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/2305.05189v3
SUR-adapter: Enhancing Text-to-Image Pre-trained Diffusion Models with Large Language Models
Shanshan Zhong, Zhongzhan Huang, Wushao Wen, Jinghui Qin, Liang Lin
cs.CL, cs.CV
Diffusion models, which have emerged to become popular text-to-image generation models, can produce high-quality and content-rich images guided by textual prompts. However, there are limitations to semantic understanding and commonsense reasoning in existing models when the input prompts are concise narrative, resultin...
http://arxiv.org/pdf/2305.05189v3
2023-08-18T09:13:46
2023-05-09T05:48:38
{ "title": [ -0.0022182465, 0.005542755, -0.025924683, -0.026748657, 0.00894928, -0.0074920654, -0.016967773, -0.025665283, 0.025177002, 0.027954102, -0.01889038, 0.008956909, 0.000518322, -0.02217102, -0.027801514, 0.0062789917, -0.001627922, 0....
http://arxiv.org/abs/2305.05672v1
$2 * n$ is better than $n^2$: Decomposing Event Coreference Resolution into Two Tractable Problems
Shafiuddin Rehan Ahmed, Abhijnan Nath, James H. Martin, Nikhil Krishnaswamy
cs.CL
Event Coreference Resolution (ECR) is the task of linking mentions of the same event either within or across documents. Most mention pairs are not coreferent, yet many that are coreferent can be identified through simple techniques such as lemma matching of the event triggers or the sentences in which they appear. Exis...
http://arxiv.org/pdf/2305.05672v1
2023-05-09T05:33:32
2023-05-09T05:33:32
{ "title": [ 0.016418457, -0.01436615, -0.03274536, -0.092163086, 0.0063438416, -0.018096924, -0.04324341, -0.00274086, -0.018722534, 0.024047852, -0.0034122467, 0.017196655, 0.025924683, -0.02633667, -0.0012874603, 0.011634827, -0.00030326843, 0...
http://arxiv.org/abs/2305.05183v1
CSED: A Chinese Semantic Error Diagnosis Corpus
Bo Sun, Baoxin Wang, Yixuan Wang, Wanxiang Che, Dayong Wu, Shijin Wang, Ting Liu
cs.CL, cs.AI
Recently, much Chinese text error correction work has focused on Chinese Spelling Check (CSC) and Chinese Grammatical Error Diagnosis (CGED). In contrast, little attention has been paid to the complicated problem of Chinese Semantic Error Diagnosis (CSED), which lacks relevant datasets. The study of semantic errors is ...
http://arxiv.org/pdf/2305.05183v1
2023-05-09T05:33:31
2023-05-09T05:33:31
{ "title": [ -0.014320374, -0.010040283, -0.03955078, -0.06530762, 0.004398346, -0.0025119781, -0.044555664, 0.0039596558, 0.033355713, 0.031173706, -0.038848877, -0.001947403, 0.026367188, -0.011001587, -0.04586792, 0.009002686, -0.029052734, 0....
http://arxiv.org/abs/2305.05181v2
MoT: Memory-of-Thought Enables ChatGPT to Self-Improve
Xiaonan Li, Xipeng Qiu
cs.CL, cs.AI
Large Language Models (LLMs) have shown impressive abilities in various tasks. However, fundamentally improving them depends on high-quality datasets or computationally expensive fine-tuning. On the contrary, humans can easily improve themselves by self-thinking and memory, without external resources. In this paper, we...
http://arxiv.org/pdf/2305.05181v2
2023-10-09T02:44:12
2023-05-09T05:25:05
{ "title": [ 0.036132812, -0.02923584, -0.005027771, -0.020874023, -0.034332275, -0.018295288, 0.013450623, -0.0066986084, -0.026977539, 0.025817871, -0.0012016296, 0.029449463, 0.011177063, -0.038482666, 0.0072669983, 0.0105896, -0.007926941, 0....
http://arxiv.org/abs/2305.05176v1
FrugalGPT: How to Use Large Language Models While Reducing Cost and Improving Performance
Lingjiao Chen, Matei Zaharia, James Zou
cs.LG, cs.AI, cs.CL, cs.SE
There is a rapidly growing number of large language models (LLMs) that users can query for a fee. We review the cost associated with querying popular LLM APIs, e.g. GPT-4, ChatGPT, J1-Jumbo, and find that these models have heterogeneous pricing structures, with fees that can differ by two orders of magnitude. In partic...
http://arxiv.org/pdf/2305.05176v1
2023-05-09T05:11:02
2023-05-09T05:11:02
{ "title": [ -0.016845703, -0.016204834, -0.054992676, -0.072021484, 0.0104599, -0.012611389, -0.04437256, 0.010124207, -0.025497437, 0.019714355, 0.00036787987, -0.008842468, 0.010429382, -0.045440674, -0.0030155182, 0.00042366982, 0.007572174, ...
http://arxiv.org/abs/2305.05171v1
Summarization with Precise Length Control
Lesly Miculicich, Yujia Xie, Song Wang, Pengcheng He
cs.CL
Many applications of text generation such as summarization benefit from accurately controlling the text length. Existing approaches on length-controlled summarization either result in degraded performance or can only control the length approximately. In this work, we present a framework to generate summaries with preci...
http://arxiv.org/pdf/2305.05171v1
2023-05-09T04:45:24
2023-05-09T04:45:24
{ "title": [ -0.006958008, -0.0012578964, -0.021911621, -0.0680542, -0.041381836, -0.006942749, -0.06958008, -0.009521484, -0.014465332, 0.06341553, 0.0063056946, 0.00756073, -0.0138549805, -0.012916565, -0.035491943, 0.020233154, 0.010604858, 0....
http://arxiv.org/abs/2305.05166v2
E2TIMT: Efficient and Effective Modal Adapter for Text Image Machine Translation
Cong Ma, Yaping Zhang, Mei Tu, Yang Zhao, Yu Zhou, Chengqing Zong
cs.CL
Text image machine translation (TIMT) aims to translate texts embedded in images from one source language to another target language. Existing methods, both two-stage cascade and one-stage end-to-end architectures, suffer from different issues. The cascade models can benefit from the large-scale optical character recog...
http://arxiv.org/pdf/2305.05166v2
2023-05-10T02:37:32
2023-05-09T04:25:52
{ "title": [ -0.017730713, -0.0042686462, 0.001657486, -0.038513184, -0.0103302, 0.009223938, -0.037261963, -0.023849487, 0.026306152, 0.064941406, -0.016098022, 0.0016269684, 0.07525635, -0.0209198, -0.044067383, 0.03579712, 0.0073928833, 0.0063...
http://arxiv.org/abs/2305.05162v1
Effective Medical Code Prediction via Label Internal Alignment
Guodong Liu
cs.LG, cs.CL
The clinical notes are usually typed into the system by physicians. They are typically required to be marked by standard medical codes, and each code represents a diagnosis or medical treatment procedure. Annotating these notes is time consuming and prone to error. In this paper, we proposed a multi-view attention base...
http://arxiv.org/pdf/2305.05162v1
2023-05-09T04:14:20
2023-05-09T04:14:20
{ "title": [ -0.046936035, -0.01890564, -0.047180176, -0.050842285, -0.0146484375, -0.002445221, -0.053466797, 0.010009766, -0.007820129, 0.03744507, 0.008369446, -0.0044937134, 0.016693115, -0.031097412, -0.025390625, 0.0053749084, 0.005897522, ...
http://arxiv.org/abs/2305.05138v1
Read, Diagnose and Chat: Towards Explainable and Interactive LLMs-Augmented Depression Detection in Social Media
Wei Qin, Zetong Chen, Lei Wang, Yunshi Lan, Weijieying Ren, Richang Hong
cs.CL
This paper proposes a new depression detection system based on LLMs that is both interpretable and interactive. It not only provides a diagnosis, but also diagnostic evidence and personalized recommendations based on natural language dialogue with the user. We address challenges such as the processing of large amounts ...
http://arxiv.org/pdf/2305.05138v1
2023-05-09T02:49:09
2023-05-09T02:49:09
{ "title": [ 0.013153076, -0.005252838, 0.0051574707, -0.04458618, -0.010902405, 0.01058197, -0.026794434, -0.026046753, -0.00004208088, 0.015930176, -0.0440979, 0.021484375, -0.0035209656, -0.016921997, -0.031555176, -0.029129028, -0.010215759, ...
http://arxiv.org/abs/2305.05133v1
Generating Phishing Attacks using ChatGPT
Sayak Saha Roy, Krishna Vamsi Naragam, Shirin Nilizadeh
cs.CR, cs.CL
The ability of ChatGPT to generate human-like responses and understand context has made it a popular tool for conversational agents, content creation, data analysis, and research and innovation. However, its effectiveness and ease of accessibility makes it a prime target for generating malicious content, such as phishi...
http://arxiv.org/pdf/2305.05133v1
2023-05-09T02:38:05
2023-05-09T02:38:05
{ "title": [ 0.012161255, -0.0068359375, -0.020095825, -0.019180298, -0.0064468384, -0.007888794, -0.005214691, -0.024887085, -0.05392456, 0.046020508, 0.014152527, 0.031341553, -0.014549255, -0.028900146, -0.017440796, -0.0022029877, -0.02607727, ...
http://arxiv.org/abs/2305.05098v1
Who Needs Decoders? Efficient Estimation of Sequence-level Attributes
Yassir Fathullah, Puria Radmard, Adian Liusie, Mark J. F. Gales
cs.LG, cs.AI, cs.CL
State-of-the-art sequence-to-sequence models often require autoregressive decoding, which can be highly expensive. However, for some downstream tasks such as out-of-distribution (OOD) detection and resource allocation, the actual decoding output is not needed just a scalar attribute of this sequence. In these scenarios...
http://arxiv.org/pdf/2305.05098v1
2023-05-09T00:01:32
2023-05-09T00:01:32
{ "title": [ -0.045135498, -0.011146545, -0.054138184, -0.059326172, -0.0006685257, -0.004886627, -0.042053223, -0.0038585663, 0.024383545, 0.03225708, -0.011138916, -0.015205383, 0.01991272, -0.022399902, -0.039245605, 0.02494812, 0.009010315, 0...
http://arxiv.org/abs/2305.06358v1
Accessible Instruction-Following Agent
Kairui Zhou
cs.AI, cs.CL
Humans can collaborate and complete tasks based on visual signals and instruction from the environment. Training such a robot is difficult especially due to the understanding of the instruction and the complicated environment. Previous instruction-following agents are biased to English-centric corpus, making it unreali...
http://arxiv.org/pdf/2305.06358v1
2023-05-08T23:57:26
2023-05-08T23:57:26
{ "title": [ 0.023086548, -0.0048065186, -0.043640137, -0.05621338, -0.039031982, -0.022750854, -0.049957275, 0.0036907196, 0.018417358, 0.058288574, -0.045928955, -0.015296936, 0.062438965, -0.015510559, -0.0025424957, 0.010009766, 0.000213027, ...
http://arxiv.org/abs/2305.05094v1
Interactive Concept Learning for Uncovering Latent Themes in Large Text Collections
Maria Leonor Pacheco, Tunazzina Islam, Lyle Ungar, Ming Yin, Dan Goldwasser
cs.CL, cs.HC
Experts across diverse disciplines are often interested in making sense of large text collections. Traditionally, this challenge is approached either by noisy unsupervised techniques such as topic models, or by following a manual theme discovery process. In this paper, we expand the definition of a theme to account for...
http://arxiv.org/pdf/2305.05094v1
2023-05-08T23:43:15
2023-05-08T23:43:15
{ "title": [ 0.010444641, -0.004573822, -0.030380249, -0.04006958, 0.009300232, -0.01675415, 0.0058555603, -0.006954193, 0.024612427, 0.01701355, -0.038513184, 0.021652222, -0.036102295, -0.0022068024, -0.057769775, -0.019485474, -0.021255493, 0....
http://arxiv.org/abs/2305.05091v1
Knowledge-enhanced Agents for Interactive Text Games
Prateek Chhikara, Jiarui Zhang, Filip Ilievski, Jonathan Francis, Kaixin Ma
cs.CL, cs.AI, cs.HC
Communication via natural language is a crucial aspect of intelligence, and it requires computational models to learn and reason about world concepts, with varying levels of supervision. While there has been significant progress made on fully-supervised non-interactive tasks, such as question-answering and procedural t...
http://arxiv.org/pdf/2305.05091v1
2023-05-08T23:31:39
2023-05-08T23:31:39
{ "title": [ -0.018356323, -0.014625549, -0.016189575, -0.042175293, -0.0068511963, -0.008605957, -0.005443573, 0.00945282, 0.013542175, 0.037597656, -0.03967285, 0.0028896332, -0.011314392, 0.002948761, -0.06021118, -0.011962891, -0.011383057, 0...
http://arxiv.org/abs/2305.06218v1
Multi-Task End-to-End Training Improves Conversational Recommendation
Naveen Ram, Dima Kuzmin, Ellie Ka In Chio, Moustafa Farid Alzantot, Santiago Ontanon, Ambarish Jash, Judith Yue Li
cs.CL, cs.AI, cs.IR
In this paper, we analyze the performance of a multitask end-to-end transformer model on the task of conversational recommendations, which aim to provide recommendations based on a user's explicit preferences expressed in dialogue. While previous works in this area adopt complex multi-component approaches where the dia...
http://arxiv.org/pdf/2305.06218v1
2023-05-08T22:42:48
2023-05-08T22:42:48
{ "title": [ 0.032073975, 0.009124756, -0.039215088, -0.07196045, -0.009605408, -0.03086853, -0.045410156, 0.00667572, -0.023452759, 0.034698486, 0.01184845, 0.01638794, -0.0021133423, -0.042877197, -0.0026454926, 0.0027198792, -0.013267517, 0.03...
http://arxiv.org/abs/2305.05079v1
A Unified Evaluation Framework for Novelty Detection and Accommodation in NLP with an Instantiation in Authorship Attribution
Neeraj Varshney, Himanshu Gupta, Eric Robertson, Bing Liu, Chitta Baral
cs.CL
State-of-the-art natural language processing models have been shown to achieve remarkable performance in 'closed-world' settings where all the labels in the evaluation set are known at training time. However, in real-world settings, 'novel' instances that do not belong to any known class are often observed. This render...
http://arxiv.org/pdf/2305.05079v1
2023-05-08T22:37:30
2023-05-08T22:37:30
{ "title": [ -0.006011963, -0.0025901794, -0.025726318, -0.037506104, 0.00957489, -0.0074920654, -0.02128601, -0.010513306, -0.004497528, 0.005027771, -0.016098022, 0.03149414, -0.023361206, -0.015213013, -0.011367798, -0.009536743, -0.052612305, ...
http://arxiv.org/abs/2305.05061v1
Coherent Wave Dynamics and Language Generation of a Generative Pre-trained Transformer
Tao Hong
cs.CL, nlin.PS, 68T07, I.2.7
Large Language Models (LLMs), such as the Generative Pretrained Transformer (GPT), have achieved tremendous success in various language tasks, but their emergent abilities have also raised many questions, concerns, and challenges that need to be addressed. To gain a better understanding of the models' inner mechanisms,...
http://arxiv.org/pdf/2305.05061v1
2023-05-08T21:35:12
2023-05-08T21:35:12
{ "title": [ 0.0022468567, 0.0015897751, -0.020614624, -0.036895752, -0.012519836, -0.0017499924, -0.030029297, -0.028213501, -0.0033130646, 0.055541992, 0.008972168, 0.0070228577, 0.04272461, -0.022384644, -0.039886475, 0.015464783, 0.01626587, ...
http://arxiv.org/abs/2305.05054v1
Dreams Are More "Predictable'' Than You Think
Lorenzo Bertolini
cs.CL
A consistent body of evidence suggests that dream reports significantly vary from other types of textual transcripts with respect to semantic content. Furthermore, it appears to be a widespread belief in the dream/sleep research community that dream reports constitute rather ``unique'' strings of text. This might be a ...
http://arxiv.org/pdf/2305.05054v1
2023-05-08T21:24:12
2023-05-08T21:24:12
{ "title": [ 0.020217896, -0.009971619, 0.022262573, 0.019332886, -0.039398193, -0.0044174194, 0.021087646, -0.030349731, -0.017288208, 0.037384033, -0.020263672, -0.016830444, -0.011955261, -0.014060974, 0.008972168, -0.017349243, 0.034484863, 0...
http://arxiv.org/abs/2305.05050v3
ANALOGICAL -- A Novel Benchmark for Long Text Analogy Evaluation in Large Language Models
Thilini Wijesiriwardene, Ruwan Wickramarachchi, Bimal G. Gajera, Shreeyash Mukul Gowaikar, Chandan Gupta, Aman Chadha, Aishwarya Naresh Reganti, Amit Sheth, Amitava Das
cs.CL, cs.AI
Over the past decade, analogies, in the form of word-level analogies, have played a significant role as an intrinsic measure of evaluating the quality of word embedding methods such as word2vec. Modern large language models (LLMs), however, are primarily evaluated on extrinsic measures based on benchmarks such as GLUE ...
http://arxiv.org/pdf/2305.05050v3
2023-05-25T20:38:17
2023-05-08T21:12:20
{ "title": [ -0.024887085, -0.0014591217, -0.007091522, -0.047973633, -0.001162529, 0.013961792, -0.041931152, -0.028884888, 0.03100586, 0.0368042, -0.0129852295, 0.018859863, 0.02319336, -0.0055236816, -0.057678223, 0.023742676, -0.024780273, 0....
http://arxiv.org/abs/2305.05027v2
Web Content Filtering through knowledge distillation of Large Language Models
Tamás Vörös, Sean Paul Bergeron, Konstantin Berlin
cs.LG, cs.CL, cs.IR
We introduce a state-of-the-art approach for URL categorization that leverages the power of Large Language Models (LLMs) to address the primary objectives of web content filtering: safeguarding organizations from legal and ethical risks, limiting access to high-risk or suspicious websites, and fostering a secure and pr...
http://arxiv.org/pdf/2305.05027v2
2023-05-10T08:36:57
2023-05-08T20:09:27
{ "title": [ -0.0018701553, 0.0011510849, -0.037841797, -0.07299805, 0.026504517, -0.012481689, -0.03765869, -0.0008120537, 0.046966553, 0.020843506, -0.03326416, -0.030776978, 0.026184082, -0.02810669, -0.046844482, -0.007701874, -0.014793396, 0...
http://arxiv.org/abs/2305.05010v1
Do Not Blindly Imitate the Teacher: Using Perturbed Loss for Knowledge Distillation
Rongzhi Zhang, Jiaming Shen, Tianqi Liu, Jialu Liu, Michael Bendersky, Marc Najork, Chao Zhang
cs.LG, cs.CL
Knowledge distillation is a popular technique to transfer knowledge from large teacher models to a small student model. Typically, the student learns to imitate the teacher by minimizing the KL divergence of its output distribution with the teacher's output distribution. In this work, we argue that such a learning obje...
http://arxiv.org/pdf/2305.05010v1
2023-05-08T19:31:09
2023-05-08T19:31:09
{ "title": [ -0.008506775, -0.004524231, -0.027832031, -0.05130005, -0.012176514, -0.0038852692, -0.032806396, -0.031051636, 0.021942139, -0.0040512085, -0.045715332, 0.0020523071, -0.03805542, -0.021209717, -0.035461426, -0.032104492, 0.009681702, ...
http://arxiv.org/abs/2305.06223v1
ComputeGPT: A computational chat model for numerical problems
Ryan Hardesty Lewis, Junfeng Jiao
cs.PL, cs.AI, cs.CL, 68T50, 68N18, 97R50, I.2.7; I.2.6; H.5.2
Language models are not accurate in numerical problems. Their architecture does not allow for anything less than a probabilistic next word. This paper introduces ComputeGPT: an approach of creating a chat model able to answer computational problems through running on-demand code. ComputeGPT converts each question to re...
http://arxiv.org/pdf/2305.06223v1
2023-05-08T19:21:41
2023-05-08T19:21:41
{ "title": [ -0.033294678, -0.01663208, 0.00035238266, -0.05291748, -0.0052452087, 0.024871826, -0.037506104, -0.0012073517, 0.024627686, 0.04031372, 0.005443573, 0.0024776459, 0.0068473816, -0.025817871, -0.050323486, 0.017578125, -0.001162529, ...
http://arxiv.org/abs/2305.05003v1
Revisiting Relation Extraction in the era of Large Language Models
Somin Wadhwa, Silvio Amir, Byron C. Wallace
cs.CL
Relation extraction (RE) is the core NLP task of inferring semantic relationships between entities from text. Standard supervised RE techniques entail training modules to tag tokens comprising entity spans and then predict the relationship between them. Recent work has instead treated the problem as a \emph{sequence-to...
http://arxiv.org/pdf/2305.05003v1
2023-05-08T19:19:07
2023-05-08T19:19:07
{ "title": [ -0.01499176, -0.013496399, -0.047546387, -0.08325195, 0.010246277, 0.005695343, -0.05630493, 0.011665344, 0.027496338, 0.009384155, -0.023910522, -0.0068473816, -0.0031738281, -0.0059432983, -0.053619385, 0.0046958923, -0.020523071, ...
http://arxiv.org/abs/2305.05001v1
GersteinLab at MEDIQA-Chat 2023: Clinical Note Summarization from Doctor-Patient Conversations through Fine-tuning and In-context Learning
Xiangru Tang, Andrew Tran, Jeffrey Tan, Mark Gerstein
cs.CL
This paper presents our contribution to the MEDIQA-2023 Dialogue2Note shared task, encompassing both subtask A and subtask B. We approach the task as a dialogue summarization problem and implement two distinct pipelines: (a) a fine-tuning of a pre-trained dialogue summarization model and GPT-3, and (b) few-shot in-cont...
http://arxiv.org/pdf/2305.05001v1
2023-05-08T19:16:26
2023-05-08T19:16:26
{ "title": [ 0.0052452087, -0.034118652, 0.008529663, -0.03543091, -0.066223145, 0.007980347, -0.053344727, 0.0039596558, 0.0033416748, 0.06591797, -0.025222778, 0.015449524, 0.04067993, 0.0026073456, -0.023010254, -0.01424408, 0.027496338, 0.024...
http://arxiv.org/abs/2305.04990v3
Explanation-based Finetuning Makes Models More Robust to Spurious Cues
Josh Magnus Ludan, Yixuan Meng, Tai Nguyen, Saurabh Shah, Qing Lyu, Marianna Apidianaki, Chris Callison-Burch
cs.CL, cs.LG
Large Language Models (LLMs) are so powerful that they sometimes learn correlations between labels and features that are irrelevant to the task, leading to poor generalization on out-of-distribution data. We propose explanation-based finetuning as a general approach to mitigate LLMs' reliance on spurious correlations. ...
http://arxiv.org/pdf/2305.04990v3
2023-06-06T15:31:33
2023-05-08T18:53:45
{ "title": [ 0.076660156, -0.0015258789, 0.006576538, -0.09753418, -0.012466431, -0.024627686, -0.03945923, -0.03527832, -0.026916504, 0.011268616, -0.030899048, 0.008010864, -0.0124053955, -0.031707764, 0.00484848, -0.019363403, -0.007293701, 0....
http://arxiv.org/abs/2305.04989v1
Knowledge Graph Guided Semantic Evaluation of Language Models For User Trust
Kaushik Roy, Tarun Garg, Vedant Palit, Yuxin Zi, Vignesh Narayanan, Amit Sheth
cs.CL, cs.AI
A fundamental question in natural language processing is - what kind of language structure and semantics is the language model capturing? Graph formats such as knowledge graphs are easy to evaluate as they explicitly express language semantics and structure. This study evaluates the semantics encoded in the self-attent...
http://arxiv.org/pdf/2305.04989v1
2023-05-08T18:53:14
2023-05-08T18:53:14
{ "title": [ -0.05303955, -0.009651184, -0.08135986, -0.04498291, -0.014503479, -0.011421204, -0.054138184, 0.014625549, -0.013130188, 0.038238525, -0.019454956, 0.006248474, 0.01586914, -0.010917664, -0.043548584, -0.00083589554, -0.024475098, 0...
http://arxiv.org/abs/2305.07666v1
Imitation versus Innovation: What children can do that large language and language-and-vision models cannot (yet)?
Eunice Yiu, Eliza Kosoy, Alison Gopnik
cs.AI, cs.CL
Much discussion about large language models and language-and-vision models has focused on whether these models are intelligent agents. We present an alternative perspective. We argue that these artificial intelligence models are cultural technologies that enhance cultural transmission in the modern world, and are effic...
http://arxiv.org/pdf/2305.07666v1
2023-05-08T18:26:39
2023-05-08T18:26:39
{ "title": [ -0.031799316, -0.0075302124, -0.014968872, -0.0063323975, -0.014389038, -0.012382507, -0.010643005, -0.012046814, -0.04159546, 0.008140564, -0.014671326, 0.002117157, -0.031311035, -0.020202637, -0.03326416, -0.051940918, -0.009414673, ...
http://arxiv.org/abs/2305.04978v2
NeuroComparatives: Neuro-Symbolic Distillation of Comparative Knowledge
Phillip Howard, Junlin Wang, Vasudev Lal, Gadi Singer, Yejin Choi, Swabha Swayamdipta
cs.CL
Comparative knowledge (e.g., steel is stronger and heavier than styrofoam) is an essential component of our world knowledge, yet understudied in prior literature. In this paper, we study the task of comparative knowledge acquisition, motivated by the dramatic improvements in the capabilities of extreme-scale language m...
http://arxiv.org/pdf/2305.04978v2
2023-11-15T17:34:56
2023-05-08T18:20:36
{ "title": [ 0.000043928623, -0.01939392, 0.0029792786, -0.036712646, -0.010360718, -0.00054216385, -0.007774353, -0.040161133, -0.0027103424, 0.036499023, -0.021087646, -0.011367798, -0.01235199, -0.008956909, -0.070251465, -0.014205933, -0.00280761...
http://arxiv.org/abs/2305.04971v1
LABO: Towards Learning Optimal Label Regularization via Bi-level Optimization
Peng Lu, Ahmad Rashid, Ivan Kobyzev, Mehdi Rezagholizadeh, Philippe Langlais
cs.LG, cs.CL
Regularization techniques are crucial to improving the generalization performance and training efficiency of deep neural networks. Many deep learning algorithms rely on weight decay, dropout, batch/layer normalization to converge faster and generalize. Label Smoothing (LS) is another simple, versatile and efficient reg...
http://arxiv.org/pdf/2305.04971v1
2023-05-08T18:04:18
2023-05-08T18:04:18
{ "title": [ -0.029296875, 0.0056533813, -0.034118652, -0.056121826, 0.012771606, 0.012771606, -0.035491943, -0.0047798157, -0.0044059753, 0.014091492, -0.005554199, 0.027572632, -0.02809143, -0.024841309, -0.013000488, 0.0046958923, -0.0016155243, ...
http://arxiv.org/abs/2305.04961v1
Joint Moment Retrieval and Highlight Detection Via Natural Language Queries
Richard Luo, Austin Peng, Heidi Yap, Koby Beard
cs.CV, cs.CL, cs.LG
Video summarization has become an increasingly important task in the field of computer vision due to the vast amount of video content available on the internet. In this project, we propose a new method for natural language query based joint video summarization and highlight detection using multi-modal transformers. Thi...
http://arxiv.org/pdf/2305.04961v1
2023-05-08T18:00:33
2023-05-08T18:00:33
{ "title": [ 0.008071899, -0.001332283, -0.0154418945, -0.026260376, 0.010429382, -0.010818481, -0.00001347065, -0.029953003, 0.0287323, 0.045684814, -0.014266968, -0.02961731, 0.025512695, -0.013000488, -0.08294678, -0.0029201508, -0.005748749, ...
http://arxiv.org/abs/2305.18086v1
The impact and applications of ChatGPT: a systematic review of literature reviews
Irene S. Gabashvili
cs.CY, cs.CL
The conversational artificial-intelligence (AI) technology ChatGPT has become one of the most widely used natural language processing tools. With thousands of published papers demonstrating its applications across various industries and fields, ChatGPT has sparked significant interest in the research community. Reviews...
http://arxiv.org/pdf/2305.18086v1
2023-05-08T17:57:34
2023-05-08T17:57:34
{ "title": [ -0.003408432, -0.024917603, -0.009315491, -0.032318115, -0.007030487, 0.012397766, -0.029693604, -0.02520752, -0.0026245117, 0.020706177, 0.010360718, 0.03112793, -0.01878357, -0.028030396, -0.021881104, 0.029647827, -0.033111572, 0....
http://arxiv.org/abs/2305.04905v1
What Do Patients Say About Their Disease Symptoms? Deep Multilabel Text Classification With Human-in-the-Loop Curation for Automatic Labeling of Patient Self Reports of Problems
Lakshmi Arbatti, Abhishek Hosamath, Vikram Ramanarayanan, Ira Shoulson
cs.CL, cs.LG, eess.AS
The USA Food and Drug Administration has accorded increasing importance to patient-reported problems in clinical and research settings. In this paper, we explore one of the largest online datasets comprising 170,141 open-ended self-reported responses (called "verbatims") from patients with Parkinson's (PwPs) to questio...
http://arxiv.org/pdf/2305.04905v1
2023-05-08T17:42:23
2023-05-08T17:42:23
{ "title": [ 0.013046265, 0.019165039, -0.015777588, -0.03466797, -0.025405884, -0.019042969, -0.041412354, -0.01184082, -0.006706238, 0.031188965, -0.05444336, 0.023513794, -0.022384644, -0.014045715, -0.011550903, -0.052490234, -0.0036792755, 0...
http://arxiv.org/abs/2305.04859v1
A Frustratingly Easy Improvement for Position Embeddings via Random Padding
Mingxu Tao, Yansong Feng, Dongyan Zhao
cs.CL, cs.AI
Position embeddings, encoding the positional relationships among tokens in text sequences, make great contributions to modeling local context features in Transformer-based pre-trained language models. However, in Extractive Question Answering, position embeddings trained with instances of varied context lengths may not...
http://arxiv.org/pdf/2305.04859v1
2023-05-08T17:08:14
2023-05-08T17:08:14
{ "title": [ 0.033843994, -0.0012044907, 0.0013895035, -0.011566162, -0.015991211, -0.022262573, 0.010620117, -0.041412354, 0.04660034, 0.02458191, -0.049835205, 0.006755829, 0.000667572, -0.027389526, -0.010856628, -0.001411438, 0.009811401, 0.0...
http://arxiv.org/abs/2305.04853v2
The Current State of Summarization
Fabian Retkowski
cs.CL, cs.AI
With the explosive growth of textual information, summarization systems have become increasingly important. This work aims to concisely indicate the current state of the art in abstractive text summarization. As part of this, we outline the current paradigm shifts towards pre-trained encoder-decoder models and large au...
http://arxiv.org/pdf/2305.04853v2
2023-08-01T17:42:59
2023-05-08T17:00:49
{ "title": [ 0.043121338, -0.041931152, 0.014854431, -0.07220459, -0.07336426, 0.0047569275, -0.029174805, 0.039123535, 0.019607544, 0.043151855, -0.037139893, 0.0021247864, -0.008476257, -0.009712219, -0.020431519, 0.020935059, 0.04034424, -0.00...
http://arxiv.org/abs/2305.04940v2
The EarlyBIRD Catches the Bug: On Exploiting Early Layers of Encoder Models for More Efficient Code Classification
Anastasiia Grishina, Max Hort, Leon Moonen
cs.SE, cs.AI, cs.CL, cs.LG
The use of modern Natural Language Processing (NLP) techniques has shown to be beneficial for software engineering tasks, such as vulnerability detection and type inference. However, training deep NLP models requires significant computational resources. This paper explores techniques that aim at achieving the best usag...
http://arxiv.org/pdf/2305.04940v2
2023-09-11T17:56:10
2023-05-08T16:47:28
{ "title": [ -0.02003479, 0.0018978119, -0.024047852, -0.029556274, 0.004085541, -0.022994995, -0.018966675, -0.025390625, 0.0084991455, 0.0368042, -0.009361267, 0.0077934265, -0.0044822693, -0.0211792, -0.043945312, 0.0031604767, -0.011260986, 0...
http://arxiv.org/abs/2305.04843v1
Reinforcement Learning for Topic Models
Jeremy Costello, Marek Z. Reformat
cs.CL, cs.LG
We apply reinforcement learning techniques to topic modeling by replacing the variational autoencoder in ProdLDA with a continuous action space reinforcement learning policy. We train the system with a policy gradient algorithm REINFORCE. Additionally, we introduced several modifications: modernize the neural network a...
http://arxiv.org/pdf/2305.04843v1
2023-05-08T16:41:08
2023-05-08T16:41:08
{ "title": [ -0.008407593, 0.01071167, -0.019180298, -0.032348633, 0.017227173, -0.0020542145, -0.017211914, -0.04425049, -0.006389618, 0.009437561, -0.020339966, 0.032714844, -0.051086426, -0.016967773, -0.0068588257, -0.0049476624, -0.0061454773, ...
http://arxiv.org/abs/2305.04835v3
How Do In-Context Examples Affect Compositional Generalization?
Shengnan An, Zeqi Lin, Qiang Fu, Bei Chen, Nanning Zheng, Jian-Guang Lou, Dongmei Zhang
cs.CL, cs.AI
Compositional generalization--understanding unseen combinations of seen primitives--is an essential reasoning capability in human intelligence. The AI community mainly studies this capability by fine-tuning neural networks on lots of training samples, while it is still unclear whether and how in-context learning--the p...
http://arxiv.org/pdf/2305.04835v3
2023-06-09T02:25:29
2023-05-08T16:32:18
{ "title": [ 0.079956055, 0.0076065063, -0.00522995, -0.052520752, -0.009353638, -0.012336731, -0.023361206, -0.020446777, -0.007106781, 0.047729492, -0.040527344, 0.023162842, 0.0008864403, -0.009223938, -0.009254456, -0.021224976, -0.0073547363, ...
http://arxiv.org/abs/2305.04824v1
Learning Summary-Worthy Visual Representation for Abstractive Summarization in Video
Zenan Xu, Xiaojun Meng, Yasheng Wang, Qinliang Su, Zexuan Qiu, Xin Jiang, Qun Liu
cs.CL
Multimodal abstractive summarization for videos (MAS) requires generating a concise textual summary to describe the highlights of a video according to multimodal resources, in our case, the video content and its transcript. Inspired by the success of the large-scale generative pre-trained language model (GPLM) in gener...
http://arxiv.org/pdf/2305.04824v1
2023-05-08T16:24:46
2023-05-08T16:24:46
{ "title": [ 0.010787964, 0.014518738, -0.035705566, -0.066467285, 0.0024795532, -0.008399963, -0.0357666, -0.017120361, 0.008865356, 0.05783081, -0.038970947, 0.013847351, -0.0146865845, 0.010818481, -0.068359375, -0.012107849, 0.006061554, 0.03...
http://arxiv.org/abs/2305.04812v3
Influence of External Information on Large Language Models Mirrors Social Cognitive Patterns
Ning Bian, Hongyu Lin, Peilin Liu, Yaojie Lu, Chunkang Zhang, Ben He, Xianpei Han, Le Sun
cs.CL
Social cognitive theory explains how people learn and acquire knowledge through observing others. Recent years have witnessed the rapid development of large language models (LLMs), which suggests their potential significance as agents in the society. LLMs, as AI agents, can observe external information, which shapes th...
http://arxiv.org/pdf/2305.04812v3
2023-10-20T10:18:44
2023-05-08T16:10:18
{ "title": [ 0.004047394, 0.0018959045, -0.02192688, -0.056243896, -0.015098572, -0.033233643, -0.029327393, -0.00491333, -0.028503418, 0.020233154, -0.02468872, -0.0015478134, 0.02053833, -0.035308838, 0.011764526, -0.012176514, 0.0038776398, 0....
http://arxiv.org/abs/2305.04808v2
CAT: A Contextualized Conceptualization and Instantiation Framework for Commonsense Reasoning
Weiqi Wang, Tianqing Fang, Baixuan Xu, Chun Yi Louis Bo, Yangqiu Song, Lei Chen
cs.CL
Commonsense reasoning, aiming at endowing machines with a human-like ability to make situational presumptions, is extremely challenging to generalize. For someone who barely knows about "meditation," while is knowledgeable about "singing," he can still infer that "meditation makes people relaxed" from the existing know...
http://arxiv.org/pdf/2305.04808v2
2023-05-10T11:40:27
2023-05-08T16:08:42
{ "title": [ 0.013389587, -0.02609253, -0.02420044, -0.08306885, 0.0076522827, 0.014678955, -0.038909912, -0.0068893433, 0.025146484, 0.01737976, -0.022888184, 0.0005340576, -0.004901886, -0.03451538, -0.02281189, 0.00022625923, -0.023147583, 0.0...
http://arxiv.org/abs/2305.06163v1
Algebra Error Classification with Large Language Models
Hunter McNichols, Mengxue Zhang, Andrew Lan
cs.CL, cs.AI
Automated feedback as students answer open-ended math questions has significant potential in improving learning outcomes at large scale. A key part of automated feedback systems is an error classification component, which identifies student errors and enables appropriate, predefined feedback to be deployed. Most existi...
http://arxiv.org/pdf/2305.06163v1
2023-05-08T15:51:38
2023-05-08T15:51:38
{ "title": [ -0.030685425, -0.0006542206, -0.03881836, -0.048950195, -0.006137848, -0.0059814453, -0.05532837, 0.0047073364, 0.022979736, 0.028701782, -0.010246277, -0.008018494, 0.0119018555, -0.025894165, -0.033813477, 0.016067505, -0.032043457, ...
http://arxiv.org/abs/2305.04790v3
MultiModal-GPT: A Vision and Language Model for Dialogue with Humans
Tao Gong, Chengqi Lyu, Shilong Zhang, Yudong Wang, Miao Zheng, Qian Zhao, Kuikun Liu, Wenwei Zhang, Ping Luo, Kai Chen
cs.CV, cs.CL
We present a vision and language model named MultiModal-GPT to conduct multi-round dialogue with humans. MultiModal-GPT can follow various instructions from humans, such as generating a detailed caption, counting the number of interested objects, and answering general questions from users. MultiModal-GPT is parameter-e...
http://arxiv.org/pdf/2305.04790v3
2023-06-13T13:31:12
2023-05-08T15:45:42
{ "title": [ -0.015403748, 0.00033903122, -0.048736572, -0.045654297, 0.003921509, -0.016418457, -0.050750732, -0.01586914, 0.015213013, 0.01525116, -0.011932373, -0.0143585205, -0.008125305, -0.029754639, -0.049621582, -0.0088272095, -0.006706238, ...
http://arxiv.org/abs/2305.04782v1
HistAlign: Improving Context Dependency in Language Generation by Aligning with History
David Wan, Shiyue Zhang, Mohit Bansal
cs.CL, cs.AI, cs.LG
Language models (LMs) can generate hallucinations and incoherent outputs, which highlights their weak context dependency. Cache-LMs, which augment LMs with a memory of recent history, can increase context dependency and have shown remarkable performance in diverse language generation tasks. However, we find that even w...
http://arxiv.org/pdf/2305.04782v1
2023-05-08T15:34:56
2023-05-08T15:34:56
{ "title": [ -0.02456665, -0.005542755, -0.05557251, -0.07501221, 0.013557434, 0.010139465, -0.057861328, 0.01727295, 0.007335663, 0.034851074, 0.014007568, 0.019729614, 0.0134887695, -0.027770996, -0.029556274, 0.027664185, -0.0024318695, 0.0202...
http://arxiv.org/abs/2305.04757v2
Augmented Large Language Models with Parametric Knowledge Guiding
Ziyang Luo, Can Xu, Pu Zhao, Xiubo Geng, Chongyang Tao, Jing Ma, Qingwei Lin, Daxin Jiang
cs.CL, cs.AI, cs.LG
Large Language Models (LLMs) have significantly advanced natural language processing (NLP) with their impressive language understanding and generation capabilities. However, their performance may be suboptimal for domain-specific tasks that require specialized knowledge due to limited exposure to the related data. Addi...
http://arxiv.org/pdf/2305.04757v2
2023-05-18T08:14:08
2023-05-08T15:05:16
{ "title": [ -0.029953003, -0.026901245, -0.043395996, -0.092041016, -0.022659302, 0.016326904, -0.09136963, 0.02545166, -0.0026893616, 0.03778076, -0.029342651, -0.01651001, 0.023666382, -0.009933472, -0.035827637, -0.0022640228, 0.00086545944, ...
http://arxiv.org/abs/2305.15323v1
ChatGPT: Vision and Challenges
Sukhpal Singh Gill, Rupinder Kaur
cs.CY, cs.CL
Artificial intelligence (AI) and machine learning have changed the nature of scientific inquiry in recent years. Of these, the development of virtual assistants has accelerated greatly in the past few years, with ChatGPT becoming a prominent AI language model. In this study, we examine the foundations, vision, research...
http://arxiv.org/pdf/2305.15323v1
2023-05-08T14:54:44
2023-05-08T14:54:44
{ "title": [ -0.0012578964, -0.015541077, -0.01776123, -0.033447266, 0.0072631836, 0.011444092, -0.013771057, -0.012962341, 0.0021915436, 0.015129089, -0.009529114, 0.016036987, -0.008636475, -0.028274536, -0.014564514, 0.023284912, -0.018356323, ...
http://arxiv.org/abs/2305.04749v1
Toeplitz Neural Network for Sequence Modeling
Zhen Qin, Xiaodong Han, Weixuan Sun, Bowen He, Dong Li, Dongxu Li, Yuchao Dai, Lingpeng Kong, Yiran Zhong
cs.CL, cs.CV
Sequence modeling has important applications in natural language processing and computer vision. Recently, the transformer-based models have shown strong performance on various sequence modeling tasks, which rely on attention to capture pairwise token relations, and position embedding to inject positional information. ...
http://arxiv.org/pdf/2305.04749v1
2023-05-08T14:49:01
2023-05-08T14:49:01
{ "title": [ -0.028457642, -0.005340576, -0.04159546, -0.053710938, 0.00081825256, -0.0036697388, -0.04244995, 0.002418518, 0.021255493, 0.038726807, -0.01828003, -0.020523071, 0.012611389, -0.024261475, -0.059906006, 0.018066406, -0.0063323975, ...
http://arxiv.org/abs/2305.04737v1
SkillQG: Learning to Generate Question for Reading Comprehension Assessment
Xiaoqiang Wang, Bang Liu, Siliang Tang, Lingfei Wu
cs.CL
We present $\textbf{$\texttt{SkillQG}$}$: a question generation framework with controllable comprehension types for assessing and improving machine reading comprehension models. Existing question generation systems widely differentiate questions by $\textit{literal}$ information such as question words and answer types ...
http://arxiv.org/pdf/2305.04737v1
2023-05-08T14:40:48
2023-05-08T14:40:48
{ "title": [ -0.025863647, -0.013977051, -0.015449524, -0.01751709, -0.026168823, 0.003479004, -0.019210815, -0.02571106, 0.022781372, 0.030075073, -0.01486969, 0.010269165, -0.012428284, -0.0010957718, -0.07312012, 0.003255844, -0.014350891, 0.0...
http://arxiv.org/abs/2305.04720v2
DEnsity: Open-domain Dialogue Evaluation Metric using Density Estimation
ChaeHun Park, Seungil Chad Lee, Daniel Rim, Jaegul Choo
cs.CL
Despite the recent advances in open-domain dialogue systems, building a reliable evaluation metric is still a challenging problem. Recent studies proposed learnable metrics based on classification models trained to distinguish the correct response. However, neural classifiers are known to make overly confident predicti...
http://arxiv.org/pdf/2305.04720v2
2023-05-25T11:40:59
2023-05-08T14:10:40
{ "title": [ -0.0029087067, 0.0010232925, -0.03414917, -0.049591064, 0.012878418, -0.012840271, -0.01965332, -0.022155762, 0.0041656494, 0.019165039, -0.003063202, -0.0057792664, -0.005672455, -0.032226562, -0.05230713, -0.0072898865, -0.024368286, ...
http://arxiv.org/abs/2305.04676v1
Enhancing Knowledge Graph Construction Using Large Language Models
Milena Trajanoska, Riste Stojanov, Dimitar Trajanov
cs.CL
The growing trend of Large Language Models (LLM) development has attracted significant attention, with models for various applications emerging consistently. However, the combined application of Large Language Models with semantic technologies for reasoning and inference is still a challenging task. This paper analyzes...
http://arxiv.org/pdf/2305.04676v1
2023-05-08T12:53:06
2023-05-08T12:53:06
{ "title": [ -0.047576904, -0.020706177, -0.066711426, -0.088012695, -0.010169983, 0.020645142, -0.09100342, 0.02017212, 0.00026798248, 0.023513794, -0.008956909, -0.0004696846, -0.009429932, -0.014251709, -0.03137207, -0.0011053085, -0.0065956116, ...
http://arxiv.org/abs/2305.04673v2
PreCog: Exploring the Relation between Memorization and Performance in Pre-trained Language Models
Leonardo Ranaldi, Elena Sofia Ruzzetti, Fabio Massimo Zanzotto
cs.CL
Pre-trained Language Models such as BERT are impressive machines with the ability to memorize, possibly generalized learning examples. We present here a small, focused contribution to the analysis of the interplay between memorization and performance of BERT in downstream tasks. We propose PreCog, a measure for evaluat...
http://arxiv.org/pdf/2305.04673v2
2023-05-09T05:37:53
2023-05-08T12:51:00
{ "title": [ -0.030380249, -0.01499176, -0.037139893, -0.051330566, -0.015060425, -0.009269714, -0.04660034, 0.011657715, -0.024002075, 0.04763794, -0.011161804, 0.016677856, 0.0037384033, -0.015258789, -0.042541504, -0.009361267, -0.00819397, 0....
http://arxiv.org/abs/2305.04636v1
Enhancing Continual Relation Extraction via Classifier Decomposition
Heming Xia, Peiyi Wang, Tianyu Liu, Binghuai Lin, Yunbo Cao, Zhifang Sui
cs.CL
Continual relation extraction (CRE) models aim at handling emerging new relations while avoiding catastrophically forgetting old ones in the streaming data. Though improvements have been shown by previous CRE studies, most of them only adopt a vanilla strategy when models first learn representations of new relations. I...
http://arxiv.org/pdf/2305.04636v1
2023-05-08T11:29:33
2023-05-08T11:29:33
{ "title": [ 0.010467529, 0.009933472, -0.054718018, -0.09564209, 0.032348633, -0.006942749, -0.04446411, 0.00831604, 0.04724121, 0.011413574, -0.013183594, -0.0013170242, -0.011505127, -0.007534027, -0.054260254, 0.0074768066, -0.0057754517, 0.0...
http://arxiv.org/abs/2305.04631v1
XAI in Computational Linguistics: Understanding Political Leanings in the Slovenian Parliament
Bojan Evkoski, Senja Pollak
cs.CL, cs.AI
The work covers the development and explainability of machine learning models for predicting political leanings through parliamentary transcriptions. We concentrate on the Slovenian parliament and the heated debate on the European migrant crisis, with transcriptions from 2014 to 2020. We develop both classical machine ...
http://arxiv.org/pdf/2305.04631v1
2023-05-08T11:19:21
2023-05-08T11:19:21
{ "title": [ -0.017166138, -0.0052452087, -0.044952393, 0.0065841675, -0.031234741, -0.015731812, -0.040100098, -0.018737793, -0.024520874, 0.02973938, -0.042236328, 0.0016670227, 0.028213501, -0.029434204, -0.0052452087, -0.018844604, -0.012107849, ...
http://arxiv.org/abs/2305.04599v1
Cone: Unsupervised Contrastive Opinion Extraction
Runcong Zhao, Lin Gui, Yulan He
cs.CL
Contrastive opinion extraction aims to extract a structured summary or key points organised as positive and negative viewpoints towards a common aspect or topic. Most recent works for unsupervised key point extraction is largely built on sentence clustering or opinion summarisation based on the popularity of opinions e...
http://arxiv.org/pdf/2305.04599v1
2023-05-08T10:18:30
2023-05-08T10:18:30
{ "title": [ 0.0124053955, -0.000096321106, -0.016815186, -0.047912598, 0.015670776, -0.0073623657, -0.013183594, -0.038726807, 0.028900146, 0.026138306, -0.038208008, 0.0033226013, -0.017181396, 0.008338928, -0.06402588, -0.009559631, -0.027160645, ...
http://arxiv.org/abs/2305.05383v1
Code Execution with Pre-trained Language Models
Chenxiao Liu, Shuai Lu, Weizhu Chen, Daxin Jiang, Alexey Svyatkovskiy, Shengyu Fu, Neel Sundaresan, Nan Duan
cs.PL, cs.AI, cs.CL, cs.SE
Code execution is a fundamental aspect of programming language semantics that reflects the exact behavior of the code. However, most pre-trained models for code intelligence ignore the execution trace and only rely on source code and syntactic structures. In this paper, we investigate how well pre-trained models can un...
http://arxiv.org/pdf/2305.05383v1
2023-05-08T10:00:05
2023-05-08T10:00:05
{ "title": [ -0.028442383, -0.019500732, -0.03488159, -0.06732178, -0.003479004, -0.0039901733, -0.06335449, 0.010894775, -0.012626648, 0.030395508, 0.0049362183, -0.006198883, -0.001288414, -0.02923584, -0.031402588, 0.011199951, -0.0062942505, ...
http://arxiv.org/abs/2305.04582v2
MultiTACRED: A Multilingual Version of the TAC Relation Extraction Dataset
Leonhard Hennig, Philippe Thomas, Sebastian Möller
cs.CL
Relation extraction (RE) is a fundamental task in information extraction, whose extension to multilingual settings has been hindered by the lack of supervised resources comparable in size to large English datasets such as TACRED (Zhang et al., 2017). To address this gap, we introduce the MultiTACRED dataset, covering 1...
http://arxiv.org/pdf/2305.04582v2
2023-05-15T07:24:58
2023-05-08T09:48:21
{ "title": [ -0.017303467, -0.03643799, -0.035247803, -0.0770874, -0.017486572, 0.023223877, -0.085998535, -0.010307312, 0.037353516, 0.0357666, 0.00018107891, -0.0031337738, 0.0030117035, -0.02960205, -0.053710938, 0.026947021, -0.0055656433, 0....
http://arxiv.org/abs/2305.04573v1
HiFi: High-Information Attention Heads Hold for Parameter-Efficient Model Adaptation
Anchun Gui, Han Xiao
cs.CL
To fully leverage the advantages of large-scale pre-trained language models (PLMs) on downstream tasks, it has become a ubiquitous adaptation paradigm to fine-tune the entire parameters of PLMs. However, this paradigm poses issues of inefficient updating and resource over-consuming for fine-tuning in data-scarce and re...
http://arxiv.org/pdf/2305.04573v1
2023-05-08T09:31:13
2023-05-08T09:31:13
{ "title": [ 0.0014343262, -0.012306213, -0.017303467, -0.05038452, -0.01525116, -0.0042495728, -0.025665283, -0.024642944, 0.04046631, 0.03643799, -0.018997192, 0.0039100647, 0.042816162, -0.024414062, -0.02293396, 0.039733887, 0.019058228, 0.00...
http://arxiv.org/abs/2305.04572v2
Putting Natural in Natural Language Processing
Grzegorz Chrupała
cs.CL, cs.AI, eess.AS
Human language is firstly spoken and only secondarily written. Text, however, is a very convenient and efficient representation of language, and modern civilization has made it ubiquitous. Thus the field of NLP has overwhelmingly focused on processing written rather than spoken language. Work on spoken language, on the...
http://arxiv.org/pdf/2305.04572v2
2023-05-23T14:15:00
2023-05-08T09:29:31
{ "title": [ -0.01612854, -0.013633728, -0.024459839, -0.055236816, -0.010177612, 0.005874634, -0.02822876, -0.007911682, -0.005241394, 0.032562256, -0.04425049, 0.0008573532, -0.008377075, -0.00573349, -0.036865234, -0.0048980713, -0.0062828064, ...
http://arxiv.org/abs/2305.04561v2
Boosting Radiology Report Generation by Infusing Comparison Prior
Sanghwan Kim, Farhad Nooralahzadeh, Morteza Rohanian, Koji Fujimoto, Mizuho Nishio, Ryo Sakamoto, Fabio Rinaldi, Michael Krauthammer
cs.CL
Recent transformer-based models have made significant strides in generating radiology reports from chest X-ray images. However, a prominent challenge remains: these models often lack prior knowledge, resulting in the generation of synthetic reports that mistakenly reference non-existent prior exams. This discrepancy ca...
http://arxiv.org/pdf/2305.04561v2
2023-06-05T10:28:11
2023-05-08T09:12:44
{ "title": [ 0.007171631, 0.0012054443, -0.06201172, -0.08111572, 0.014282227, -0.0078125, -0.044128418, 0.008216858, -0.01725769, 0.02229309, -0.020202637, 0.023422241, -0.033447266, -0.024963379, -0.0259552, -0.016479492, 0.001625061, 0.0156402...
http://arxiv.org/abs/2305.04557v1
Toward Adversarial Training on Contextualized Language Representation
Hongqiu Wu, Yongxiang Liu, Hanwen Shi, Hai Zhao, Min Zhang
cs.CL
Beyond the success story of adversarial training (AT) in the recent text domain on top of pre-trained language models (PLMs), our empirical study showcases the inconsistent gains from AT on some tasks, e.g. commonsense reasoning, named entity recognition. This paper investigates AT from the perspective of the contextua...
http://arxiv.org/pdf/2305.04557v1
2023-05-08T08:56:51
2023-05-08T08:56:51
{ "title": [ 0.003917694, 0.0056266785, -0.016159058, -0.052001953, 0.008605957, 0.0015649796, -0.02166748, -0.009384155, -0.036254883, 0.03869629, 0.008880615, 0.01751709, -0.013198853, -0.016311646, -0.026123047, -0.0029582977, -0.007003784, 0....
http://arxiv.org/abs/2305.04556v2
Non-Autoregressive Math Word Problem Solver with Unified Tree Structure
Yi Bin, Mengqun Han, Wenhao Shi, Lei Wang, Yang Yang, See-Kiong Ng, Heng Tao Shen
cs.CL, cs.AI
Existing MWP solvers employ sequence or binary tree to present the solution expression and decode it from given problem description. However, such structures fail to handle the variants that can be derived via mathematical manipulation, e.g., $(a_1+a_2) * a_3$ and $a_1 * a_3+a_2 * a_3$ can both be possible valid soluti...
http://arxiv.org/pdf/2305.04556v2
2023-10-28T07:09:38
2023-05-08T08:53:37
{ "title": [ -0.019836426, -0.002708435, -0.042999268, -0.03253174, -0.030960083, -0.014259338, -0.046417236, 0.0071868896, -0.0066375732, 0.03463745, -0.02508545, 0.004890442, 0.010322571, -0.024108887, -0.02168274, -0.01260376, 0.010482788, -0....
http://arxiv.org/abs/2305.04547v1
Diffusion Theory as a Scalpel: Detecting and Purifying Poisonous Dimensions in Pre-trained Language Models Caused by Backdoor or Bias
Zhiyuan Zhang, Deli Chen, Hao Zhou, Fandong Meng, Jie Zhou, Xu Sun
cs.CL
Pre-trained Language Models (PLMs) may be poisonous with backdoors or bias injected by the suspicious attacker during the fine-tuning process. A core challenge of purifying potentially poisonous PLMs is precisely finding poisonous dimensions. To settle this issue, we propose the Fine-purifying approach, which utilizes ...
http://arxiv.org/pdf/2305.04547v1
2023-05-08T08:40:30
2023-05-08T08:40:30
{ "title": [ -0.006000519, 0.0027103424, -0.036956787, -0.033813477, 0.0051574707, -0.017852783, -0.021484375, -0.027511597, -0.01739502, 0.007881165, -0.0045166016, 0.013923645, 0.008590698, -0.04360962, 0.010261536, -0.004573822, -0.012359619, ...
http://arxiv.org/abs/2305.04533v1
Prompted LLMs as Chatbot Modules for Long Open-domain Conversation
Gibbeum Lee, Volker Hartmann, Jongho Park, Dimitris Papailiopoulos, Kangwook Lee
cs.CL, cs.AI, cs.LG
In this paper, we propose MPC (Modular Prompted Chatbot), a new approach for creating high-quality conversational agents without the need for fine-tuning. Our method utilizes pre-trained large language models (LLMs) as individual modules for long-term consistency and flexibility, by using techniques such as few-shot pr...
http://arxiv.org/pdf/2305.04533v1
2023-05-08T08:09:00
2023-05-08T08:09:00
{ "title": [ -0.004272461, -0.031188965, -0.02708435, -0.053985596, -0.04611206, -0.012565613, -0.0390625, 0.019348145, -0.02859497, 0.008399963, -0.025650024, 0.017837524, -0.001704216, -0.044128418, 0.022247314, -0.031951904, -0.007850647, 0.02...
http://arxiv.org/abs/2305.04530v1
A Multi-Modal Context Reasoning Approach for Conditional Inference on Joint Textual and Visual Clues
Yunxin Li, Baotian Hu, Xinyu Chen, Yuxin Ding, Lin Ma, Min Zhang
cs.CL
Conditional inference on joint textual and visual clues is a multi-modal reasoning task that textual clues provide prior permutation or external knowledge, which are complementary with visual content and pivotal to deducing the correct option. Previous methods utilizing pretrained vision-language models (VLMs) have ach...
http://arxiv.org/pdf/2305.04530v1
2023-05-08T08:05:40
2023-05-08T08:05:40
{ "title": [ -0.017715454, 0.0029125214, -0.043182373, -0.060791016, 0.01184082, 0.010910034, -0.050598145, -0.0158844, 0.024978638, 0.017837524, -0.04663086, 0.007659912, 0.014770508, -0.013191223, -0.025848389, -0.014350891, -0.0005130768, 0.02...
http://arxiv.org/abs/2305.04522v1
Event Knowledge Incorporation with Posterior Regularization for Event-Centric Question Answering
Junru Lu, Gabriele Pergola, Lin Gui, Yulan He
cs.CL
We propose a simple yet effective strategy to incorporate event knowledge extracted from event trigger annotations via posterior regularization to improve the event reasoning capability of mainstream question-answering (QA) models for event-centric QA. In particular, we define event-related knowledge constraints based ...
http://arxiv.org/pdf/2305.04522v1
2023-05-08T07:45:12
2023-05-08T07:45:12
{ "title": [ -0.015655518, 0.005886078, -0.033111572, -0.09039307, 0.017089844, 0.011451721, -0.04815674, 0.011917114, 0.03012085, 0.02696228, -0.000012099743, 0.0149002075, -0.0013084412, -0.012939453, -0.055633545, 0.017791748, 0.0181427, 0.024...
http://arxiv.org/abs/2305.04518v1
Sparks of Artificial General Recommender (AGR): Early Experiments with ChatGPT
Guo Lin, Yongfeng Zhang
cs.IR, cs.CL, cs.LG
This study investigates the feasibility of developing an Artificial General Recommender (AGR), facilitated by recent advancements in Large Language Models (LLMs). An AGR comprises both conversationality and universality to engage in natural dialogues and generate recommendations across various domains. We propose ten f...
http://arxiv.org/pdf/2305.04518v1
2023-05-08T07:28:16
2023-05-08T07:28:16
{ "title": [ 0.0071525574, 0.008117676, -0.034484863, -0.014152527, 0.020996094, -0.020812988, 0.020904541, -0.024124146, -0.009895325, 0.024749756, -0.019851685, 0.033447266, -0.009513855, -0.011550903, -0.020874023, 0.020080566, -0.022079468, 0...
http://arxiv.org/abs/2305.04505v2
Target-Side Augmentation for Document-Level Machine Translation
Guangsheng Bao, Zhiyang Teng, Yue Zhang
cs.CL
Document-level machine translation faces the challenge of data sparsity due to its long input length and a small amount of training data, increasing the risk of learning spurious patterns. To address this challenge, we propose a target-side augmentation method, introducing a data augmentation (DA) model to generate man...
http://arxiv.org/pdf/2305.04505v2
2023-06-04T11:05:03
2023-05-08T07:01:18
{ "title": [ -0.024032593, -0.012901306, -0.015235901, -0.060028076, -0.006538391, 0.015068054, -0.039489746, 0.014816284, 0.014472961, 0.048431396, -0.0077705383, 0.008049011, 0.031341553, -0.016937256, -0.039733887, 0.025817871, -0.0047683716, ...
http://arxiv.org/abs/2305.04493v2
Token-Level Fitting Issues of Seq2seq Models
Guangsheng Bao, Zhiyang Teng, Yue Zhang
cs.CL
Sequence-to-sequence (seq2seq) models have been widely used for natural language processing, computer vision, and other deep learning tasks. We find that seq2seq models trained with early-stopping suffer from issues at the token level. In particular, while some tokens in the vocabulary demonstrate overfitting, others u...
http://arxiv.org/pdf/2305.04493v2
2023-06-22T07:42:08
2023-05-08T06:40:24
{ "title": [ -0.029052734, -0.0013475418, -0.0335083, -0.03817749, -0.010421753, -0.0048980713, -0.05166626, -0.028457642, -0.011787415, 0.012550354, -0.0059394836, -0.00095939636, -0.027877808, -0.045898438, -0.022460938, 0.009559631, -0.014289856, ...
http://arxiv.org/abs/2305.04492v8
MGR: Multi-generator Based Rationalization
Wei Liu, Haozhao Wang, Jun Wang, Ruixuan Li, Xinyang Li, Yuankai Zhang, Yang Qiu
cs.LG, cs.CL
Rationalization is to employ a generator and a predictor to construct a self-explaining NLP model in which the generator selects a subset of human-intelligible pieces of the input text to the following predictor. However, rationalization suffers from two key challenges, i.e., spurious correlation and degeneration, wher...
http://arxiv.org/pdf/2305.04492v8
2023-07-23T08:54:43
2023-05-08T06:36:46
{ "title": [ -0.02468872, 0.008766174, -0.046020508, -0.076293945, 0.013656616, 0.008659363, -0.0501709, -0.012031555, -0.03375244, 0.029647827, 0.0149002075, 0.0071640015, 0.013397217, -0.050567627, -0.004840851, 0.019424438, -0.007709503, 0.043...
http://arxiv.org/abs/2305.04476v4
AlignSTS: Speech-to-Singing Conversion via Cross-Modal Alignment
Ruiqi Li, Rongjie Huang, Lichao Zhang, Jinglin Liu, Zhou Zhao
eess.AS, cs.CL, cs.MM, cs.SD
The speech-to-singing (STS) voice conversion task aims to generate singing samples corresponding to speech recordings while facing a major challenge: the alignment between the target (singing) pitch contour and the source (speech) content is difficult to learn in a text-free situation. This paper proposes AlignSTS, an ...
http://arxiv.org/pdf/2305.04476v4
2023-05-24T16:37:42
2023-05-08T06:02:10
{ "title": [ -0.027938843, -0.016723633, -0.056640625, -0.06945801, -0.002811432, 0.0054473877, -0.060760498, 0.014381409, 0.010246277, 0.051605225, -0.024368286, 0.0131073, 0.041900635, -0.021072388, -0.04840088, 0.012245178, -0.014572144, 0.048...
http://arxiv.org/abs/2305.04465v1
Can Diffusion Model Achieve Better Performance in Text Generation? Bridging the Gap between Training and Inference!
Zecheng Tang, Pinzheng Wang, Keyan Zhou, Juntao Li, Ziqiang Cao, Min Zhang
cs.CL
Diffusion models have been successfully adapted to text generation tasks by mapping the discrete text into the continuous space. However, there exist nonnegligible gaps between training and inference, owing to the absence of the forward process during inference. Thus, the model only predicts based on the previously gen...
http://arxiv.org/pdf/2305.04465v1
2023-05-08T05:32:22
2023-05-08T05:32:22
{ "title": [ -0.03265381, 0.018707275, -0.047576904, -0.012481689, -0.0042800903, -0.015808105, -0.027374268, -0.01651001, 0.02015686, 0.014785767, -0.061706543, 0.010932922, 0.017074585, -0.019195557, -0.0008292198, -0.025436401, -0.02607727, 0....
http://arxiv.org/abs/2305.04460v1
Language Independent Neuro-Symbolic Semantic Parsing for Form Understanding
Bhanu Prakash Voutharoja, Lizhen Qu, Fatemeh Shiri
cs.CL, cs.AI
Recent works on form understanding mostly employ multimodal transformers or large-scale pre-trained language models. These models need ample data for pre-training. In contrast, humans can usually identify key-value pairings from a form only by looking at layouts, even if they don't comprehend the language used. No prio...
http://arxiv.org/pdf/2305.04460v1
2023-05-08T05:03:07
2023-05-08T05:03:07
{ "title": [ -0.019927979, -0.008468628, -0.002571106, -0.056793213, -0.017807007, -0.0028381348, -0.045288086, -0.0025501251, 0.00894928, 0.027130127, -0.009231567, -0.02406311, 0.006126404, -0.026672363, -0.050842285, 0.002002716, 0.00031995773, ...
http://arxiv.org/abs/2305.04446v1
Facilitating Fine-grained Detection of Chinese Toxic Language: Hierarchical Taxonomy, Resources, and Benchmarks
Junyu Lu, Bo Xu, Xiaokun Zhang, Changrong Min, Liang Yang, Hongfei Lin
cs.CL, cs.AI
The widespread dissemination of toxic online posts is increasingly damaging to society. However, research on detecting toxic language in Chinese has lagged significantly. Existing datasets lack fine-grained annotation of toxic types and expressions, and ignore the samples with indirect toxicity. In addition, it is cruc...
http://arxiv.org/pdf/2305.04446v1
2023-05-08T03:50:38
2023-05-08T03:50:38
{ "title": [ 0.012786865, 0.0051002502, -0.02331543, -0.07269287, 0.02633667, -0.00381279, -0.015991211, -0.0034332275, -0.014877319, 0.022216797, -0.015037537, 0.025222778, -0.010017395, -0.021377563, -0.026321411, -0.009216309, -0.016433716, 0....
http://arxiv.org/abs/2305.04429v1
Improving Cross-Task Generalization with Step-by-Step Instructions
Yang Wu, Yanyan Zhao, Zhongyang Li, Bing Qin, Kai Xiong
cs.CL
Instruction tuning has been shown to be able to improve cross-task generalization of language models. However, it is still challenging for language models to complete the target tasks following the instructions, as the instructions are general and lack intermediate steps. To address this problem, we propose to incorpor...
http://arxiv.org/pdf/2305.04429v1
2023-05-08T02:50:41
2023-05-08T02:50:41
{ "title": [ -0.01625061, -0.00046300888, -0.036102295, -0.05505371, 0.007507324, -0.0036182404, -0.035980225, -0.014389038, 0.012176514, 0.032165527, -0.028808594, 0.035125732, -0.018508911, 0.0007419586, -0.028274536, 0.012397766, 0.010681152, ...
http://arxiv.org/abs/2305.04417v1
Unlocking Practical Applications in Legal Domain: Evaluation of GPT for Zero-Shot Semantic Annotation of Legal Texts
Jaromir Savelka
cs.CL, cs.AI
We evaluated the capability of a state-of-the-art generative pre-trained transformer (GPT) model to perform semantic annotation of short text snippets (one to few sentences) coming from legal documents of various types. Discussions of potential uses (e.g., document drafting, summarization) of this emerging technology i...
http://arxiv.org/pdf/2305.04417v1
2023-05-08T01:55:53
2023-05-08T01:55:53
{ "title": [ -0.018966675, -0.0029716492, -0.031585693, -0.04425049, -0.014663696, -0.00434494, -0.040283203, -0.00014328957, 0.024536133, 0.038146973, -0.032928467, 0.015823364, 0.019012451, -0.00856781, -0.053771973, 0.010681152, -0.015312195, ...
http://arxiv.org/abs/2305.04400v1
Do Large Language Models Show Decision Heuristics Similar to Humans? A Case Study Using GPT-3.5
Gaurav Suri, Lily R. Slater, Ali Ziaee, Morgan Nguyen
cs.AI, cs.CL, q-bio.NC
A Large Language Model (LLM) is an artificial intelligence system that has been trained on vast amounts of natural language data, enabling it to generate human-like responses to written or spoken language input. GPT-3.5 is an example of an LLM that supports a conversational agent called ChatGPT. In this work, we used a...
http://arxiv.org/pdf/2305.04400v1
2023-05-08T01:02:52
2023-05-08T01:02:52
{ "title": [ -0.017150879, 0.013038635, -0.015655518, -0.007472992, -0.023635864, -0.018920898, -0.03189087, -0.03213501, -0.0095825195, 0.020599365, -0.040649414, 0.017333984, -0.0009937286, -0.0011310577, -0.029006958, -0.0016727448, -0.020385742, ...
http://arxiv.org/abs/2305.04388v1
Language Models Don't Always Say What They Think: Unfaithful Explanations in Chain-of-Thought Prompting
Miles Turpin, Julian Michael, Ethan Perez, Samuel R. Bowman
cs.CL, cs.AI
Large Language Models (LLMs) can achieve strong performance on many tasks by producing step-by-step reasoning before giving a final output, often referred to as chain-of-thought reasoning (CoT). It is tempting to interpret these CoT explanations as the LLM's process for solving a task. However, we find that CoT explana...
http://arxiv.org/pdf/2305.04388v1
2023-05-07T22:44:25
2023-05-07T22:44:25
{ "title": [ -0.015991211, -0.016845703, 0.010955811, -0.036315918, -0.028076172, -0.0018920898, -0.0289917, -0.03277588, -0.0112838745, 0.0023422241, -0.034088135, 0.0019359589, 0.0004930496, -0.027496338, -0.026260376, -0.018615723, -0.013458252, ...
http://arxiv.org/abs/2305.04365v1
LatinCy: Synthetic Trained Pipelines for Latin NLP
Patrick J. Burns
cs.CL
This paper introduces LatinCy, a set of trained general purpose Latin-language "core" pipelines for use with the spaCy natural language processing framework. The models are trained on a large amount of available Latin data, including all five of the Latin Universal Dependency treebanks, which have been preprocessed to ...
http://arxiv.org/pdf/2305.04365v1
2023-05-07T19:59:01
2023-05-07T19:59:01
{ "title": [ -0.05053711, -0.029342651, -0.026504517, -0.049957275, -0.018539429, 0.03262329, -0.04748535, 0.006450653, 0.0063819885, 0.016784668, -0.024276733, -0.002029419, -0.011543274, -0.017166138, -0.05606079, -0.0015296936, -0.024047852, 0...
http://arxiv.org/abs/2305.04356v1
Stanford MLab at SemEval-2023 Task 10: Exploring GloVe- and Transformer-Based Methods for the Explainable Detection of Online Sexism
Hee Jung Choi, Trevor Chow, Aaron Wan, Hong Meng Yam, Swetha Yogeswaran, Beining Zhou
cs.CL, cs.LG
In this paper, we discuss the methods we applied at SemEval-2023 Task 10: Towards the Explainable Detection of Online Sexism. Given an input text, we perform three classification tasks to predict whether the text is sexist and classify the sexist text into subcategories in order to provide an additional explanation as ...
http://arxiv.org/pdf/2305.04356v1
2023-05-07T18:58:54
2023-05-07T18:58:54
{ "title": [ 0.011413574, 0.004776001, -0.017242432, -0.0011291504, -0.004787445, -0.025039673, 0.011665344, -0.0158844, 0.00415802, 0.029190063, -0.024475098, 0.022476196, 0.016845703, -0.014732361, -0.005290985, -0.007843018, -0.029190063, 0.06...
http://arxiv.org/abs/2305.04346v1
Laziness Is a Virtue When It Comes to Compositionality in Neural Semantic Parsing
Maxwell Crouse, Pavan Kapanipathi, Subhajit Chaudhury, Tahira Naseem, Ramon Astudillo, Achille Fokoue, Tim Klinger
cs.CL, cs.AI
Nearly all general-purpose neural semantic parsers generate logical forms in a strictly top-down autoregressive fashion. Though such systems have achieved impressive results across a variety of datasets and domains, recent works have called into question whether they are ultimately limited in their ability to compositi...
http://arxiv.org/pdf/2305.04346v1
2023-05-07T17:53:08
2023-05-07T17:53:08
{ "title": [ 0.021270752, 0.014854431, -0.034576416, -0.061340332, -0.0024852753, -0.018432617, -0.022354126, -0.017242432, -0.03225708, 0.052215576, -0.009780884, -0.022399902, 0.007270813, -0.026245117, -0.040496826, -0.032806396, -0.021575928, ...
http://arxiv.org/abs/2305.04344v1
Empowering Language Model with Guided Knowledge Fusion for Biomedical Document Re-ranking
Deepak Gupta, Dina Demner-Fushman
cs.CL
Pre-trained language models (PLMs) have proven to be effective for document re-ranking task. However, they lack the ability to fully interpret the semantics of biomedical and health-care queries and often rely on simplistic patterns for retrieving documents. To address this challenge, we propose an approach that integr...
http://arxiv.org/pdf/2305.04344v1
2023-05-07T17:45:47
2023-05-07T17:45:47
{ "title": [ -0.013061523, -0.0009150505, -0.027999878, -0.05267334, -0.018814087, 0.0112838745, -0.061706543, -0.01701355, 0.03289795, 0.04650879, -0.022155762, -0.01071167, 0.048950195, -0.0038223267, -0.053710938, 0.0152282715, -0.007686615, 0...
http://arxiv.org/abs/2305.04329v2
FACTIFY-5WQA: 5W Aspect-based Fact Verification through Question Answering
Anku Rani, S. M Towhidul Islam Tonmoy, Dwip Dalal, Shreya Gautam, Megha Chakraborty, Aman Chadha, Amit Sheth, Amitava Das
cs.CL
Automatic fact verification has received significant attention recently. Contemporary automatic fact-checking systems focus on estimating truthfulness using numerical scores which are not human-interpretable. A human fact-checker generally follows several logical steps to verify a verisimilitude claim and conclude whet...
http://arxiv.org/pdf/2305.04329v2
2023-05-28T16:33:06
2023-05-07T16:52:21
{ "title": [ -0.014564514, -0.013336182, -0.04031372, -0.09979248, 0.008644104, 0.021377563, -0.051239014, 0.027145386, -0.0041046143, 0.03842163, 0.010246277, 0.0046195984, 0.025283813, -0.033111572, -0.041259766, 0.0115356445, -0.010482788, 0.0...
http://arxiv.org/abs/2305.04320v2
Unified Demonstration Retriever for In-Context Learning
Xiaonan Li, Kai Lv, Hang Yan, Tianyang Lin, Wei Zhu, Yuan Ni, Guotong Xie, Xiaoling Wang, Xipeng Qiu
cs.CL
In-context learning is a new learning paradigm where a language model conditions on a few input-output pairs (demonstrations) and a test input, and directly outputs the prediction. It has been shown highly dependent on the provided demonstrations and thus promotes the research of demonstration retrieval: given a test i...
http://arxiv.org/pdf/2305.04320v2
2023-05-16T05:59:25
2023-05-07T16:07:11
{ "title": [ 0.018997192, 0.018112183, -0.04638672, -0.08984375, 0.033233643, -0.012786865, -0.02961731, -0.011627197, -0.015457153, 0.043121338, -0.026519775, 0.017944336, 0.008644104, -0.020553589, 0.0006132126, -0.0033016205, 0.0012073517, 0.0...
http://arxiv.org/abs/2305.06273v1
Learning Robust Self-attention Features for Speech Emotion Recognition with Label-adaptive Mixup
Lei Kang, Lichao Zhang, Dazhi Jiang
cs.CL, cs.SD, eess.AS
Speech Emotion Recognition (SER) is to recognize human emotions in a natural verbal interaction scenario with machines, which is considered as a challenging problem due to the ambiguous human emotions. Despite the recent progress in SER, state-of-the-art models struggle to achieve a satisfactory performance. We propose...
http://arxiv.org/pdf/2305.06273v1
2023-05-07T15:10:59
2023-05-07T15:10:59
{ "title": [ 0.0008826256, 0.027145386, -0.060668945, -0.05859375, 0.03353882, -0.014404297, -0.027801514, 0.000702858, 0.013519287, 0.044555664, -0.014579773, 0.0018014908, 0.024398804, -0.027023315, -0.05596924, 0.0065345764, -0.008834839, 0.03...
http://arxiv.org/abs/2305.04297v1
HIORE: Leveraging High-order Interactions for Unified Entity Relation Extraction
Yijun Wang, Changzhi Sun, Yuanbin Wu, Lei Li, Junchi Yan, Hao Zhou
cs.CL
Entity relation extraction consists of two sub-tasks: entity recognition and relation extraction. Existing methods either tackle these two tasks separately or unify them with word-by-word interactions. In this paper, we propose HIORE, a new method for unified entity relation extraction. The key insight is to leverage t...
http://arxiv.org/pdf/2305.04297v1
2023-05-07T14:57:42
2023-05-07T14:57:42
{ "title": [ 0.011398315, -0.004676819, -0.052124023, -0.103515625, 0.03756714, -0.0031986237, -0.036132812, 0.009857178, 0.032440186, 0.027893066, -0.008178711, 0.02029419, 0.022125244, -0.005558014, -0.035736084, 0.02204895, -0.029022217, 0.050...
http://arxiv.org/abs/2305.04265v1
An Investigation on Word Embedding Offset Clustering as Relationship Classification
Didier Gohourou, Kazuhiro Kuwabara
cs.CL
Vector representations obtained from word embedding are the source of many groundbreaking advances in natural language processing. They yield word representations that are capable of capturing semantics and analogies of words within a text corpus. This study is an investigation in an attempt to elicit a vector represen...
http://arxiv.org/pdf/2305.04265v1
2023-05-07T13:03:17
2023-05-07T13:03:17
{ "title": [ -0.011894226, -0.022094727, -0.041656494, -0.0368042, -0.007972717, -0.015586853, -0.029251099, -0.0054473877, 0.021713257, 0.008094788, -0.024108887, -0.011978149, 0.008369446, -0.025466919, -0.029571533, -0.0075950623, -0.014984131, ...
http://arxiv.org/abs/2305.04934v2
Generative Pretrained Autoregressive Transformer Graph Neural Network applied to the Analysis and Discovery of Novel Proteins
Markus J. Buehler
q-bio.BM, cond-mat.dis-nn, cond-mat.soft, cs.CL, cs.LG
We report a flexible language-model based deep learning strategy, applied here to solve complex forward and inverse problems in protein modeling, based on an attention neural network that integrates transformer and graph convolutional architectures in a causal multi-headed graph mechanism, to realize a generative pretr...
http://arxiv.org/pdf/2305.04934v2
2023-07-11T12:41:39
2023-05-07T12:30:24
{ "title": [ -0.01852417, -0.0078125, -0.05203247, -0.05419922, -0.0007276535, -0.02861023, -0.04776001, 0.012283325, -0.002790451, 0.02331543, 0.011802673, 0.005947113, 0.0029621124, -0.04006958, -0.009681702, 0.012504578, 0.021835327, 0.0514831...
http://arxiv.org/abs/2305.04241v2
Vcc: Scaling Transformers to 128K Tokens or More by Prioritizing Important Tokens
Zhanpeng Zeng, Cole Hawkins, Mingyi Hong, Aston Zhang, Nikolaos Pappas, Vikas Singh, Shuai Zheng
cs.CL, cs.LG
Transformers are central in modern natural language processing and computer vision applications. Despite recent works devoted to reducing the quadratic cost of such models (as a function of the sequence length), dealing with ultra long sequences (e.g., with more than 16K tokens) remains challenging. Applications such a...
http://arxiv.org/pdf/2305.04241v2
2023-05-27T04:17:13
2023-05-07T10:32:18
{ "title": [ 0.011108398, -0.0095825195, -0.054779053, -0.06567383, -0.004295349, -0.0049438477, -0.01474762, 0.05041504, -0.031829834, 0.040100098, -0.06451416, 0.00015473366, 0.008674622, -0.017608643, -0.024032593, -0.056762695, 0.017105103, 0...
http://arxiv.org/abs/2305.06155v1
Leveraging Synthetic Targets for Machine Translation
Sarthak Mittal, Oleksii Hrinchuk, Oleksii Kuchaiev
cs.CL, cs.AI, cs.LG
In this work, we provide a recipe for training machine translation models in a limited resource setting by leveraging synthetic target data generated using a large pre-trained model. We show that consistently across different benchmarks in bilingual, multilingual, and speech translation setups, training models on synth...
http://arxiv.org/pdf/2305.06155v1
2023-05-07T07:42:22
2023-05-07T07:42:22
{ "title": [ -0.044403076, -0.016204834, -0.00983429, -0.048736572, -0.020401001, 0.024597168, -0.04675293, 0.0023555756, -0.020843506, 0.046722412, -0.007965088, 0.017318726, 0.00004130602, -0.008842468, -0.040008545, 0.01651001, -0.008178711, 0...
http://arxiv.org/abs/2305.06162v3
Interpretable multimodal sentiment analysis based on textual modality descriptions by using large-scale language models
Sixia Li, Shogo Okada
cs.CL, cs.LG, cs.MM
Multimodal sentiment analysis is an important area for understanding the user's internal states. Deep learning methods were effective, but the problem of poor interpretability has gradually gained attention. Previous works have attempted to use attention weights or vector distributions to provide interpretability. Howe...
http://arxiv.org/pdf/2305.06162v3
2023-05-12T00:39:13
2023-05-07T06:48:06
{ "title": [ -0.024246216, 0.0050315857, -0.040161133, -0.05227661, 0.003932953, 0.00025177002, -0.059020996, -0.015640259, 0.024734497, 0.04473877, -0.016479492, -0.011123657, 0.0211792, -0.005203247, -0.070739746, 0.0023021698, -0.015045166, 0....
http://arxiv.org/abs/2305.04195v3
Cross-Modal Retrieval for Motion and Text via DopTriple Loss
Sheng Yan, Yang Liu, Haoqiang Wang, Xin Du, Mengyuan Liu, Hong Liu
cs.CV, cs.CL
Cross-modal retrieval of image-text and video-text is a prominent research area in computer vision and natural language processing. However, there has been insufficient attention given to cross-modal retrieval between human motion and text, despite its wide-ranging applicability. To address this gap, we utilize a conci...
http://arxiv.org/pdf/2305.04195v3
2023-10-03T04:42:09
2023-05-07T05:40:48
{ "title": [ -0.0124053955, 0.025299072, -0.029373169, -0.010978699, 0.016601562, -0.02218628, 0.0126953125, -0.056640625, 0.015731812, 0.06542969, -0.028137207, 0.022216797, 0.033599854, -0.005596161, -0.04586792, 0.010795593, 0.0010519028, 0.00...
http://arxiv.org/abs/2305.04183v1
OpenViVQA: Task, Dataset, and Multimodal Fusion Models for Visual Question Answering in Vietnamese
Nghia Hieu Nguyen, Duong T. D. Vo, Kiet Van Nguyen, Ngan Luu-Thuy Nguyen
cs.CL
In recent years, visual question answering (VQA) has attracted attention from the research community because of its highly potential applications (such as virtual assistance on intelligent cars, assistant devices for blind people, or information retrieval from document images using natural language as queries) and chal...
http://arxiv.org/pdf/2305.04183v1
2023-05-07T03:59:31
2023-05-07T03:59:31
{ "title": [ -0.019851685, -0.0033493042, -0.022613525, -0.030761719, -0.01675415, 0.0061302185, -0.02949524, -0.013221741, 0.022888184, 0.028060913, -0.03543091, -0.01637268, -0.0027484894, -0.017227173, -0.06970215, -0.02658081, 0.0043754578, 0...
http://arxiv.org/abs/2305.04181v1
Shall We Trust All Relational Tuples by Open Information Extraction? A Study on Speculation Detection
Kuicai Dong, Aixin Sun, Jung-Jae Kim, Xiaoli Li
cs.CL, cs.AI
Open Information Extraction (OIE) aims to extract factual relational tuples from open-domain sentences. Downstream tasks use the extracted OIE tuples as facts, without examining the certainty of these facts. However, uncertainty/speculation is a common linguistic phenomenon. Existing studies on speculation detection ar...
http://arxiv.org/pdf/2305.04181v1
2023-05-07T03:47:05
2023-05-07T03:47:05
{ "title": [ -0.011634827, 0.012245178, -0.03881836, -0.05441284, 0.006465912, -0.0104522705, -0.04220581, -0.023803711, -0.02619934, 0.014152527, -0.031677246, 0.0013780594, -0.0035324097, -0.017471313, -0.016784668, -0.039093018, -0.02494812, 0...
http://arxiv.org/abs/2305.04177v1
MIReAD: Simple Method for Learning High-quality Representations from Scientific Documents
Anastasia Razdaibiedina, Alexander Brechalov
cs.CL, cs.AI
Learning semantically meaningful representations from scientific documents can facilitate academic literature search and improve performance of recommendation systems. Pre-trained language models have been shown to learn rich textual representations, yet they cannot provide powerful document-level representations for s...
http://arxiv.org/pdf/2305.04177v1
2023-05-07T03:29:55
2023-05-07T03:29:55
{ "title": [ -0.017852783, -0.0031967163, -0.037994385, -0.056518555, 0.019165039, -0.0023880005, -0.03111267, -0.01953125, 0.018157959, 0.02305603, -0.0027046204, -0.0033226013, -0.0076026917, -0.026779175, -0.050354004, 0.012641907, -0.0021858215, ...
http://arxiv.org/abs/2305.04166v2
UIT-OpenViIC: A Novel Benchmark for Evaluating Image Captioning in Vietnamese
Doanh C. Bui, Nghia Hieu Nguyen, Khang Nguyen
cs.CV, cs.CL
Image Captioning is one of the vision-language tasks that still interest the research community worldwide in the 2020s. MS-COCO Caption benchmark is commonly used to evaluate the performance of advanced captioning models, although it was published in 2015. Recent captioning models trained on the MS-COCO Caption dataset...
http://arxiv.org/pdf/2305.04166v2
2023-05-09T12:46:06
2023-05-07T02:48:47
{ "title": [ 0.0038871765, -0.0072631836, 0.014678955, -0.0076522827, -0.026489258, -0.006313324, -0.012336731, -0.044830322, -0.008522034, 0.05706787, -0.03805542, 0.017822266, -0.0034999847, -0.0042648315, -0.05834961, -0.024032593, -0.016921997, ...
http://arxiv.org/abs/2305.04160v3
X-LLM: Bootstrapping Advanced Large Language Models by Treating Multi-Modalities as Foreign Languages
Feilong Chen, Minglun Han, Haozhi Zhao, Qingyang Zhang, Jing Shi, Shuang Xu, Bo Xu
cs.CL, cs.AI, cs.CV, eess.AS
Large language models (LLMs) have demonstrated remarkable language abilities. GPT-4, based on advanced LLMs, exhibits extraordinary multimodal capabilities beyond previous visual language models. We attribute this to the use of more advanced LLMs compared with previous multimodal models. Unfortunately, the model archit...
http://arxiv.org/pdf/2305.04160v3
2023-05-22T02:37:02
2023-05-07T02:25:42
{ "title": [ -0.05758667, -0.0055351257, -0.06958008, -0.064208984, 0.0042648315, 0.0009756088, -0.092285156, -0.0011005402, -0.025909424, 0.020629883, -0.00048732758, 0.014587402, 0.016281128, -0.04168701, 0.0049476624, 0.016616821, -0.025787354, ...
http://arxiv.org/abs/2305.10408v1
Extracting Blockchain Concepts from Text
Rodrigo Veiga, Markus Endler, Valeria de Paiva
cs.IR, cs.CL, cs.CR
Blockchains provide a mechanism through which mutually distrustful remote parties can reach consensus on the state of a ledger of information. With the great acceleration with which this space is developed, the demand for those seeking to learn about blockchain also grows. Being a technical subject, it can be quite int...
http://arxiv.org/pdf/2305.10408v1
2023-05-07T00:16:30
2023-05-07T00:16:30
{ "title": [ -0.03250122, 0.0034751892, -0.046966553, -0.047729492, 0.005508423, -0.00080156326, -0.022521973, 0.002155304, 0.0030326843, 0.041015625, -0.039642334, 0.0074539185, -0.007007599, 0.0058403015, -0.06756592, -0.015655518, -0.016784668, ...
http://arxiv.org/abs/2305.04147v1
Controllable Mixed-Initiative Dialogue Generation through Prompting
Maximillian Chen, Xiao Yu, Weiyan Shi, Urvi Awasthi, Zhou Yu
cs.CL, cs.AI, cs.HC
Mixed-initiative dialogue tasks involve repeated exchanges of information and conversational control. Conversational agents gain control by generating responses that follow particular dialogue intents or strategies, prescribed by a policy planner. The standard approach has been fine-tuning pre-trained language models t...
http://arxiv.org/pdf/2305.04147v1
2023-05-06T23:11:25
2023-05-06T23:11:25
{ "title": [ -0.017028809, -0.008071899, -0.045318604, -0.05908203, 0.006931305, -0.001701355, -0.030670166, 0.011924744, -0.024642944, 0.016937256, -0.008460999, 0.02267456, -0.034423828, -0.022064209, -0.029586792, -0.0131073, -0.016296387, 0.0...
http://arxiv.org/abs/2305.04118v2
Exploring Human-Like Translation Strategy with Large Language Models
Zhiwei He, Tian Liang, Wenxiang Jiao, Zhuosheng Zhang, Yujiu Yang, Rui Wang, Zhaopeng Tu, Shuming Shi, Xing Wang
cs.CL
Large language models (LLMs) have demonstrated impressive capabilities in general scenarios, exhibiting a level of aptitude that approaches, in some aspects even surpasses, human-level intelligence. Among their numerous skills, the translation abilities of LLMs have received considerable attention. In contrast to tradi...
http://arxiv.org/pdf/2305.04118v2
2023-06-22T15:05:16
2023-05-06T19:03:12
{ "title": [ -0.05050659, -0.0039596558, -0.026641846, -0.026779175, -0.014823914, -0.0018358231, -0.035491943, -0.0033226013, -0.024749756, 0.055786133, 0.0014476776, 0.020690918, 0.024414062, -0.008171082, -0.044158936, 0.030776978, -0.0052948, ...
http://arxiv.org/abs/2305.04105v2
"When Words Fail, Emojis Prevail": Generating Sarcastic Utterances with Emoji Using Valence Reversal and Semantic Incongruity
Faria Binte Kader, Nafisa Hossain Nujat, Tasmia Binte Sogir, Mohsinul Kabir, Hasan Mahmud, Kamrul Hasan
cs.CL
Sarcasm is a form of figurative language that serves as a humorous tool for mockery and ridicule. We present a novel architecture for sarcasm generation with emoji from a non-sarcastic input sentence in English. We divide the generation task into two sub tasks: one for generating textual sarcasm and another for collect...
http://arxiv.org/pdf/2305.04105v2
2023-06-16T15:11:03
2023-05-06T17:49:41
{ "title": [ 0.0069351196, -0.010475159, -0.008148193, -0.02470398, -0.029525757, -0.0060577393, -0.021240234, -0.014465332, 0.015388489, 0.051086426, -0.02305603, -0.012321472, 0.0552063, -0.01626587, -0.046936035, -0.0037231445, -0.033477783, 0...
http://arxiv.org/abs/2305.04100v1
Rhetorical Role Labeling of Legal Documents using Transformers and Graph Neural Networks
Anshika Gupta, Shaz Furniturewala, Vijay Kumari, Yashvardhan Sharma
cs.CL
A legal document is usually long and dense requiring human effort to parse it. It also contains significant amounts of jargon which make deriving insights from it using existing models a poor approach. This paper presents the approaches undertaken to perform the task of rhetorical role labelling on Indian Court Judgeme...
http://arxiv.org/pdf/2305.04100v1
2023-05-06T17:04:51
2023-05-06T17:04:51
{ "title": [ -0.022140503, -0.010131836, -0.026824951, -0.04168701, -0.032928467, 0.008354187, -0.06726074, -0.002462387, -0.022399902, 0.023956299, -0.021438599, 0.016494751, -0.003921509, -0.012481689, -0.024398804, -0.014663696, -0.029586792, ...