanding in a Collaborative Environment by Interacting with Agents Through Help Feedback
Nikhil Mehta, Milagro Teruel, Patricio Figueroa Sanz, Xin Deng, Ahmed Hassan Awadallah, Julia Kiseleva
cs.CL, cs.AI
Many approaches to Natural Language Processing (NLP) tasks often treat them as single-step problems, where an agent receives an instruction, executes it, and is evaluated based on the final outcome. However, human language is inherently interactive, as evidenced by the back-and-forth nature of human conversations. In l...
http://arxiv.org/pdf/2304.10750v1
2023-04-21T05:37:59
2023-04-21T05:37:59
{ "title": [ -0.031188965, -0.00025820732, -0.048797607, -0.048828125, -0.00006866455, -0.008369446, -0.04272461, 0.004398346, 0.0029659271, 0.024932861, -0.027069092, 0.0067977905, 0.009422302, -0.009117126, -0.045318604, -0.005180359, -0.0158844, ...
http://arxiv.org/abs/2304.10739v1
KitchenScale: Learning to predict ingredient quantities from recipe contexts
Donghee Choi, Mogan Gim, Samy Badreddine, Hajung Kim, Donghyeon Park, Jaewoo Kang
cs.CL, cs.AI, cs.LG, cs.NA, math.NA, H.4.m
Determining proper quantities for ingredients is an essential part of cooking practice from the perspective of enriching tastiness and promoting healthiness. We introduce KitchenScale, a fine-tuned Pre-trained Language Model (PLM) that predicts a target ingredient's quantity and measurement unit given its recipe contex...
http://arxiv.org/pdf/2304.10739v1
2023-04-21T04:28:16
2023-04-21T04:28:16
{ "title": [ -0.017044067, -0.008399963, -0.06616211, -0.07330322, 0.0101623535, -0.013031006, -0.046020508, 0.01474762, -0.0020008087, 0.03274536, -0.044128418, 0.034362793, -0.0057525635, 0.00063085556, -0.024627686, -0.0029335022, -0.005504608, ...
http://arxiv.org/abs/2304.10706v1
TC-GAT: Graph Attention Network for Temporal Causality Discovery
Xiaosong Yuan, Ke Chen, Wanli Zuo, Yijia Zhang
cs.CL
The present study explores the intricacies of causal relationship extraction, a vital component in the pursuit of causality knowledge. Causality is frequently intertwined with temporal elements, as the progression from cause to effect is not instantaneous but rather ensconced in a temporal dimension. Thus, the extracti...
http://arxiv.org/pdf/2304.10706v1
2023-04-21T02:26:42
2023-04-21T02:26:42
{ "title": [ 0.008522034, -0.004261017, -0.05819702, -0.09289551, 0.011878967, -0.018310547, -0.05038452, 0.00818634, 0.020645142, 0.022125244, 0.006084442, 0.014373779, 0.011810303, -0.026779175, -0.030334473, 0.018920898, -0.014831543, 0.051666...
http://arxiv.org/abs/2304.10703v1
ReCEval: Evaluating Reasoning Chains via Correctness and Informativeness
Archiki Prasad, Swarnadeep Saha, Xiang Zhou, Mohit Bansal
cs.CL, cs.AI, cs.LG
Multi-step reasoning ability is fundamental to many natural language tasks, yet it is unclear what constitutes a good reasoning chain and how to evaluate them. Most existing methods focus solely on whether the reasoning chain leads to the correct conclusion, but this answer-oriented view may confound the quality of rea...
http://arxiv.org/pdf/2304.10703v1
2023-04-21T02:19:06
2023-04-21T02:19:06
{ "title": [ -0.024978638, -0.0031604767, -0.055511475, -0.06982422, 0.010307312, 0.00092315674, -0.046295166, -0.011810303, 0.008148193, 0.011154175, -0.04083252, -0.008354187, -0.022857666, -0.015365601, -0.03640747, -0.02458191, -0.012016296, ...
http://arxiv.org/abs/2304.10663v1
Meta Semantics: Towards better natural language understanding and reasoning
Xiaolin Hu
cs.CL, cs.AI, 03B65(Primary) 68T50(Secondary), I.2.4; I.2.7
Natural language understanding is one of the most challenging topics in artificial intelligence. Deep neural network methods, particularly large language module (LLM) methods such as ChatGPT and GPT-3, have powerful flexibility to adopt informal text but are weak on logical deduction and suffer from the out-of-vocabula...
http://arxiv.org/pdf/2304.10663v1
2023-04-20T22:16:16
2023-04-20T22:16:16
{ "title": [ -0.022994995, -0.02583313, -0.025848389, -0.07727051, 0.016418457, 0.026107788, -0.050445557, -0.012619019, 0.020065308, 0.029785156, -0.014015198, 0.0033950806, 0.00087976456, -0.013832092, -0.047668457, 0.014907837, -0.010856628, 0...
http://arxiv.org/abs/2304.10642v1
Word Sense Induction with Knowledge Distillation from BERT
Anik Saha, Alex Gittens, Bulent Yener
cs.CL
Pre-trained contextual language models are ubiquitously employed for language understanding tasks, but are unsuitable for resource-constrained systems. Noncontextual word embeddings are an efficient alternative in these settings. Such methods typically use one vector to encode multiple different meanings of a word, and...
http://arxiv.org/pdf/2304.10642v1
2023-04-20T21:05:35
2023-04-20T21:05:35
{ "title": [ -0.017791748, -0.02192688, -0.044830322, -0.057403564, -0.016998291, -0.0049438477, -0.055786133, 0.012290955, 0.02204895, 0.02494812, -0.060638428, 0.0006184578, 0.011054993, -0.017089844, -0.0262146, -0.0022125244, -0.011100769, 0....
http://arxiv.org/abs/2304.10637v3
IXA/Cogcomp at SemEval-2023 Task 2: Context-enriched Multilingual Named Entity Recognition using Knowledge Bases
Iker García-Ferrero, Jon Ander Campos, Oscar Sainz, Ander Salaberria, Dan Roth
cs.CL
Named Entity Recognition (NER) is a core natural language processing task in which pre-trained language models have shown remarkable performance. However, standard benchmarks like CoNLL 2003 do not address many of the challenges that deployed NER systems face, such as having to classify emerging or complex entities in ...
http://arxiv.org/pdf/2304.10637v3
2023-04-27T20:51:36
2023-04-20T20:30:34
{ "title": [ -0.013153076, -0.00051259995, -0.04724121, -0.078186035, 0.02357483, 0.016921997, -0.05819702, 0.0014410019, 0.016540527, 0.033813477, -0.037017822, 0.023361206, 0.017974854, 0.0030136108, -0.018310547, 0.020141602, -0.011444092, 0.0...
http://arxiv.org/abs/2304.10619v1
"HOT" ChatGPT: The promise of ChatGPT in detecting and discriminating hateful, offensive, and toxic comments on social media
Lingyao Li, Lizhou Fan, Shubham Atreja, Libby Hemphill
cs.CL, cs.AI, cs.HC
Harmful content is pervasive on social media, poisoning online communities and negatively impacting participation. A common approach to address this issue is to develop detection models that rely on human annotations. However, the tasks required to build such models expose annotators to harmful and offensive content an...
http://arxiv.org/pdf/2304.10619v1
2023-04-20T19:40:51
2023-04-20T19:40:51
{ "title": [ 0.027801514, 0.002614975, -0.0309906, -0.008850098, 0.010154724, -0.030700684, 0.021469116, -0.0075531006, -0.023010254, 0.05886841, -0.05255127, 0.032928467, 0.04083252, -0.00028133392, -0.004234314, -0.019454956, -0.01687622, 0.087...
http://arxiv.org/abs/2304.10611v2
Joint Repetition Suppression and Content Moderation of Large Language Models
Minghui Zhang, Alex Sokolov, Weixin Cai, Si-Qing Chen
cs.CL, cs.LG
Natural language generation (NLG) is one of the most impactful fields in NLP, and recent years have witnessed its evolution brought about by large language models (LLMs). As the key instrument for writing assistance applications, they are generally prone to replicating or extending offensive content provided in the inp...
http://arxiv.org/pdf/2304.10611v2
2023-06-05T18:16:29
2023-04-20T19:17:49
{ "title": [ -0.0073394775, 0.008102417, -0.033203125, -0.054382324, 0.014411926, 0.00007545948, -0.03994751, -0.010108948, 0.010513306, 0.04937744, -0.002401352, -0.012001038, 0.042388916, -0.027755737, -0.031143188, 0.03439331, -0.027374268, 0....
http://arxiv.org/abs/2304.14339v1
MarsEclipse at SemEval-2023 Task 3: Multi-Lingual and Multi-Label Framing Detection with Contrastive Learning
Qisheng Liao, Meiting Lai, Preslav Nakov
cs.CL, cs.AI, cs.LG, cs.NE, 68T50, F.2.2; I.2.7
This paper describes our system for SemEval-2023 Task 3 Subtask 2 on Framing Detection. We used a multi-label contrastive loss for fine-tuning large pre-trained language models in a multi-lingual setting, achieving very competitive results: our system was ranked first on the official test set and on the official shared...
http://arxiv.org/pdf/2304.14339v1
2023-04-20T18:42:23
2023-04-20T18:42:23
{ "title": [ 0.018615723, -0.0016365051, -0.018325806, -0.026626587, 0.0005226135, -0.031951904, -0.025619507, -0.0340271, -0.0057373047, 0.028915405, -0.02519226, 0.015823364, 0.006511688, -0.025863647, -0.016052246, 0.0027713776, -0.020355225, ...
http://arxiv.org/abs/2304.11040v1
Emotional Expression Detection in Spoken Language Employing Machine Learning Algorithms
Mehrab Hosain, Most. Yeasmin Arafat, Gazi Zahirul Islam, Jia Uddin, Md. Mobarak Hossain, Fatema Alam
cs.SD, cs.CL, cs.LG, eess.AS
There are a variety of features of the human voice that can be classified as pitch, timbre, loudness, and vocal tone. It is observed in numerous incidents that human expresses their feelings using different vocal qualities when they are speaking. The primary objective of this research is to recognize different emotions...
http://arxiv.org/pdf/2304.11040v1
2023-04-20T17:57:08
2023-04-20T17:57:08
{ "title": [ -0.015838623, -0.0054473877, -0.038146973, -0.02432251, -0.0005478859, -0.027175903, -0.01727295, -0.008522034, 0.009727478, 0.030197144, -0.021911621, -0.02406311, 0.010673523, -0.02784729, -0.049835205, -0.005256653, -0.00026774406, ...
http://arxiv.org/abs/2304.10513v2
Why Does ChatGPT Fall Short in Providing Truthful Answers?
Shen Zheng, Jie Huang, Kevin Chen-Chuan Chang
cs.CL, cs.AI
Recent advancements in Large Language Models, such as ChatGPT, have demonstrated significant potential to impact various aspects of human life. However, ChatGPT still faces challenges in aspects like truthfulness, e.g. providing accurate and reliable outputs. Therefore, in this paper, we seek to understand why ChatGPT ...
http://arxiv.org/pdf/2304.10513v2
2023-05-24T04:58:47
2023-04-20T17:48:43
{ "title": [ -0.04510498, -0.0048675537, 0.0019893646, 0.049346924, -0.055389404, 0.018066406, -0.005847931, -0.054260254, -0.03744507, 0.021621704, -0.0093688965, 0.007698059, -0.03817749, -0.017440796, 0.00894165, -0.01499939, -0.011405945, 0.0...
http://arxiv.org/abs/2304.10512v1
"Can We Detect Substance Use Disorder?": Knowledge and Time Aware Classification on Social Media from Darkweb
Usha Lokala, Orchid Chetia Phukan, Triyasha Ghosh Dastidar, Francois Lamy, Raminta Daniulaityte, Amit Sheth
cs.LG, cs.CL, cs.SI
Opioid and substance misuse is rampant in the United States today, with the phenomenon known as the "opioid crisis". The relationship between substance use and mental health has been extensively studied, with one possible relationship being: substance misuse causes poor mental health. However, the lack of evidence on t...
http://arxiv.org/pdf/2304.10512v1
2023-04-20T17:47:13
2023-04-20T17:47:13
{ "title": [ 0.0440979, 0.032165527, -0.04232788, -0.066467285, 0.05496216, -0.031555176, -0.014533997, 0.0004196167, -0.0012140274, 0.0033092499, -0.041992188, 0.028686523, -0.02520752, -0.018295288, -0.027053833, -0.032318115, -0.0031967163, 0....
http://arxiv.org/abs/2304.10464v3
Learning to Program with Natural Language
Yiduo Guo, Yaobo Liang, Chenfei Wu, Wenshan Wu, Dongyan Zhao, Nan Duan
cs.CL
Large Language Models (LLMs) have shown remarkable performance in various basic natural language tasks, which raises hope for achieving Artificial General Intelligence. For completing the complex task, we still need a program for the task first and then ask LLMs to follow the program to generate the specific solution. ...
http://arxiv.org/pdf/2304.10464v3
2023-05-29T01:12:18
2023-04-20T17:09:12
{ "title": [ -0.032714844, -0.024841309, -0.024368286, -0.04547119, -0.012870789, 0.007247925, -0.028259277, 0.0065689087, -0.008232117, 0.016342163, -0.043273926, 0.00052404404, -0.040039062, -0.012817383, -0.032348633, -0.022476196, -0.0051994324, ...
http://arxiv.org/abs/2304.10453v1
Phoenix: Democratizing ChatGPT across Languages
Zhihong Chen, Feng Jiang, Junying Chen, Tiannan Wang, Fei Yu, Guiming Chen, Hongbo Zhang, Juhao Liang, Chen Zhang, Zhiyi Zhang, Jianquan Li, Xiang Wan, Benyou Wang, Haizhou Li
cs.CL, cs.AI
This paper presents our efforts to democratize ChatGPT across language. We release a large language model "Phoenix", achieving competitive performance among open-source English and Chinese models while excelling in languages with limited resources (covering both Latin and non-Latin languages). We believe this work will...
http://arxiv.org/pdf/2304.10453v1
2023-04-20T16:50:04
2023-04-20T16:50:04
{ "title": [ -0.03286743, -0.046295166, -0.016403198, -0.029800415, -0.040649414, 0.025024414, -0.058044434, 0.0025787354, -0.0043945312, 0.02168274, -0.015930176, 0.021255493, 0.0025482178, -0.026153564, -0.01991272, 0.028625488, -0.025268555, 0...
http://arxiv.org/abs/2304.10447v1
Domain-specific Continued Pretraining of Language Models for Capturing Long Context in Mental Health
Shaoxiong Ji, Tianlin Zhang, Kailai Yang, Sophia Ananiadou, Erik Cambria, Jörg Tiedemann
cs.CL
Pretrained language models have been used in various natural language processing applications. In the mental health domain, domain-specific language models are pretrained and released, which facilitates the early detection of mental health conditions. Social posts, e.g., on Reddit, are usually long documents. However, ...
http://arxiv.org/pdf/2304.10447v1
2023-04-20T16:43:56
2023-04-20T16:43:56
{ "title": [ 0.025787354, 0.0059127808, -0.02748108, -0.10443115, 0.015289307, -0.014755249, -0.06793213, 0.0033073425, -0.0014352798, 0.04159546, 0.0076560974, 0.0079574585, 0.03012085, -0.04598999, -0.016983032, 0.03036499, -0.0023441315, 0.020...
http://arxiv.org/abs/2304.10436v1
Safety Assessment of Chinese Large Language Models
Hao Sun, Zhexin Zhang, Jiawen Deng, Jiale Cheng, Minlie Huang
cs.CL
With the rapid popularity of large language models such as ChatGPT and GPT-4, a growing amount of attention is paid to their safety concerns. These models may generate insulting and discriminatory content, reflect incorrect social values, and may be used for malicious purposes such as fraud and dissemination of mislead...
http://arxiv.org/pdf/2304.10436v1
2023-04-20T16:27:35
2023-04-20T16:27:35
{ "title": [ -0.024337769, 0.0034122467, -0.03366089, -0.07965088, 0.0154418945, 0.0063095093, -0.05569458, 0.009086609, -0.014930725, 0.03817749, -0.011222839, -0.0018529892, 0.014137268, -0.030471802, -0.020080566, 0.009460449, -0.011268616, 0....
http://arxiv.org/abs/2304.10428v4
GPT-NER: Named Entity Recognition via Large Language Models
Shuhe Wang, Xiaofei Sun, Xiaoya Li, Rongbin Ouyang, Fei Wu, Tianwei Zhang, Jiwei Li, Guoyin Wang
cs.CL
Despite the fact that large-scale Language Models (LLM) have achieved SOTA performances on a variety of NLP tasks, its performance on NER is still significantly below supervised baselines. This is due to the gap between the two tasks the NER and LLMs: the former is a sequence labeling task in nature while the latter is...
http://arxiv.org/pdf/2304.10428v4
2023-10-07T14:25:28
2023-04-20T16:17:26
{ "title": [ -0.018218994, -0.004184723, -0.047088623, -0.081848145, 0.025543213, 0.007575989, -0.05029297, 0.0027256012, 0.016983032, 0.01637268, -0.03475952, 0.009414673, 0.006477356, -0.0035953522, -0.030899048, 0.006095886, -0.011695862, 0.04...
http://arxiv.org/abs/2304.10392v1
CKBP v2: An Expert-Annotated Evaluation Set for Commonsense Knowledge Base Population
Tianqing Fang, Quyet V. Do, Sehyun Choi, Weiqi Wang, Yangqiu Song
cs.CL
Populating Commonsense Knowledge Bases (CSKB) is an important yet hard task in NLP, as it tackles knowledge from external sources with unseen events and entities. Fang et al. (2021a) proposed a CSKB Population benchmark with an evaluation set CKBP v1. However, CKBP v1 adopts crowdsourced annotations that suffer from a ...
http://arxiv.org/pdf/2304.10392v1
2023-04-20T15:27:29
2023-04-20T15:27:29
{ "title": [ -0.013038635, -0.016204834, -0.03930664, -0.062438965, 0.009025574, 0.013504028, -0.037719727, -0.010322571, 0.03729248, 0.017288208, -0.014511108, 0.0018091202, -0.03111267, -0.01701355, -0.045532227, 0.012466431, -0.011543274, 0.03...
http://arxiv.org/abs/2304.10354v1
Prompt-Learning for Cross-Lingual Relation Extraction
Chiaming Hsu, Changtong Zan, Liang Ding, Longyue Wang, Xiaoting Wang, Weifeng Liu, Fu Lin, Wenbin Hu
cs.CL
Relation Extraction (RE) is a crucial task in Information Extraction, which entails predicting relationships between entities within a given sentence. However, extending pre-trained RE models to other languages is challenging, particularly in real-world scenarios where Cross-Lingual Relation Extraction (XRE) is require...
http://arxiv.org/pdf/2304.10354v1
2023-04-20T14:52:06
2023-04-20T14:52:06
{ "title": [ -0.0060157776, -0.004219055, -0.045043945, -0.07141113, 0.015388489, 0.0017433167, -0.0552063, 0.0041389465, 0.025238037, 0.011253357, -0.023147583, 0.0010070801, -0.0048599243, -0.012069702, -0.038970947, -0.0016527176, -0.010887146, ...
http://arxiv.org/abs/2304.11075v2
Spaiche: Extending State-of-the-Art ASR Models to Swiss German Dialects
Clement Sicard, Kajetan Pyszkowski, Victor Gillioz
cs.CL, cs.LG, cs.SD, eess.AS
Recent breakthroughs in NLP largely increased the presence of ASR systems in our daily lives. However, for many low-resource languages, ASR models still need to be improved due in part to the difficulty of acquiring pertinent data. This project aims to help advance research in ASR models for Swiss German dialects, by p...
http://arxiv.org/pdf/2304.11075v2
2023-09-13T16:12:56
2023-04-20T14:42:54
{ "title": [ -0.015586853, 0.0034160614, -0.062347412, -0.06561279, 0.013450623, 0.0069351196, -0.054260254, 0.026443481, -0.024978638, 0.05041504, 0.009094238, -0.018234253, 0.062408447, -0.044311523, -0.02784729, 0.018722534, -0.009475708, 0.05...
http://arxiv.org/abs/2304.10346v1
Interventional Probing in High Dimensions: An NLI Case Study
Julia Rozanova, Marco Valentino, Lucas Cordeiro, Andre Freitas
cs.CL
Probing strategies have been shown to detect the presence of various linguistic features in large language models; in particular, semantic features intermediate to the "natural logic" fragment of the Natural Language Inference task (NLI). In the case of natural logic, the relation between the intermediate features and ...
http://arxiv.org/pdf/2304.10346v1
2023-04-20T14:34:31
2023-04-20T14:34:31
{ "title": [ -0.021743774, -0.01335144, -0.019165039, 0.003194809, -0.03717041, -0.006263733, -0.024291992, -0.022415161, -0.0051116943, 0.024047852, -0.008155823, -0.015609741, 0.022323608, -0.040283203, -0.015579224, -0.009277344, -0.011917114, ...
http://arxiv.org/abs/2304.10327v2
Towards a Benchmark for Scientific Understanding in Humans and Machines
Kristian Gonzalez Barman, Sascha Caron, Tom Claassen, Henk de Regt
cs.AI, cs.CL, cs.HC, hep-ph, physics.hist-ph
Scientific understanding is a fundamental goal of science, allowing us to explain the world. There is currently no good way to measure the scientific understanding of agents, whether these be humans or Artificial Intelligence systems. Without a clear benchmark, it is challenging to evaluate and compare different levels...
http://arxiv.org/pdf/2304.10327v2
2023-04-21T08:57:06
2023-04-20T14:05:53
{ "title": [ -0.019470215, -0.0032482147, -0.024169922, -0.028686523, 0.003704071, -0.025772095, -0.018249512, -0.025543213, 0.0021152496, 0.02885437, -0.007381439, 0.001159668, -0.019378662, -0.023605347, -0.06713867, 0.005584717, 0.005279541, 0...
http://arxiv.org/abs/2304.11061v1
CEIL: A General Classification-Enhanced Iterative Learning Framework for Text Clustering
Mingjun Zhao, Mengzhen Wang, Yinglong Ma, Di Niu, Haijiang Wu
cs.CL, cs.AI
Text clustering, as one of the most fundamental challenges in unsupervised learning, aims at grouping semantically similar text segments without relying on human annotations. With the rapid development of deep learning, deep clustering has achieved significant advantages over traditional clustering methods. Despite the...
http://arxiv.org/pdf/2304.11061v1
2023-04-20T14:04:31
2023-04-20T14:04:31
{ "title": [ -0.025558472, 0.0005850792, -0.04446411, -0.04083252, 0.013069153, -0.0029773712, -0.030700684, -0.018936157, 0.020843506, 0.035705566, -0.022109985, 0.041046143, -0.007385254, 0.0073165894, -0.0440979, 0.013069153, -0.015914917, 0.0...
http://arxiv.org/abs/2304.10321v1
DropDim: A Regularization Method for Transformer Networks
Hao Zhang, Dan Qu, Keji Shao, Xukui Yang
cs.CL, cs.CV
We introduceDropDim, a structured dropout method designed for regularizing the self-attention mechanism, which is a key component of the transformer. In contrast to the general dropout method, which randomly drops neurons, DropDim drops part of the embedding dimensions. In this way, the semantic information can be comp...
http://arxiv.org/pdf/2304.10321v1
2023-04-20T13:54:18
2023-04-20T13:54:18
{ "title": [ -0.030136108, 0.013214111, -0.0501709, -0.049316406, 0.03451538, 0.008255005, -0.03665161, -0.013916016, -0.011116028, 0.0289917, 0.0010766983, 0.017623901, 0.012557983, -0.03048706, -0.02078247, 0.025924683, 0.011528015, 0.02470398,...
http://arxiv.org/abs/2304.10309v1
Improving Speech Translation by Cross-Modal Multi-Grained Contrastive Learning
Hao Zhang, Nianwen Si, Yaqi Chen, Wenlin Zhang, Xukui Yang, Dan Qu, Wei-Qiang Zhang
cs.CL, cs.CV
The end-to-end speech translation (E2E-ST) model has gradually become a mainstream paradigm due to its low latency and less error propagation. However, it is non-trivial to train such a model well due to the task complexity and data scarcity. The speech-and-text modality differences result in the E2E-ST model performan...
http://arxiv.org/pdf/2304.10309v1
2023-04-20T13:41:56
2023-04-20T13:41:56
{ "title": [ -0.006603241, 0.012008667, -0.01991272, -0.039489746, -0.0048332214, -0.004299164, -0.040771484, -0.015045166, 0.0070724487, 0.04837036, -0.017166138, 0.019042969, 0.016693115, -0.01927185, -0.050628662, 0.002067566, -0.00983429, 0.0...
http://arxiv.org/abs/2304.10295v1
Decouple Non-parametric Knowledge Distillation For End-to-end Speech Translation
Hao Zhang, Nianwen Si, Yaqi Chen, Wenlin Zhang, Xukui Yang, Dan Qu, Zhen Li
cs.CL, cs.SD, eess.AS
Existing techniques often attempt to make knowledge transfer from a powerful machine translation (MT) to speech translation (ST) model with some elaborate techniques, which often requires transcription as extra input during training. However, transcriptions are not always available, and how to improve the ST model perf...
http://arxiv.org/pdf/2304.10295v1
2023-04-20T13:20:03
2023-04-20T13:20:03
{ "title": [ -0.019729614, -0.0038604736, -0.045837402, -0.09338379, -0.014572144, -0.0047569275, -0.082214355, 0.032440186, 0.00844574, 0.046020508, -0.013755798, -0.0206604, 0.055267334, -0.03527832, -0.028396606, 0.003255844, 0.008811951, 0.00...
http://arxiv.org/abs/2304.10283v1
Is augmentation effective to improve prediction in imbalanced text datasets?
Gabriel O. Assunção, Rafael Izbicki, Marcos O. Prates
cs.CL, stat.ML
Imbalanced datasets present a significant challenge for machine learning models, often leading to biased predictions. To address this issue, data augmentation techniques are widely used in natural language processing (NLP) to generate new samples for the minority class. However, in this paper, we challenge the common a...
http://arxiv.org/pdf/2304.10283v1
2023-04-20T13:07:31
2023-04-20T13:07:31
{ "title": [ 0.010437012, 0.016067505, -0.035949707, -0.0287323, -0.005378723, -0.019134521, -0.021408081, -0.018173218, 0.014663696, 0.042388916, -0.053466797, 0.023727417, 0.037841797, 0.003282547, -0.031463623, -0.0047302246, -0.027679443, 0.0...
http://arxiv.org/abs/2304.10256v1
Indian Sign Language Recognition Using Mediapipe Holistic
Dr. Velmathi G, Kaushal Goyal
cs.CV, cs.CL, cs.LG
Deaf individuals confront significant communication obstacles on a daily basis. Their inability to hear makes it difficult for them to communicate with those who do not understand sign language. Moreover, it presents difficulties in educational, occupational, and social contexts. By providing alternative communication ...
http://arxiv.org/pdf/2304.10256v1
2023-04-20T12:25:47
2023-04-20T12:25:47
{ "title": [ -0.025863647, 0.016998291, -0.059417725, -0.028808594, -0.024551392, -0.013122559, -0.053894043, -0.013900757, 0.0073280334, 0.014778137, -0.051696777, -0.0067749023, -0.017318726, -0.031280518, -0.019714355, -0.032562256, -0.018600464, ...
http://arxiv.org/abs/2304.10220v1
Effective Open Intent Classification with K-center Contrastive Learning and Adjustable Decision Boundary
Xiaokang Liu, Jianquan Li, Jingjing Mu, Min Yang, Ruifeng Xu, Benyou Wang
cs.CL, 68T01, I.2.1
Open intent classification, which aims to correctly classify the known intents into their corresponding classes while identifying the new unknown (open) intents, is an essential but challenging task in dialogue systems. In this paper, we introduce novel K-center contrastive learning and adjustable decision boundary lea...
http://arxiv.org/pdf/2304.10220v1
2023-04-20T11:35:06
2023-04-20T11:35:06
{ "title": [ 0.0048332214, 0.019821167, -0.026611328, -0.03286743, 0.014122009, -0.021636963, -0.0027275085, -0.040100098, 0.025054932, 0.021377563, -0.040740967, 0.015914917, -0.013458252, -0.020462036, -0.03289795, -0.012527466, -0.014602661, 0...
http://arxiv.org/abs/2304.10216v1
Exploring Paracrawl for Document-level Neural Machine Translation
Yusser Al Ghussin, Jingyi Zhang, Josef van Genabith
cs.CL
Document-level neural machine translation (NMT) has outperformed sentence-level NMT on a number of datasets. However, document-level NMT is still not widely adopted in real-world translation systems mainly due to the lack of large-scale general-domain training data for document-level NMT. We examine the effectiveness o...
http://arxiv.org/pdf/2304.10216v1
2023-04-20T11:21:34
2023-04-20T11:21:34
{ "title": [ -0.018234253, -0.01499939, -0.010307312, -0.06347656, -0.0073928833, 0.013381958, -0.030654907, 0.014953613, 0.0017137527, 0.045532227, -0.023132324, -0.000019311905, 0.012794495, -0.004425049, -0.0435791, 0.0014162064, -0.0066604614, ...
http://arxiv.org/abs/2304.10195v1
CoT-MoTE: Exploring ConTextual Masked Auto-Encoder Pre-training with Mixture-of-Textual-Experts for Passage Retrieval
Guangyuan Ma, Xing Wu, Peng Wang, Songlin Hu
cs.CL, cs.IR
Passage retrieval aims to retrieve relevant passages from large collections of the open-domain corpus. Contextual Masked Auto-Encoding has been proven effective in representation bottleneck pre-training of a monolithic dual-encoder for passage retrieval. Siamese or fully separated dual-encoders are often adopted as bas...
http://arxiv.org/pdf/2304.10195v1
2023-04-20T10:12:09
2023-04-20T10:12:09
{ "title": [ -0.007633209, 0.013191223, -0.033050537, -0.026062012, -0.0059928894, -0.016708374, -0.0143966675, -0.024032593, 0.01285553, 0.06518555, -0.014038086, 0.031433105, 0.061279297, 0.00030303, -0.031402588, 0.030761719, -0.00048589706, 0...
http://arxiv.org/abs/2304.11073v3
OLISIA: a Cascade System for Spoken Dialogue State Tracking
Léo Jacqmin, Lucas Druart, Yannick Estève, Benoît Favre, Lina Maria Rojas-Barahona, Valentin Vielzeuf
eess.AS, cs.AI, cs.CL, cs.SD
Though Dialogue State Tracking (DST) is a core component of spoken dialogue systems, recent work on this task mostly deals with chat corpora, disregarding the discrepancies between spoken and written language.In this paper, we propose OLISIA, a cascade system which integrates an Automatic Speech Recognition (ASR) model...
http://arxiv.org/pdf/2304.11073v3
2023-08-31T08:51:33
2023-04-20T09:30:50
{ "title": [ -0.008514404, -0.010185242, -0.0440979, -0.08618164, 0.006126404, 0.00012588501, -0.062316895, 0.010444641, 0.0043144226, 0.008361816, -0.033111572, 0.0008149147, -0.010658264, -0.020980835, -0.046569824, -0.01737976, -0.007713318, 0...
http://arxiv.org/abs/2304.10164v1
Analyzing FOMC Minutes: Accuracy and Constraints of Language Models
Wonseong Kim, Jan Frederic Spörer, Siegfried Handschuh
cs.CL
This research article analyzes the language used in the official statements released by the Federal Open Market Committee (FOMC) after its scheduled meetings to gain insights into the impact of FOMC official statements on financial markets and economic forecasting. The study reveals that the FOMC is careful to avoid ex...
http://arxiv.org/pdf/2304.10164v1
2023-04-20T08:54:00
2023-04-20T08:54:00
{ "title": [ -0.03048706, -0.0025405884, -0.047454834, -0.025878906, -0.011405945, -0.01309967, -0.04071045, -0.014083862, -0.011474609, 0.029373169, -0.021102905, -0.006752014, -0.0012683868, -0.021316528, -0.037475586, -0.0032596588, -0.022888184, ...
http://arxiv.org/abs/2304.10158v1
Does Manipulating Tokenization Aid Cross-Lingual Transfer? A Study on POS Tagging for Non-Standardized Languages
Verena Blaschke, Hinrich Schütze, Barbara Plank
cs.CL
One of the challenges with finetuning pretrained language models (PLMs) is that their tokenizer is optimized for the language(s) it was pretrained on, but brittle when it comes to previously unseen variations in the data. This can for instance be observed when finetuning PLMs on one language and evaluating them on data...
http://arxiv.org/pdf/2304.10158v1
2023-04-20T08:32:34
2023-04-20T08:32:34
{ "title": [ -0.0055503845, 0.011390686, -0.016815186, -0.033081055, 0.002714157, -0.01171875, -0.02519226, -0.034729004, -0.018875122, 0.019302368, -0.036834717, 0.035736084, 0.01071167, -0.0098724365, -0.016571045, -0.0093688965, -0.012489319, ...
http://arxiv.org/abs/2304.10153v1
On the Independence of Association Bias and Empirical Fairness in Language Models
Laura Cabello, Anna Katrine Jørgensen, Anders Søgaard
cs.CL, cs.HC
The societal impact of pre-trained language models has prompted researchers to probe them for strong associations between protected attributes and value-loaded terms, from slur to prestigious job titles. Such work is said to probe models for bias or fairness-or such probes 'into representational biases' are said to be ...
http://arxiv.org/pdf/2304.10153v1
2023-04-20T08:27:21
2023-04-20T08:27:21
{ "title": [ -0.031555176, -0.010108948, -0.061523438, -0.047332764, 0.0033435822, -0.010559082, -0.04510498, 0.0029296875, 0.0037231445, 0.031433105, -0.02848816, -0.004497528, 0.02406311, -0.014892578, -0.046142578, 0.016143799, -0.036865234, 0...
http://arxiv.org/abs/2304.10145v2
Can ChatGPT Reproduce Human-Generated Labels? A Study of Social Computing Tasks
Yiming Zhu, Peixian Zhang, Ehsan-Ul Haq, Pan Hui, Gareth Tyson
cs.AI, cs.CL
The release of ChatGPT has uncovered a range of possibilities whereby large language models (LLMs) can substitute human intelligence. In this paper, we seek to understand whether ChatGPT has the potential to reproduce human-generated label annotations in social computing tasks. Such an achievement could significantly r...
http://arxiv.org/pdf/2304.10145v2
2023-04-22T08:55:33
2023-04-20T08:08:12
{ "title": [ -0.0024299622, -0.0048065186, 0.0024814606, -0.004676819, -0.008773804, -0.0034255981, -0.00093460083, -0.029769897, -0.02482605, 0.027633667, -0.0077323914, 0.031677246, -0.007850647, -0.008888245, -0.01184845, 0.011047363, -0.036346436...
http://arxiv.org/abs/2304.11095v1
Is Cross-modal Information Retrieval Possible without Training?
Hyunjin Choi, Hyunjae Lee, Seongho Joe, Youngjune L. Gwon
cs.LG, cs.AI, cs.CL, cs.CV, cs.IR
Encoded representations from a pretrained deep learning model (e.g., BERT text embeddings, penultimate CNN layer activations of an image) convey a rich set of features beneficial for information retrieval. Embeddings for a particular modality of data occupy a high-dimensional space of its own, but it can be semanticall...
http://arxiv.org/pdf/2304.11095v1
2023-04-20T02:36:18
2023-04-20T02:36:18
{ "title": [ 0.023269653, 0.016952515, -0.0047683716, 0.015380859, -0.018066406, -0.032409668, 0.008361816, -0.048858643, 0.011680603, 0.023834229, -0.031951904, -0.018798828, 0.0033283234, -0.02192688, 0.0036373138, -0.02229309, 0.004436493, 0.0...
http://arxiv.org/abs/2304.09991v2
Supporting Human-AI Collaboration in Auditing LLMs with LLMs
Charvi Rastogi, Marco Tulio Ribeiro, Nicholas King, Saleema Amershi
cs.HC, cs.AI, cs.CL
Large language models are becoming increasingly pervasive and ubiquitous in society via deployment in sociotechnical systems. Yet these language models, be it for classification or generation, have been shown to be biased and behave irresponsibly, causing harm to people at scale. It is crucial to audit these language m...
http://arxiv.org/pdf/2304.09991v2
2023-08-18T20:09:46
2023-04-19T21:59:04
{ "title": [ -0.052703857, -0.014122009, -0.06048584, -0.0042381287, -0.060760498, -0.011161804, -0.03942871, 0.0018005371, -0.0064430237, 0.025543213, -0.079833984, 0.0036907196, 0.012901306, -0.0058517456, -0.016952515, -0.043914795, -0.013008118, ...
http://arxiv.org/abs/2304.09982v1
Radar de Parité: An NLP system to measure gender representation in French news stories
Valentin-Gabriel Soumah, Prashanth Rao, Philipp Eibl, Maite Taboada
cs.CL
We present the Radar de Parit\'e, an automated Natural Language Processing (NLP) system that measures the proportion of women and men quoted daily in six Canadian French-language media outlets. We outline the system's architecture and detail the challenges we overcame to address French-specific issues, in particular re...
http://arxiv.org/pdf/2304.09982v1
2023-04-19T21:33:59
2023-04-19T21:33:59
{ "title": [ -0.05734253, -0.020095825, -0.054718018, -0.010368347, -0.020645142, -0.010032654, -0.030334473, 0.0008625984, -0.009407043, 0.029159546, -0.046722412, 0.001455307, 0.021133423, -0.0007047653, -0.043914795, -0.021759033, -0.039398193, ...
http://arxiv.org/abs/2304.09972v2
MasakhaNEWS: News Topic Classification for African languages
David Ifeoluwa Adelani, Marek Masiak, Israel Abebe Azime, Jesujoba Alabi, Atnafu Lambebo Tonja, Christine Mwase, Odunayo Ogundepo, Bonaventure F. P. Dossou, Akintunde Oladipo, Doreen Nixdorf, Chris Chinenye Emezue, sana al-azzawi, Blessing Sibanda, Davis David, Lolwethu Ndolela, Jonathan Mukiibi, Tunde Ajayi, Tatiana M...
cs.CL
African languages are severely under-represented in NLP research due to lack of datasets covering several NLP tasks. While there are individual language specific datasets that are being expanded to different tasks, only a handful of NLP tasks (e.g. named entity recognition and machine translation) have standardized ben...
http://arxiv.org/pdf/2304.09972v2
2023-09-20T17:14:40
2023-04-19T21:12:23
{ "title": [ -0.029830933, -0.035003662, -0.010322571, -0.019317627, -0.045410156, 0.011146545, -0.042175293, -0.016021729, -0.025970459, 0.04269409, 0.016494751, 0.036376953, 0.022384644, -0.04626465, -0.011619568, 0.027694702, -0.0064735413, 0....
http://arxiv.org/abs/2304.09960v3
A Latent Space Theory for Emergent Abilities in Large Language Models
Hui Jiang
cs.CL, cs.AI, cs.LG
Languages are not created randomly but rather to communicate information. There is a strong association between languages and their underlying meanings, resulting in a sparse joint distribution that is heavily peaked according to their correlations. Moreover, these peak values happen to match with the marginal distribu...
http://arxiv.org/pdf/2304.09960v3
2023-09-13T18:52:02
2023-04-19T20:45:01
{ "title": [ -0.022491455, -0.010818481, -0.039886475, -0.03756714, -0.010795593, -0.018463135, -0.0062942505, 0.015350342, -0.0044441223, 0.03286743, -0.018081665, 0.016113281, 0.003698349, -0.008171082, -0.04864502, 0.004890442, -0.018951416, 0...
http://arxiv.org/abs/2304.09957v1
Low-resource Bilingual Dialect Lexicon Induction with Large Language Models
Ekaterina Artemova, Barbara Plank
cs.CL
Bilingual word lexicons are crucial tools for multilingual natural language understanding and machine translation tasks, as they facilitate the mapping of words in one language to their synonyms in another language. To achieve this, numerous papers have explored bilingual lexicon induction (BLI) in high-resource scenar...
http://arxiv.org/pdf/2304.09957v1
2023-04-19T20:20:41
2023-04-19T20:20:41
{ "title": [ -0.029052734, -0.01247406, -0.026916504, -0.060394287, -0.010398865, 0.025024414, -0.07324219, 0.013969421, 0.0005736351, 0.043640137, -0.021270752, 0.00207901, 0.0041618347, -0.01247406, -0.04815674, 0.0032978058, -0.011505127, 0.05...
http://arxiv.org/abs/2304.09948v1
Catch Me If You Can: Identifying Fraudulent Physician Reviews with Large Language Models Using Generative Pre-Trained Transformers
Aishwarya Deep Shukla, Laksh Agarwal, Jie Mein, Goh, Guodong, Gao, Ritu Agarwal
cs.CL, cs.AI
The proliferation of fake reviews of doctors has potentially detrimental consequences for patient well-being and has prompted concern among consumer protection groups and regulatory bodies. Yet despite significant advancements in the fields of machine learning and natural language processing, there remains limited comp...
http://arxiv.org/pdf/2304.09948v1
2023-04-19T19:59:26
2023-04-19T19:59:26
{ "title": [ 0.020980835, -0.009841919, -0.0071105957, -0.040496826, -0.013031006, -0.0074539185, -0.03152466, -0.039001465, -0.021453857, 0.012962341, -0.036499023, 0.026290894, -0.016113281, -0.014877319, -0.009269714, -0.026519775, -0.0178833, ...
http://arxiv.org/abs/2304.09919v1
The eBible Corpus: Data and Model Benchmarks for Bible Translation for Low-Resource Languages
Vesa Akerman, David Baines, Damien Daspit, Ulf Hermjakob, Taeho Jang, Colin Leong, Michael Martin, Joel Mathew, Jonathan Robie, Marcus Schwarting
cs.CL, cs.AI
Efficiently and accurately translating a corpus into a low-resource language remains a challenge, regardless of the strategies employed, whether manual, automated, or a combination of the two. Many Christian organizations are dedicated to the task of translating the Holy Bible into languages that lack a modern translat...
http://arxiv.org/pdf/2304.09919v1
2023-04-19T18:52:49
2023-04-19T18:52:49
{ "title": [ -0.032318115, 0.002822876, -0.032104492, -0.040740967, -0.0061798096, -0.0037059784, -0.047332764, -0.018585205, 0.0062675476, 0.05532837, -0.013008118, 0.0053253174, 0.06213379, -0.016448975, -0.044189453, 0.02494812, 0.005580902, 0...
http://arxiv.org/abs/2304.11046v1
Affective social anthropomorphic intelligent system
Md. Adyelullahil Mamun, Hasnat Md. Abdullah, Md. Golam Rabiul Alam, Muhammad Mehedi Hassan, Md. Zia Uddin
cs.SD, cs.AI, cs.CL, cs.HC, cs.LG
Human conversational styles are measured by the sense of humor, personality, and tone of voice. These characteristics have become essential for conversational intelligent virtual assistants. However, most of the state-of-the-art intelligent virtual assistants (IVAs) are failed to interpret the affective semantics of hu...
http://arxiv.org/pdf/2304.11046v1
2023-04-19T18:24:57
2023-04-19T18:24:57
{ "title": [ -0.009147644, 0.009277344, -0.021743774, -0.012825012, -0.012550354, -0.023635864, -0.0025157928, -0.03237915, 0.01576233, 0.035339355, -0.033996582, -0.018341064, 0.033294678, -0.024932861, -0.02268982, 0.01058197, 0.0021800995, 0.0...
http://arxiv.org/abs/2304.11053v1
A Comparison of Semi-Supervised Learning Techniques for Streaming ASR at Scale
Cal Peyser, Michael Picheny, Kyunghyun Cho, Rohit Prabhavalkar, Ronny Huang, Tara Sainath
cs.CL, cs.SD, eess.AS
Unpaired text and audio injection have emerged as dominant methods for improving ASR performance in the absence of a large labeled corpus. However, little guidance exists on deploying these methods to improve production ASR systems that are trained on very large supervised corpora and with realistic requirements like a...
http://arxiv.org/pdf/2304.11053v1
2023-04-19T18:09:27
2023-04-19T18:09:27
{ "title": [ -0.008712769, 0.0039138794, -0.06750488, -0.086364746, 0.010635376, -0.0053596497, -0.05670166, 0.040771484, -0.025283813, 0.030258179, -0.037506104, 0.002916336, -0.003929138, -0.016647339, -0.023986816, -0.024002075, 0.0042152405, ...
http://arxiv.org/abs/2304.09848v2
Evaluating Verifiability in Generative Search Engines
Nelson F. Liu, Tianyi Zhang, Percy Liang
cs.CL, cs.IR
Generative search engines directly generate responses to user queries, along with in-line citations. A prerequisite trait of a trustworthy generative search engine is verifiability, i.e., systems should cite comprehensively (high citation recall; all statements are fully supported by citations) and accurately (high cit...
http://arxiv.org/pdf/2304.09848v2
2023-10-23T19:11:38
2023-04-19T17:56:12
{ "title": [ 0.02368164, -0.005542755, -0.029693604, -0.066711426, 0.018981934, -0.020065308, -0.015670776, -0.0048294067, -0.0127334595, 0.018295288, -0.009651184, 0.005428314, -0.026107788, -0.024551392, -0.04522705, -0.012023926, -0.040405273, ...
http://arxiv.org/abs/2304.11082v4
Fundamental Limitations of Alignment in Large Language Models
Yotam Wolf, Noam Wies, Oshri Avnery, Yoav Levine, Amnon Shashua
cs.CL, cs.AI
An important aspect in developing language models that interact with humans is aligning their behavior to be useful and unharmful for their human users. This is usually achieved by tuning the model in a way that enhances desired behaviors and inhibits undesired ones, a process referred to as alignment. In this paper, w...
http://arxiv.org/pdf/2304.11082v4
2023-10-11T09:45:15
2023-04-19T17:50:09
{ "title": [ -0.040374756, -0.0049552917, -0.0440979, -0.058624268, -0.0029506683, 0.005962372, -0.06530762, 0.011383057, 0.018081665, 0.04748535, -0.015197754, -0.014060974, 0.025177002, -0.01259613, -0.05581665, 0.023788452, -0.006259918, 0.040...
http://arxiv.org/abs/2304.09842v3
Chameleon: Plug-and-Play Compositional Reasoning with Large Language Models
Pan Lu, Baolin Peng, Hao Cheng, Michel Galley, Kai-Wei Chang, Ying Nian Wu, Song-Chun Zhu, Jianfeng Gao
cs.CL, cs.AI, cs.CV, cs.LG
Large language models (LLMs) have achieved remarkable progress in solving various natural language processing tasks due to emergent reasoning abilities. However, LLMs have inherent limitations as they are incapable of accessing up-to-date information (stored on the Web or in task-specific knowledge bases), using extern...
http://arxiv.org/pdf/2304.09842v3
2023-10-31T17:43:39
2023-04-19T17:47:47
{ "title": [ -0.03414917, -0.026016235, -0.048339844, -0.11248779, 0.0033302307, 0.032073975, -0.11303711, 0.028335571, -0.0013523102, 0.013206482, -0.009643555, 0.0077705383, -0.0075187683, -0.025634766, -0.019973755, 0.0027618408, -0.02229309, ...
http://arxiv.org/abs/2304.09805v1
A Survey of Corpora for Germanic Low-Resource Languages and Dialects
Verena Blaschke, Hinrich Schütze, Barbara Plank
cs.CL
Despite much progress in recent years, the vast majority of work in natural language processing (NLP) is on standard languages with many speakers. In this work, we instead focus on low-resource languages and in particular non-standardized low-resource languages. Even within branches of major language families, often co...
http://arxiv.org/pdf/2304.09805v1
2023-04-19T16:45:16
2023-04-19T16:45:16
{ "title": [ -0.0103302, -0.0043792725, -0.03753662, -0.018173218, -0.015914917, -0.012763977, -0.031463623, -0.0017986298, -0.004512787, 0.05117798, -0.025238037, 0.0041618347, 0.05368042, -0.026565552, -0.010368347, 0.019744873, 0.009429932, 0....
http://arxiv.org/abs/2304.09797v5
Progressive-Hint Prompting Improves Reasoning in Large Language Models
Chuanyang Zheng, Zhengying Liu, Enze Xie, Zhenguo Li, Yu Li
cs.CL, cs.LG
The performance of Large Language Models (LLMs) in reasoning tasks depends heavily on prompt design, with Chain-of-Thought (CoT) and self-consistency being critical methods that enhance this ability. However, these methods do not fully exploit the answers generated by the LLM to guide subsequent responses. This paper p...
http://arxiv.org/pdf/2304.09797v5
2023-08-10T03:41:04
2023-04-19T16:29:48
{ "title": [ -0.010108948, -0.004611969, -0.038238525, -0.09844971, -0.00009447336, -0.000018179417, -0.0713501, 0.022216797, 0.019226074, 0.0015487671, -0.03274536, -0.006340027, 0.012580872, -0.023498535, -0.01525116, -0.0006008148, 0.006702423, ...
http://arxiv.org/abs/2304.11062v1
Scaling Transformer to 1M tokens and beyond with RMT
Aydar Bulatov, Yuri Kuratov, Mikhail S. Burtsev
cs.CL, cs.AI, cs.LG
This technical report presents the application of a recurrent memory to extend the context length of BERT, one of the most effective Transformer-based models in natural language processing. By leveraging the Recurrent Memory Transformer architecture, we have successfully increased the model's effective context length t...
http://arxiv.org/pdf/2304.11062v1
2023-04-19T16:18:54
2023-04-19T16:18:54
{ "title": [ 0.026321411, -0.0037269592, -0.039794922, -0.046844482, -0.021591187, -0.016601562, -0.044708252, -0.006931305, -0.02268982, 0.015434265, -0.030975342, 0.00051116943, 0.045654297, -0.04650879, 0.04397583, -0.02067566, 0.0037670135, 0...
http://arxiv.org/abs/2304.09667v3
GeneGPT: Augmenting Large Language Models with Domain Tools for Improved Access to Biomedical Information
Qiao Jin, Yifan Yang, Qingyu Chen, Zhiyong Lu
cs.CL, cs.AI, q-bio.GN
While large language models (LLMs) have been successfully applied to various tasks, they still face challenges with hallucinations. Augmenting LLMs with domain-specific tools such as database utilities can facilitate easier and more precise access to specialized knowledge. In this paper, we present GeneGPT, a novel met...
http://arxiv.org/pdf/2304.09667v3
2023-05-16T13:24:53
2023-04-19T13:53:19
{ "title": [ -0.022979736, -0.021347046, -0.046417236, -0.07745361, -0.000037789345, 0.010444641, -0.06317139, 0.003227234, 0.022476196, 0.02746582, 0.00818634, -0.016067505, 0.0184021, -0.031463623, -0.043182373, 0.027938843, -0.010108948, 0.022...
http://arxiv.org/abs/2304.09660v1
MPMQA: Multimodal Question Answering on Product Manuals
Liang Zhang, Anwen Hu, Jing Zhang, Shuo Hu, Qin Jin
cs.CL
Visual contents, such as illustrations and images, play a big role in product manual understanding. Existing Product Manual Question Answering (PMQA) datasets tend to ignore visual contents and only retain textual parts. In this work, to emphasize the importance of multimodal contents, we propose a Multimodal Product M...
http://arxiv.org/pdf/2304.09660v1
2023-04-19T13:48:14
2023-04-19T13:48:14
{ "title": [ -0.05734253, -0.011764526, -0.04849243, -0.032165527, -0.012184143, -0.005382538, -0.044952393, 0.00045895576, 0.014533997, 0.01838684, -0.01083374, 0.0057640076, -0.0012702942, -0.009971619, -0.039764404, -0.008560181, -0.014694214, ...
http://arxiv.org/abs/2304.09649v1
BRENT: Bidirectional Retrieval Enhanced Norwegian Transformer
Lucas Georges Gabriel Charpentier, Sondre Wold, David Samuel, Egil Rønningstad
cs.CL
Retrieval-based language models are increasingly employed in question-answering tasks. These models search in a corpus of documents for relevant information instead of having all factual knowledge stored in its parameters, thereby enhancing efficiency, transparency, and adaptability. We develop the first Norwegian retr...
http://arxiv.org/pdf/2304.09649v1
2023-04-19T13:40:47
2023-04-19T13:40:47
{ "title": [ 0.034606934, 0.011459351, -0.00856781, -0.035003662, -0.005558014, -0.015167236, -0.0075035095, -0.061798096, -0.0011014938, 0.0501709, -0.0015144348, -0.010635376, 0.057403564, -0.045440674, -0.008415222, 0.02279663, -0.008201599, 0...
http://arxiv.org/abs/2304.09616v2
Bridging Natural Language Processing and Psycholinguistics: computationally grounded semantic similarity datasets for Basque and Spanish
J. Goikoetxea, M. Arantzeta, I. San Martin
cs.CL
We present a computationally-grounded word similarity dataset based on two well-known Natural Language Processing resources; text corpora and knowledge bases. This dataset aims to fulfil a gap in psycholinguistic research by providing a variety of quantifications of semantic similarity in an extensive set of noun pairs...
http://arxiv.org/pdf/2304.09616v2
2023-04-20T08:23:21
2023-04-19T12:47:51
{ "title": [ -0.027114868, -0.003583908, -0.01739502, -0.045562744, -0.009010315, 0.011383057, -0.0513916, -0.0040130615, 0.008956909, 0.037597656, -0.03387451, 0.017852783, 0.030212402, -0.00881958, -0.035705566, 0.0064201355, -0.014137268, 0.04...
http://arxiv.org/abs/2304.09607v2
CB-Conformer: Contextual biasing Conformer for biased word recognition
Yaoxun Xu, Baiji Liu, Qiaochu Huang and, Xingchen Song, Zhiyong Wu, Shiyin Kang, Helen Meng
cs.SD, cs.CL, eess.AS
Due to the mismatch between the source and target domains, how to better utilize the biased word information to improve the performance of the automatic speech recognition model in the target domain becomes a hot research topic. Previous approaches either decode with a fixed external language model or introduce a sizea...
http://arxiv.org/pdf/2304.09607v2
2023-04-25T07:45:40
2023-04-19T12:26:04
{ "title": [ -0.022323608, 0.012084961, -0.056854248, -0.04473877, -0.0064735413, -0.011802673, -0.034210205, -0.00013899803, 0.016525269, 0.06286621, -0.049591064, 0.0009870529, 0.030975342, 0.0053138733, -0.054840088, -0.003829956, -0.023117065, ...
http://arxiv.org/abs/2304.09582v1
Is ChatGPT Equipped with Emotional Dialogue Capabilities?
Weixiang Zhao, Yanyan Zhao, Xin Lu, Shilong Wang, Yanpeng Tong, Bing Qin
cs.CL
This report presents a study on the emotional dialogue capability of ChatGPT, an advanced language model developed by OpenAI. The study evaluates the performance of ChatGPT on emotional dialogue understanding and generation through a series of experiments on several downstream tasks. Our findings indicate that while Ch...
http://arxiv.org/pdf/2304.09582v1
2023-04-19T11:42:40
2023-04-19T11:42:40
{ "title": [ -0.004508972, -0.014465332, 0.0025253296, 0.019882202, -0.038635254, -0.0042152405, -0.0015668869, -0.024337769, 0.0062332153, 0.017593384, 0.00096035004, -0.0018672943, 0.00067281723, -0.042236328, 0.006416321, 0.017166138, -0.009590149...
http://arxiv.org/abs/2304.09563v1
On the Robustness of Aspect-based Sentiment Analysis: Rethinking Model, Data, and Training
Hao Fei, Tat-Seng Chua, Chenliang Li, Donghong Ji, Meishan Zhang, Yafeng Ren
cs.CL
Aspect-based sentiment analysis (ABSA) aims at automatically inferring the specific sentiment polarities toward certain aspects of products or services behind the social media texts or reviews, which has been a fundamental application to the real-world society. Since the early 2010s, ABSA has achieved extraordinarily h...
http://arxiv.org/pdf/2304.09563v1
2023-04-19T11:07:43
2023-04-19T11:07:43
{ "title": [ 0.007709503, -0.0061950684, -0.017852783, -0.053527832, 0.020233154, 0.00013267994, -0.014472961, -0.011779785, 0.023086548, 0.045776367, -0.010665894, -0.011474609, -0.0011587143, -0.006515503, -0.08166504, 0.008232117, -0.010536194, ...
http://arxiv.org/abs/2304.09548v3
SemEval 2023 Task 6: LegalEval - Understanding Legal Texts
Ashutosh Modi, Prathamesh Kalamkar, Saurabh Karn, Aman Tiwari, Abhinav Joshi, Sai Kiran Tanikella, Shouvik Kumar Guha, Sachin Malhan, Vivek Raghavan
cs.CL, cs.AI, cs.LG
In populous countries, pending legal cases have been growing exponentially. There is a need for developing NLP-based techniques for processing and automatically understanding legal documents. To promote research in the area of Legal NLP we organized the shared task LegalEval - Understanding Legal Texts at SemEval 2023....
http://arxiv.org/pdf/2304.09548v3
2023-05-01T11:33:08
2023-04-19T10:28:32
{ "title": [ -0.01751709, 0.0011587143, -0.043060303, -0.01739502, 0.0009832382, -0.009788513, 0.0012226105, -0.0073432922, 0.0044059753, 0.048919678, -0.009239197, 0.005970001, 0.044281006, -0.03781128, 0.0037670135, 0.00022053719, 0.0069732666, ...
http://arxiv.org/abs/2304.09542v2
Is ChatGPT Good at Search? Investigating Large Language Models as Re-Ranking Agents
Weiwei Sun, Lingyong Yan, Xinyu Ma, Shuaiqiang Wang, Pengjie Ren, Zhumin Chen, Dawei Yin, Zhaochun Ren
cs.CL, cs.IR
Large Language Models (LLMs) have demonstrated remarkable zero-shot generalization across various language-related tasks, including search engines. However, existing work utilizes the generative ability of LLMs for Information Retrieval (IR) rather than direct passage ranking. The discrepancy between the pre-training o...
http://arxiv.org/pdf/2304.09542v2
2023-10-27T12:11:16
2023-04-19T10:16:03
{ "title": [ -0.0075302124, -0.010902405, 0.005958557, -0.010383606, -0.016525269, 0.00919342, -0.013832092, -0.030807495, -0.001455307, 0.042541504, -0.024475098, 0.010520935, 0.015899658, 0.007965088, -0.0423584, 0.00970459, -0.04196167, 0.0337...
http://arxiv.org/abs/2304.09516v2
Controlling keywords and their positions in text generation
Yuichi Sasazawa, Terufumi Morishita, Hiroaki Ozaki, Osamu Imaichi, Yasuhiro Sogawa
cs.CL
One of the challenges in text generation is to control text generation as intended by the user. Previous studies proposed specifying the keywords that should be included in the generated text. However, this approach is insufficient to generate text that reflect the user's intent. For example, placing an important keywo...
http://arxiv.org/pdf/2304.09516v2
2023-10-31T09:34:30
2023-04-19T09:11:45
{ "title": [ -0.015975952, -0.005393982, -0.022277832, -0.010604858, -0.032989502, -0.0118255615, -0.009529114, -0.023513794, -0.011985779, 0.07556152, -0.014556885, 0.016815186, 0.02861023, -0.009300232, -0.037231445, 0.014312744, -0.006801605, ...
http://arxiv.org/abs/2304.09493v1
Emotion fusion for mental illness detection from social media: A survey
Tianlin Zhang, Kailai Yang, Shaoxiong Ji, Sophia Ananiadou
cs.CL, cs.AI
Mental illnesses are one of the most prevalent public health problems worldwide, which negatively influence people's lives and society's health. With the increasing popularity of social media, there has been a growing research interest in the early detection of mental illness by analysing user-generated posts on social...
http://arxiv.org/pdf/2304.09493v1
2023-04-19T08:28:34
2023-04-19T08:28:34
{ "title": [ 0.02281189, 0.019180298, -0.010238647, -0.040374756, 0.039916992, -0.006790161, 0.009513855, -0.028366089, 0.0025596619, 0.03765869, -0.028717041, 0.011108398, 0.028320312, -0.020965576, -0.028778076, -0.00022387505, -0.009819031, 0....
http://arxiv.org/abs/2304.11065v1
Conversational Process Modelling: State of the Art, Applications, and Implications in Practice
Nataliia Klievtsova, Janik-Vasily Benzin, Timotheus Kampik, Juergen Mangler, Stefanie Rinderle-Ma
cs.CL, cs.AI
Chatbots such as ChatGPT have caused a tremendous hype lately. For BPM applications, it is often not clear how to apply chatbots to generate business value. Hence, this work aims at the systematic analysis of existing chatbots for their support of conversational process modelling as process-oriented capability. Applica...
http://arxiv.org/pdf/2304.11065v1
2023-04-19T06:54:14
2023-04-19T06:54:14
{ "title": [ -0.013282776, 0.003835678, -0.04647827, -0.07366943, 0.012611389, 0.009513855, -0.078430176, -0.0181427, 0.007972717, 0.036010742, 0.014640808, -0.008644104, 0.009986877, -0.039276123, -0.05581665, 0.023406982, -0.022750854, 0.031555...
http://arxiv.org/abs/2304.09448v1
EC^2: Emergent Communication for Embodied Control
Yao Mu, Shunyu Yao, Mingyu Ding, Ping Luo, Chuang Gan
cs.LG, cs.CL, cs.CV
Embodied control requires agents to leverage multi-modal pre-training to quickly learn how to act in new environments, where video demonstrations contain visual and motion details needed for low-level perception and control, and language instructions support generalization with abstract, symbolic structures. While rece...
http://arxiv.org/pdf/2304.09448v1
2023-04-19T06:36:02
2023-04-19T06:36:02
{ "title": [ -0.04208374, -0.008766174, -0.015464783, -0.01184845, -0.0041923523, 0.006416321, -0.0051994324, -0.03845215, -0.00983429, 0.03289795, 0.0027942657, -0.001742363, 0.022598267, -0.028060913, -0.029678345, 0.006412506, 0.0040245056, 0....
http://arxiv.org/abs/2304.09433v2
Language Models Enable Simple Systems for Generating Structured Views of Heterogeneous Data Lakes
Simran Arora, Brandon Yang, Sabri Eyuboglu, Avanika Narayan, Andrew Hojel, Immanuel Trummer, Christopher Ré
cs.CL
A long standing goal of the data management community is to develop general, automated systems that ingest semi-structured documents and output queryable tables without human effort or domain specific customization. Given the sheer variety of potential documents, state-of-the art systems make simplifying assumptions an...
http://arxiv.org/pdf/2304.09433v2
2023-04-20T04:12:38
2023-04-19T06:00:26
{ "title": [ 0.00040006638, -0.04046631, -0.06591797, -0.11767578, -0.045196533, 0.011543274, -0.13049316, 0.04727173, 0.038085938, 0.044433594, -0.03062439, -0.015609741, 0.0040359497, -0.029434204, -0.033691406, -0.026672363, 0.0062713623, -0.0...
http://arxiv.org/abs/2304.09421v1
TieFake: Title-Text Similarity and Emotion-Aware Fake News Detection
Quanjiang Guo, Zhao Kang, Ling Tian, Zhouguo Chen
cs.CL, cs.CV, cs.LG, cs.SI
Fake news detection aims to detect fake news widely spreading on social media platforms, which can negatively influence the public and the government. Many approaches have been developed to exploit relevant information from news images, text, or videos. However, these methods may suffer from the following limitations: ...
http://arxiv.org/pdf/2304.09421v1
2023-04-19T04:47:36
2023-04-19T04:47:36
{ "title": [ -0.0127334595, 0.001917839, -0.018234253, -0.009742737, 0.004802704, -0.010047913, 0.021896362, -0.023773193, -0.022140503, 0.034423828, -0.023956299, 0.03753662, -0.0009241104, -0.0039787292, -0.03475952, -0.008308411, -0.043518066, ...
http://arxiv.org/abs/2304.09406v1
How to Do Things with Deep Learning Code
Minh Hua, Rita Raley
cs.CL
The premise of this article is that a basic understanding of the composition and functioning of large language models is critically urgent. To that end, we extract a representational map of OpenAI's GPT-2 with what we articulate as two classes of deep learning code, that which pertains to the model and that which under...
http://arxiv.org/pdf/2304.09406v1
2023-04-19T03:46:12
2023-04-19T03:46:12
{ "title": [ 0.010215759, -0.008720398, -0.024749756, -0.03463745, -0.03994751, -0.02619934, -0.03201294, -0.0023040771, 0.0032291412, 0.034698486, -0.024902344, -0.020187378, -0.0051116943, -0.032226562, -0.0181427, -0.020721436, 0.034118652, -0...
http://arxiv.org/abs/2304.13140v1
ESimCSE Unsupervised Contrastive Learning Jointly with UDA Semi-Supervised Learning for Large Label System Text Classification Mode
Ruan Lu, Zhou HangCheng, Ran Meng, Zhao Jin, Qin JiaoYu, Wei Feng, Wang ChenZi
cs.LG, cs.CL
The challenges faced by text classification with large tag systems in natural language processing tasks include multiple tag systems, uneven data distribution, and high noise. To address these problems, the ESimCSE unsupervised comparative learning and UDA semi-supervised comparative learning models are combined throug...
http://arxiv.org/pdf/2304.13140v1
2023-04-19T03:44:23
2023-04-19T03:44:23
{ "title": [ -0.028015137, 0.016113281, -0.04324341, -0.011161804, 0.0075263977, -0.026016235, -0.0033550262, -0.014129639, -0.0019207001, 0.024963379, -0.027008057, -0.0048828125, 0.018005371, -0.042663574, -0.011444092, -0.028137207, -0.018722534, ...
http://arxiv.org/abs/2304.09402v2
MixPro: Simple yet Effective Data Augmentation for Prompt-based Learning
Bohan Li, Longxu Dou, Yutai Hou, Yunlong Feng, Honglin Mu, Qingfu Zhu, Qinghua Sun, Wanxiang Che
cs.CL, cs.LG
Prompt-based learning has shown considerable promise in reformulating various downstream tasks as cloze problems by combining original input with a predetermined template. This approach demonstrates its effectiveness, especially in few-shot learning scenarios, where the model is trained on a scarce amount of data. Desp...
http://arxiv.org/pdf/2304.09402v2
2023-11-11T15:15:26
2023-04-19T03:38:25
{ "title": [ -0.021697998, 0.0052261353, -0.08508301, -0.05355835, 0.027633667, -0.029830933, -0.018157959, -0.011543274, -0.009918213, 0.008850098, -0.04473877, 0.02178955, -0.018722534, -0.026809692, -0.011444092, -0.007801056, -0.014907837, 0....
http://arxiv.org/abs/2304.09388v1
An Empirical Study of Leveraging Knowledge Distillation for Compressing Multilingual Neural Machine Translation Models
Varun Gumma, Raj Dabre, Pratyush Kumar
cs.CL, cs.AI
Knowledge distillation (KD) is a well-known method for compressing neural models. However, works focusing on distilling knowledge from large multilingual neural machine translation (MNMT) models into smaller ones are practically nonexistent, despite the popularity and superiority of MNMT. This paper bridges this gap by...
http://arxiv.org/pdf/2304.09388v1
2023-04-19T02:57:55
2023-04-19T02:57:55
{ "title": [ -0.04849243, -0.01864624, -0.031555176, -0.044677734, -0.018051147, 0.024383545, -0.06237793, 0.019927979, 0.012008667, 0.04675293, -0.004512787, -0.024780273, 0.01184845, -0.01979065, -0.06732178, -0.005004883, 0.016433716, -0.00479...
http://arxiv.org/abs/2304.09374v1
Shuffle & Divide: Contrastive Learning for Long Text
Joonseok Lee, Seongho Joe, Kyoungwon Park, Bogun Kim, Hoyoung Kang, Jaeseon Park, Youngjune Gwon
cs.CL, cs.LG
We propose a self-supervised learning method for long text documents based on contrastive learning. A key to our method is Shuffle and Divide (SaD), a simple text augmentation algorithm that sets up a pretext task required for contrastive updates to BERT-based document embedding. SaD splits a document into two sub-docu...
http://arxiv.org/pdf/2304.09374v1
2023-04-19T02:02:29
2023-04-19T02:02:29
{ "title": [ -0.013114929, -0.0062294006, -0.02180481, -0.03857422, 0.0027618408, -0.01109314, -0.0066070557, -0.016708374, 0.023925781, 0.02041626, -0.044677734, 0.01902771, -0.011871338, -0.0052223206, -0.042999268, -0.011512756, -0.020446777, ...
http://arxiv.org/abs/2304.09349v4
LLM as A Robotic Brain: Unifying Egocentric Memory and Control
Jinjie Mai, Jun Chen, Bing Li, Guocheng Qian, Mohamed Elhoseiny, Bernard Ghanem
cs.AI, cs.CL, cs.RO
Embodied AI focuses on the study and development of intelligent systems that possess a physical or virtual embodiment (i.e. robots) and are able to dynamically interact with their environment. Memory and control are the two essential parts of an embodied system and usually require separate frameworks to model each of t...
http://arxiv.org/pdf/2304.09349v4
2023-06-12T14:07:42
2023-04-19T00:08:48
{ "title": [ -0.025817871, -0.0041007996, -0.03277588, -0.006454468, -0.026824951, -0.020370483, -0.006969452, -0.035705566, -0.029876709, 0.0317688, -0.019104004, 0.0073890686, 0.00049448013, -0.027038574, -0.022735596, -0.008102417, 0.0072517395, ...
http://arxiv.org/abs/2304.09339v1
The Unintended Consequences of Censoring Digital Technology -- Evidence from Italy's ChatGPT Ban
David H. Kreitmeir, Paul A. Raschky
econ.GN, cs.CL, q-fin.EC
We analyse the effects of the ban of ChatGPT, a generative pre-trained transformer chatbot, on individual productivity. We first compile data on the hourly coding output of over 8,000 professional GitHub users in Italy and other European countries to analyse the impact of the ban on individual productivity. Combining t...
http://arxiv.org/pdf/2304.09339v1
2023-04-18T23:11:04
2023-04-18T23:11:04
{ "title": [ 0.0011949539, -0.003698349, -0.018356323, 0.009223938, 0.003768921, -0.024765015, 0.022521973, -0.029159546, -0.051849365, 0.042266846, -0.011619568, 0.024841309, -0.00674057, -0.016784668, -0.024551392, -0.015960693, -0.013999939, 0...
http://arxiv.org/abs/2304.09185v1
Token Imbalance Adaptation for Radiology Report Generation
Yuexin Wu, I-Chan Huang, Xiaolei Huang
cs.CL, cs.AI
Imbalanced token distributions naturally exist in text documents, leading neural language models to overfit on frequent tokens. The token imbalance may dampen the robustness of radiology report generators, as complex medical terms appear less frequently but reflect more medical information. In this study, we demonstrat...
http://arxiv.org/pdf/2304.09185v1
2023-04-18T23:09:36
2023-04-18T23:09:36
{ "title": [ -0.0041542053, 0.0008878708, -0.04849243, -0.04925537, -0.00090646744, 0.0027542114, -0.033966064, -0.0042800903, 0.021606445, 0.033325195, -0.031433105, -0.020126343, 0.0020885468, -0.027679443, -0.05899048, -0.01171875, -0.018081665, ...
http://arxiv.org/abs/2304.09333v1
BIM-GPT: a Prompt-Based Virtual Assistant Framework for BIM Information Retrieval
Junwen Zheng, Martin Fischer
cs.CL, cs.HC
Efficient information retrieval (IR) from building information models (BIMs) poses significant challenges due to the necessity for deep BIM knowledge or extensive engineering efforts for automation. We introduce BIM-GPT, a prompt-based virtual assistant (VA) framework integrating BIM and generative pre-trained transfor...
http://arxiv.org/pdf/2304.09333v1
2023-04-18T22:46:02
2023-04-18T22:46:02
{ "title": [ -0.028427124, -0.021957397, -0.05883789, -0.049194336, -0.015655518, -0.0028877258, -0.032409668, 0.022247314, 0.033843994, 0.04837036, -0.035186768, -0.013259888, 0.040527344, -0.028625488, -0.06317139, 0.0021076202, -0.011680603, 0...
http://arxiv.org/abs/2304.09276v1
A Neural Lambda Calculus: Neurosymbolic AI meets the foundations of computing and functional programming
João Flach, Luis C. Lamb
cs.LG, cs.AI, cs.CL, cs.LO
Over the last decades, deep neural networks based-models became the dominant paradigm in machine learning. Further, the use of artificial neural networks in symbolic learning has been seen as increasingly relevant recently. To study the capabilities of neural networks in the symbolic AI domain, researchers have explore...
http://arxiv.org/pdf/2304.09276v1
2023-04-18T20:30:16
2023-04-18T20:30:16
{ "title": [ -0.018493652, -0.0065689087, -0.018798828, -0.042266846, -0.0013608932, -0.015823364, -0.008636475, -0.0052986145, -0.0073165894, 0.025222778, -0.027130127, 0.008270264, -0.0015926361, -0.020904541, -0.026885986, -0.002380371, 0.00083589...
http://arxiv.org/abs/2304.12203v1
Creating Large Language Model Resistant Exams: Guidelines and Strategies
Simon kaare Larsen
cs.CL
The proliferation of Large Language Models (LLMs), such as ChatGPT, has raised concerns about their potential impact on academic integrity, prompting the need for LLM-resistant exam designs. This article investigates the performance of LLMs on exams and their implications for assessment, focusing on ChatGPT's abilities...
http://arxiv.org/pdf/2304.12203v1
2023-04-18T18:01:32
2023-04-18T18:01:32
{ "title": [ -0.0011739731, -0.004184723, -0.004501343, -0.042297363, -0.0131073, -0.0021419525, -0.01776123, -0.02041626, -0.020950317, 0.026153564, 0.020858765, -0.012939453, -0.01109314, -0.04449463, -0.02720642, 0.0066375732, -0.011604309, 0....
http://arxiv.org/abs/2304.09151v1
UniMax: Fairer and more Effective Language Sampling for Large-Scale Multilingual Pretraining
Hyung Won Chung, Noah Constant, Xavier Garcia, Adam Roberts, Yi Tay, Sharan Narang, Orhan Firat
cs.CL
Pretrained multilingual large language models have typically used heuristic temperature-based sampling to balance between different languages. However previous work has not systematically evaluated the efficacy of different pretraining language distributions across model scales. In this paper, we propose a new sampling...
http://arxiv.org/pdf/2304.09151v1
2023-04-18T17:45:50
2023-04-18T17:45:50
{ "title": [ -0.026504517, -0.009895325, -0.045196533, -0.062072754, -0.0028686523, -0.0021438599, -0.057128906, -0.0016174316, -0.026168823, 0.024032593, -0.020736694, 0.04144287, 0.004863739, -0.023544312, -0.005760193, 0.005004883, -0.009399414, ...
http://arxiv.org/abs/2304.09145v3
Outlier Suppression+: Accurate quantization of large language models by equivalent and optimal shifting and scaling
Xiuying Wei, Yunchen Zhang, Yuhang Li, Xiangguo Zhang, Ruihao Gong, Jinyang Guo, Xianglong Liu
cs.CL
Post-training quantization~(PTQ) of transformer language models faces significant challenges due to the existence of detrimental outliers in activations. We observe that these outliers are concentrated in specific channels and are asymmetric across channels. To address this issue, we propose the Outlier Suppression+~(O...
http://arxiv.org/pdf/2304.09145v3
2023-10-23T08:48:31
2023-04-18T17:34:23
{ "title": [ -0.038604736, 0.0043945312, -0.049560547, -0.036499023, 0.0062065125, -0.0061035156, -0.048461914, -0.02859497, -0.013244629, 0.06008911, -0.0038032532, 0.004711151, 0.018218994, -0.022155762, -0.051116943, 0.00712204, -0.0070724487, ...
http://arxiv.org/abs/2304.09138v1
Exploring the Trade-Offs: Unified Large Language Models vs Local Fine-Tuned Models for Highly-Specific Radiology NLI Task
Zihao Wu, Lu Zhang, Chao Cao, Xiaowei Yu, Haixing Dai, Chong Ma, Zhengliang Liu, Lin Zhao, Gang Li, Wei Liu, Quanzheng Li, Dinggang Shen, Xiang Li, Dajiang Zhu, Tianming Liu
cs.CL
Recently, ChatGPT and GPT-4 have emerged and gained immense global attention due to their unparalleled performance in language processing. Despite demonstrating impressive capability in various open-domain tasks, their adequacy in highly specific fields like radiology remains untested. Radiology presents unique linguis...
http://arxiv.org/pdf/2304.09138v1
2023-04-18T17:21:48
2023-04-18T17:21:48
{ "title": [ -0.01789856, -0.015541077, -0.00894165, -0.045928955, -0.03604126, 0.009086609, -0.05899048, -0.004463196, -0.00598526, 0.028533936, -0.0052986145, -0.0146484375, -0.014022827, -0.034576416, -0.024932861, -0.011070251, 0.0015268326, ...
http://arxiv.org/abs/2304.09116v3
NaturalSpeech 2: Latent Diffusion Models are Natural and Zero-Shot Speech and Singing Synthesizers
Kai Shen, Zeqian Ju, Xu Tan, Yanqing Liu, Yichong Leng, Lei He, Tao Qin, Sheng Zhao, Jiang Bian
eess.AS, cs.AI, cs.CL, cs.LG, cs.SD
Scaling text-to-speech (TTS) to large-scale, multi-speaker, and in-the-wild datasets is important to capture the diversity in human speech such as speaker identities, prosodies, and styles (e.g., singing). Current large TTS systems usually quantize speech into discrete tokens and use language models to generate these t...
http://arxiv.org/pdf/2304.09116v3
2023-05-30T16:09:10
2023-04-18T16:31:59
{ "title": [ -0.03479004, -0.011947632, -0.048858643, 0.006164551, -0.03390503, -0.015823364, -0.03387451, 0.014846802, -0.004119873, 0.04006958, -0.019851685, 0.035339355, -0.029159546, -0.008880615, -0.012168884, -0.013282776, -0.0063209534, 0....
http://arxiv.org/abs/2304.11063v1
Think Before You Act: Unified Policy for Interleaving Language Reasoning with Actions
Lina Mezghani, Piotr Bojanowski, Karteek Alahari, Sainbayar Sukhbaatar
cs.CL, cs.AI
The success of transformer models trained with a language modeling objective brings a promising opportunity to the reinforcement learning framework. Decision Transformer is a step towards this direction, showing how to train transformers with a similar next-step prediction objective on offline data. Another important d...
http://arxiv.org/pdf/2304.11063v1
2023-04-18T16:12:38
2023-04-18T16:12:38
{ "title": [ -0.010566711, -0.024307251, -0.037719727, -0.07989502, 0.008979797, -0.0015659332, -0.03466797, 0.0066452026, -0.017837524, 0.008056641, -0.02507019, 0.017974854, -0.019744873, -0.019515991, -0.012397766, -0.020248413, -0.011360168, ...
http://arxiv.org/abs/2304.09058v2
Revisiting k-NN for Fine-tuning Pre-trained Language Models
Lei Li, Jing Chen, Bozhong Tian, Ningyu Zhang
cs.CL, cs.AI, cs.DB, cs.IR, cs.LG
Pre-trained Language Models (PLMs), as parametric-based eager learners, have become the de-facto choice for current paradigms of Natural Language Processing (NLP). In contrast, k-Nearest-Neighbor (kNN) classifiers, as the lazy learning paradigm, tend to mitigate over-fitting and isolated noise. In this paper, we revisi...
http://arxiv.org/pdf/2304.09058v2
2023-06-18T02:51:29
2023-04-18T15:28:47
{ "title": [ -0.036193848, -0.02571106, -0.017562866, -0.062805176, -0.026016235, 0.017547607, -0.06866455, 0.0034770966, -0.01184082, 0.039764404, -0.017166138, 0.008132935, 0.009346008, -0.007572174, -0.040863037, 0.010215759, -0.0054092407, 0....
http://arxiv.org/abs/2304.09048v1
CodeKGC: Code Language Model for Generative Knowledge Graph Construction
Zhen Bi, Jing Chen, Yinuo Jiang, Feiyu Xiong, Wei Guo, Huajun Chen, Ningyu Zhang
cs.CL, cs.AI, cs.IR, cs.LG, cs.SE
Current generative knowledge graph construction approaches usually fail to capture structural knowledge by simply flattening natural language into serialized texts or a specification language. However, large generative language model trained on structured data such as code has demonstrated impressive capability in unde...
http://arxiv.org/pdf/2304.09048v1
2023-04-18T15:12:34
2023-04-18T15:12:34
{ "title": [ -0.05819702, -0.05001831, -0.060058594, -0.07507324, -0.032470703, 0.024765015, -0.07940674, 0.024108887, 0.014419556, 0.024597168, 0.003572464, 0.008705139, -0.009529114, -0.027374268, -0.03277588, 0.006580353, -0.0016498566, 0.0256...
http://arxiv.org/abs/2304.08999v1
A Biomedical Entity Extraction Pipeline for Oncology Health Records in Portuguese
Hugo Sousa, Arian Pasquali, Alípio Jorge, Catarina Sousa Santos, Mário Amorim Lopes
cs.CL, cs.AI
Textual health records of cancer patients are usually protracted and highly unstructured, making it very time-consuming for health professionals to get a complete overview of the patient's therapeutic course. As such limitations can lead to suboptimal and/or inefficient treatment procedures, healthcare providers would ...
http://arxiv.org/pdf/2304.08999v1
2023-04-18T14:10:34
2023-04-18T14:10:34
{ "title": [ -0.023513794, -0.0010213852, -0.05871582, -0.07885742, 0.014091492, 0.010231018, -0.047576904, 0.008491516, -0.018829346, 0.034820557, -0.010757446, -0.00052785873, 0.03201294, -0.020553589, -0.027145386, 0.008712769, -0.007820129, 0...
http://arxiv.org/abs/2304.08991v1
D2CSE: Difference-aware Deep continuous prompts for Contrastive Sentence Embeddings
Hyunjae Lee
cs.CL, cs.AI
This paper describes Difference-aware Deep continuous prompt for Contrastive Sentence Embeddings (D2CSE) that learns sentence embeddings. Compared to state-of-the-art approaches, D2CSE computes sentence vectors that are exceptional to distinguish a subtle difference in similar sentences by employing a simple neural arc...
http://arxiv.org/pdf/2304.08991v1
2023-04-18T13:45:07
2023-04-18T13:45:07
{ "title": [ 0.015808105, 0.008781433, -0.028579712, -0.05734253, 0.013626099, -0.012550354, -0.014045715, -0.00548172, 0.024642944, 0.015457153, -0.04071045, 0.03111267, 0.0043792725, 0.00029969215, -0.03475952, 0.0005335808, -0.020690918, 0.048...
http://arxiv.org/abs/2304.08981v2
MER 2023: Multi-label Learning, Modality Robustness, and Semi-Supervised Learning
Zheng Lian, Haiyang Sun, Licai Sun, Kang Chen, Mingyu Xu, Kexin Wang, Ke Xu, Yu He, Ying Li, Jinming Zhao, Ye Liu, Bin Liu, Jiangyan Yi, Meng Wang, Erik Cambria, Guoying Zhao, Björn W. Schuller, Jianhua Tao
cs.CL, cs.CV
The first Multimodal Emotion Recognition Challenge (MER 2023) was successfully held at ACM Multimedia. The challenge focuses on system robustness and consists of three distinct tracks: (1) MER-MULTI, where participants are required to recognize both discrete and dimensional emotions; (2) MER-NOISE, in which noise is ad...
http://arxiv.org/pdf/2304.08981v2
2023-09-14T04:03:28
2023-04-18T13:23:42
{ "title": [ -0.011375427, -0.0084991455, -0.072265625, -0.0569458, 0.019561768, -0.03111267, -0.043426514, 0.0015096664, -0.0143966675, 0.031097412, -0.004714966, 0.023086548, 0.0064697266, -0.039367676, -0.0071144104, 0.021102905, -0.010002136, ...
http://arxiv.org/abs/2304.08968v1
Stochastic Parrots Looking for Stochastic Parrots: LLMs are Easy to Fine-Tune and Hard to Detect with other LLMs
Da Silva Gameiro Henrique, Andrei Kucharavy, Rachid Guerraoui
cs.CL, cs.AI, cs.CR, cs.LG, I.2.7; K.6.5
The self-attention revolution allowed generative language models to scale and achieve increasingly impressive abilities. Such models - commonly referred to as Large Language Models (LLMs) - have recently gained prominence with the general public, thanks to conversational fine-tuning, putting their behavior in line with...
http://arxiv.org/pdf/2304.08968v1
2023-04-18T13:05:01
2023-04-18T13:05:01
{ "title": [ -0.0063667297, 0.00040912628, 0.012931824, 0.0005340576, -0.014701843, 0.00365448, -0.014335632, -0.031707764, -0.013420105, -0.005680084, -0.035980225, 0.032989502, -0.031982422, -0.031921387, 0.020370483, -0.033050537, -0.009468079, ...
http://arxiv.org/abs/2304.08931v2
Enhancing Textbooks with Visuals from the Web for Improved Learning
Janvijay Singh, Vilém Zouhar, Mrinmaya Sachan
cs.CV, cs.CL
Textbooks are one of the main mediums for delivering high-quality education to students. In particular, explanatory and illustrative visuals play a key role in retention, comprehension and general transfer of knowledge. However, many textbooks lack these interesting visuals to support student learning. In this paper, w...
http://arxiv.org/pdf/2304.08931v2
2023-10-20T11:06:14
2023-04-18T12:16:39
{ "title": [ 0.04260254, 0.0154800415, -0.03994751, -0.037322998, -0.007331848, -0.036071777, -0.00042247772, -0.0146865845, -0.015205383, 0.023849487, -0.04623413, 0.018875122, -0.012290955, -0.0262146, 0.008262634, -0.027832031, 0.018707275, 0....
http://arxiv.org/abs/2304.08911v1
Towards Zero-Shot Personalized Table-to-Text Generation with Contrastive Persona Distillation
Haolan Zhan, Xuming Lin, Shaobo Cui, Zhongzhou Zhao, Wei Zhou, Haiqing Chen
cs.CL
Existing neural methods have shown great potentials towards generating informative text from structured tabular data as well as maintaining high content fidelity. However, few of them shed light on generating personalized expressions, which often requires well-aligned persona-table-text datasets that are difficult to o...
http://arxiv.org/pdf/2304.08911v1
2023-04-18T11:32:33
2023-04-18T11:32:33
{ "title": [ 0.011886597, 0.0039749146, -0.035949707, -0.06488037, -0.003824234, 0.0035934448, -0.035614014, -0.0042419434, 0.022567749, 0.064208984, -0.013870239, 0.025894165, 0.03213501, 0.00016474724, -0.047332764, 0.014587402, -0.019485474, 0...
http://arxiv.org/abs/2304.08891v2
Tailoring Domain Adaptation for Machine Translation Quality Estimation
Javad Pourmostafa Roshan Sharami, Dimitar Shterionov, Frédéric Blain, Eva Vanmassenhove, Mirella De Sisto, Chris Emmery, Pieter Spronck
cs.CL
While quality estimation (QE) can play an important role in the translation process, its effectiveness relies on the availability and quality of training data. For QE in particular, high-quality labeled data is often lacking due to the high cost and effort associated with labeling such data. Aside from the data scarcit...
http://arxiv.org/pdf/2304.08891v2
2023-05-09T08:34:19
2023-04-18T10:36:50
{ "title": [ -0.05154419, -0.0096588135, -0.02229309, -0.032836914, -0.021591187, 0.004638672, -0.0362854, 0.00024080276, -0.007106781, 0.037719727, 0.011749268, 0.0014753342, 0.021942139, -0.024520874, -0.04296875, 0.021194458, -0.021102905, 0.0...
http://arxiv.org/abs/2304.08865v1
Romanization-based Large-scale Adaptation of Multilingual Language Models
Sukannya Purkayastha, Sebastian Ruder, Jonas Pfeiffer, Iryna Gurevych, Ivan Vulić
cs.CL, cs.LG
Large multilingual pretrained language models (mPLMs) have become the de facto state of the art for cross-lingual transfer in NLP. However, their large-scale deployment to many languages, besides pretraining data scarcity, is also hindered by the increase in vocabulary size and limitations in their parameter budget. In...
http://arxiv.org/pdf/2304.08865v1
2023-04-18T09:58:34
2023-04-18T09:58:34
{ "title": [ -0.043884277, -0.007980347, -0.016830444, -0.030654907, -0.008369446, 0.009819031, -0.047302246, -0.008003235, -0.0017681122, 0.06750488, 0.011909485, -0.008300781, 0.044525146, -0.02104187, -0.048950195, 0.02722168, -0.010437012, 0....
http://arxiv.org/abs/2304.08862v2
Approximate Nearest Neighbour Phrase Mining for Contextual Speech Recognition
Maurits Bleeker, Pawel Swietojanski, Stefan Braun, Xiaodan Zhuang
cs.CL, eess.AS
This paper presents an extension to train end-to-end Context-Aware Transformer Transducer ( CATT ) models by using a simple, yet efficient method of mining hard negative phrases from the latent space of the context encoder. During training, given a reference query, we mine a number of similar phrases using approximate ...
http://arxiv.org/pdf/2304.08862v2
2023-08-16T10:48:12
2023-04-18T09:52:11
{ "title": [ -0.00051116943, -0.004737854, -0.02960205, -0.051605225, 0.014305115, 0.0034389496, -0.039215088, -0.0048294067, -0.021575928, 0.060668945, -0.051208496, 0.005744934, 0.0065727234, 0.0034313202, -0.06964111, -0.015594482, -0.017929077, ...
http://arxiv.org/abs/2304.08823v1
Transfer to a Low-Resource Language via Close Relatives: The Case Study on Faroese
Vésteinn Snæbjarnarson, Annika Simonsen, Goran Glavaš, Ivan Vulić
cs.CL
Multilingual language models have pushed state-of-the-art in cross-lingual NLP transfer. The majority of zero-shot cross-lingual transfer, however, use one and the same massively multilingual transformer (e.g., mBERT or XLM-R) to transfer to all target languages, irrespective of their typological, etymological, and phy...
http://arxiv.org/pdf/2304.08823v1
2023-04-18T08:42:38
2023-04-18T08:42:38
{ "title": [ 0.017547607, 0.0018396378, -0.030731201, -0.042053223, 0.00073337555, -0.022003174, -0.0262146, -0.0066871643, -0.0077056885, 0.029724121, -0.0040283203, 0.006252289, 0.022750854, -0.04156494, -0.014915466, 0.000035643578, -0.00057792664...