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/2306.11932v1
Opportunities and Risks of LLMs for Scalable Deliberation with Polis
Christopher T. Small, Ivan Vendrov, Esin Durmus, Hadjar Homaei, Elizabeth Barry, Julien Cornebise, Ted Suzman, Deep Ganguli, Colin Megill
cs.SI, cs.CL, cs.CY, cs.HC
Polis is a platform that leverages machine intelligence to scale up deliberative processes. In this paper, we explore the opportunities and risks associated with applying Large Language Models (LLMs) towards challenges with facilitating, moderating and summarizing the results of Polis engagements. In particular, we dem...
http://arxiv.org/pdf/2306.11932v1
2023-06-20T22:52:51
2023-06-20T22:52:51
{ "title": [ 0.01537323, -0.0049209595, -0.031341553, -0.03375244, -0.032562256, -0.014877319, -0.048828125, 0.000053465366, -0.032592773, 0.009178162, -0.072265625, 0.030075073, -0.010482788, -0.023971558, 0.03277588, -0.04949951, 0.00002092123, ...
http://arxiv.org/abs/2306.11900v1
Evaluation of Chinese-English Machine Translation of Emotion-Loaded Microblog Texts: A Human Annotated Dataset for the Quality Assessment of Emotion Translation
Shenbin Qian, Constantin Orasan, Felix do Carmo, Qiuliang Li, Diptesh Kanojia
cs.CL
In this paper, we focus on how current Machine Translation (MT) tools perform on the translation of emotion-loaded texts by evaluating outputs from Google Translate according to a framework proposed in this paper. We propose this evaluation framework based on the Multidimensional Quality Metrics (MQM) and perform a det...
http://arxiv.org/pdf/2306.11900v1
2023-06-20T21:22:45
2023-06-20T21:22:45
{ "title": [ -0.03451538, 0.0041007996, 0.010063171, 0.00069618225, -0.016937256, 0.006954193, -0.015899658, -0.019866943, 0.014328003, 0.05496216, -0.012237549, 0.013298035, 0.02999878, -0.024398804, -0.05947876, 0.021652222, -0.0054969788, 0.01...
http://arxiv.org/abs/2306.11892v1
Exploring New Frontiers in Agricultural NLP: Investigating the Potential of Large Language Models for Food Applications
Saed Rezayi, Zhengliang Liu, Zihao Wu, Chandra Dhakal, Bao Ge, Haixing Dai, Gengchen Mai, Ninghao Liu, Chen Zhen, Tianming Liu, Sheng Li
cs.CL
This paper explores new frontiers in agricultural natural language processing by investigating the effectiveness of using food-related text corpora for pretraining transformer-based language models. In particular, we focus on the task of semantic matching, which involves establishing mappings between food descriptions ...
http://arxiv.org/pdf/2306.11892v1
2023-06-20T21:12:16
2023-06-20T21:12:16
{ "title": [ -0.02935791, -0.011367798, -0.027130127, -0.030441284, -0.0049819946, -0.0037765503, -0.03878784, 0.002696991, -0.0011816025, 0.028137207, -0.019851685, -0.0113220215, 0.026657104, -0.02281189, -0.03277588, -0.007904053, -0.031311035, ...
http://arxiv.org/abs/2306.11879v1
Open-Domain Text Evaluation via Meta Distribution Modeling
Sidi Lu, Asli Celikyilmaz, Tianlu Wang, Nanyun Peng
cs.CL
Recent advances in open-domain text generation models powered by large pre-trained language models (LLMs) have achieved remarkable performance. However, evaluating and controlling these models for desired attributes remains a challenge, as traditional reference-based metrics such as BLEU, ROUGE, and METEOR are insuffic...
http://arxiv.org/pdf/2306.11879v1
2023-06-20T20:37:54
2023-06-20T20:37:54
{ "title": [ -0.030548096, 0.022735596, -0.025161743, -0.033447266, 0.0033988953, 0.009033203, -0.039123535, -0.051940918, 0.0036716461, 0.046173096, 0.0013103485, 0.0012426376, -0.01348114, -0.013893127, -0.06567383, -0.0006222725, -0.035583496, ...
http://arxiv.org/abs/2306.11843v1
Retrieval-Based Transformer for Table Augmentation
Michael Glass, Xueqing Wu, Ankita Rajaram Naik, Gaetano Rossiello, Alfio Gliozzo
cs.CL, cs.AI, cs.DB, cs.IR
Data preparation, also called data wrangling, is considered one of the most expensive and time-consuming steps when performing analytics or building machine learning models. Preparing data typically involves collecting and merging data from complex heterogeneous, and often large-scale data sources, such as data lakes. ...
http://arxiv.org/pdf/2306.11843v1
2023-06-20T18:51:21
2023-06-20T18:51:21
{ "title": [ 0.028274536, 0.030044556, -0.0541687, -0.057403564, 0.021453857, -0.015449524, -0.032958984, -0.016784668, 0.008743286, 0.07098389, -0.032714844, -0.0046081543, 0.045684814, -0.006095886, -0.046539307, 0.012702942, 0.01776123, 0.0182...
http://arxiv.org/abs/2306.11838v1
Efficient Machine Translation Corpus Generation
Kamer Ali Yuksel, Ahmet Gunduz, Shreyas Sharma, Hassan Sawaf
cs.CL
This paper proposes an efficient and semi-automated method for human-in-the-loop post-editing for machine translation (MT) corpus generation. The method is based on online training of a custom MT quality estimation metric on-the-fly as linguists perform post-edits. The online estimator is used to prioritize worse hypot...
http://arxiv.org/pdf/2306.11838v1
2023-06-20T18:46:47
2023-06-20T18:46:47
{ "title": [ -0.03326416, -0.015914917, -0.013435364, -0.06335449, -0.0018959045, 0.021453857, -0.04925537, 0.0050735474, 0.011947632, 0.043395996, -0.017166138, 0.0015077591, 0.023483276, -0.011749268, -0.04525757, 0.01777649, -0.010002136, 0.04...
http://arxiv.org/abs/2306.11832v1
QuOTeS: Query-Oriented Technical Summarization
Juan Ramirez-Orta, Eduardo Xamena, Ana Maguitman, Axel J. Soto, Flavia P. Zanoto, Evangelos Milios
cs.IR, cs.CL
Abstract. When writing an academic paper, researchers often spend considerable time reviewing and summarizing papers to extract relevant citations and data to compose the Introduction and Related Work sections. To address this problem, we propose QuOTeS, an interactive system designed to retrieve sentences related to a...
http://arxiv.org/pdf/2306.11832v1
2023-06-20T18:43:24
2023-06-20T18:43:24
{ "title": [ 0.033569336, 0.0053520203, -0.031341553, -0.0597229, -0.019699097, -0.013458252, -0.044189453, 0.0050468445, -0.050323486, 0.07635498, -0.015472412, 0.006084442, 0.03591919, -0.02986145, -0.010467529, -0.03753662, 0.011154175, 0.0107...
http://arxiv.org/abs/2306.11825v1
On Compositionality and Improved Training of NADO
Sidi Lu, Wenbo Zhao, Chenyang Tao, Arpit Gupta, Shanchan Wu, Tagyoung Chung, Nanyun Peng
cs.CL
NeurAlly-Decomposed Oracle (NADO) is a powerful approach for controllable generation with large language models. Differentiating from finetuning/prompt tuning, it has the potential to avoid catastrophic forgetting of the large base model and achieve guaranteed convergence to an entropy-maximized closed-form solution wi...
http://arxiv.org/pdf/2306.11825v1
2023-06-20T18:36:52
2023-06-20T18:36:52
{ "title": [ 0.015449524, 0.0127334595, -0.03387451, -0.10534668, -0.011962891, 0.014457703, -0.10235596, -0.008018494, -0.022735596, 0.053131104, 0.020370483, 0.001326561, -0.0034694672, -0.037841797, -0.022109985, 0.004497528, -0.017501831, 0.0...
http://arxiv.org/abs/2306.11823v1
EvolveMT: an Ensemble MT Engine Improving Itself with Usage Only
Kamer Ali Yuksel, Ahmet Gunduz, Mohamed Al-Badrashiny, Shreyas Sharma, Hassan Sawaf
cs.CL
This paper presents EvolveMT for efficiently combining multiple machine translation (MT) engines. The proposed system selects the output from a single engine for each segment by utilizing online learning techniques to predict the most suitable system for every translation request. A neural quality estimation metric sup...
http://arxiv.org/pdf/2306.11823v1
2023-06-20T18:32:30
2023-06-20T18:32:30
{ "title": [ 0.00030851364, 0.0015707016, -0.047851562, -0.058746338, 0.0049705505, -0.015823364, -0.026779175, 0.0048065186, -0.005016327, 0.049743652, -0.02557373, 0.008911133, 0.008880615, -0.020202637, -0.045074463, -0.0009493828, -0.0031795502, ...
http://arxiv.org/abs/2306.11816v2
Learning to Generate Better Than Your LLM
Jonathan D. Chang, Kiante Brantley, Rajkumar Ramamurthy, Dipendra Misra, Wen Sun
cs.LG, cs.AI, cs.CL
Reinforcement learning (RL) has emerged as a powerful paradigm for fine-tuning Large Language Models (LLMs) for text generation. In particular, recent LLMs such as ChatGPT and GPT-4 can engage in fluent conversations with users after finetuning with RL. Capitalizing on key properties of text generation, we seek to inve...
http://arxiv.org/pdf/2306.11816v2
2023-11-13T18:51:42
2023-06-20T18:19:17
{ "title": [ -0.015670776, -0.0289917, -0.030319214, -0.025390625, -0.074645996, -0.0023097992, -0.036315918, 0.013000488, -0.035736084, 0.0049934387, -0.031829834, 0.016067505, -0.035095215, -0.03543091, 0.046966553, -0.03024292, 0.00248909, 0.0...
http://arxiv.org/abs/2306.11702v2
Lingua Manga: A Generic Large Language Model Centric System for Data Curation
Zui Chen, Lei Cao, Sam Madden
cs.DB, cs.CL
Data curation is a wide-ranging area which contains many critical but time-consuming data processing tasks. However, the diversity of such tasks makes it challenging to develop a general-purpose data curation system. To address this issue, we present Lingua Manga, a user-friendly and versatile system that utilizes pre-...
http://arxiv.org/pdf/2306.11702v2
2023-09-01T15:40:40
2023-06-20T17:30:02
{ "title": [ -0.044677734, -0.028289795, -0.0692749, -0.07067871, -0.0036888123, 0.006603241, -0.05407715, 0.029708862, 0.03274536, 0.05947876, -0.01777649, -0.030136108, 0.029006958, -0.006717682, -0.08099365, 0.0021152496, -0.007411957, 0.04101...
http://arxiv.org/abs/2306.11698v1
DecodingTrust: A Comprehensive Assessment of Trustworthiness in GPT Models
Boxin Wang, Weixin Chen, Hengzhi Pei, Chulin Xie, Mintong Kang, Chenhui Zhang, Chejian Xu, Zidi Xiong, Ritik Dutta, Rylan Schaeffer, Sang T. Truong, Simran Arora, Mantas Mazeika, Dan Hendrycks, Zinan Lin, Yu Cheng, Sanmi Koyejo, Dawn Song, Bo Li
cs.CL, cs.AI, cs.CR
Generative Pre-trained Transformer (GPT) models have exhibited exciting progress in capabilities, capturing the interest of practitioners and the public alike. Yet, while the literature on the trustworthiness of GPT models remains limited, practitioners have proposed employing capable GPT models for sensitive applicati...
http://arxiv.org/pdf/2306.11698v1
2023-06-20T17:24:23
2023-06-20T17:24:23
{ "title": [ -0.020385742, 0.006690979, -0.042877197, -0.038360596, 0.016998291, -0.00919342, -0.021957397, -0.030151367, -0.016433716, 0.015914917, -0.010276794, 0.0129470825, -0.019241333, -0.026000977, -0.03137207, 0.0028247833, -0.027130127, ...
http://arxiv.org/abs/2306.11695v2
A Simple and Effective Pruning Approach for Large Language Models
Mingjie Sun, Zhuang Liu, Anna Bair, J. Zico Kolter
cs.CL, cs.AI, cs.LG
As their size increases, Large Languages Models (LLMs) are natural candidates for network pruning methods: approaches that drop a subset of network weights while striving to preserve performance. Existing methods, however, require either retraining, which is rarely affordable for billion-scale LLMs, or solving a weight...
http://arxiv.org/pdf/2306.11695v2
2023-10-06T17:11:09
2023-06-20T17:18:20
{ "title": [ -0.028259277, -0.021469116, -0.031341553, -0.08654785, 0.010368347, 0.024978638, -0.072509766, 0.030029297, 0.024978638, 0.020965576, -0.01940918, -0.005050659, 0.011741638, -0.016555786, -0.049591064, 0.006980896, -0.014793396, 0.04...
http://arxiv.org/abs/2306.11648v1
Harnessing the Power of Adversarial Prompting and Large Language Models for Robust Hypothesis Generation in Astronomy
Ioana Ciucă, Yuan-Sen Ting, Sandor Kruk, Kartheik Iyer
astro-ph.IM, astro-ph.GA, cs.AI, cs.CL
This study investigates the application of Large Language Models (LLMs), specifically GPT-4, within Astronomy. We employ in-context prompting, supplying the model with up to 1000 papers from the NASA Astrophysics Data System, to explore the extent to which performance can be improved by immersing the model in domain-sp...
http://arxiv.org/pdf/2306.11648v1
2023-06-20T16:16:56
2023-06-20T16:16:56
{ "title": [ -0.019241333, -0.012451172, -0.019638062, -0.025375366, -0.013626099, 0.004917145, -0.025146484, -0.02494812, -0.007843018, 0.02558899, 0.000006198883, 0.01272583, -0.013313293, -0.023803711, -0.032806396, 0.00028157234, 0.018859863, ...
http://arxiv.org/abs/2306.11646v1
Recent Advances in Direct Speech-to-text Translation
Chen Xu, Rong Ye, Qianqian Dong, Chengqi Zhao, Tom Ko, Mingxuan Wang, Tong Xiao, Jingbo Zhu
cs.CL, eess.AS
Recently, speech-to-text translation has attracted more and more attention and many studies have emerged rapidly. In this paper, we present a comprehensive survey on direct speech translation aiming to summarize the current state-of-the-art techniques. First, we categorize the existing research work into three directio...
http://arxiv.org/pdf/2306.11646v1
2023-06-20T16:14:27
2023-06-20T16:14:27
{ "title": [ -0.01689148, -0.0062408447, -0.00969696, -0.083740234, 0.00061130524, 0.017807007, -0.05493164, 0.012168884, 0.010559082, 0.059692383, -0.010307312, 0.013626099, 0.025726318, -0.019470215, -0.059387207, 0.028533936, 0.00055503845, 0....
http://arxiv.org/abs/2306.11644v2
Textbooks Are All You Need
Suriya Gunasekar, Yi Zhang, Jyoti Aneja, Caio César Teodoro Mendes, Allie Del Giorno, Sivakanth Gopi, Mojan Javaheripi, Piero Kauffmann, Gustavo de Rosa, Olli Saarikivi, Adil Salim, Shital Shah, Harkirat Singh Behl, Xin Wang, Sébastien Bubeck, Ronen Eldan, Adam Tauman Kalai, Yin Tat Lee, Yuanzhi Li
cs.CL, cs.AI, cs.LG
We introduce phi-1, a new large language model for code, with significantly smaller size than competing models: phi-1 is a Transformer-based model with 1.3B parameters, trained for 4 days on 8 A100s, using a selection of ``textbook quality" data from the web (6B tokens) and synthetically generated textbooks and exercis...
http://arxiv.org/pdf/2306.11644v2
2023-10-02T06:12:30
2023-06-20T16:14:25
{ "title": [ 0.026184082, -0.00036621094, 0.0035324097, -0.07147217, -0.037353516, 0.010673523, -0.075927734, 0.0043411255, -0.0006017685, 0.047302246, -0.023803711, 0.0043525696, -0.018966675, -0.004261017, -0.0036792755, -0.011123657, 0.041107178, ...
http://arxiv.org/abs/2307.07513v1
An empirical study of using radiology reports and images to improve ICU mortality prediction
Mingquan Lin, Song Wang, Ying Ding, Lihui Zhao, Fei Wang, Yifan Peng
cs.AI, cs.CL, cs.CV, cs.LG, eess.IV
Background: The predictive Intensive Care Unit (ICU) scoring system plays an important role in ICU management because it predicts important outcomes, especially mortality. Many scoring systems have been developed and used in the ICU. These scoring systems are primarily based on the structured clinical data in the elect...
http://arxiv.org/pdf/2307.07513v1
2023-06-20T15:43:28
2023-06-20T15:43:28
{ "title": [ 0.0121536255, 0.015808105, -0.025253296, -0.034118652, 0.016616821, -0.023880005, 0.026473999, -0.027008057, -0.039611816, 0.031982422, -0.024856567, 0.034973145, -0.027862549, -0.016494751, -0.03652954, -0.020339966, 0.0054626465, 0...
http://arxiv.org/abs/2306.11593v1
Improving Image Captioning Descriptiveness by Ranking and LLM-based Fusion
Simone Bianco, Luigi Celona, Marco Donzella, Paolo Napoletano
cs.CV, cs.AI, cs.CL, cs.DB, cs.LG
State-of-The-Art (SoTA) image captioning models often rely on the Microsoft COCO (MS-COCO) dataset for training. This dataset contains annotations provided by human annotators, who typically produce captions averaging around ten tokens. However, this constraint presents a challenge in effectively capturing complex scen...
http://arxiv.org/pdf/2306.11593v1
2023-06-20T15:13:02
2023-06-20T15:13:02
{ "title": [ 0.008338928, 0.020965576, -0.007522583, -0.019744873, 0.014945984, -0.014030457, -0.0012722015, -0.04397583, -0.006401062, 0.048461914, -0.033935547, 0.048614502, -0.006576538, 0.012420654, -0.03781128, -0.005290985, -0.020996094, 0....
http://arxiv.org/abs/2306.11585v2
FAIR: A Causal Framework for Accurately Inferring Judgments Reversals
Minghua He, Nanfei Gu, Yuntao Shi, Qionghui Zhang, Yaying Chen
cs.CL
Artificial intelligence researchers have made significant advances in legal intelligence in recent years. However, the existing studies have not focused on the important value embedded in judgments reversals, which limits the improvement of the efficiency of legal intelligence. In this paper, we propose a causal Framew...
http://arxiv.org/pdf/2306.11585v2
2023-07-20T14:31:10
2023-06-20T15:02:25
{ "title": [ -0.011886597, -0.018753052, -0.033172607, -0.043395996, -0.009422302, -0.0112838745, -0.023483276, -0.028503418, 0.027511597, 0.013130188, -0.022399902, -0.020904541, 0.005039215, -0.030563354, -0.038970947, -0.00017940998, -0.008026123,...
http://arxiv.org/abs/2306.11559v1
The Ecological Fallacy in Annotation: Modelling Human Label Variation goes beyond Sociodemographics
Matthias Orlikowski, Paul Röttger, Philipp Cimiano, Dirk Hovy
cs.CL
Many NLP tasks exhibit human label variation, where different annotators give different labels to the same texts. This variation is known to depend, at least in part, on the sociodemographics of annotators. Recent research aims to model individual annotator behaviour rather than predicting aggregated labels, and we wou...
http://arxiv.org/pdf/2306.11559v1
2023-06-20T14:23:32
2023-06-20T14:23:32
{ "title": [ -0.006839752, 0.0049438477, -0.008705139, -0.024032593, -0.016799927, -0.023376465, 0.000729084, -0.027404785, -0.015113831, 0.01473999, -0.0068359375, 0.0061798096, -0.042266846, -0.028182983, -0.044830322, -0.024993896, -0.0154953, ...
http://arxiv.org/abs/2306.11520v1
Hallucination is the last thing you need
Shawn Curran, Sam Lansley, Oliver Bethell
cs.CL
The legal profession necessitates a multidimensional approach that involves synthesizing an in-depth comprehension of a legal issue with insightful commentary based on personal experience, combined with a comprehensive understanding of pertinent legislation, regulation, and case law, in order to deliver an informed leg...
http://arxiv.org/pdf/2306.11520v1
2023-06-20T13:14:15
2023-06-20T13:14:15
{ "title": [ 0.06274414, -0.019073486, 0.017044067, -0.03503418, -0.03933716, -0.0022735596, -0.0012903214, -0.024536133, -0.015090942, 0.057525635, 0.02168274, -0.02319336, -0.026229858, -0.010009766, -0.018112183, 0.014732361, 0.017684937, 0.04...
http://arxiv.org/abs/2306.11518v2
One model to rule them all: ranking Slovene summarizers
Aleš Žagar, Marko Robnik-Šikonja
cs.CL
Text summarization is an essential task in natural language processing, and researchers have developed various approaches over the years, ranging from rule-based systems to neural networks. However, there is no single model or approach that performs well on every type of text. We propose a system that recommends the mo...
http://arxiv.org/pdf/2306.11518v2
2023-08-07T09:17:43
2023-06-20T13:12:58
{ "title": [ 0.01927185, -0.0011224747, -0.013130188, -0.06225586, -0.014556885, 0.004295349, -0.057922363, -0.025741577, -0.026550293, 0.06890869, -0.010063171, 0.01171875, 0.025543213, -0.0107803345, -0.048797607, 0.0058021545, -0.009735107, 0....
http://arxiv.org/abs/2306.11507v1
TrustGPT: A Benchmark for Trustworthy and Responsible Large Language Models
Yue Huang, Qihui Zhang, Philip S. Y, Lichao Sun
cs.CL, cs.AI
Large Language Models (LLMs) such as ChatGPT, have gained significant attention due to their impressive natural language processing capabilities. It is crucial to prioritize human-centered principles when utilizing these models. Safeguarding the ethical and moral compliance of LLMs is of utmost importance. However, ind...
http://arxiv.org/pdf/2306.11507v1
2023-06-20T12:53:39
2023-06-20T12:53:39
{ "title": [ -0.021911621, -0.017059326, -0.05581665, -0.08502197, 0.012992859, 0.0067977905, -0.062805176, 0.013267517, -0.004989624, 0.028244019, 0.0021152496, 0.008422852, 0.008766174, -0.024261475, -0.03503418, 0.018447876, -0.019821167, 0.03...
http://arxiv.org/abs/2306.11489v1
ChatGPT is not Enough: Enhancing Large Language Models with Knowledge Graphs for Fact-aware Language Modeling
Linyao Yang, Hongyang Chen, Zhao Li, Xiao Ding, Xindong Wu
cs.CL, cs.AI
Recently, ChatGPT, a representative large language model (LLM), has gained considerable attention due to its powerful emergent abilities. Some researchers suggest that LLMs could potentially replace structured knowledge bases like knowledge graphs (KGs) and function as parameterized knowledge bases. However, while LLMs...
http://arxiv.org/pdf/2306.11489v1
2023-06-20T12:21:06
2023-06-20T12:21:06
{ "title": [ -0.020843506, -0.021896362, -0.04019165, -0.06744385, -0.018096924, 0.009155273, -0.06677246, 0.018157959, -0.007282257, 0.023986816, -0.0072288513, 0.011245728, 0.019958496, -0.019714355, -0.021408081, 0.009292603, -0.028015137, 0.0...
http://arxiv.org/abs/2306.12443v1
DEPAC: a Corpus for Depression and Anxiety Detection from Speech
Mashrura Tasnim, Malikeh Ehghaghi, Brian Diep, Jekaterina Novikova
eess.AS, cs.CL, cs.LG, cs.SD
Mental distress like depression and anxiety contribute to the largest proportion of the global burden of diseases. Automated diagnosis systems of such disorders, empowered by recent innovations in Artificial Intelligence, can pave the way to reduce the sufferings of the affected individuals. Development of such systems...
http://arxiv.org/pdf/2306.12443v1
2023-06-20T12:21:06
2023-06-20T12:21:06
{ "title": [ 0.010932922, 0.000044465065, -0.019470215, -0.07489014, 0.016357422, 0.0012521744, -0.04675293, -0.0062294006, 0.023391724, 0.03475952, -0.020629883, -0.018707275, 0.03869629, -0.026138306, -0.05987549, -0.0023269653, -0.01727295, 0....
http://arxiv.org/abs/2306.11485v2
Explicit Syntactic Guidance for Neural Text Generation
Yafu Li, Leyang Cui, Jianhao Yan, Yongjing Yin, Wei Bi, Shuming Shi, Yue Zhang
cs.CL
Most existing text generation models follow the sequence-to-sequence paradigm. Generative Grammar suggests that humans generate natural language texts by learning language grammar. We propose a syntax-guided generation schema, which generates the sequence guided by a constituency parse tree in a top-down direction. The...
http://arxiv.org/pdf/2306.11485v2
2023-06-25T13:32:14
2023-06-20T12:16:31
{ "title": [ -0.03353882, -0.015083313, -0.025283813, -0.052520752, -0.020690918, 0.018981934, -0.05810547, 0.0040016174, -0.0011081696, 0.0463562, -0.010192871, 0.017242432, 0.0026626587, -0.008544922, -0.046020508, 0.016113281, -0.01689148, 0.0...
http://arxiv.org/abs/2306.11477v1
CATS: A Pragmatic Chinese Answer-to-Sequence Dataset with Large Scale and High Quality
Liang Li, Ruiying Geng, Chengyang Fang, Bing Li, Can Ma, Rongyu Cao, Binhua Li, Fei Huang, Yongbin Li
cs.CL
There are three problems existing in the popular data-to-text datasets. First, the large-scale datasets either contain noise or lack real application scenarios. Second, the datasets close to real applications are relatively small in size. Last, current datasets bias in the English language while leaving other languages...
http://arxiv.org/pdf/2306.11477v1
2023-06-20T12:02:26
2023-06-20T12:02:26
{ "title": [ -0.016647339, -0.005344391, -0.022460938, -0.075683594, 0.009681702, -0.0054969788, -0.048797607, 0.015945435, 0.045532227, 0.011489868, -0.018676758, -0.005344391, 0.0078048706, -0.02720642, -0.049926758, 0.020965576, -0.0005469322, ...
http://arxiv.org/abs/2306.11473v1
Timestamped Embedding-Matching Acoustic-to-Word CTC ASR
Woojay Jeon
cs.CL, eess.AS
In this work, we describe a novel method of training an embedding-matching word-level connectionist temporal classification (CTC) automatic speech recognizer (ASR) such that it directly produces word start times and durations, required by many real-world applications, in addition to the transcription. The word timestam...
http://arxiv.org/pdf/2306.11473v1
2023-06-20T11:53:43
2023-06-20T11:53:43
{ "title": [ -0.009811401, -0.012794495, -0.04928589, -0.046325684, -0.027145386, -0.015403748, -0.052215576, 0.027893066, -0.0000012516975, 0.043945312, -0.044647217, -0.006839752, 0.030319214, -0.024871826, -0.015274048, 0.004524231, 0.011299133, ...
http://arxiv.org/abs/2306.11444v1
Blackbird language matrices (BLM), a new task for rule-like generalization in neural networks: Motivations and Formal Specifications
Paola Merlo
cs.CL
We motivate and formally define a new task for fine-tuning rule-like generalization in large language models. It is conjectured that the shortcomings of current LLMs are due to a lack of ability to generalize. It has been argued that, instead, humans are better at generalization because they have a tendency at extracti...
http://arxiv.org/pdf/2306.11444v1
2023-06-20T10:45:56
2023-06-20T10:45:56
{ "title": [ -0.019165039, -0.034332275, -0.03390503, -0.0017004013, -0.06274414, -0.014419556, -0.03930664, 0.01637268, -0.010177612, 0.018005371, -0.004360199, 0.019577026, -0.016479492, -0.029968262, 0.009063721, -0.026977539, -0.0050315857, 0...
http://arxiv.org/abs/2306.17184v1
Why can neural language models solve next-word prediction? A mathematical perspective
Vinoth Nandakumar, Peng Mi, Tongliang Liu
cs.CL, cs.LG, I.2.7
Recently, deep learning has revolutionized the field of natural language processing, with neural language models proving to be very effective for next-word prediction. However, a rigorous theoretical explanation for their success in the context of formal language theory has not yet been developed, as it is unclear why ...
http://arxiv.org/pdf/2306.17184v1
2023-06-20T10:41:23
2023-06-20T10:41:23
{ "title": [ -0.018112183, 0.0024871826, -0.0129470825, -0.034729004, -0.010467529, -0.0019893646, -0.042419434, -0.00944519, -0.010414124, 0.046295166, -0.02532959, 0.004787445, 0.014907837, 0.008453369, -0.049591064, 0.0046539307, -0.0082092285, ...
http://arxiv.org/abs/2306.11432v1
Towards Theory-based Moral AI: Moral AI with Aggregating Models Based on Normative Ethical Theory
Masashi Takeshita, Rzepka Rafal, Kenji Araki
cs.AI, cs.CL
Moral AI has been studied in the fields of philosophy and artificial intelligence. Although most existing studies are only theoretical, recent developments in AI have made it increasingly necessary to implement AI with morality. On the other hand, humans are under the moral uncertainty of not knowing what is morally ri...
http://arxiv.org/pdf/2306.11432v1
2023-06-20T10:22:24
2023-06-20T10:22:24
{ "title": [ -0.007911682, 0.01600647, -0.04437256, -0.023330688, 0.026947021, -0.025360107, 0.020095825, -0.020019531, -0.016860962, 0.039245605, -0.035614014, 0.019195557, -0.0016078949, -0.00008517504, -0.040924072, -0.0011930466, -0.023529053, ...
http://arxiv.org/abs/2306.11426v1
Exploring the Performance and Efficiency of Transformer Models for NLP on Mobile Devices
Ioannis Panopoulos, Sokratis Nikolaidis, Stylianos I. Venieris, Iakovos S. Venieris
cs.LG, cs.CL
Deep learning (DL) is characterised by its dynamic nature, with new deep neural network (DNN) architectures and approaches emerging every few years, driving the field's advancement. At the same time, the ever-increasing use of mobile devices (MDs) has resulted in a surge of DNN-based mobile applications. Although tradi...
http://arxiv.org/pdf/2306.11426v1
2023-06-20T10:15:01
2023-06-20T10:15:01
{ "title": [ -0.02609253, 0.005039215, -0.016998291, -0.022033691, -0.0012397766, -0.0031738281, -0.040222168, -0.027679443, -0.01966858, 0.02178955, -0.046875, 0.015975952, 0.024795532, -0.0054969788, -0.009300232, -0.0018672943, -0.029724121, 0...
http://arxiv.org/abs/2306.11420v1
On Evaluating Multilingual Compositional Generalization with Translated Datasets
Zi Wang, Daniel Hershcovich
cs.CL
Compositional generalization allows efficient learning and human-like inductive biases. Since most research investigating compositional generalization in NLP is done on English, important questions remain underexplored. Do the necessary compositional generalization abilities differ across languages? Can models composit...
http://arxiv.org/pdf/2306.11420v1
2023-06-20T10:03:57
2023-06-20T10:03:57
{ "title": [ -0.023422241, -0.0035591125, -0.016983032, -0.06951904, 0.0019006729, 0.010948181, -0.074401855, 0.0010299683, -0.0064430237, 0.046539307, -0.010826111, 0.017822266, -0.029159546, -0.0048332214, -0.06286621, -0.009429932, -0.025009155, ...
http://arxiv.org/abs/2306.11400v1
MuDPT: Multi-modal Deep-symphysis Prompt Tuning for Large Pre-trained Vision-Language Models
Yongzhu Miao, Shasha Li, Jintao Tang, Ting Wang
cs.CV, cs.CL
Prompt tuning, like CoOp, has recently shown promising vision recognizing and transfer learning ability on various downstream tasks with the emergence of large pre-trained vision-language models like CLIP. However, we identify that existing uni-modal prompt tuning approaches may result in sub-optimal performance since ...
http://arxiv.org/pdf/2306.11400v1
2023-06-20T09:15:52
2023-06-20T09:15:52
{ "title": [ -0.011810303, -0.017669678, -0.01953125, -0.06137085, -0.02305603, 0.001964569, -0.07232666, -0.009132385, 0.0077285767, 0.012901306, -0.02192688, -0.011413574, 0.021759033, -0.03845215, -0.013427734, -0.009567261, 0.021911621, 0.003...
http://arxiv.org/abs/2306.11386v1
Did the Models Understand Documents? Benchmarking Models for Language Understanding in Document-Level Relation Extraction
Haotian Chen, Bingsheng Chen, Xiangdong Zhou
cs.CL
Document-level relation extraction (DocRE) attracts more research interest recently. While models achieve consistent performance gains in DocRE, their underlying decision rules are still understudied: Do they make the right predictions according to rationales? In this paper, we take the first step toward answering this...
http://arxiv.org/pdf/2306.11386v1
2023-06-20T08:52:05
2023-06-20T08:52:05
{ "title": [ -0.025161743, -0.00015938282, -0.027755737, -0.05569458, -0.018798828, -0.002614975, -0.068481445, -0.013450623, 0.037322998, 0.014778137, -0.035614014, -0.014579773, 0.0047836304, -0.003025055, -0.06365967, -0.010124207, -0.010986328, ...
http://arxiv.org/abs/2306.11372v1
Democratizing LLMs for Low-Resource Languages by Leveraging their English Dominant Abilities with Linguistically-Diverse Prompts
Xuan-Phi Nguyen, Sharifah Mahani Aljunied, Shafiq Joty, Lidong Bing
cs.CL, cs.AI
Large language models (LLMs) are known to effectively perform tasks by simply observing few exemplars. However, in low-resource languages, obtaining such hand-picked exemplars can still be challenging, where unsupervised techniques may be necessary. Moreover, competent generative capabilities of LLMs are observed only ...
http://arxiv.org/pdf/2306.11372v1
2023-06-20T08:27:47
2023-06-20T08:27:47
{ "title": [ -0.031341553, -0.028533936, -0.047912598, -0.0036182404, -0.052825928, -0.013587952, -0.06970215, 0.0037670135, -0.01373291, 0.035827637, -0.016860962, 0.025024414, 0.015419006, -0.032470703, 0.009048462, 0.0020580292, -0.015655518, ...
http://arxiv.org/abs/2306.11371v2
Visually grounded few-shot word learning in low-resource settings
Leanne Nortje, Dan Oneata, Herman Kamper
eess.AS, cs.CL
We propose a visually grounded speech model that learns new words and their visual depictions from just a few word-image example pairs. Given a set of test images and a spoken query, we ask the model which image depicts the query word. Previous work has simplified this few-shot learning problem by either using an artif...
http://arxiv.org/pdf/2306.11371v2
2023-06-21T07:22:08
2023-06-20T08:27:42
{ "title": [ -0.019119263, -0.018096924, -0.02355957, -0.0077323914, -0.039398193, -0.016708374, -0.004924774, -0.0038375854, -0.001964569, 0.02772522, -0.04916382, 0.020751953, -0.013427734, -0.006122589, -0.04272461, -0.022338867, -0.0017404556, ...
http://arxiv.org/abs/2306.11345v1
KiUT: Knowledge-injected U-Transformer for Radiology Report Generation
Zhongzhen Huang, Xiaofan Zhang, Shaoting Zhang
cs.CV, cs.CL
Radiology report generation aims to automatically generate a clinically accurate and coherent paragraph from the X-ray image, which could relieve radiologists from the heavy burden of report writing. Although various image caption methods have shown remarkable performance in the natural image field, generating accurate...
http://arxiv.org/pdf/2306.11345v1
2023-06-20T07:27:28
2023-06-20T07:27:28
{ "title": [ -0.005542755, -0.0049362183, -0.043304443, -0.07269287, -0.0037517548, 0.008392334, -0.05368042, -0.00003939867, -0.012565613, 0.03186035, -0.019897461, 0.004501343, 0.016220093, -0.037994385, -0.022506714, 0.0013790131, 0.0028800964, ...
http://arxiv.org/abs/2306.11341v1
MSVD-Indonesian: A Benchmark for Multimodal Video-Text Tasks in Indonesian
Willy Fitra Hendria
cs.MM, cs.CL, cs.CV, cs.LG, eess.IV
Multimodal learning on video and text data has been receiving growing attention from many researchers in various research tasks, including text-to-video retrieval, video-to-text retrieval, and video captioning. Although many algorithms have been proposed for those challenging tasks, most of them are developed on Englis...
http://arxiv.org/pdf/2306.11341v1
2023-06-20T07:19:36
2023-06-20T07:19:36
{ "title": [ -0.00881958, -0.014549255, -0.01737976, -0.0395813, -0.029937744, -0.004623413, -0.053009033, -0.011917114, 0.007850647, 0.035308838, -0.01586914, 0.017745972, -0.014793396, -0.010932922, -0.064697266, -0.013549805, -0.013305664, 0.0...
http://arxiv.org/abs/2306.11309v1
Multi-pass Training and Cross-information Fusion for Low-resource End-to-end Accented Speech Recognition
Xuefei Wang, Yanhua Long, Yijie Li, Haoran Wei
cs.SD, cs.CL, eess.AS, eess.SP
Low-resource accented speech recognition is one of the important challenges faced by current ASR technology in practical applications. In this study, we propose a Conformer-based architecture, called Aformer, to leverage both the acoustic information from large non-accented and limited accented training data. Specifica...
http://arxiv.org/pdf/2306.11309v1
2023-06-20T06:08:09
2023-06-20T06:08:09
{ "title": [ 0.00081157684, -0.0025367737, -0.019805908, -0.053741455, -0.0027179718, -0.0010032654, -0.023040771, 0.002664566, -0.014678955, 0.042755127, -0.019927979, 0.01537323, 0.012962341, -0.03201294, -0.018966675, -0.0055236816, -0.012992859, ...
http://arxiv.org/abs/2306.11300v2
RS5M: A Large Scale Vision-Language Dataset for Remote Sensing Vision-Language Foundation Model
Zilun Zhang, Tiancheng Zhao, Yulong Guo, Jianwei Yin
cs.CV, cs.AI, cs.CL, cs.MM
Pre-trained Vision-Language Foundation Models utilizing extensive image-text paired data have demonstrated unprecedented image-text association capabilities, achieving remarkable results across various downstream tasks. A critical challenge is how to make use of existing large-scale pre-trained VLMs, which are trained ...
http://arxiv.org/pdf/2306.11300v2
2023-08-31T22:33:54
2023-06-20T05:30:59
{ "title": [ -0.031555176, -0.022201538, -0.058502197, -0.054016113, -0.012527466, -0.019119263, -0.057800293, 0.012413025, 0.016830444, 0.021774292, -0.030853271, -0.01222229, -0.017547607, -0.016693115, -0.035095215, -0.014183044, 0.025970459, ...
http://arxiv.org/abs/2306.11296v2
ChatGPT Chemistry Assistant for Text Mining and Prediction of MOF Synthesis
Zhiling Zheng, Oufan Zhang, Christian Borgs, Jennifer T. Chayes, Omar M. Yaghi
cs.IR, cond-mat.mtrl-sci, cs.CL, physics.chem-ph
We use prompt engineering to guide ChatGPT in the automation of text mining of metal-organic frameworks (MOFs) synthesis conditions from diverse formats and styles of the scientific literature. This effectively mitigates ChatGPT's tendency to hallucinate information -- an issue that previously made the use of Large Lan...
http://arxiv.org/pdf/2306.11296v2
2023-07-20T02:20:35
2023-06-20T05:20:29
{ "title": [ -0.031341553, -0.016021729, -0.08148193, -0.06616211, -0.031234741, -0.0042419434, -0.0848999, 0.04437256, -0.037078857, 0.04269409, -0.00762558, 0.049102783, 0.0116119385, -0.037841797, 0.028182983, 0.008392334, -0.034210205, 0.0443...
http://arxiv.org/abs/2306.11270v2
Evaluating the Zero-shot Robustness of Instruction-tuned Language Models
Jiuding Sun, Chantal Shaib, Byron C. Wallace
cs.CL, cs.LG
Instruction fine-tuning has recently emerged as a promising approach for improving the zero-shot capabilities of Large Language Models (LLMs) on new tasks. This technique has shown particular strength in improving the performance of modestly sized LLMs, sometimes inducing performance competitive with much larger model ...
http://arxiv.org/pdf/2306.11270v2
2023-07-09T00:39:30
2023-06-20T03:48:51
{ "title": [ -0.037994385, -0.019134521, -0.017990112, -0.050720215, -0.02003479, 0.00086832047, -0.074279785, 0.001455307, -0.003408432, 0.037322998, -0.015350342, 0.017715454, -0.011291504, -0.005809784, -0.064331055, 0.010597229, -0.010848999, ...
http://arxiv.org/abs/2306.11260v3
A Novel Counterfactual Data Augmentation Method for Aspect-Based Sentiment Analysis
Dongming Wu, Lulu Wen, Chao Chen, Zhaoshu Shi
cs.CL
Aspect-based-sentiment-analysis (ABSA) is a fine-grained sentiment evaluation task, which analyzes the emotional polarity of the evaluation aspects. Generally, the emotional polarity of an aspect exists in the corresponding opinion expression, whose diversity has great impact on model's performance. To mitigate this pr...
http://arxiv.org/pdf/2306.11260v3
2023-10-07T02:23:14
2023-06-20T03:25:51
{ "title": [ 0.024414062, -0.0030212402, -0.023132324, -0.04550171, 0.016052246, -0.0076408386, -0.0011978149, -0.0036315918, 0.04724121, 0.043121338, -0.03213501, -0.006263733, 0.0044136047, -0.0068206787, -0.08331299, 0.0076675415, -0.013313293, ...
http://arxiv.org/abs/2306.11256v1
GUMSum: Multi-Genre Data and Evaluation for English Abstractive Summarization
Yang Janet Liu, Amir Zeldes
cs.CL
Automatic summarization with pre-trained language models has led to impressively fluent results, but is prone to 'hallucinations', low performance on non-news genres, and outputs which are not exactly summaries. Targeting ACL 2023's 'Reality Check' theme, we present GUMSum, a small but carefully crafted dataset of Engl...
http://arxiv.org/pdf/2306.11256v1
2023-06-20T03:21:10
2023-06-20T03:21:10
{ "title": [ -0.013282776, 0.0029277802, -0.049682617, -0.06896973, 0.0068130493, 0.00004005432, -0.051605225, -0.011306763, -0.0013923645, 0.037872314, 0.0008363724, 0.03152466, -0.017364502, -0.0054626465, -0.054534912, 0.00819397, -0.019851685, ...
http://arxiv.org/abs/2306.11252v1
HK-LegiCoST: Leveraging Non-Verbatim Transcripts for Speech Translation
Cihan Xiao, Henry Li Xinyuan, Jinyi Yang, Dongji Gao, Matthew Wiesner, Kevin Duh, Sanjeev Khudanpur
cs.CL, cs.LG
We introduce HK-LegiCoST, a new three-way parallel corpus of Cantonese-English translations, containing 600+ hours of Cantonese audio, its standard traditional Chinese transcript, and English translation, segmented and aligned at the sentence level. We describe the notable challenges in corpus preparation: segmentation...
http://arxiv.org/pdf/2306.11252v1
2023-06-20T03:09:32
2023-06-20T03:09:32
{ "title": [ -0.041503906, -0.022521973, -0.035064697, -0.058441162, -0.03363037, 0.013313293, -0.0635376, 0.025863647, -0.0032138824, 0.04421997, -0.025634766, 0.018798828, 0.011108398, -0.0075683594, -0.055603027, 0.011108398, -0.007083893, 0.0...
http://arxiv.org/abs/2306.11232v1
Eight challenges in developing theory of intelligence
Haiping Huang
q-bio.NC, cond-mat.stat-mech, cs.AI, cs.CL
A good theory of mathematical beauty is more practical than any current observation, as new predictions of physical reality can be verified self-consistently. This belief applies to the current status of understanding deep neural networks including large language models and even the biological intelligence. Toy models ...
http://arxiv.org/pdf/2306.11232v1
2023-06-20T01:45:42
2023-06-20T01:45:42
{ "title": [ 0.030807495, 0.018997192, -0.02583313, 0.02180481, 0.014137268, -0.045074463, 0.06124878, -0.026245117, -0.057434082, 0.04547119, -0.011077881, 0.0028152466, -0.001209259, -0.037384033, -0.017044067, -0.015045166, -0.0059928894, 0.03...
http://arxiv.org/abs/2306.11222v2
LoSparse: Structured Compression of Large Language Models based on Low-Rank and Sparse Approximation
Yixiao Li, Yifan Yu, Qingru Zhang, Chen Liang, Pengcheng He, Weizhu Chen, Tuo Zhao
cs.LG, cs.CL
Transformer models have achieved remarkable results in various natural language tasks, but they are often prohibitively large, requiring massive memories and computational resources. To reduce the size and complexity of these models, we propose LoSparse (Low-Rank and Sparse approximation), a novel model compression tec...
http://arxiv.org/pdf/2306.11222v2
2023-06-26T15:34:57
2023-06-20T01:16:11
{ "title": [ -0.06958008, -0.022659302, -0.04360962, -0.06866455, -0.012992859, 0.028747559, -0.08392334, 0.022537231, 0.0023822784, 0.010528564, -0.003929138, -0.0013628006, -0.0016260147, -0.034851074, -0.014190674, 0.0060768127, 0.014984131, -...
http://arxiv.org/abs/2306.11207v3
Quilt-1M: One Million Image-Text Pairs for Histopathology
Wisdom Oluchi Ikezogwo, Mehmet Saygin Seyfioglu, Fatemeh Ghezloo, Dylan Stefan Chan Geva, Fatwir Sheikh Mohammed, Pavan Kumar Anand, Ranjay Krishna, Linda Shapiro
cs.CV, cs.CL, cs.LG
Recent accelerations in multi-modal applications have been made possible with the plethora of image and text data available online. However, the scarcity of analogous data in the medical field, specifically in histopathology, has slowed comparable progress. To enable similar representation learning for histopathology, ...
http://arxiv.org/pdf/2306.11207v3
2023-10-27T23:56:33
2023-06-20T00:14:47
{ "title": [ -0.016845703, -0.00920105, -0.049346924, -0.0501709, -0.001996994, -0.013725281, -0.031280518, -0.00491333, 0.0049095154, 0.022064209, -0.021362305, 0.009857178, -0.009498596, -0.02130127, -0.015716553, 0.00447464, 0.012840271, 0.018...
http://arxiv.org/abs/2306.11197v4
Sparse Modular Activation for Efficient Sequence Modeling
Liliang Ren, Yang Liu, Shuohang Wang, Yichong Xu, Chenguang Zhu, ChengXiang Zhai
cs.LG, cs.CL
Recent hybrid models combining Linear State Space Models (SSMs) with self-attention mechanisms have demonstrated impressive results across a range of sequence modeling tasks. However, current approaches apply attention modules statically and uniformly to all elements in the input sequences, leading to sub-optimal quali...
http://arxiv.org/pdf/2306.11197v4
2023-11-04T21:26:03
2023-06-19T23:10:02
{ "title": [ -0.023208618, -0.012138367, -0.061920166, -0.11071777, 0.0146865845, 0.0039863586, -0.07897949, 0.041229248, 0.007835388, 0.044830322, 0.0020198822, 0.009063721, 0.019836426, -0.019210815, -0.047576904, 0.026687622, 0.0048217773, 0.0...
http://arxiv.org/abs/2306.11189v1
BioREx: Improving Biomedical Relation Extraction by Leveraging Heterogeneous Datasets
Po-Ting Lai, Chih-Hsuan Wei, Ling Luo, Qingyu Chen, Zhiyong Lu
cs.CL
Biomedical relation extraction (RE) is the task of automatically identifying and characterizing relations between biomedical concepts from free text. RE is a central task in biomedical natural language processing (NLP) research and plays a critical role in many downstream applications, such as literature-based discover...
http://arxiv.org/pdf/2306.11189v1
2023-06-19T22:48:18
2023-06-19T22:48:18
{ "title": [ -0.012878418, -0.008644104, -0.05529785, -0.07873535, 0.02017212, 0.012031555, -0.049926758, -0.006214142, 0.02658081, 0.015335083, -0.013198853, 0.0000078082085, -0.008384705, -0.012893677, -0.04296875, 0.004295349, -0.014854431, 0....
http://arxiv.org/abs/2306.11174v1
Evaluating Privacy Questions From Stack Overflow: Can ChatGPT Compete?
Zack Delile, Sean Radel, Joe Godinez, Garrett Engstrom, Theo Brucker, Kenzie Young, Sepideh Ghanavati
cs.SE, cs.AI, cs.CL
Stack Overflow and other similar forums are used commonly by developers to seek answers for their software development as well as privacy-related concerns. Recently, ChatGPT has been used as an alternative to generate code or produce responses to developers' questions. In this paper, we aim to understand developers' pr...
http://arxiv.org/pdf/2306.11174v1
2023-06-19T21:33:04
2023-06-19T21:33:04
{ "title": [ 0.004173279, -0.0051727295, -0.023010254, 0.000017940998, 0.009315491, -0.016067505, 0.024093628, -0.02607727, -0.03555298, 0.009315491, -0.015052795, 0.017944336, -0.030212402, -0.013183594, -0.015174866, -0.022247314, -0.02218628, ...
http://arxiv.org/abs/2306.11167v4
Large Language Models are Fixated by Red Herrings: Exploring Creative Problem Solving and Einstellung Effect using the Only Connect Wall Dataset
Saeid Naeini, Raeid Saqur, Mozhgan Saeidi, John Giorgi, Babak Taati
cs.CL, cs.AI, cs.LG, I.2.7
The quest for human imitative AI has been an enduring topic in AI research since its inception. The technical evolution and emerging capabilities of the latest cohort of large language models (LLMs) have reinvigorated the subject beyond academia to the cultural zeitgeist. While recent NLP evaluation benchmark tasks tes...
http://arxiv.org/pdf/2306.11167v4
2023-11-08T20:47:00
2023-06-19T21:14:57
{ "title": [ -0.0005712509, -0.013298035, -0.010734558, 0.033081055, -0.054870605, -0.036590576, -0.0028953552, -0.032928467, -0.0060310364, 0.026382446, 0.016418457, -0.029403687, 0.018447876, -0.042663574, -0.014854431, -0.028015137, -0.024734497, ...
http://arxiv.org/abs/2306.11066v2
Adversarial Robustness of Prompt-based Few-Shot Learning for Natural Language Understanding
Venkata Prabhakara Sarath Nookala, Gaurav Verma, Subhabrata Mukherjee, Srijan Kumar
cs.CL, cs.LG
State-of-the-art few-shot learning (FSL) methods leverage prompt-based fine-tuning to obtain remarkable results for natural language understanding (NLU) tasks. While much of the prior FSL methods focus on improving downstream task performance, there is a limited understanding of the adversarial robustness of such metho...
http://arxiv.org/pdf/2306.11066v2
2023-06-21T03:56:39
2023-06-19T17:01:13
{ "title": [ 0.0074157715, -0.0069351196, -0.014762878, -0.053009033, -0.0031642914, 0.0052757263, -0.03793335, -0.0056877136, -0.012046814, -0.006137848, -0.025650024, 0.007820129, -0.036193848, -0.025390625, -0.036315918, -0.039855957, 0.0072288513...
http://arxiv.org/abs/2306.11065v1
Cross-Modal Attribute Insertions for Assessing the Robustness of Vision-and-Language Learning
Shivaen Ramshetty, Gaurav Verma, Srijan Kumar
cs.CL, cs.AI, cs.CV
The robustness of multimodal deep learning models to realistic changes in the input text is critical for their applicability to important tasks such as text-to-image retrieval and cross-modal entailment. To measure robustness, several existing approaches edit the text data, but do so without leveraging the cross-modal ...
http://arxiv.org/pdf/2306.11065v1
2023-06-19T17:00:03
2023-06-19T17:00:03
{ "title": [ 0.013946533, 0.016479492, -0.013465881, -0.0044136047, 0.005542755, -0.036102295, 0.00060224533, -0.059814453, 0.0046691895, 0.020370483, -0.030303955, 0.010719299, 0.008628845, -0.017044067, -0.04525757, -0.012184143, -0.015563965, ...
http://arxiv.org/abs/2306.11044v1
Frequency effects in Linear Discriminative Learning
Maria Heitmeier, Yu-Ying Chuang, Seth D. Axen, R. Harald Baayen
cs.CL
Word frequency is a strong predictor in most lexical processing tasks. Thus, any model of word recognition needs to account for how word frequency effects arise. The Discriminative Lexicon Model (DLM; Baayen et al., 2018a, 2019) models lexical processing with linear mappings between words' forms and their meanings. So ...
http://arxiv.org/pdf/2306.11044v1
2023-06-19T16:15:46
2023-06-19T16:15:46
{ "title": [ 0.008728027, 0.0052223206, -0.0234375, -0.026916504, 0.017181396, -0.03250122, 0.027664185, -0.024032593, -0.01927185, 0.025177002, -0.029418945, 0.018615723, -0.030029297, -0.023330688, -0.033050537, -0.009391785, -0.010276794, 0.04...
http://arxiv.org/abs/2306.14913v1
FSUIE: A Novel Fuzzy Span Mechanism for Universal Information Extraction
Tianshuo Peng, Zuchao Li, Lefei Zhang, Bo Du, Hai Zhao
cs.CL, cs.AI
Universal Information Extraction (UIE) has been introduced as a unified framework for various Information Extraction (IE) tasks and has achieved widespread success. Despite this, UIE models have limitations. For example, they rely heavily on span boundaries in the data during training, which does not reflect the realit...
http://arxiv.org/pdf/2306.14913v1
2023-06-19T15:59:28
2023-06-19T15:59:28
{ "title": [ -0.042785645, -0.00749588, -0.055541992, -0.03805542, -0.0034599304, 0.0007915497, -0.03475952, -0.0032615662, 0.009880066, 0.040496826, -0.025222778, 0.010154724, -0.0012207031, -0.0065727234, -0.039611816, 0.0018367767, -0.009178162, ...
http://arxiv.org/abs/2306.11027v1
JiuZhang 2.0: A Unified Chinese Pre-trained Language Model for Multi-task Mathematical Problem Solving
Wayne Xin Zhao, Kun Zhou, Beichen Zhang, Zheng Gong, Zhipeng Chen, Yuanhang Zhou, Ji-Rong Wen, Jing Sha, Shijin Wang, Cong Liu, Guoping Hu
cs.CL, cs.AI
Although pre-trained language models~(PLMs) have recently advanced the research progress in mathematical reasoning, they are not specially designed as a capable multi-task solver, suffering from high cost for multi-task deployment (\eg a model copy for a task) and inferior performance on complex mathematical problems i...
http://arxiv.org/pdf/2306.11027v1
2023-06-19T15:45:36
2023-06-19T15:45:36
{ "title": [ -0.010185242, -0.01953125, -0.016860962, -0.038208008, -0.005821228, -0.013420105, -0.02067566, 0.0019273758, -0.0012254715, 0.046173096, 0.020980835, -0.014129639, 0.017120361, -0.030456543, -0.0335083, 0.0034980774, -0.011695862, 0...
http://arxiv.org/abs/2306.11025v1
Temporal Data Meets LLM -- Explainable Financial Time Series Forecasting
Xinli Yu, Zheng Chen, Yuan Ling, Shujing Dong, Zongyi Liu, Yanbin Lu
cs.LG, cs.AI, cs.CL, q-fin.ST, F.2.2; I.2.7; I.2.1
This paper presents a novel study on harnessing Large Language Models' (LLMs) outstanding knowledge and reasoning abilities for explainable financial time series forecasting. The application of machine learning models to financial time series comes with several challenges, including the difficulty in cross-sequence rea...
http://arxiv.org/pdf/2306.11025v1
2023-06-19T15:42:02
2023-06-19T15:42:02
{ "title": [ -0.01902771, 0.004547119, -0.06726074, -0.07183838, 0.011390686, -0.020446777, -0.03237915, 0.017654419, -0.029144287, 0.011787415, -0.052246094, 0.017654419, -0.03262329, -0.014976501, -0.023361206, -0.0473938, 0.006298065, 0.036163...
http://arxiv.org/abs/2306.11020v1
Dual-Gated Fusion with Prefix-Tuning for Multi-Modal Relation Extraction
Qian Li, Shu Guo, Cheng Ji, Xutan Peng, Shiyao Cui, Jianxin Li
cs.CL
Multi-Modal Relation Extraction (MMRE) aims at identifying the relation between two entities in texts that contain visual clues. Rich visual content is valuable for the MMRE task, but existing works cannot well model finer associations among different modalities, failing to capture the truly helpful visual information ...
http://arxiv.org/pdf/2306.11020v1
2023-06-19T15:31:34
2023-06-19T15:31:34
{ "title": [ -0.007709503, 0.008857727, -0.05014038, -0.07434082, 0.010940552, 0.00497818, -0.060760498, -0.0058555603, 0.027557373, 0.019683838, 0.0023822784, 0.0076065063, 0.01777649, -0.030731201, -0.034240723, 0.009819031, -0.026321411, 0.038...
http://arxiv.org/abs/2306.10985v1
LARG, Language-based Automatic Reward and Goal Generation
Julien Perez, Denys Proux, Claude Roux, Michael Niemaz
cs.CL, cs.LG, cs.RO
Goal-conditioned and Multi-Task Reinforcement Learning (GCRL and MTRL) address numerous problems related to robot learning, including locomotion, navigation, and manipulation scenarios. Recent works focusing on language-defined robotic manipulation tasks have led to the tedious production of massive human annotations t...
http://arxiv.org/pdf/2306.10985v1
2023-06-19T14:52:39
2023-06-19T14:52:39
{ "title": [ -0.00819397, -0.012794495, -0.032714844, -0.014808655, -0.01210022, -0.016586304, 0.00045323372, -0.016159058, -0.03060913, 0.029678345, 0.003019333, -0.0014209747, -0.0020046234, -0.04006958, 0.009307861, -0.00579834, 0.011543274, 0...
http://arxiv.org/abs/2306.10974v2
Fine-Tuning Language Models for Scientific Writing Support
Justin Mücke, Daria Waldow, Luise Metzger, Philipp Schauz, Marcel Hoffman, Nicolas Lell, Ansgar Scherp
cs.CL
We support scientific writers in determining whether a written sentence is scientific, to which section it belongs, and suggest paraphrasings to improve the sentence. Firstly, we propose a regression model trained on a corpus of scientific sentences extracted from peer-reviewed scientific papers and non-scientific text...
http://arxiv.org/pdf/2306.10974v2
2023-06-21T08:31:48
2023-06-19T14:34:49
{ "title": [ -0.054748535, -0.011909485, -0.02949524, -0.03451538, -0.01864624, 0.017227173, -0.05444336, 0.00023424625, 0.0004632473, 0.041381836, 0.0037765503, -0.009399414, -0.0009317398, -0.014282227, -0.064331055, 0.019836426, -0.01309967, 0...
http://arxiv.org/abs/2306.10968v2
BayLing: Bridging Cross-lingual Alignment and Instruction Following through Interactive Translation for Large Language Models
Shaolei Zhang, Qingkai Fang, Zhuocheng Zhang, Zhengrui Ma, Yan Zhou, Langlin Huang, Mengyu Bu, Shangtong Gui, Yunji Chen, Xilin Chen, Yang Feng
cs.CL, cs.AI
Large language models (LLMs) have demonstrated remarkable prowess in language understanding and generation. Advancing from foundation LLMs to instructionfollowing LLMs, instruction tuning plays a vital role in aligning LLMs to human preferences. However, the existing LLMs are usually focused on English, leading to infe...
http://arxiv.org/pdf/2306.10968v2
2023-06-21T11:31:50
2023-06-19T14:30:52
{ "title": [ -0.041503906, -0.009651184, -0.026123047, -0.06347656, -0.0118637085, 0.0054969788, -0.07055664, -0.011253357, -0.0015144348, 0.03781128, -0.0149383545, 0.02960205, 0.01701355, -0.018585205, -0.01965332, 0.015716553, -0.018493652, 0....
http://arxiv.org/abs/2306.10964v1
Multilingual Few-Shot Learning via Language Model Retrieval
Genta Indra Winata, Liang-Kang Huang, Soumya Vadlamannati, Yash Chandarana
cs.CL
Transformer-based language models have achieved remarkable success in few-shot in-context learning and drawn a lot of research interest. However, these models' performance greatly depends on the choice of the example prompts and also has high variability depending on how samples are chosen. In this paper, we conduct a ...
http://arxiv.org/pdf/2306.10964v1
2023-06-19T14:27:21
2023-06-19T14:27:21
{ "title": [ -0.017730713, -0.027496338, -0.0127334595, -0.036712646, -0.03024292, 0.014572144, -0.052734375, 0.012771606, 0.0009822845, 0.03704834, -0.01878357, 0.010856628, -0.0068092346, -0.011161804, -0.040771484, -0.004928589, 0.0033454895, ...
http://arxiv.org/abs/2306.10869v1
Grammatical gender in Swedish is predictable using recurrent neural networks
Edvin Listo Zec, Olof Mogren
cs.CL, cs.LG
The grammatical gender of Swedish nouns is a mystery. While there are few rules that can indicate the gender with some certainty, it does in general not depend on either meaning or the structure of the word. In this paper we demonstrate the surprising fact that grammatical gender for Swedish nouns can be predicted with...
http://arxiv.org/pdf/2306.10869v1
2023-06-19T11:42:47
2023-06-19T11:42:47
{ "title": [ 0.01965332, -0.0115737915, -0.04156494, -0.051330566, -0.015655518, -0.029373169, -0.013885498, 0.025878906, -0.006454468, 0.053009033, -0.0024757385, -0.0022144318, 0.064086914, -0.045074463, 0.0044136047, 0.014297485, -0.0058403015, ...
http://arxiv.org/abs/2306.17181v4
Unsupervised Text Embedding Space Generation Using Generative Adversarial Networks for Text Synthesis
Jun-Min Lee, Tae-Bin Ha
cs.CL, cs.LG
Generative Adversarial Networks (GAN) is a model for data synthesis, which creates plausible data through the competition of generator and discriminator. Although GAN application to image synthesis is extensively studied, it has inherent limitations to natural language generation. Because natural language is composed o...
http://arxiv.org/pdf/2306.17181v4
2023-10-17T10:41:12
2023-06-19T10:22:12
{ "title": [ -0.03918457, -0.0135269165, -0.03024292, -0.013442993, -0.025009155, 0.0010595322, -0.031021118, -0.016784668, 0.0021038055, 0.032043457, -0.033569336, 0.0110321045, 0.008758545, -0.010528564, -0.033477783, -0.011184692, -0.012367249, ...
http://arxiv.org/abs/2306.10821v1
Comparison of L2 Korean pronunciation error patterns from five L1 backgrounds by using automatic phonetic transcription
Eun Jung Yeo, Hyungshin Ryu, Jooyoung Lee, Sunhee Kim, Minhwa Chung
cs.CL, cs.SD, eess.AS
This paper presents a large-scale analysis of L2 Korean pronunciation error patterns from five different language backgrounds, Chinese, Vietnamese, Japanese, Thai, and English, by using automatic phonetic transcription. For the analysis, confusion matrices are generated for each L1, by aligning canonical phone sequence...
http://arxiv.org/pdf/2306.10821v1
2023-06-19T10:10:46
2023-06-19T10:10:46
{ "title": [ -0.023025513, 0.0020713806, -0.03942871, -0.009757996, -0.042053223, -0.003955841, -0.04449463, 0.006752014, -0.02507019, 0.055725098, -0.020050049, 0.022720337, 0.029373169, -0.030593872, -0.016586304, -0.0034122467, -0.020462036, 0...
http://arxiv.org/abs/2306.17180v1
Replace and Report: NLP Assisted Radiology Report Generation
Kaveri Kale, pushpak Bhattacharyya, Kshitij Jadhav
cs.CL, cs.AI, cs.CV
Clinical practice frequently uses medical imaging for diagnosis and treatment. A significant challenge for automatic radiology report generation is that the radiology reports are long narratives consisting of multiple sentences for both abnormal and normal findings. Therefore, applying conventional image captioning app...
http://arxiv.org/pdf/2306.17180v1
2023-06-19T10:04:42
2023-06-19T10:04:42
{ "title": [ -0.013290405, -0.005355835, -0.05380249, -0.051116943, -0.014144897, -0.006549835, -0.03250122, 0.004753113, -0.0029525757, 0.011161804, -0.03543091, 0.00207901, -0.000647068, -0.028182983, -0.0063591003, 0.0007543564, -0.0178833, 0....
http://arxiv.org/abs/2306.10790v1
Preserving Commonsense Knowledge from Pre-trained Language Models via Causal Inference
Junhao Zheng, Qianli Ma, Shengjie Qiu, Yue Wu, Peitian Ma, Junlong Liu, Huawen Feng, Xichen Shang, Haibin Chen
cs.CL, cs.AI
Fine-tuning has been proven to be a simple and effective technique to transfer the learned knowledge of Pre-trained Language Models (PLMs) to downstream tasks. However, vanilla fine-tuning easily overfits the target data and degrades the generalization ability. Most existing studies attribute it to catastrophic forgett...
http://arxiv.org/pdf/2306.10790v1
2023-06-19T09:06:44
2023-06-19T09:06:44
{ "title": [ -0.0126953125, -0.037475586, -0.027557373, -0.0826416, -0.035949707, 0.007873535, -0.06402588, 0.016052246, 0.0015249252, 0.014968872, -0.008308411, -0.0070266724, 0.0020503998, -0.029190063, -0.024459839, -0.0035305023, -0.008430481, ...
http://arxiv.org/abs/2306.10787v1
Adaptive Ordered Information Extraction with Deep Reinforcement Learning
Wenhao Huang, Jiaqing Liang, Zhixu Li, Yanghua Xiao, Chuanjun Ji
cs.CL
Information extraction (IE) has been studied extensively. The existing methods always follow a fixed extraction order for complex IE tasks with multiple elements to be extracted in one instance such as event extraction. However, we conduct experiments on several complex IE datasets and observe that different extraction...
http://arxiv.org/pdf/2306.10787v1
2023-06-19T08:58:56
2023-06-19T08:58:56
{ "title": [ -0.010223389, 0.0011959076, -0.049713135, -0.066589355, 0.019348145, -0.014793396, -0.029205322, -0.015342712, -0.026489258, 0.014060974, -0.0065460205, 0.012939453, -0.039764404, -0.00881958, -0.030426025, -0.011260986, -0.011734009, ...
http://arxiv.org/abs/2306.10786v1
AMRs Assemble! Learning to Ensemble with Autoregressive Models for AMR Parsing
Abelardo Carlos Martínez Lorenzo, Pere-Lluís Huguet Cabot, Roberto Navigli
cs.CL, cs.AI
In this paper, we examine the current state-of-the-art in AMR parsing, which relies on ensemble strategies by merging multiple graph predictions. Our analysis reveals that the present models often violate AMR structural constraints. To address this issue, we develop a validation method, and show how ensemble models can...
http://arxiv.org/pdf/2306.10786v1
2023-06-19T08:58:47
2023-06-19T08:58:47
{ "title": [ -0.029800415, -0.00030183792, -0.055267334, -0.08319092, -0.0044937134, 0.011894226, -0.0925293, 0.017333984, -0.034118652, 0.028442383, -0.013183594, 0.009666443, -0.03250122, -0.019119263, -0.06021118, -0.022567749, -0.014823914, 0...
http://arxiv.org/abs/2306.10769v1
Gender Differences in Abuse: The Case of Dutch Politicians on Twitter
Isabelle van der Vegt
cs.CL, cs.SI
Online abuse and threats towards politicians have become a significant concern in the Netherlands, like in many other countries across the world. This paper analyses gender differences in abuse received by Dutch politicians on Twitter, while taking into account the possible additional impact of ethnic minority status. ...
http://arxiv.org/pdf/2306.10769v1
2023-06-19T08:23:24
2023-06-19T08:23:24
{ "title": [ 0.012512207, 0.00034332275, -0.032196045, -0.061309814, 0.013679504, -0.020690918, -0.0007596016, 0.012306213, -0.036224365, 0.036254883, -0.026229858, 0.028152466, 0.0046577454, -0.010559082, -0.0413208, -0.014816284, -0.009178162, ...
http://arxiv.org/abs/2306.10765v1
Path to Medical AGI: Unify Domain-specific Medical LLMs with the Lowest Cost
Juexiao Zhou, Xiuying Chen, Xin Gao
cs.AI, cs.CL, cs.CV
Medical artificial general intelligence (AGI) is an emerging field that aims to develop systems specifically designed for medical applications that possess the ability to understand, learn, and apply knowledge across a wide range of tasks and domains. Large language models (LLMs) represent a significant step towards AG...
http://arxiv.org/pdf/2306.10765v1
2023-06-19T08:15:14
2023-06-19T08:15:14
{ "title": [ -0.011497498, -0.007881165, -0.055603027, -0.040222168, -0.039520264, -0.0047302246, -0.077697754, 0.0049209595, -0.068603516, 0.025878906, 0.0051841736, 0.01838684, 0.035369873, -0.060424805, 0.034851074, -0.015220642, 0.01399231, 0...
http://arxiv.org/abs/2306.10763v2
Guiding Language Models of Code with Global Context using Monitors
Lakshya A Agrawal, Aditya Kanade, Navin Goyal, Shuvendu K. Lahiri, Sriram K. Rajamani
cs.CL, cs.AI, cs.LG, cs.PL, cs.SE, I.2.2; I.2.7; I.2.5
Language models of code (LMs) work well when the surrounding code provides sufficient context. This is not true when it becomes necessary to use types, functionality or APIs defined elsewhere in the repository or a linked library, especially those not seen during training. LMs suffer from limited awareness of such glob...
http://arxiv.org/pdf/2306.10763v2
2023-11-03T11:13:15
2023-06-19T08:13:50
{ "title": [ -0.038970947, -0.033966064, -0.041778564, -0.07867432, -0.020950317, 0.013404846, -0.068359375, 0.03036499, 0.018966675, 0.031982422, -0.03161621, -0.010482788, 0.014572144, -0.024475098, -0.03012085, 0.0006351471, 0.011230469, -0.00...
http://arxiv.org/abs/2306.10755v1
Unsupervised Open-domain Keyphrase Generation
Lam Thanh Do, Pritom Saha Akash, Kevin Chen-Chuan Chang
cs.CL
In this work, we study the problem of unsupervised open-domain keyphrase generation, where the objective is a keyphrase generation model that can be built without using human-labeled data and can perform consistently across domains. To solve this problem, we propose a seq2seq model that consists of two modules, namely ...
http://arxiv.org/pdf/2306.10755v1
2023-06-19T07:57:13
2023-06-19T07:57:13
{ "title": [ -0.04473877, -0.0002272129, -0.033111572, -0.037841797, -0.003458023, 0.010650635, -0.0473938, -0.010772705, -0.00058078766, 0.041503906, 0.002254486, 0.019760132, 0.012031555, -0.016433716, -0.02709961, 0.015579224, -0.027664185, 0....
http://arxiv.org/abs/2306.10727v1
Jamp: Controlled Japanese Temporal Inference Dataset for Evaluating Generalization Capacity of Language Models
Tomoki Sugimoto, Yasumasa Onoe, Hitomi Yanaka
cs.CL
Natural Language Inference (NLI) tasks involving temporal inference remain challenging for pre-trained language models (LMs). Although various datasets have been created for this task, they primarily focus on English and do not address the need for resources in other languages. It is unclear whether current LMs realize...
http://arxiv.org/pdf/2306.10727v1
2023-06-19T07:00:14
2023-06-19T07:00:14
{ "title": [ -0.025222778, -0.009727478, -0.054504395, -0.08972168, -0.0049819946, -0.0022182465, -0.07196045, 0.025436401, 0.019317627, 0.025787354, -0.032043457, 0.0059165955, 0.017166138, -0.008148193, -0.058258057, 0.0014362335, -0.012184143, ...
http://arxiv.org/abs/2306.10723v2
Fine-tuning Large Enterprise Language Models via Ontological Reasoning
Teodoro Baldazzi, Luigi Bellomarini, Stefano Ceri, Andrea Colombo, Andrea Gentili, Emanuel Sallinger
cs.CL, cs.DB, cs.LO
Large Language Models (LLMs) exploit fine-tuning as a technique to adapt to diverse goals, thanks to task-specific training data. Task specificity should go hand in hand with domain orientation, that is, the specialization of an LLM to accurately address the tasks of a given realm of interest. However, models are usual...
http://arxiv.org/pdf/2306.10723v2
2023-09-18T21:37:37
2023-06-19T06:48:45
{ "title": [ -0.043304443, -0.029144287, -0.052703857, -0.076660156, -0.0035247803, 0.015823364, -0.0713501, 0.02458191, -0.011764526, 0.019973755, -0.02305603, 0.0028419495, 0.008087158, -0.01473999, -0.011123657, -0.0018196106, -0.015281677, 0....
http://arxiv.org/abs/2307.05493v1
Cases of EFL Secondary Students' Prompt Engineering Pathways to Complete a Writing Task with ChatGPT
David James Woo, Kai Guo, Hengky Susanto
cs.HC, cs.AI, cs.CL
ChatGPT is a state-of-the-art (SOTA) chatbot. Although it has potential to support English as a foreign language (EFL) students' writing, to effectively collaborate with it, a student must learn to engineer prompts, that is, the skill of crafting appropriate instructions so that ChatGPT produces desired outputs. Howeve...
http://arxiv.org/pdf/2307.05493v1
2023-06-19T06:45:04
2023-06-19T06:45:04
{ "title": [ 0.013801575, -0.02355957, 0.012489319, 0.047210693, -0.0340271, -0.008422852, 0.031677246, -0.032562256, -0.03164673, 0.016983032, 0.009109497, 0.04119873, -0.030731201, -0.046325684, 0.0050849915, -0.0087509155, -0.04901123, 0.02688...
http://arxiv.org/abs/2306.10658v1
Distributed Marker Representation for Ambiguous Discourse Markers and Entangled Relations
Dongyu Ru, Lin Qiu, Xipeng Qiu, Yue Zhang, Zheng Zhang
cs.CL
Discourse analysis is an important task because it models intrinsic semantic structures between sentences in a document. Discourse markers are natural representations of discourse in our daily language. One challenge is that the markers as well as pre-defined and human-labeled discourse relations can be ambiguous when ...
http://arxiv.org/pdf/2306.10658v1
2023-06-19T00:49:51
2023-06-19T00:49:51
{ "title": [ -0.00018525124, 0.013198853, -0.023071289, -0.043640137, 0.000787735, -0.002023697, -0.01689148, -0.027328491, 0.0039253235, 0.025924683, 0.021331787, -0.00969696, 0.024597168, -0.033294678, -0.020202637, 0.005191803, -0.01638794, 0....
http://arxiv.org/abs/2307.00009v2
Comparison of Machine Learning Methods for Assigning Software Issues to Team Members
Büşra Tabak, Fatma Başak Aydemir
cs.CL, cs.LG, cs.SE
Software issues contain units of work to fix, improve, or create new threads during the development and facilitate communication among the team members. Assigning an issue to the most relevant team member and determining a category of an issue is a tedious and challenging task. Wrong classifications cause delays and re...
http://arxiv.org/pdf/2307.00009v2
2023-08-23T12:24:02
2023-06-18T20:06:58
{ "title": [ -0.00077962875, -0.0058174133, -0.0003244877, 0.002243042, -0.017532349, -0.030181885, 0.014404297, 0.010398865, -0.04336548, 0.04006958, -0.03878784, 0.008666992, 0.008529663, -0.027267456, 0.022125244, -0.026550293, 0.00018596649, ...
http://arxiv.org/abs/2306.10555v1
Summarization from Leaderboards to Practice: Choosing A Representation Backbone and Ensuring Robustness
David Demeter, Oshin Agarwal, Simon Ben Igeri, Marko Sterbentz, Neil Molino, John M. Conroy, Ani Nenkova
cs.CL
Academic literature does not give much guidance on how to build the best possible customer-facing summarization system from existing research components. Here we present analyses to inform the selection of a system backbone from popular models; we find that in both automatic and human evaluation, BART performs better t...
http://arxiv.org/pdf/2306.10555v1
2023-06-18T13:35:41
2023-06-18T13:35:41
{ "title": [ 0.031982422, 0.0070991516, -0.018859863, -0.088134766, -0.011001587, -0.0030384064, -0.06817627, -0.014656067, -0.014389038, 0.05328369, -0.021026611, 0.049591064, -0.0041389465, -0.002998352, -0.034332275, 0.008308411, 0.0010852814, ...
http://arxiv.org/abs/2306.14911v1
"You might think about slightly revising the title": identifying hedges in peer-tutoring interactions
Yann Raphalen, Chloé Clavel, Justine Cassell
cs.CL, cs.AI
Hedges play an important role in the management of conversational interaction. In peer tutoring, they are notably used by tutors in dyads (pairs of interlocutors) experiencing low rapport to tone down the impact of instructions and negative feedback. Pursuing the objective of building a tutoring agent that manages rapp...
http://arxiv.org/pdf/2306.14911v1
2023-06-18T12:47:54
2023-06-18T12:47:54
{ "title": [ 0.015304565, 0.016357422, -0.012931824, 0.01689148, -0.0037174225, -0.030975342, 0.02633667, -0.057434082, -0.052337646, 0.034698486, 0.010604858, 0.028427124, -0.017440796, -0.028961182, -0.025894165, -0.016784668, -0.021347046, 0.0...
http://arxiv.org/abs/2306.10543v1
UniMC: A Unified Framework for Long-Term Memory Conversation via Relevance Representation Learning
Kang Zhao, Wei Liu, Jian Luan, Minglei Gao, Li Qian, Hanlin Teng, Bin Wang
cs.CL
Open-domain long-term memory conversation can establish long-term intimacy with humans, and the key is the ability to understand and memorize long-term dialogue history information. Existing works integrate multiple models for modelling through a pipeline, which ignores the coupling between different stages. In this pa...
http://arxiv.org/pdf/2306.10543v1
2023-06-18T12:30:50
2023-06-18T12:30:50
{ "title": [ 0.008621216, 0.0077667236, -0.0362854, -0.07678223, 0.013694763, -0.00042009354, -0.05166626, 0.0062408447, 0.008575439, 0.037261963, -0.003921509, 0.0073165894, -0.0004310608, -0.017211914, -0.046295166, 0.0010757446, -0.0085372925, ...
http://arxiv.org/abs/2306.17177v1
Leveraging ChatGPT As Text Annotation Tool For Sentiment Analysis
Mohammad Belal, James She, Simon Wong
cs.CL
Sentiment analysis is a well-known natural language processing task that involves identifying the emotional tone or polarity of a given piece of text. With the growth of social media and other online platforms, sentiment analysis has become increasingly crucial for businesses and organizations seeking to monitor and co...
http://arxiv.org/pdf/2306.17177v1
2023-06-18T12:20:42
2023-06-18T12:20:42
{ "title": [ -0.0011644363, -0.011238098, -0.020263672, -0.03164673, 0.0016613007, 0.0035743713, -0.010276794, -0.011131287, 0.0066223145, 0.04525757, -0.021865845, 0.023040771, -0.013343811, 0.007247925, -0.06732178, 0.0077705383, -0.02267456, 0...
http://arxiv.org/abs/2306.14910v1
The Importance of Human-Labeled Data in the Era of LLMs
Yang Liu
cs.CL, cs.AI, cs.CY, cs.LG
The advent of large language models (LLMs) has brought about a revolution in the development of tailored machine learning models and sparked debates on redefining data requirements. The automation facilitated by the training and implementation of LLMs has led to discussions and aspirations that human-level labeling int...
http://arxiv.org/pdf/2306.14910v1
2023-06-18T12:12:03
2023-06-18T12:12:03
{ "title": [ 0.025344849, -0.012763977, -0.03451538, -0.05001831, -0.047546387, -0.033996582, -0.036193848, 0.022201538, -0.040985107, 0.015853882, -0.047027588, 0.005264282, -0.0309906, -0.03161621, 0.023513794, -0.04147339, -0.003145218, 0.0330...
http://arxiv.org/abs/2306.10530v1
Gender Bias in Transformer Models: A comprehensive survey
Praneeth Nemani, Yericherla Deepak Joel, Palla Vijay, Farhana Ferdousi Liza
cs.CL, cs.AI
Gender bias in artificial intelligence (AI) has emerged as a pressing concern with profound implications for individuals' lives. This paper presents a comprehensive survey that explores gender bias in Transformer models from a linguistic perspective. While the existence of gender bias in language models has been acknow...
http://arxiv.org/pdf/2306.10530v1
2023-06-18T11:40:47
2023-06-18T11:40:47
{ "title": [ -0.020446777, 0.011108398, -0.04598999, -0.01461792, 0.006439209, -0.024719238, -0.010437012, -0.03164673, -0.022491455, 0.056396484, -0.0093688965, -0.010536194, 0.030441284, -0.020233154, -0.027954102, 0.008453369, -0.022979736, 0....
http://arxiv.org/abs/2306.10514v1
Evolutionary Verbalizer Search for Prompt-based Few Shot Text Classification
Tongtao Ling, Lei Chen, Yutao Lai, Hai-Lin Liu
cs.CL
Recent advances for few-shot text classification aim to wrap textual inputs with task-specific prompts to cloze questions. By processing them with a masked language model to predict the masked tokens and using a verbalizer that constructs the mapping between predicted words and target labels. This approach of using pre...
http://arxiv.org/pdf/2306.10514v1
2023-06-18T10:03:11
2023-06-18T10:03:11
{ "title": [ -0.0061454773, 0.0099487305, -0.012817383, -0.026794434, -0.0032634735, 0.00016701221, -0.0043945312, -0.0138168335, 0.012435913, 0.017166138, -0.041900635, 0.026290894, -0.029647827, -0.005607605, -0.057250977, -0.039764404, -0.00939178...
http://arxiv.org/abs/2306.10512v2
Efficiently Measuring the Cognitive Ability of LLMs: An Adaptive Testing Perspective
Yan Zhuang, Qi Liu, Yuting Ning, Weizhe Huang, Rui Lv, Zhenya Huang, Guanhao Zhao, Zheng Zhang, Qingyang Mao, Shijin Wang, Enhong Chen
cs.CL
Large language models (LLMs), like ChatGPT, have shown some human-like cognitive abilities. For comparing these abilities of different models, several benchmarks (i.e. sets of standard test questions) from different fields (e.g., Literature, Biology and Psychology) are often adopted and the test results under tradition...
http://arxiv.org/pdf/2306.10512v2
2023-10-28T13:02:24
2023-06-18T09:54:33
{ "title": [ -0.053100586, 0.0015678406, -0.05126953, -0.018341064, -0.011245728, -0.00051021576, -0.0418396, -0.010192871, 0.01209259, 0.005748749, -0.022598267, -0.01184082, -0.026977539, -0.027526855, -0.054351807, -0.018066406, -0.017608643, ...
http://arxiv.org/abs/2306.10493v1
MOSPC: MOS Prediction Based on Pairwise Comparison
Kexin Wang, Yunlong Zhao, Qianqian Dong, Tom Ko, Mingxuan Wang
cs.SD, cs.CL, eess.AS
As a subjective metric to evaluate the quality of synthesized speech, Mean opinion score~(MOS) usually requires multiple annotators to score the same speech. Such an annotation approach requires a lot of manpower and is also time-consuming. MOS prediction model for automatic evaluation can significantly reduce labor co...
http://arxiv.org/pdf/2306.10493v1
2023-06-18T07:38:17
2023-06-18T07:38:17
{ "title": [ -0.014427185, -0.010406494, -0.04434204, -0.06585693, 0.002632141, -0.0070762634, -0.047698975, -0.016448975, 0.012489319, 0.030349731, -0.0025959015, -0.016860962, 0.01285553, -0.053588867, -0.022338867, -0.005001068, -0.010597229, ...
http://arxiv.org/abs/2306.17176v1
News Verifiers Showdown: A Comparative Performance Evaluation of ChatGPT 3.5, ChatGPT 4.0, Bing AI, and Bard in News Fact-Checking
Kevin Matthe Caramancion
cs.CL, cs.AI, cs.HC
This study aimed to evaluate the proficiency of prominent Large Language Models (LLMs), namely OpenAI's ChatGPT 3.5 and 4.0, Google's Bard(LaMDA), and Microsoft's Bing AI in discerning the truthfulness of news items using black box testing. A total of 100 fact-checked news items, all sourced from independent fact-check...
http://arxiv.org/pdf/2306.17176v1
2023-06-18T04:30:29
2023-06-18T04:30:29
{ "title": [ 0.0054779053, -0.020248413, -0.0062332153, -0.008033752, -0.04586792, -0.014389038, -0.007698059, -0.014961243, -0.02192688, 0.058441162, -0.034729004, 0.018676758, 0.008232117, 0.00617218, -0.04147339, -0.025527954, -0.048034668, 0....
http://arxiv.org/abs/2306.10469v1
Transferring Neural Potentials For High Order Dependency Parsing
Farshad Noravesh
cs.CL
High order dependency parsing leverages high order features such as siblings or grandchildren to improve state of the art accuracy of current first order dependency parsers. The present paper uses biaffine scores to provide an estimate of the arc scores and is then propagated into a graphical model. The inference insid...
http://arxiv.org/pdf/2306.10469v1
2023-06-18T03:58:41
2023-06-18T03:58:41
{ "title": [ -0.0007715225, 0.009719849, -0.035308838, -0.08239746, 0.020736694, 0.007865906, -0.044830322, 0.011016846, -0.011741638, 0.026138306, -0.002872467, 0.0017490387, -0.002855301, -0.0234375, -0.023834229, -0.0016765594, -0.0149002075, ...
http://arxiv.org/abs/2306.10452v1
MISMATCH: Fine-grained Evaluation of Machine-generated Text with Mismatch Error Types
Keerthiram Murugesan, Sarathkrishna Swaminathan, Soham Dan, Subhajit Chaudhury, Chulaka Gunasekara, Maxwell Crouse, Diwakar Mahajan, Ibrahim Abdelaziz, Achille Fokoue, Pavan Kapanipathi, Salim Roukos, Alexander Gray
cs.CL
With the growing interest in large language models, the need for evaluating the quality of machine text compared to reference (typically human-generated) text has become focal attention. Most recent works focus either on task-specific evaluation metrics or study the properties of machine-generated text captured by the ...
http://arxiv.org/pdf/2306.10452v1
2023-06-18T01:38:53
2023-06-18T01:38:53
{ "title": [ -0.0309906, 0.0064849854, -0.036224365, -0.030334473, 0.011650085, -0.0072631836, -0.019927979, -0.013442993, 0.0039863586, 0.030426025, -0.024902344, 0.0079956055, 0.0037002563, -0.018875122, -0.048950195, -0.015144348, -0.024642944, ...
http://arxiv.org/abs/2306.10448v1
Generation of Radiology Findings in Chest X-Ray by Leveraging Collaborative Knowledge
Manuela Daniela Danu, George Marica, Sanjeev Kumar Karn, Bogdan Georgescu, Awais Mansoor, Florin Ghesu, Lucian Mihai Itu, Constantin Suciu, Sasa Grbic, Oladimeji Farri, Dorin Comaniciu
cs.CV, cs.CL
Among all the sub-sections in a typical radiology report, the Clinical Indications, Findings, and Impression often reflect important details about the health status of a patient. The information included in Impression is also often covered in Findings. While Findings and Impression can be deduced by inspecting the imag...
http://arxiv.org/pdf/2306.10448v1
2023-06-18T00:51:28
2023-06-18T00:51:28
{ "title": [ 0.0010595322, 0.0034809113, -0.064819336, -0.06335449, 0.006679535, -0.0335083, -0.029144287, 0.008712769, -0.0032844543, 0.021148682, -0.034698486, 0.007598877, -0.039398193, -0.01878357, -0.04949951, -0.020385742, 0.02331543, -0.00...
http://arxiv.org/abs/2306.10444v1
Universal Information Extraction with Meta-Pretrained Self-Retrieval
Xin Cong. Bowen Yu, Mengcheng Fang, Tingwen Liu, Haiyang Yu, Zhongkai Hu, Fei Huang, Yongbin Li, Bin Wang
cs.CL, cs.AI
Universal Information Extraction~(Universal IE) aims to solve different extraction tasks in a uniform text-to-structure generation manner. Such a generation procedure tends to struggle when there exist complex information structures to be extracted. Retrieving knowledge from external knowledge bases may help models to ...
http://arxiv.org/pdf/2306.10444v1
2023-06-18T00:16:00
2023-06-18T00:16:00
{ "title": [ -0.013793945, 0.0053520203, -0.04284668, -0.066711426, 0.0071144104, 0.0073890686, -0.045440674, -0.016494751, 0.0066566467, 0.050354004, 0.012031555, 0.029510498, -0.018661499, -0.010810852, -0.050445557, 0.028747559, 0.0006222725, ...
http://arxiv.org/abs/2306.17175v2
RECAP-KG: Mining Knowledge Graphs from Raw GP Notes for Remote COVID-19 Assessment in Primary Care
Rakhilya Lee Mekhtieva, Brandon Forbes, Dalal Alrajeh, Brendan Delaney, Alessandra Russo
cs.CL, cs.AI
Clinical decision-making is a fundamental stage in delivering appropriate care to patients. In recent years several decision-making systems designed to aid the clinician in this process have been developed. However, technical solutions currently in use are based on simple regression models and are only able to take int...
http://arxiv.org/pdf/2306.17175v2
2023-10-09T13:23:53
2023-06-17T23:35:51
{ "title": [ -0.031021118, -0.046173096, -0.029006958, -0.081848145, -0.037750244, 0.02331543, -0.07635498, 0.025299072, -0.0151901245, 0.018585205, -0.05050659, 0.011154175, -0.011192322, -0.017105103, -0.028457642, -0.039154053, 0.019226074, 0....
http://arxiv.org/abs/2306.10419v1
Multilingual Multiword Expression Identification Using Lateral Inhibition and Domain Adaptation
Andrei-Marius Avram, Verginica Barbu Mititelu, Vasile Păiş, Dumitru-Clementin Cercel, Ştefan Trăuşan-Matu
cs.CL
Correctly identifying multiword expressions (MWEs) is an important task for most natural language processing systems since their misidentification can result in ambiguity and misunderstanding of the underlying text. In this work, we evaluate the performance of the mBERT model for MWE identification in a multilingual co...
http://arxiv.org/pdf/2306.10419v1
2023-06-17T20:28:32
2023-06-17T20:28:32
{ "title": [ -0.020751953, 0.014427185, -0.03552246, -0.019714355, 0.0016546249, -0.005203247, -0.031799316, -0.010040283, 0.005958557, 0.04547119, 0.029251099, -0.0129470825, 0.020721436, -0.0309906, -0.060760498, 0.02116394, -0.013183594, 0.047...
http://arxiv.org/abs/2306.10414v1
KEST: Kernel Distance Based Efficient Self-Training for Improving Controllable Text Generation
Yuxi Feng, Xiaoyuan Yi, Laks V. S. Lakshmanan, Xing Xie
cs.CL
Self-training (ST) has come to fruition in language understanding tasks by producing pseudo labels, which reduces the labeling bottleneck of language model fine-tuning. Nevertheless, in facilitating semi-supervised controllable language generation, ST faces two key challenges. First, augmented by self-generated pseudo ...
http://arxiv.org/pdf/2306.10414v1
2023-06-17T19:40:57
2023-06-17T19:40:57
{ "title": [ -0.032165527, -0.0043907166, -0.0231781, -0.021331787, -0.018844604, 0.012451172, -0.054473877, -0.027496338, 0.008583069, 0.05117798, -0.008636475, 0.035064697, 0.015914917, -0.006225586, -0.03363037, 0.020477295, -0.013069153, 0.03...
http://arxiv.org/abs/2306.10354v1
LLMVA-GEBC: Large Language Model with Video Adapter for Generic Event Boundary Captioning
Yunlong Tang, Jinrui Zhang, Xiangchen Wang, Teng Wang, Feng Zheng
cs.CV, cs.CL
Our winning entry for the CVPR 2023 Generic Event Boundary Captioning (GEBC) competition is detailed in this paper. Unlike conventional video captioning tasks, GEBC demands that the captioning model possess an understanding of immediate changes in status around the designated video boundary, making it a difficult task....
http://arxiv.org/pdf/2306.10354v1
2023-06-17T13:55:54
2023-06-17T13:55:54
{ "title": [ 0.006942749, -0.015151978, -0.03286743, -0.07647705, -0.012207031, -0.0030918121, -0.058044434, 0.0029449463, 0.0012359619, 0.04949951, -0.0131073, 0.0030059814, 0.027862549, -0.020736694, -0.04360962, 0.016616821, -0.0077323914, 0.0...
http://arxiv.org/abs/2306.10348v1
Typo-Robust Representation Learning for Dense Retrieval
Panuthep Tasawong, Wuttikorn Ponwitayarat, Peerat Limkonchotiwat, Can Udomcharoenchaikit, Ekapol Chuangsuwanich, Sarana Nutanong
cs.IR, cs.CL, I.2.7
Dense retrieval is a basic building block of information retrieval applications. One of the main challenges of dense retrieval in real-world settings is the handling of queries containing misspelled words. A popular approach for handling misspelled queries is minimizing the representations discrepancy between misspelle...
http://arxiv.org/pdf/2306.10348v1
2023-06-17T13:48:30
2023-06-17T13:48:30
{ "title": [ 0.011161804, 0.012062073, -0.030380249, -0.05126953, 0.021591187, -0.0037879944, -0.033721924, -0.03225708, 0.057739258, 0.040527344, -0.023483276, 0.00003504753, 0.005302429, -0.00045895576, -0.06347656, 0.016845703, -0.0032691956, ...
http://arxiv.org/abs/2306.17174v1
Empowering NLG: Offline Reinforcement Learning for Informal Summarization in Online Domains
Zhi-Xuan Tai, Po-Chuan Chen
cs.CL, cs.AI
Our research introduces an innovative Natural Language Generation (NLG) approach that aims to optimize user experience and alleviate the workload of human customer support agents. Our primary objective is to generate informal summaries for online articles and posts using an offline reinforcement learning technique. In ...
http://arxiv.org/pdf/2306.17174v1
2023-06-17T13:00:54
2023-06-17T13:00:54
{ "title": [ 0.013061523, -0.017089844, -0.027114868, -0.06689453, -0.0071411133, -0.012634277, -0.030258179, 0.006538391, -0.022994995, 0.025039673, -0.018127441, 0.018356323, -0.005302429, -0.011253357, -0.013549805, -0.019607544, -0.0178833, 0...