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/2310.16803v1
Language Agnostic Code Embeddings
Saiteja Utpala, Alex Gu, Pin Yu Chen
cs.CL, cs.LG
Recently, code language models have achieved notable advancements in addressing a diverse array of essential code comprehension and generation tasks. Yet, the field lacks a comprehensive deep dive and understanding of the code embeddings of multilingual code models. In this paper, we present a comprehensive study on mu...
http://arxiv.org/pdf/2310.16803v1
2023-10-25T17:34:52
2023-10-25T17:34:52
{ "title": [ -0.008972168, -0.016021729, -0.04257202, -0.060699463, -0.008155823, -0.0064468384, -0.037719727, 0.023742676, 0.012428284, 0.034179688, -0.018936157, 0.002746582, -0.01512146, -0.026367188, -0.009254456, 0.019989014, 0.002380371, 0....
http://arxiv.org/abs/2310.16790v1
Improving a Named Entity Recognizer Trained on Noisy Data with a Few Clean Instances
Zhendong Chu, Ruiyi Zhang, Tong Yu, Rajiv Jain, Vlad I Morariu, Jiuxiang Gu, Ani Nenkova
cs.CL, cs.AI, cs.LG
To achieve state-of-the-art performance, one still needs to train NER models on large-scale, high-quality annotated data, an asset that is both costly and time-intensive to accumulate. In contrast, real-world applications often resort to massive low-quality labeled data through non-expert annotators via crowdsourcing a...
http://arxiv.org/pdf/2310.16790v1
2023-10-25T17:23:37
2023-10-25T17:23:37
{ "title": [ -0.0018405914, 0.014282227, -0.0513916, -0.058258057, 0.030578613, -0.01902771, -0.026687622, -0.014533997, 0.022140503, 0.008804321, -0.047180176, 0.018295288, -0.003162384, -0.001461029, -0.020355225, 0.0013456345, -0.0030574799, 0...
http://arxiv.org/abs/2310.16789v2
Detecting Pretraining Data from Large Language Models
Weijia Shi, Anirudh Ajith, Mengzhou Xia, Yangsibo Huang, Daogao Liu, Terra Blevins, Danqi Chen, Luke Zettlemoyer
cs.CL, cs.CR, cs.LG
Although large language models (LLMs) are widely deployed, the data used to train them is rarely disclosed. Given the incredible scale of this data, up to trillions of tokens, it is all but certain that it includes potentially problematic text such as copyrighted materials, personally identifiable information, and test...
http://arxiv.org/pdf/2310.16789v2
2023-11-03T05:27:37
2023-10-25T17:21:23
{ "title": [ 0.0049705505, 0.010620117, -0.03918457, -0.061431885, 0.014076233, -0.015029907, -0.041992188, -0.018829346, -0.021652222, 0.025131226, -0.014160156, 0.010749817, -0.006450653, -0.022857666, -0.0259552, -0.0046043396, -0.009986877, 0...
http://arxiv.org/abs/2310.16787v3
The Data Provenance Initiative: A Large Scale Audit of Dataset Licensing & Attribution in AI
Shayne Longpre, Robert Mahari, Anthony Chen, Naana Obeng-Marnu, Damien Sileo, William Brannon, Niklas Muennighoff, Nathan Khazam, Jad Kabbara, Kartik Perisetla, Xinyi Wu, Enrico Shippole, Kurt Bollacker, Tongshuang Wu, Luis Villa, Sandy Pentland, Sara Hooker
cs.CL, cs.AI, cs.LG
The race to train language models on vast, diverse, and inconsistently documented datasets has raised pressing concerns about the legal and ethical risks for practitioners. To remedy these practices threatening data transparency and understanding, we convene a multi-disciplinary effort between legal and machine learnin...
http://arxiv.org/pdf/2310.16787v3
2023-11-04T19:10:06
2023-10-25T17:20:26
{ "title": [ -0.031311035, -0.007019043, -0.06951904, -0.037322998, 0.00818634, -0.026779175, -0.007472992, 0.014152527, -0.008224487, 0.019119263, -0.05834961, 0.019485474, 0.015403748, -0.02571106, 0.004600525, -0.023406982, 0.0007543564, 0.042...
http://arxiv.org/abs/2310.16781v1
Kiki or Bouba? Sound Symbolism in Vision-and-Language Models
Morris Alper, Hadar Averbuch-Elor
cs.CV, cs.CL, cs.LG
Although the mapping between sound and meaning in human language is assumed to be largely arbitrary, research in cognitive science has shown that there are non-trivial correlations between particular sounds and meanings across languages and demographic groups, a phenomenon known as sound symbolism. Among the many dimen...
http://arxiv.org/pdf/2310.16781v1
2023-10-25T17:15:55
2023-10-25T17:15:55
{ "title": [ -0.04067993, -0.020431519, -0.0059432983, -0.006641388, -0.03942871, -0.0010213852, -0.0423584, -0.030944824, -0.0021781921, 0.04345703, -0.013824463, -0.03564453, 0.035614014, -0.031829834, -0.036987305, -0.004749298, -0.01008606, 0...
http://arxiv.org/abs/2310.16776v2
DEFT: Data Efficient Fine-Tuning for Large Language Models via Unsupervised Core-Set Selection
Devleena Das, Vivek Khetan
cs.CL, cs.AI
Recent advances have led to the availability of many pre-trained language models (PLMs); however, a question that remains is how much data is truly needed to fine-tune PLMs for downstream tasks? In this work, we introduce DEFT, a data-efficient fine-tuning framework that leverages unsupervised core-set selection to min...
http://arxiv.org/pdf/2310.16776v2
2023-10-26T03:13:26
2023-10-25T17:06:42
{ "title": [ -0.03765869, -0.021057129, -0.03994751, -0.072509766, -0.0052948, 0.008766174, -0.062438965, 0.018798828, 0.014793396, 0.023757935, -0.017364502, 0.015533447, 0.014305115, -0.015266418, -0.035003662, 0.0056381226, -0.016647339, 0.049...
http://arxiv.org/abs/2310.16763v1
SuperHF: Supervised Iterative Learning from Human Feedback
Gabriel Mukobi, Peter Chatain, Su Fong, Robert Windesheim, Gitta Kutyniok, Kush Bhatia, Silas Alberti
cs.CL, cs.AI, cs.LG
While large language models demonstrate remarkable capabilities, they often present challenges in terms of safety, alignment with human values, and stability during training. Here, we focus on two prevalent methods used to align these models, Supervised Fine-Tuning (SFT) and Reinforcement Learning from Human Feedback (...
http://arxiv.org/pdf/2310.16763v1
2023-10-25T16:52:00
2023-10-25T16:52:00
{ "title": [ -0.024475098, 0.008766174, -0.039093018, -0.05569458, 0.007709503, -0.010070801, -0.06561279, -0.03842163, 0.012390137, 0.02319336, -0.041381836, 0.012756348, -0.00548172, 0.0034389496, -0.036956787, -0.006374359, -0.00080251694, 0.0...
http://arxiv.org/abs/2310.16761v1
IntenDD: A Unified Contrastive Learning Approach for Intent Detection and Discovery
Bhavuk Singhal, Ashim Gupta, Shivasankaran V P, Amrith Krishna
cs.CL
Identifying intents from dialogue utterances forms an integral component of task-oriented dialogue systems. Intent-related tasks are typically formulated either as a classification task, where the utterances are classified into predefined categories or as a clustering task when new and previously unknown intent categor...
http://arxiv.org/pdf/2310.16761v1
2023-10-25T16:50:24
2023-10-25T16:50:24
{ "title": [ 0.004814148, 0.011688232, -0.03189087, -0.035308838, 0.03668213, -0.023727417, 0.0044784546, -0.03640747, 0.0209198, 0.011543274, -0.031921387, 0.010139465, -0.012313843, -0.03149414, -0.02696228, -0.0024604797, -0.025009155, 0.06094...
http://arxiv.org/abs/2310.16755v1
HI-TOM: A Benchmark for Evaluating Higher-Order Theory of Mind Reasoning in Large Language Models
Yinghui He, Yufan Wu, Yilin Jia, Rada Mihalcea, Yulong Chen, Naihao Deng
cs.CL, cs.AI
Theory of Mind (ToM) is the ability to reason about one's own and others' mental states. ToM plays a critical role in the development of intelligence, language understanding, and cognitive processes. While previous work has primarily focused on first and second-order ToM, we explore higher-order ToM, which involves rec...
http://arxiv.org/pdf/2310.16755v1
2023-10-25T16:41:15
2023-10-25T16:41:15
{ "title": [ -0.008392334, -0.026657104, -0.002243042, -0.07104492, -0.0078125, 0.011009216, -0.041137695, 0.0014266968, 0.014953613, 0.024551392, -0.007369995, 0.0060768127, 0.02520752, -0.02619934, -0.038269043, 0.011230469, -0.024871826, 0.015...
http://arxiv.org/abs/2310.16753v1
PROMINET: Prototype-based Multi-View Network for Interpretable Email Response Prediction
Yuqing Wang, Prashanth Vijayaraghavan, Ehsan Degan
cs.CL, cs.AI, cs.LG
Email is a widely used tool for business communication, and email marketing has emerged as a cost-effective strategy for enterprises. While previous studies have examined factors affecting email marketing performance, limited research has focused on understanding email response behavior by considering email content and...
http://arxiv.org/pdf/2310.16753v1
2023-10-25T16:39:00
2023-10-25T16:39:00
{ "title": [ -0.017318726, -0.029037476, -0.03475952, -0.08648682, -0.031402588, 0.016174316, -0.0892334, 0.02458191, -0.00015950203, 0.03543091, -0.0007901192, -0.009887695, -0.0043296814, -0.013061523, -0.066101074, -0.015571594, -0.010299683, ...
http://arxiv.org/abs/2310.16749v1
DISCO: A Large Scale Human Annotated Corpus for Disfluency Correction in Indo-European Languages
Vineet Bhat, Preethi Jyothi, Pushpak Bhattacharyya
cs.CL, cs.HC
Disfluency correction (DC) is the process of removing disfluent elements like fillers, repetitions and corrections from spoken utterances to create readable and interpretable text. DC is a vital post-processing step applied to Automatic Speech Recognition (ASR) outputs, before subsequent processing by downstream langua...
http://arxiv.org/pdf/2310.16749v1
2023-10-25T16:32:02
2023-10-25T16:32:02
{ "title": [ 0.010810852, 0.012519836, -0.045135498, -0.055541992, 0.02482605, -0.0073890686, -0.023086548, 0.0043296814, 0.046051025, 0.045440674, -0.03488159, -0.0184021, 0.08081055, -0.023208618, -0.056732178, 0.020584106, -0.019638062, 0.0529...
http://arxiv.org/abs/2310.16746v1
HANSEN: Human and AI Spoken Text Benchmark for Authorship Analysis
Nafis Irtiza Tripto, Adaku Uchendu, Thai Le, Mattia Setzu, Fosca Giannotti, Dongwon Lee
cs.CL
Authorship Analysis, also known as stylometry, has been an essential aspect of Natural Language Processing (NLP) for a long time. Likewise, the recent advancement of Large Language Models (LLMs) has made authorship analysis increasingly crucial for distinguishing between human-written and AI-generated texts. However, t...
http://arxiv.org/pdf/2310.16746v1
2023-10-25T16:23:17
2023-10-25T16:23:17
{ "title": [ -0.059051514, -0.009895325, -0.05734253, -0.04940796, -0.011421204, -0.0031909943, -0.054901123, 0.012435913, -0.005203247, 0.035461426, -0.01259613, 0.02861023, 0.023452759, -0.0037956238, -0.035125732, 0.014915466, -0.023361206, 0....
http://arxiv.org/abs/2310.16738v1
Improving Conversational Recommendation Systems via Bias Analysis and Language-Model-Enhanced Data Augmentation
Xi Wang, Hossein A. Rahmani, Jiqun Liu, Emine Yilmaz
cs.CL, cs.IR
Conversational Recommendation System (CRS) is a rapidly growing research area that has gained significant attention alongside advancements in language modelling techniques. However, the current state of conversational recommendation faces numerous challenges due to its relative novelty and limited existing contribution...
http://arxiv.org/pdf/2310.16738v1
2023-10-25T16:11:55
2023-10-25T16:11:55
{ "title": [ 0.018722534, 0.0017929077, -0.042816162, -0.04824829, 0.0021381378, -0.013366699, -0.036956787, -0.0032634735, 0.019470215, 0.032196045, -0.034210205, 0.019210815, 0.0061302185, -0.0057144165, -0.03955078, 0.00025582314, -0.021316528, ...
http://arxiv.org/abs/2310.16731v1
Disentangling Extraction and Reasoning in Multi-hop Spatial Reasoning
Roshanak Mirzaee, Parisa Kordjamshidi
cs.CL
Spatial reasoning over text is challenging as the models not only need to extract the direct spatial information from the text but also reason over those and infer implicit spatial relations. Recent studies highlight the struggles even large language models encounter when it comes to performing spatial reasoning over t...
http://arxiv.org/pdf/2310.16731v1
2023-10-25T16:00:47
2023-10-25T16:00:47
{ "title": [ -0.009773254, -0.025619507, -0.03829956, -0.0256958, -0.008964539, -0.016906738, 0.0041770935, -0.00043916702, 0.021728516, 0.023422241, -0.03640747, -0.014381409, 0.0018024445, -0.020706177, -0.033355713, -0.009346008, 0.0068473816, ...
http://arxiv.org/abs/2310.16713v2
SkyMath: Technical Report
Liu Yang, Haihua Yang, Wenjun Cheng, Lei Lin, Chenxia Li, Yifu Chen, Lunan Liu, Jianfei Pan, Tianwen Wei, Biye Li, Liang Zhao, Lijie Wang, Bo Zhu, Guoliang Li, Xuejie Wu, Xilin Luo, Rui Hu
cs.CL, cs.AI
Large language models (LLMs) have shown great potential to solve varieties of natural language processing (NLP) tasks, including mathematical reasoning. In this work, we present SkyMath, a large language model for mathematics with 13 billion parameters. By applying self-compare fine-tuning, we have enhanced mathematica...
http://arxiv.org/pdf/2310.16713v2
2023-10-26T09:01:16
2023-10-25T15:34:55
{ "title": [ 0.011383057, 0.0027923584, -0.021438599, -0.04156494, -0.020095825, -0.00642395, -0.040008545, -0.013885498, -0.052215576, 0.05117798, -0.015686035, -0.014434814, 0.07055664, -0.028640747, 0.042785645, -0.02331543, 0.017837524, 0.003...
http://arxiv.org/abs/2310.16712v1
LLM Performance Predictors are good initializers for Architecture Search
Ganesh Jawahar, Muhammad Abdul-Mageed, Laks V. S. Lakshmanan, Dujian Ding
cs.CL
Large language models (LLMs) have become an integral component in solving a wide range of NLP tasks. In this work, we explore a novel use case of using LLMs to build performance predictors (PP): models that, given a specific deep neural network architecture, predict its performance on a downstream task. We design PP pr...
http://arxiv.org/pdf/2310.16712v1
2023-10-25T15:34:30
2023-10-25T15:34:30
{ "title": [ -0.03289795, -0.0013856888, -0.009887695, -0.009559631, -0.04925537, 0.004043579, -0.04925537, -0.024749756, -0.037200928, 0.01739502, -0.015792847, 0.011360168, 0.012428284, -0.046020508, 0.05819702, -0.02017212, -0.00014603138, 0.0...
http://arxiv.org/abs/2310.16685v1
Detection of news written by the ChatGPT through authorship attribution performed by a Bidirectional LSTM model
Amanda Ferrari Iaquinta, Gustavo Voltani von Atzingen
cs.CL, cs.AI
The large language based-model chatbot ChatGPT gained a lot of popularity since its launch and has been used in a wide range of situations. This research centers around a particular situation, when the ChatGPT is used to produce news that will be consumed by the population, causing the facilitation in the production of...
http://arxiv.org/pdf/2310.16685v1
2023-10-25T14:48:58
2023-10-25T14:48:58
{ "title": [ -0.00283432, 0.0069618225, -0.014053345, -0.0036029816, -0.018936157, 0.0036201477, -0.021224976, -0.048187256, -0.051879883, 0.03152466, -0.0075798035, 0.028793335, -0.017807007, -0.012802124, -0.006790161, -0.013214111, -0.047424316, ...
http://arxiv.org/abs/2310.16681v1
BabyStories: Can Reinforcement Learning Teach Baby Language Models to Write Better Stories?
Xingmeng Zhao, Tongnian Wang, Sheri Osborn, Anthony Rios
cs.CL
Language models have seen significant growth in the size of their corpus, leading to notable performance improvements. Yet, there has been limited progress in developing models that handle smaller, more human-like datasets. As part of the BabyLM shared task, this study explores the impact of reinforcement learning from...
http://arxiv.org/pdf/2310.16681v1
2023-10-25T14:45:48
2023-10-25T14:45:48
{ "title": [ -0.046325684, -0.0038757324, -0.035461426, 0.012763977, -0.035064697, -0.013412476, -0.009284973, -0.012916565, -0.051605225, 0.02267456, -0.03277588, 0.019332886, -0.021209717, -0.012435913, 0.0008883476, -0.05331421, -0.014595032, ...
http://arxiv.org/abs/2310.16676v1
SSLCL: An Efficient Model-Agnostic Supervised Contrastive Learning Framework for Emotion Recognition in Conversations
Tao Shi, Xiao Liang, Yaoyuan Liang, Xinyi Tong, Shao-Lun Huang
cs.CL
Emotion recognition in conversations (ERC) is a rapidly evolving task within the natural language processing community, which aims to detect the emotions expressed by speakers during a conversation. Recently, a growing number of ERC methods have focused on leveraging supervised contrastive learning (SCL) to enhance the...
http://arxiv.org/pdf/2310.16676v1
2023-10-25T14:41:14
2023-10-25T14:41:14
{ "title": [ 0.00017082691, 0.009757996, -0.026657104, -0.05496216, 0.0009598732, -0.0034370422, -0.04824829, -0.015304565, 0.044036865, 0.022323608, -0.03213501, -0.011665344, 0.011680603, -0.02116394, -0.05807495, -0.0064468384, -0.017807007, 0...
http://arxiv.org/abs/2310.16654v1
ChatGPT is a Potential Zero-Shot Dependency Parser
Boda Lin, Xinyi Zhou, Binghao Tang, Xiaocheng Gong, Si Li
cs.CL
Pre-trained language models have been widely used in dependency parsing task and have achieved significant improvements in parser performance. However, it remains an understudied question whether pre-trained language models can spontaneously exhibit the ability of dependency parsing without introducing additional parse...
http://arxiv.org/pdf/2310.16654v1
2023-10-25T14:08:39
2023-10-25T14:08:39
{ "title": [ 0.020523071, -0.020187378, -0.030548096, -0.06781006, -0.04586792, 0.017150879, -0.06976318, 0.02418518, -0.00982666, 0.01965332, -0.000579834, 0.005748749, -0.015556335, -0.039001465, 0.012741089, -0.002313614, -0.0056533813, 0.0411...
http://arxiv.org/abs/2310.16621v1
ArTST: Arabic Text and Speech Transformer
Hawau Olamide Toyin, Amirbek Djanibekov, Ajinkya Kulkarni, Hanan Aldarmaki
cs.CL, cs.AI, cs.SD, eess.AS
We present ArTST, a pre-trained Arabic text and speech transformer for supporting open-source speech technologies for the Arabic language. The model architecture follows the unified-modal framework, SpeechT5, that was recently released for English, and is focused on Modern Standard Arabic (MSA), with plans to extend th...
http://arxiv.org/pdf/2310.16621v1
2023-10-25T13:20:54
2023-10-25T13:20:54
{ "title": [ 0.012321472, -0.013504028, -0.010398865, -0.04522705, -0.028137207, 0.003856659, -0.03543091, -0.0029087067, -0.02748108, 0.04626465, 0.0158844, -0.0037574768, 0.042144775, -0.06109619, 0.0041618347, 0.00044059753, 0.03387451, -0.002...
http://arxiv.org/abs/2310.16616v1
Context Does Matter: End-to-end Panoptic Narrative Grounding with Deformable Attention Refined Matching Network
Yiming Lin, Xiao-Bo Jin, Qiufeng Wang, Kaizhu Huang
cs.CV, cs.CL
Panoramic Narrative Grounding (PNG) is an emerging visual grounding task that aims to segment visual objects in images based on dense narrative captions. The current state-of-the-art methods first refine the representation of phrase by aggregating the most similar $k$ image pixels, and then match the refined text repre...
http://arxiv.org/pdf/2310.16616v1
2023-10-25T13:12:39
2023-10-25T13:12:39
{ "title": [ -0.013664246, 0.013633728, -0.041870117, -0.05834961, 0.0090789795, -0.0055618286, -0.02482605, 0.007461548, 0.0010509491, 0.026657104, 0.0010232925, -0.0082473755, 0.028640747, -0.030807495, -0.025253296, 0.008117676, -0.0075531006, ...
http://arxiv.org/abs/2310.16609v1
Back Transcription as a Method for Evaluating Robustness of Natural Language Understanding Models to Speech Recognition Errors
Marek Kubis, Paweł Skórzewski, Marcin Sowański, Tomasz Ziętkiewicz
cs.CL, cs.AI, cs.SD, eess.AS
In a spoken dialogue system, an NLU model is preceded by a speech recognition system that can deteriorate the performance of natural language understanding. This paper proposes a method for investigating the impact of speech recognition errors on the performance of natural language understanding models. The proposed me...
http://arxiv.org/pdf/2310.16609v1
2023-10-25T13:07:07
2023-10-25T13:07:07
{ "title": [ -0.006855011, -0.0031642914, -0.01889038, -0.06726074, -0.029281616, -0.00033688545, -0.09820557, -0.015533447, 0.012580872, 0.019622803, -0.046722412, -0.0068359375, 0.00031375885, -0.010688782, -0.060638428, -0.028900146, -0.017425537,...
http://arxiv.org/abs/2310.16607v2
On the Interplay between Fairness and Explainability
Stephanie Brandl, Emanuele Bugliarello, Ilias Chalkidis
cs.CL
In order to build reliable and trustworthy NLP applications, models need to be both fair across different demographics and explainable. Usually these two objectives, fairness and explainability, are optimized and/or examined independently of each other. Instead, we argue that forthcoming, trustworthy NLP systems should...
http://arxiv.org/pdf/2310.16607v2
2023-11-13T15:20:43
2023-10-25T12:59:51
{ "title": [ 0.010940552, 0.005783081, -0.017837524, -0.068847656, 0.008895874, -0.000323534, -0.023880005, -0.024780273, -0.030578613, 0.0057144165, -0.037628174, -0.0013818741, -0.021026611, -0.023971558, -0.016830444, -0.020996094, -0.015541077, ...
http://arxiv.org/abs/2310.16582v1
Tailoring Personality Traits in Large Language Models via Unsupervisedly-Built Personalized Lexicons
Tianlong Li, Xiaoqing Zheng, Xuanjing Huang
cs.CL
Personality plays a pivotal role in shaping human expression patterns, and empowering and manipulating large language models (LLMs) with personality traits holds significant promise in enhancing the user experience of LLMs. However, prior approaches either rely on fine-tuning LLMs on a corpus enriched with personalized...
http://arxiv.org/pdf/2310.16582v1
2023-10-25T12:16:33
2023-10-25T12:16:33
{ "title": [ -0.03930664, -0.008781433, -0.032409668, -0.056030273, -0.005405426, 0.01676941, -0.04928589, 0.013008118, -0.0071907043, 0.032958984, -0.009628296, 0.015792847, 0.0115737915, -0.011756897, -0.027618408, 0.017486572, -0.022567749, 0....
http://arxiv.org/abs/2310.16579v1
WSDMS: Debunk Fake News via Weakly Supervised Detection of Misinforming Sentences with Contextualized Social Wisdom
Ruichao Yang, Wei Gao, Jing Ma, Hongzhan Lin, Zhiwei Yang
cs.CL
In recent years, we witness the explosion of false and unconfirmed information (i.e., rumors) that went viral on social media and shocked the public. Rumors can trigger versatile, mostly controversial stance expressions among social media users. Rumor verification and stance detection are different yet relevant tasks. ...
http://arxiv.org/pdf/2310.16579v1
2023-10-25T12:06:55
2023-10-25T12:06:55
{ "title": [ 0.01525116, 0.028244019, -0.03451538, -0.047180176, 0.016021729, -0.01309967, -0.020126343, -0.04525757, -0.029205322, 0.014785767, -0.022567749, 0.021240234, 0.012039185, -0.019363403, -0.0012769699, -0.020721436, -0.022201538, 0.07...
http://arxiv.org/abs/2310.16570v1
Give Me the Facts! A Survey on Factual Knowledge Probing in Pre-trained Language Models
Paul Youssef, Osman Alperen Koraş, Meijie Li, Jörg Schlötterer, Christin Seifert
cs.CL
Pre-trained Language Models (PLMs) are trained on vast unlabeled data, rich in world knowledge. This fact has sparked the interest of the community in quantifying the amount of factual knowledge present in PLMs, as this explains their performance on downstream tasks, and potentially justifies their use as knowledge bas...
http://arxiv.org/pdf/2310.16570v1
2023-10-25T11:57:13
2023-10-25T11:57:13
{ "title": [ -0.018844604, -0.020843506, -0.024856567, -0.07116699, -0.017486572, -0.018234253, -0.05038452, 0.0016937256, -0.009208679, 0.031158447, -0.01209259, 0.004245758, 0.016433716, -0.025177002, -0.03970337, 0.0036239624, -0.013847351, 0....
http://arxiv.org/abs/2310.16568v1
1-PAGER: One Pass Answer Generation and Evidence Retrieval
Palak Jain, Livio Baldini Soares, Tom Kwiatkowski
cs.CL
We present 1-Pager the first system that answers a question and retrieves evidence using a single Transformer-based model and decoding process. 1-Pager incrementally partitions the retrieval corpus using constrained decoding to select a document and answer string, and we show that this is competitive with comparable re...
http://arxiv.org/pdf/2310.16568v1
2023-10-25T11:51:22
2023-10-25T11:51:22
{ "title": [ -0.013694763, -0.020584106, -0.031555176, -0.059020996, -0.007259369, 0.010643005, -0.023254395, 0.0023651123, 0.010246277, 0.014785767, -0.014541626, 0.024246216, -0.026672363, -0.025131226, -0.026443481, 0.012260437, 0.0018529892, ...
http://arxiv.org/abs/2310.18366v1
A Multilingual Virtual Guide for Self-Attachment Technique
Alicia Jiayun Law, Ruoyu Hu, Lisa Alazraki, Anandha Gopalan, Neophytos Polydorou, Abbas Edalat
cs.CL, cs.AI, cs.LG
In this work, we propose a computational framework that leverages existing out-of-language data to create a conversational agent for the delivery of Self-Attachment Technique (SAT) in Mandarin. Our framework does not require large-scale human translations, yet it achieves a comparable performance whilst also maintainin...
http://arxiv.org/pdf/2310.18366v1
2023-10-25T10:50:18
2023-10-25T10:50:18
{ "title": [ 0.022003174, -0.008575439, -0.024765015, -0.046051025, -0.03314209, -0.009254456, -0.03062439, 0.020690918, -0.05606079, 0.056030273, 0.053100586, -0.020401001, 0.021560669, -0.07531738, 0.008735657, 0.0096206665, 0.038726807, 0.0272...
http://arxiv.org/abs/2310.16538v1
FedTherapist: Mental Health Monitoring with User-Generated Linguistic Expressions on Smartphones via Federated Learning
Jaemin Shin, Hyungjun Yoon, Seungjoo Lee, Sungjoon Park, Yunxin Liu, Jinho D. Choi, Sung-Ju Lee
cs.CL, cs.AI, cs.LG
Psychiatrists diagnose mental disorders via the linguistic use of patients. Still, due to data privacy, existing passive mental health monitoring systems use alternative features such as activity, app usage, and location via mobile devices. We propose FedTherapist, a mobile mental health monitoring system that utilizes...
http://arxiv.org/pdf/2310.16538v1
2023-10-25T10:35:09
2023-10-25T10:35:09
{ "title": [ 0.018798828, 0.0121154785, -0.03378296, -0.055725098, 0.016174316, -0.0076065063, -0.03024292, -0.0055503845, -0.01576233, 0.03741455, -0.018356323, 0.008987427, 0.017944336, -0.024368286, -0.020874023, -0.0068511963, -0.0017900467, ...
http://arxiv.org/abs/2310.16535v1
R$^3$ Prompting: Review, Rephrase and Resolve for Chain-of-Thought Reasoning in Large Language Models under Noisy Context
Qingyuan Tian, Hanlun Zhu, Lei Wang, Yang Li, Yunshi Lan
cs.CL, cs.AI
With the help of Chain-of-Thought (CoT) prompting, Large Language Models (LLMs) have achieved remarkable performance on various reasoning tasks. However, most of them have been evaluated under noise-free context and the dilemma for LLMs to produce inaccurate results under the noisy context has not been fully investigat...
http://arxiv.org/pdf/2310.16535v1
2023-10-25T10:34:02
2023-10-25T10:34:02
{ "title": [ -0.01309967, -0.03225708, -0.004753113, -0.07635498, -0.0262146, 0.011497498, -0.07147217, 0.010871887, 0.0059547424, 0.0067214966, -0.07446289, 0.009292603, 0.015960693, -0.015205383, -0.011253357, -0.022125244, 0.00017726421, -0.00...
http://arxiv.org/abs/2310.16534v1
An Early Evaluation of GPT-4V(ision)
Yang Wu, Shilong Wang, Hao Yang, Tian Zheng, Hongbo Zhang, Yanyan Zhao, Bing Qin
cs.CL, cs.CV
In this paper, we evaluate different abilities of GPT-4V including visual understanding, language understanding, visual puzzle solving, and understanding of other modalities such as depth, thermal, video, and audio. To estimate GPT-4V's performance, we manually construct 656 test instances and carefully evaluate the re...
http://arxiv.org/pdf/2310.16534v1
2023-10-25T10:33:17
2023-10-25T10:33:17
{ "title": [ 0.036743164, -0.003917694, 0.011207581, -0.003364563, -0.028915405, -0.013427734, -0.01889038, -0.044952393, -0.032073975, 0.028381348, -0.0036716461, -0.018051147, -0.014297485, -0.041015625, -0.0033683777, -0.01802063, -0.013954163, ...
http://arxiv.org/abs/2310.16528v1
CUNI Submission to MRL 2023 Shared Task on Multi-lingual Multi-task Information Retrieval
Jindřich Helcl, Jindřich Libovický
cs.CL
We present the Charles University system for the MRL~2023 Shared Task on Multi-lingual Multi-task Information Retrieval. The goal of the shared task was to develop systems for named entity recognition and question answering in several under-represented languages. Our solutions to both subtasks rely on the translate-tes...
http://arxiv.org/pdf/2310.16528v1
2023-10-25T10:22:49
2023-10-25T10:22:49
{ "title": [ -0.024276733, -0.01612854, -0.044921875, -0.030334473, -0.0033302307, -0.011695862, -0.03591919, -0.010574341, -0.010482788, 0.06109619, 0.0075798035, -0.00623703, 0.04058838, -0.011634827, -0.024505615, 0.025650024, -0.0019721985, 0...
http://arxiv.org/abs/2310.16523v1
Improving Diversity of Demographic Representation in Large Language Models via Collective-Critiques and Self-Voting
Preethi Lahoti, Nicholas Blumm, Xiao Ma, Raghavendra Kotikalapudi, Sahitya Potluri, Qijun Tan, Hansa Srinivasan, Ben Packer, Ahmad Beirami, Alex Beutel, Jilin Chen
cs.CL, cs.AI
A crucial challenge for generative large language models (LLMs) is diversity: when a user's prompt is under-specified, models may follow implicit assumptions while generating a response, which may result in homogenization of the responses, as well as certain demographic groups being under-represented or even erased fro...
http://arxiv.org/pdf/2310.16523v1
2023-10-25T10:17:17
2023-10-25T10:17:17
{ "title": [ -0.01423645, 0.007820129, -0.041381836, -0.045806885, 0.014640808, -0.013198853, -0.03048706, 0.00057554245, 0.006664276, 0.038513184, -0.018676758, 0.0022277832, 0.007648468, -0.007167816, -0.04714966, 0.0070343018, -0.021057129, 0....
http://arxiv.org/abs/2310.16517v1
OccuQuest: Mitigating Occupational Bias for Inclusive Large Language Models
Mingfeng Xue, Dayiheng Liu, Kexin Yang, Guanting Dong, Wenqiang Lei, Zheng Yuan, Chang Zhou, Jingren Zhou
cs.CL
The emergence of large language models (LLMs) has revolutionized natural language processing tasks. However, existing instruction-tuning datasets suffer from occupational bias: the majority of data relates to only a few occupations, which hampers the instruction-tuned LLMs to generate helpful responses to professional ...
http://arxiv.org/pdf/2310.16517v1
2023-10-25T10:06:17
2023-10-25T10:06:17
{ "title": [ -0.033447266, -0.007549286, -0.06124878, -0.07104492, -0.004989624, -0.013771057, -0.072631836, 0.007507324, -0.021896362, 0.026885986, -0.008323669, 0.011505127, 0.022705078, -0.014984131, -0.015991211, 0.0072669983, -0.028762817, 0...
http://arxiv.org/abs/2310.16484v1
Subspace Chronicles: How Linguistic Information Emerges, Shifts and Interacts during Language Model Training
Max Müller-Eberstein, Rob van der Goot, Barbara Plank, Ivan Titov
cs.CL
Representational spaces learned via language modeling are fundamental to Natural Language Processing (NLP), however there has been limited understanding regarding how and when during training various types of linguistic information emerge and interact. Leveraging a novel information theoretic probing suite, which enabl...
http://arxiv.org/pdf/2310.16484v1
2023-10-25T09:09:55
2023-10-25T09:09:55
{ "title": [ -0.03604126, -0.00806427, -0.036621094, -0.014678955, -0.026519775, -0.0104522705, -0.03086853, -0.012023926, -0.044311523, 0.019180298, -0.032104492, 0.018417358, 0.01146698, -0.035614014, 0.0022830963, -0.017349243, -0.015098572, 0...
http://arxiv.org/abs/2310.16450v1
CLEX: Continuous Length Extrapolation for Large Language Models
Guanzheng Chen, Xin Li, Zaiqiao Meng, Shangsong Liang, Lidong Bing
cs.CL
Transformer-based Large Language Models (LLMs) are pioneering advances in many natural language processing tasks, however, their exceptional capabilities are restricted within the preset context window of Transformer. Position Embedding (PE) scaling methods, while effective in extending the context window to a specific...
http://arxiv.org/pdf/2310.16450v1
2023-10-25T08:13:02
2023-10-25T08:13:02
{ "title": [ -0.049346924, -0.026107788, -0.0463562, -0.06335449, -0.012580872, 0.00554657, -0.06555176, 0.012702942, -0.011512756, 0.022842407, -0.022506714, 0.0075263977, 0.0013027191, -0.015090942, -0.033416748, 0.00039243698, -0.01864624, 0.0...
http://arxiv.org/abs/2310.16446v1
Diversity Enhanced Narrative Question Generation for Storybooks
Hokeun Yoon, JinYeong Bak
cs.CL, cs.AI
Question generation (QG) from a given context can enhance comprehension, engagement, assessment, and overall efficacy in learning or conversational environments. Despite recent advancements in QG, the challenge of enhancing or measuring the diversity of generated questions often remains unaddressed. In this paper, we i...
http://arxiv.org/pdf/2310.16446v1
2023-10-25T08:10:04
2023-10-25T08:10:04
{ "title": [ -0.03378296, -0.008956909, -0.050933838, -0.042419434, -0.010444641, -0.010658264, -0.011047363, 0.026626587, -0.011756897, 0.006790161, -0.017333984, 0.021560669, -0.03729248, -0.02319336, -0.026351929, -0.0141067505, -0.032806396, ...
http://arxiv.org/abs/2310.16436v2
DDCoT: Duty-Distinct Chain-of-Thought Prompting for Multimodal Reasoning in Language Models
Ge Zheng, Bin Yang, Jiajin Tang, Hong-Yu Zhou, Sibei Yang
cs.CV, cs.CL
A long-standing goal of AI systems is to perform complex multimodal reasoning like humans. Recently, large language models (LLMs) have made remarkable strides in such multi-step reasoning on the language modality solely by leveraging the chain of thought (CoT) to mimic human thinking. However, the transfer of these adv...
http://arxiv.org/pdf/2310.16436v2
2023-10-26T04:16:52
2023-10-25T08:03:10
{ "title": [ -0.0033397675, -0.015167236, -0.029586792, -0.06921387, -0.024017334, 0.0022125244, -0.070617676, 0.0038585663, 0.015899658, 0.005470276, -0.026901245, -0.0033016205, 0.028961182, -0.02949524, -0.024246216, 0.00062799454, -0.018814087, ...
http://arxiv.org/abs/2310.16427v1
PromptAgent: Strategic Planning with Language Models Enables Expert-level Prompt Optimization
Xinyuan Wang, Chenxi Li, Zhen Wang, Fan Bai, Haotian Luo, Jiayou Zhang, Nebojsa Jojic, Eric P. Xing, Zhiting Hu
cs.CL
Highly effective, task-specific prompts are often heavily engineered by experts to integrate detailed instructions and domain insights based on a deep understanding of both instincts of large language models (LLMs) and the intricacies of the target task. However, automating the generation of such expert-level prompts r...
http://arxiv.org/pdf/2310.16427v1
2023-10-25T07:47:01
2023-10-25T07:47:01
{ "title": [ -0.018859863, -0.018096924, -0.027114868, -0.054504395, -0.040130615, 0.00039243698, -0.06713867, -0.0022563934, -0.034484863, 0.021453857, -0.022369385, 0.012290955, -0.01108551, -0.034240723, 0.013534546, -0.024551392, 0.014663696, ...
http://arxiv.org/abs/2310.16417v1
Enhanced Simultaneous Machine Translation with Word-level Policies
Kang Kim, Hankyu Cho
cs.CL
Recent years have seen remarkable advances in the field of Simultaneous Machine Translation (SiMT) due to the introduction of innovative policies that dictate whether to READ or WRITE at each step of the translation process. However, a common assumption in many existing studies is that operations are carried out at the...
http://arxiv.org/pdf/2310.16417v1
2023-10-25T07:10:42
2023-10-25T07:10:42
{ "title": [ -0.018310547, -0.01637268, -0.0042648315, -0.08630371, -0.01108551, 0.018341064, -0.04949951, 0.01878357, -0.021896362, 0.047821045, 0.003408432, 0.021850586, 0.009490967, -0.020187378, -0.029815674, 0.020095825, -0.0043029785, 0.030...
http://arxiv.org/abs/2310.16411v1
Decoding Stumpers: Large Language Models vs. Human Problem-Solvers
Alon Goldstein, Miriam Havin, Roi Reichart, Ariel Goldstein
cs.CL, cs.HC
This paper investigates the problem-solving capabilities of Large Language Models (LLMs) by evaluating their performance on stumpers, unique single-step intuition problems that pose challenges for human solvers but are easily verifiable. We compare the performance of four state-of-the-art LLMs (Davinci-2, Davinci-3, GP...
http://arxiv.org/pdf/2310.16411v1
2023-10-25T06:54:39
2023-10-25T06:54:39
{ "title": [ -0.026351929, -0.014480591, -0.017684937, -0.04647827, -0.012489319, -0.00434494, -0.06274414, -0.025161743, -0.015640259, 0.022384644, -0.020767212, -0.0034751892, -0.009391785, -0.011764526, -0.044799805, -0.01411438, -0.014099121, ...
http://arxiv.org/abs/2310.16402v1
Video Referring Expression Comprehension via Transformer with Content-conditioned Query
Ji Jiang, Meng Cao, Tengtao Song, Long Chen, Yi Wang, Yuexian Zou
cs.CV, cs.CL
Video Referring Expression Comprehension (REC) aims to localize a target object in videos based on the queried natural language. Recent improvements in video REC have been made using Transformer-based methods with learnable queries. However, we contend that this naive query design is not ideal given the open-world natu...
http://arxiv.org/pdf/2310.16402v1
2023-10-25T06:38:42
2023-10-25T06:38:42
{ "title": [ 0.013626099, 0.017623901, -0.030288696, -0.048614502, 0.010093689, -0.007972717, -0.020584106, -0.022003174, 0.016464233, 0.038146973, -0.039276123, -0.016174316, 0.055847168, -0.029296875, -0.012428284, -0.0124053955, -0.0030708313, ...
http://arxiv.org/abs/2311.02089v1
LlamaRec: Two-Stage Recommendation using Large Language Models for Ranking
Zhenrui Yue, Sara Rabhi, Gabriel de Souza Pereira Moreira, Dong Wang, Even Oldridge
cs.IR, cs.AI, cs.CL
Recently, large language models (LLMs) have exhibited significant progress in language understanding and generation. By leveraging textual features, customized LLMs are also applied for recommendation and demonstrate improvements across diverse recommendation scenarios. Yet the majority of existing methods perform trai...
http://arxiv.org/pdf/2311.02089v1
2023-10-25T06:23:48
2023-10-25T06:23:48
{ "title": [ -0.01134491, -0.014556885, -0.028244019, -0.03845215, 0.0052108765, 0.0042533875, -0.016174316, -0.012260437, -0.016601562, 0.027938843, -0.018569946, 0.01864624, -0.017150879, -0.011650085, -0.024795532, -0.00919342, -0.01776123, 0....
http://arxiv.org/abs/2310.16393v1
ZGUL: Zero-shot Generalization to Unseen Languages using Multi-source Ensembling of Language Adapters
Vipul Rathore, Rajdeep Dhingra, Parag Singla, Mausam
cs.CL
We tackle the problem of zero-shot cross-lingual transfer in NLP tasks via the use of language adapters (LAs). Most of the earlier works have explored training with adapter of a single source (often English), and testing either using the target LA or LA of another related language. Training target LA requires unlabeled...
http://arxiv.org/pdf/2310.16393v1
2023-10-25T06:22:29
2023-10-25T06:22:29
{ "title": [ -0.03048706, -0.018157959, -0.033081055, -0.06298828, -0.023239136, 0.01638794, -0.05899048, 0.02357483, 0.005168915, 0.03225708, -0.008766174, 0.027374268, -0.00059223175, -0.010467529, -0.03289795, 0.011528015, 0.0026168823, 0.0395...
http://arxiv.org/abs/2310.16368v1
Transformer-based Live Update Generation for Soccer Matches from Microblog Posts
Masashi Oshika, Kosuke Yamada, Ryohei Sasano, Koichi Takeda
cs.CL
It has been known to be difficult to generate adequate sports updates from a sequence of vast amounts of diverse live tweets, although the live sports viewing experience with tweets is gaining the popularity. In this paper, we focus on soccer matches and work on building a system to generate live updates for soccer mat...
http://arxiv.org/pdf/2310.16368v1
2023-10-25T05:12:35
2023-10-25T05:12:35
{ "title": [ 0.0026474, 0.011398315, -0.0067214966, -0.042816162, 0.017303467, -0.0010786057, -0.025177002, -0.009185791, -0.012916565, 0.05731201, -0.016418457, 0.012672424, 0.020690918, -0.01625061, -0.044036865, 0.0035305023, 0.019546509, 0.00...
http://arxiv.org/abs/2310.16361v1
InstructPTS: Instruction-Tuning LLMs for Product Title Summarization
Besnik Fetahu, Zhiyu Chen, Oleg Rokhlenko, Shervin Malmasi
cs.CL, cs.AI
E-commerce product catalogs contain billions of items. Most products have lengthy titles, as sellers pack them with product attributes to improve retrieval, and highlight key product aspects. This results in a gap between such unnatural products titles, and how customers refer to them. It also limits how e-commerce sto...
http://arxiv.org/pdf/2310.16361v1
2023-10-25T04:56:07
2023-10-25T04:56:07
{ "title": [ -0.07330322, -0.047180176, -0.029052734, -0.034454346, -0.060058594, 0.0032997131, -0.06976318, 0.011077881, -0.022155762, 0.03805542, -0.032470703, 0.050720215, -0.011795044, 0.0005965233, -0.020690918, -0.00869751, -0.0017461777, 0...
http://arxiv.org/abs/2310.16358v1
From Simple to Complex: A Progressive Framework for Document-level Informative Argument Extraction
Quzhe Huang, Yanxi Zhang, Dongyan Zhao
cs.CL
Document-level Event Argument Extraction (EAE) requires the model to extract arguments of multiple events from a single document. Considering the underlying dependencies between these events, recent efforts leverage the idea of "memory", where the results of already predicted events are cached and can be retrieved to h...
http://arxiv.org/pdf/2310.16358v1
2023-10-25T04:38:02
2023-10-25T04:38:02
{ "title": [ 0.03829956, -0.00034570694, -0.018081665, -0.10394287, 0.018310547, 0.003227234, -0.044525146, -0.009918213, -0.005405426, 0.016418457, 0.0007352829, 0.022460938, -0.033996582, -0.012298584, -0.041137695, -0.01008606, -0.008728027, 0...
http://arxiv.org/abs/2310.16356v1
A Multi-Modal Multilingual Benchmark for Document Image Classification
Yoshinari Fujinuma, Siddharth Varia, Nishant Sankaran, Srikar Appalaraju, Bonan Min, Yogarshi Vyas
cs.CL
Document image classification is different from plain-text document classification and consists of classifying a document by understanding the content and structure of documents such as forms, emails, and other such documents. We show that the only existing dataset for this task (Lewis et al., 2006) has several limitat...
http://arxiv.org/pdf/2310.16356v1
2023-10-25T04:35:06
2023-10-25T04:35:06
{ "title": [ -0.030273438, 0.005908966, -0.023162842, -0.029541016, -0.0018110275, -0.009712219, -0.056121826, -0.0262146, 0.03817749, 0.040771484, -0.029647827, 0.013839722, -0.0045661926, -0.001367569, -0.04953003, -0.001115799, 0.006351471, 0....
http://arxiv.org/abs/2310.16350v2
Unraveling Feature Extraction Mechanisms in Neural Networks
Xiaobing Sun, Jiaxi Li, Wei Lu
cs.CL
The underlying mechanism of neural networks in capturing precise knowledge has been the subject of consistent research efforts. In this work, we propose a theoretical approach based on Neural Tangent Kernels (NTKs) to investigate such mechanisms. Specifically, considering the infinite network width, we hypothesize the ...
http://arxiv.org/pdf/2310.16350v2
2023-10-26T03:26:30
2023-10-25T04:22:40
{ "title": [ -0.018920898, 0.00945282, -0.03805542, -0.024597168, 0.032440186, -0.0178833, 0.009010315, -0.027557373, 0.0039787292, 0.023773193, -0.00592041, -0.012535095, -0.020858765, -0.019363403, -0.057891846, 0.00843811, 0.0009202957, 0.0304...
http://arxiv.org/abs/2310.16343v1
A Comprehensive Evaluation of Constrained Text Generation for Large Language Models
Xiang Chen, Xiaojun Wan
cs.CL
Advancements in natural language generation (NLG) and large language models (LLMs) have led to proficient text generation in various tasks. However, integrating intricate constraints into neural text generation, due to LLMs' opacity, remains challenging. This study investigates constrained text generation for LLMs, whe...
http://arxiv.org/pdf/2310.16343v1
2023-10-25T03:58:49
2023-10-25T03:58:49
{ "title": [ -0.05947876, -0.02078247, -0.04837036, -0.054840088, -0.01184845, 0.024414062, -0.080200195, 0.011650085, 0.0090408325, 0.03643799, -0.018966675, 0.0064086914, 0.027770996, -0.008636475, -0.043670654, 0.017349243, -0.024414062, 0.044...
http://arxiv.org/abs/2310.16340v1
RCAgent: Cloud Root Cause Analysis by Autonomous Agents with Tool-Augmented Large Language Models
Zefan Wang, Zichuan Liu, Yingying Zhang, Aoxiao Zhong, Lunting Fan, Lingfei Wu, Qingsong Wen
cs.SE, cs.CL
Large language model (LLM) applications in cloud root cause analysis (RCA) have been actively explored recently. However, current methods are still reliant on manual workflow settings and do not unleash LLMs' decision-making and environment interaction capabilities. We present RCAgent, a tool-augmented LLM autonomous a...
http://arxiv.org/pdf/2310.16340v1
2023-10-25T03:53:31
2023-10-25T03:53:31
{ "title": [ -0.03692627, -0.016326904, -0.04763794, -0.08074951, -0.005706787, 0.010803223, -0.08050537, 0.01612854, -0.031585693, 0.029220581, 0.0031795502, 0.012191772, 0.018753052, -0.040100098, 0.0032367706, 0.013191223, 0.010131836, 0.00153...
http://arxiv.org/abs/2310.16338v1
Generative Pre-training for Speech with Flow Matching
Alexander H. Liu, Matt Le, Apoorv Vyas, Bowen Shi, Andros Tjandra, Wei-Ning Hsu
eess.AS, cs.CL, cs.LG, cs.SD
Generative models have gained more and more attention in recent years for their remarkable success in tasks that required estimating and sampling data distribution to generate high-fidelity synthetic data. In speech, text-to-speech synthesis and neural vocoder are good examples where generative models have shined. Whil...
http://arxiv.org/pdf/2310.16338v1
2023-10-25T03:40:50
2023-10-25T03:40:50
{ "title": [ -0.009590149, -0.014228821, -0.04446411, -0.09851074, -0.0034885406, -0.00034070015, -0.07281494, 0.028839111, -0.016662598, 0.05618286, -0.00035572052, 0.01940918, 0.012283325, -0.02468872, -0.040283203, 0.013221741, 0.007537842, 0....
http://arxiv.org/abs/2310.16329v1
CoheSentia: A Novel Benchmark of Incremental versus Holistic Assessment of Coherence in Generated Texts
Aviya Maimon, Reut Tsarfaty
cs.CL, cs.AI, cs.DB
Coherence is a linguistic term that refers to the relations between small textual units (sentences, propositions), which make the text logically consistent and meaningful to the reader. With the advances of generative foundational models in NLP, there is a pressing need to automatically assess the human-perceived coher...
http://arxiv.org/pdf/2310.16329v1
2023-10-25T03:21:20
2023-10-25T03:21:20
{ "title": [ -0.021148682, 0.007255554, -0.047851562, -0.058044434, -0.0039749146, -0.00995636, -0.079956055, -0.010643005, 0.0038890839, 0.046569824, 0.014289856, 0.017074585, 0.03970337, -0.023773193, -0.04208374, 0.028579712, -0.029312134, 0.0...
http://arxiv.org/abs/2310.16322v1
Samsung R&D Institute Philippines at WMT 2023
Jan Christian Blaise Cruz
cs.CL
In this paper, we describe the constrained MT systems submitted by Samsung R&D Institute Philippines to the WMT 2023 General Translation Task for two directions: en$\rightarrow$he and he$\rightarrow$en. Our systems comprise of Transformer-based sequence-to-sequence models that are trained with a mix of best practices: ...
http://arxiv.org/pdf/2310.16322v1
2023-10-25T03:10:52
2023-10-25T03:10:52
{ "title": [ 0.013328552, -0.030059814, 0.01651001, 0.010848999, -0.04864502, 0.007980347, -0.0126571655, -0.02418518, -0.02619934, 0.042388916, -0.009857178, -0.0050582886, 0.019622803, -0.031951904, 0.008155823, -0.04046631, 0.020889282, 0.0368...
http://arxiv.org/abs/2310.16319v1
DiQAD: A Benchmark Dataset for End-to-End Open-domain Dialogue Assessment
Yukun Zhao, Lingyong Yan, Weiwei Sun, Chong Meng, Shuaiqiang Wang, Zhicong Cheng, Zhaochun Ren, Dawei Yin
cs.CL
Dialogue assessment plays a critical role in the development of open-domain dialogue systems. Existing work are uncapable of providing an end-to-end and human-epistemic assessment dataset, while they only provide sub-metrics like coherence or the dialogues are conversed between annotators far from real user settings. I...
http://arxiv.org/pdf/2310.16319v1
2023-10-25T03:04:57
2023-10-25T03:04:57
{ "title": [ 0.0032291412, 0.010421753, -0.034973145, -0.08691406, 0.02053833, -0.0057411194, -0.04788208, 0.0017614365, 0.020111084, 0.01576233, 0.0101623535, -0.0012292862, -0.01940918, -0.0289917, -0.059814453, 0.0021781921, -0.024032593, 0.02...
http://arxiv.org/abs/2310.16303v1
URL-BERT: Training Webpage Representations via Social Media Engagements
Ayesha Qamar, Chetan Verma, Ahmed El-Kishky, Sumit Binnani, Sneha Mehta, Taylor Berg-Kirkpatrick
cs.CL, cs.IR
Understanding and representing webpages is crucial to online social networks where users may share and engage with URLs. Common language model (LM) encoders such as BERT can be used to understand and represent the textual content of webpages. However, these representations may not model thematic information of web doma...
http://arxiv.org/pdf/2310.16303v1
2023-10-25T02:22:50
2023-10-25T02:22:50
{ "title": [ 0.029632568, 0.00039887428, -0.007194519, -0.062408447, 0.012130737, -0.010108948, -0.014587402, -0.015731812, -0.007858276, 0.03265381, -0.011062622, 0.025665283, -0.0051193237, -0.016616821, -0.014663696, -0.00095796585, -0.025314331, ...
http://arxiv.org/abs/2310.16301v1
Is ChatGPT a Good Multi-Party Conversation Solver?
Chao-Hong Tan, Jia-Chen Gu, Zhen-Hua Ling
cs.CL
Large Language Models (LLMs) have emerged as influential instruments within the realm of natural language processing; nevertheless, their capacity to handle multi-party conversations (MPCs) -- a scenario marked by the presence of multiple interlocutors involved in intricate information exchanges -- remains uncharted. I...
http://arxiv.org/pdf/2310.16301v1
2023-10-25T02:18:40
2023-10-25T02:18:40
{ "title": [ 0.0047950745, -0.013748169, 0.0028514862, -0.01727295, -0.04159546, 0.005996704, -0.054656982, -0.017745972, -0.036132812, 0.033447266, 0.021621704, 0.016464233, -0.01348114, -0.034851074, 0.01209259, -0.009346008, -0.028076172, 0.04...
http://arxiv.org/abs/2310.16278v1
XFEVER: Exploring Fact Verification across Languages
Yi-Chen Chang, Canasai Kruengkrai, Junichi Yamagishi
cs.CL, cs.AI
This paper introduces the Cross-lingual Fact Extraction and VERification (XFEVER) dataset designed for benchmarking the fact verification models across different languages. We constructed it by translating the claim and evidence texts of the Fact Extraction and VERification (FEVER) dataset into six languages. The train...
http://arxiv.org/pdf/2310.16278v1
2023-10-25T01:20:17
2023-10-25T01:20:17
{ "title": [ -0.025283813, -0.013198853, -0.0395813, -0.07409668, 0.01373291, 0.0063972473, -0.05206299, 0.0023460388, -0.018981934, 0.035186768, -0.0023097992, 0.017669678, 0.018798828, -0.014228821, -0.031341553, 0.019927979, -0.037902832, 0.07...
http://arxiv.org/abs/2310.18365v1
Using GPT-4 to Augment Unbalanced Data for Automatic Scoring
Luyang Fang, Gyeong-Geon Lee, Xiaoming Zhai
cs.CL, cs.AI, cs.CY
Machine learning-based automatic scoring can be challenging if students' responses are unbalanced across scoring categories, as it introduces uncertainty in the machine training process. To meet this challenge, we introduce a novel text data augmentation framework leveraging GPT-4, a generative large language model, sp...
http://arxiv.org/pdf/2310.18365v1
2023-10-25T01:07:50
2023-10-25T01:07:50
{ "title": [ 0.014862061, 0.025939941, -0.020462036, -0.032928467, 0.022018433, -0.018203735, -0.008331299, -0.053741455, 0.022384644, 0.006164551, -0.028808594, -0.009750366, -0.02772522, -0.036834717, -0.040130615, -0.003955841, -0.016952515, 0...
http://arxiv.org/abs/2310.16271v1
CycleAlign: Iterative Distillation from Black-box LLM to White-box Models for Better Human Alignment
Jixiang Hong, Quan Tu, Changyu Chen, Xing Gao, Ji Zhang, Rui Yan
cs.CL, cs.AI
Language models trained on large-scale corpus often generate content that is harmful, toxic, or contrary to human preferences, making their alignment with human values a critical concern. Reinforcement learning from human feedback (RLHF) with algorithms like PPO is a prevalent approach for alignment but is often comple...
http://arxiv.org/pdf/2310.16271v1
2023-10-25T01:05:03
2023-10-25T01:05:03
{ "title": [ -0.046173096, -0.0023097992, -0.061553955, -0.054138184, -0.019927979, 0.0004196167, -0.08917236, -0.013839722, 0.0041160583, 0.040893555, -0.015716553, 0.012519836, 0.025924683, -0.022369385, -0.034729004, 0.010864258, -0.0012893677, ...
http://arxiv.org/abs/2310.16270v1
Attention Lens: A Tool for Mechanistically Interpreting the Attention Head Information Retrieval Mechanism
Mansi Sakarvadia, Arham Khan, Aswathy Ajith, Daniel Grzenda, Nathaniel Hudson, André Bauer, Kyle Chard, Ian Foster
cs.CL, cs.AI, cs.LG
Transformer-based Large Language Models (LLMs) are the state-of-the-art for natural language tasks. Recent work has attempted to decode, by reverse engineering the role of linear layers, the internal mechanisms by which LLMs arrive at their final predictions for text completion tasks. Yet little is known about the spec...
http://arxiv.org/pdf/2310.16270v1
2023-10-25T01:03:35
2023-10-25T01:03:35
{ "title": [ -0.009017944, 0.014152527, -0.039154053, -0.028930664, -0.0009908676, -0.028274536, 0.0008778572, -0.0340271, 0.009124756, 0.03427124, -0.024215698, -0.0062217712, 0.021835327, -0.0127334595, -0.050048828, 0.025375366, 0.012565613, 0...
http://arxiv.org/abs/2310.16269v1
Multilingual Coarse Political Stance Classification of Media. The Editorial Line of a ChatGPT and Bard Newspaper
Cristina España-Bonet
cs.CL, cs.AI, cs.CY
Neutrality is difficult to achieve and, in politics, subjective. Traditional media typically adopt an editorial line that can be used by their potential readers as an indicator of the media bias. Several platforms currently rate news outlets according to their political bias. The editorial line and the ratings help rea...
http://arxiv.org/pdf/2310.16269v1
2023-10-25T01:01:28
2023-10-25T01:01:28
{ "title": [ -0.0030975342, -0.0076026917, 0.005844116, 0.016647339, -0.034423828, -0.005168915, 0.020523071, -0.003528595, -0.009391785, 0.029006958, 0.003862381, 0.017715454, 0.006374359, -0.038330078, 0.019927979, -0.0067596436, -0.021972656, ...
http://arxiv.org/abs/2310.16263v1
Enhancing Large Language Models for Secure Code Generation: A Dataset-driven Study on Vulnerability Mitigation
Jiexin Wang, Liuwen Cao, Xitong Luo, Zhiping Zhou, Jiayuan Xie, Adam Jatowt, Yi Cai
cs.SE, cs.AI, cs.CL, cs.CR
Large language models (LLMs) have brought significant advancements to code generation, benefiting both novice and experienced developers. However, their training using unsanitized data from open-source repositories, like GitHub, introduces the risk of inadvertently propagating security vulnerabilities. To effectively m...
http://arxiv.org/pdf/2310.16263v1
2023-10-25T00:32:56
2023-10-25T00:32:56
{ "title": [ -0.03366089, -0.018920898, -0.05999756, -0.07104492, 0.0005283356, -0.0051498413, -0.06335449, 0.0390625, -0.022781372, 0.029510498, 0.006881714, 0.0060768127, -0.0069389343, -0.02130127, -0.030960083, -0.005428314, -0.019943237, 0.0...
http://arxiv.org/abs/2310.16261v1
The Distributional Hypothesis Does Not Fully Explain the Benefits of Masked Language Model Pretraining
Ting-Rui Chiang, Dani Yogatama
cs.CL
We analyze the masked language modeling pretraining objective function from the perspective of the distributional hypothesis. We investigate whether better sample efficiency and the better generalization capability of models pretrained with masked language modeling can be attributed to the semantic similarity encoded i...
http://arxiv.org/pdf/2310.16261v1
2023-10-25T00:31:29
2023-10-25T00:31:29
{ "title": [ -0.031402588, 0.01158905, -0.036010742, 0.019332886, -0.043273926, -0.0024757385, -0.01979065, -0.017196655, -0.054260254, 0.017562866, -0.008377075, -0.020095825, -0.0005288124, -0.07495117, 0.039642334, -0.02670288, -0.007774353, 0...
http://arxiv.org/abs/2310.16251v1
Speakerly: A Voice-based Writing Assistant for Text Composition
Dhruv Kumar, Vipul Raheja, Alice Kaiser-Schatzlein, Robyn Perry, Apurva Joshi, Justin Hugues-Nuger, Samuel Lou, Navid Chowdhury
cs.CL, cs.AI
We present Speakerly, a new real-time voice-based writing assistance system that helps users with text composition across various use cases such as emails, instant messages, and notes. The user can interact with the system through instructions or dictation, and the system generates a well-formatted and coherent documen...
http://arxiv.org/pdf/2310.16251v1
2023-10-24T23:53:15
2023-10-24T23:53:15
{ "title": [ -0.041107178, -0.02330017, -0.054779053, -0.06124878, -0.014137268, 0.003255844, -0.05899048, 0.031311035, 0.015296936, 0.050933838, -0.018432617, -0.0001449585, 0.042114258, -0.006515503, -0.057556152, 0.009346008, -0.036132812, 0.0...
http://arxiv.org/abs/2310.16248v2
GlotLID: Language Identification for Low-Resource Languages
Amir Hossein Kargaran, Ayyoob Imani, François Yvon, Hinrich Schütze
cs.CL
Several recent papers have published good solutions for language identification (LID) for about 300 high-resource and medium-resource languages. However, there is no LID available that (i) covers a wide range of low-resource languages, (ii) is rigorously evaluated and reliable and (iii) efficient and easy to use. Here,...
http://arxiv.org/pdf/2310.16248v2
2023-11-04T11:28:10
2023-10-24T23:45:57
{ "title": [ -0.020309448, 0.0065612793, -0.037597656, -0.058807373, 0.012763977, 0.0018434525, -0.05758667, -0.00793457, -0.028167725, 0.025970459, -0.0061836243, 0.0063819885, 0.004875183, -0.032287598, -0.024765015, -0.007698059, -0.014785767, ...
http://arxiv.org/abs/2310.16242v1
ZzzGPT: An Interactive GPT Approach to Enhance Sleep Quality
Yonchanok Khaokaew, Thuc Hanh Nguyen, Kaixin Ji, Hiruni Kegalle, Marwah Alaofi
cs.LG, cs.CL
In today's world, sleep quality is pivotal for overall well-being. While wearable sensors offer real-time monitoring, they often lack actionable insights, leading to user abandonment. This paper delves into the role of technology in understanding sleep patterns. We introduce a two-stage framework, utilizing Large Langu...
http://arxiv.org/pdf/2310.16242v1
2023-10-24T23:30:17
2023-10-24T23:30:17
{ "title": [ -0.016052246, -0.0077667236, -0.022750854, 0.004589081, 0.0015296936, -0.017196655, -0.003616333, -0.03656006, -0.034820557, 0.0034618378, -0.005027771, 0.0032539368, -0.043029785, -0.041412354, -0.020996094, -0.01247406, -0.0023536682, ...
http://arxiv.org/abs/2310.16240v1
Mixture-of-Linguistic-Experts Adapters for Improving and Interpreting Pre-trained Language Models
Raymond Li, Gabriel Murray, Giuseppe Carenini
cs.CL
In this work, we propose a method that combines two popular research areas by injecting linguistic structures into pre-trained language models in the parameter-efficient fine-tuning (PEFT) setting. In our approach, parallel adapter modules encoding different linguistic structures are combined using a novel Mixture-of-L...
http://arxiv.org/pdf/2310.16240v1
2023-10-24T23:29:06
2023-10-24T23:29:06
{ "title": [ -0.029846191, -0.0055732727, -0.040039062, -0.06732178, -0.010414124, -0.0074310303, -0.057800293, 0.004524231, -0.011634827, 0.043273926, 0.008903503, -0.0082092285, 0.027404785, -0.024276733, -0.041412354, 0.01777649, 0.013633728, ...
http://arxiv.org/abs/2310.16226v1
TiC-CLIP: Continual Training of CLIP Models
Saurabh Garg, Mehrdad Farajtabar, Hadi Pouransari, Raviteja Vemulapalli, Sachin Mehta, Oncel Tuzel, Vaishaal Shankar, Fartash Faghri
cs.CV, cs.CL, cs.LG
Keeping large foundation models up to date on latest data is inherently expensive. To avoid the prohibitive costs of constantly retraining, it is imperative to continually train these models. This problem is exacerbated by the lack of any large scale continual learning benchmarks or baselines. We introduce the first se...
http://arxiv.org/pdf/2310.16226v1
2023-10-24T22:41:14
2023-10-24T22:41:14
{ "title": [ -0.015129089, -0.0019416809, -0.046813965, -0.06964111, 0.0032730103, 0.0063552856, -0.07159424, -0.02029419, -0.005256653, 0.023910522, -0.0018377304, -0.00090932846, -0.005252838, -0.04168701, -0.036895752, 0.015449524, 0.0050735474, ...
http://arxiv.org/abs/2310.16225v1
CleanCoNLL: A Nearly Noise-Free Named Entity Recognition Dataset
Susanna Rücker, Alan Akbik
cs.CL, cs.AI, cs.LG
The CoNLL-03 corpus is arguably the most well-known and utilized benchmark dataset for named entity recognition (NER). However, prior works found significant numbers of annotation errors, incompleteness, and inconsistencies in the data. This poses challenges to objectively comparing NER approaches and analyzing their e...
http://arxiv.org/pdf/2310.16225v1
2023-10-24T22:34:43
2023-10-24T22:34:43
{ "title": [ -0.013900757, -0.0033454895, -0.06536865, -0.088378906, 0.035217285, -0.0042877197, -0.04360962, 0.014564514, 0.010002136, 0.028213501, -0.021713257, 0.0095825195, 0.0009384155, -0.0032958984, -0.03967285, 0.0057373047, -0.008430481, ...
http://arxiv.org/abs/2310.16218v2
Knowledge Editing for Large Language Models: A Survey
Song Wang, Yaochen Zhu, Haochen Liu, Zaiyi Zheng, Chen Chen, Jundong Li
cs.CL, cs.AI
Large language models (LLMs) have recently transformed both the academic and industrial landscapes due to their remarkable capacity to understand, analyze, and generate texts based on their vast knowledge and reasoning ability. Nevertheless, one major drawback of LLMs is their substantial computational cost for pre-tra...
http://arxiv.org/pdf/2310.16218v2
2023-10-26T00:45:42
2023-10-24T22:18:13
{ "title": [ -0.057250977, -0.03201294, -0.057250977, -0.07672119, -0.009162903, 0.017410278, -0.07220459, 0.037261963, 0.004875183, 0.031707764, -0.014976501, -0.022369385, 0.026351929, -0.015159607, -0.04650879, 0.0034255981, -0.018737793, 0.02...
http://arxiv.org/abs/2310.16197v1
Background Summarization of Event Timelines
Adithya Pratapa, Kevin Small, Markus Dreyer
cs.CL
Generating concise summaries of news events is a challenging natural language processing task. While journalists often curate timelines to highlight key sub-events, newcomers to a news event face challenges in catching up on its historical context. In this paper, we address this need by introducing the task of backgrou...
http://arxiv.org/pdf/2310.16197v1
2023-10-24T21:30:15
2023-10-24T21:30:15
{ "title": [ 0.054870605, 0.014282227, -0.023452759, -0.11846924, -0.012390137, -0.015579224, -0.077941895, 0.00995636, -0.017105103, 0.072509766, -0.015579224, 0.026519775, 0.03186035, -0.023498535, 0.0043525696, 0.022125244, 0.04269409, 0.00654...
http://arxiv.org/abs/2310.16193v1
Length is a Curse and a Blessing for Document-level Semantics
Chenghao Xiao, Yizhi Li, G Thomas Hudson, Chenghua Lin, Noura Al Moubayed
cs.CL, cs.AI
In recent years, contrastive learning (CL) has been extensively utilized to recover sentence and document-level encoding capability from pre-trained language models. In this work, we question the length generalizability of CL-based models, i.e., their vulnerability towards length-induced semantic shift. We verify not o...
http://arxiv.org/pdf/2310.16193v1
2023-10-24T21:23:53
2023-10-24T21:23:53
{ "title": [ 0.0023822784, -0.035858154, -0.033935547, -0.05303955, -0.053527832, -0.0046081543, -0.049468994, 0.030807495, -0.021575928, 0.042633057, -0.061676025, -0.016479492, 0.06756592, -0.015075684, 0.012817383, -0.009346008, -0.0018663406, ...
http://arxiv.org/abs/2310.16183v1
BLP 2023 Task 2: Sentiment Analysis
Md. Arid Hasan, Firoj Alam, Anika Anjum, Shudipta Das, Afiyat Anjum
cs.CL, cs.LG, I.2.7
We present an overview of the BLP Sentiment Shared Task, organized as part of the inaugural BLP 2023 workshop, co-located with EMNLP 2023. The task is defined as the detection of sentiment in a given piece of social media text. This task attracted interest from 71 participants, among whom 29 and 30 teams submitted syst...
http://arxiv.org/pdf/2310.16183v1
2023-10-24T21:00:41
2023-10-24T21:00:41
{ "title": [ 0.0056037903, -0.0062789917, -0.025817871, -0.035949707, -0.009124756, -0.024169922, -0.038269043, -0.027786255, -0.0024490356, 0.05203247, 0.003408432, -0.008628845, -0.0034675598, -0.042419434, -0.03186035, 0.015129089, 0.0074920654, ...
http://arxiv.org/abs/2310.16181v1
Hidden Citations Obscure True Impact in Science
Xiangyi Meng, Onur Varol, Albert-László Barabási
cs.CL, cs.DL, cs.SI, physics.soc-ph
References, the mechanism scientists rely on to signal previous knowledge, lately have turned into widely used and misused measures of scientific impact. Yet, when a discovery becomes common knowledge, citations suffer from obliteration by incorporation. This leads to the concept of hidden citation, representing a clea...
http://arxiv.org/pdf/2310.16181v1
2023-10-24T20:58:07
2023-10-24T20:58:07
{ "title": [ -0.0066375732, 0.012229919, -0.020019531, 0.041656494, -0.039001465, -0.022949219, 0.012374878, -0.051727295, -0.06365967, 0.02973938, -0.013671875, -0.048339844, -0.009567261, -0.04244995, -0.0039100647, -0.050720215, 0.008285522, 0...
http://arxiv.org/abs/2310.16176v2
Correction with Backtracking Reduces Hallucination in Summarization
Zhenzhen Liu, Chao Wan, Varsha Kishore, Jin Peng Zhou, Minmin Chen, Kilian Q. Weinberger
cs.CL, cs.AI
Abstractive summarization aims at generating natural language summaries of a source document that are succinct while preserving the important elements. Despite recent advances, neural text summarization models are known to be susceptible to hallucinating (or more correctly confabulating), that is to produce summaries w...
http://arxiv.org/pdf/2310.16176v2
2023-10-31T14:48:14
2023-10-24T20:48:11
{ "title": [ 0.029052734, -0.0014076233, 0.0032863617, -0.015586853, -0.049804688, -0.014579773, -0.051208496, -0.057037354, -0.029769897, 0.066589355, 0.01966858, -0.010192871, 0.03427124, -0.03781128, -0.02722168, 0.0073890686, 0.0012435913, 0....
http://arxiv.org/abs/2310.16153v1
WojoodNER 2023: The First Arabic Named Entity Recognition Shared Task
Mustafa Jarrar, Muhammad Abdul-Mageed, Mohammed Khalilia, Bashar Talafha, AbdelRahim Elmadany, Nagham Hamad, Alaa' Omar
cs.CL
We present WojoodNER-2023, the first Arabic Named Entity Recognition (NER) Shared Task. The primary focus of WojoodNER-2023 is on Arabic NER, offering novel NER datasets (i.e., Wojood) and the definition of subtasks designed to facilitate meaningful comparisons between different NER approaches. WojoodNER-2023 encompass...
http://arxiv.org/pdf/2310.16153v1
2023-10-24T19:50:07
2023-10-24T19:50:07
{ "title": [ -0.026245117, -0.010383606, -0.040618896, -0.035858154, -0.015296936, 0.0012865067, -0.039031982, 0.0000538826, -0.013206482, 0.036346436, -0.0059661865, 0.02418518, 0.026550293, -0.044311523, 0.03265381, 0.024520874, 0.0132369995, 0...
http://arxiv.org/abs/2310.16147v2
PreWoMe: Exploiting Presuppositions as Working Memory for Long Form Question Answering
Wookje Han, Jinsol Park, Kyungjae Lee
cs.CL, cs.AI
Information-seeking questions in long-form question answering (LFQA) often prove misleading due to ambiguity or false presupposition in the question. While many existing approaches handle misleading questions, they are tailored to limited questions, which are insufficient in a real-world setting with unpredictable inpu...
http://arxiv.org/pdf/2310.16147v2
2023-11-14T02:46:32
2023-10-24T19:47:26
{ "title": [ 0.0181427, -0.006702423, -0.037078857, -0.07324219, 0.009033203, -0.026123047, -0.0071105957, 0.0061073303, 0.029220581, 0.023895264, -0.017028809, -0.006679535, -0.009101868, -0.0062026978, -0.062072754, -0.008361816, -0.00166893, 0...
http://arxiv.org/abs/2310.18364v1
From Heuristic to Analytic: Cognitively Motivated Strategies for Coherent Physical Commonsense Reasoning
Zheyuan Zhang, Shane Storks, Fengyuan Hu, Sungryull Sohn, Moontae Lee, Honglak Lee, Joyce Chai
cs.CL, cs.AI
Pre-trained language models (PLMs) have shown impressive performance in various language tasks. However, they are prone to spurious correlations, and often generate illusory information. In real-world applications, PLMs should justify decisions with formalized, coherent reasoning chains, but this challenge remains unde...
http://arxiv.org/pdf/2310.18364v1
2023-10-24T19:46:04
2023-10-24T19:46:04
{ "title": [ 0.00818634, -0.01247406, -0.016464233, -0.047332764, -0.004722595, -0.018676758, -0.018188477, -0.02571106, -0.0033988953, 0.03225708, 0.012077332, 0.019058228, 0.003665924, -0.0335083, -0.021865845, 0.012260437, 0.003479004, -0.0012...
http://arxiv.org/abs/2310.16146v1
Clinfo.ai: An Open-Source Retrieval-Augmented Large Language Model System for Answering Medical Questions using Scientific Literature
Alejandro Lozano, Scott L Fleming, Chia-Chun Chiang, Nigam Shah
cs.IR, cs.AI, cs.CL
The quickly-expanding nature of published medical literature makes it challenging for clinicians and researchers to keep up with and summarize recent, relevant findings in a timely manner. While several closed-source summarization tools based on large language models (LLMs) now exist, rigorous and systematic evaluation...
http://arxiv.org/pdf/2310.16146v1
2023-10-24T19:43:39
2023-10-24T19:43:39
{ "title": [ -0.043701172, -0.008056641, -0.051330566, -0.048553467, -0.010322571, 0.014595032, -0.03878784, -0.0027332306, -0.008285522, 0.050628662, -0.0029182434, -0.015098572, 0.01461792, -0.027023315, -0.047576904, -0.005443573, -0.0010261536, ...
http://arxiv.org/abs/2310.16142v1
A Language Model with Limited Memory Capacity Captures Interference in Human Sentence Processing
William Timkey, Tal Linzen
cs.CL, cs.AI, cs.LG
Two of the central factors believed to underpin human sentence processing difficulty are expectations and retrieval from working memory. A recent attempt to create a unified cognitive model integrating these two factors relied on the parallels between the self-attention mechanism of transformer language models and cue-...
http://arxiv.org/pdf/2310.16142v1
2023-10-24T19:33:27
2023-10-24T19:33:27
{ "title": [ 0.033813477, 0.00207901, -0.009399414, -0.0413208, -0.040222168, -0.025756836, -0.04486084, -0.020751953, -0.027679443, 0.041748047, -0.020202637, -0.020217896, 0.044036865, -0.036956787, -0.01838684, -0.009727478, -0.0027637482, 0.0...
http://arxiv.org/abs/2310.16135v1
Can You Follow Me? Testing Situational Understanding in ChatGPT
Chenghao Yang, Allyson Ettinger
cs.CL
Understanding sentence meanings and updating information states appropriately across time -- what we call "situational understanding" (SU) -- is a critical ability for human-like AI agents. SU is essential in particular for chat models, such as ChatGPT, to enable consistent, coherent, and effective dialogue between hum...
http://arxiv.org/pdf/2310.16135v1
2023-10-24T19:22:01
2023-10-24T19:22:01
{ "title": [ 0.015411377, -0.0046920776, -0.009048462, -0.043426514, 0.0012645721, -0.0010309219, -0.032196045, -0.03161621, -0.003025055, 0.004776001, -0.03427124, 0.021209717, -0.011810303, -0.013473511, -0.032104492, -0.005367279, -0.02279663, ...
http://arxiv.org/abs/2310.16131v1
GenKIE: Robust Generative Multimodal Document Key Information Extraction
Panfeng Cao, Ye Wang, Qiang Zhang, Zaiqiao Meng
cs.CL
Key information extraction (KIE) from scanned documents has gained increasing attention because of its applications in various domains. Although promising results have been achieved by some recent KIE approaches, they are usually built based on discriminative models, which lack the ability to handle optical character r...
http://arxiv.org/pdf/2310.16131v1
2023-10-24T19:12:56
2023-10-24T19:12:56
{ "title": [ -0.07171631, -0.0158844, -0.056152344, -0.03604126, -0.019973755, 0.0010824203, -0.053741455, -0.0039978027, 0.016067505, 0.032348633, -0.010299683, 0.011329651, 0.012931824, -0.010421753, -0.04522705, 0.002401352, -0.007843018, 0.02...
http://arxiv.org/abs/2310.16127v1
Octopus: A Multitask Model and Toolkit for Arabic Natural Language Generation
AbdelRahim Elmadany, El Moatez Billah Nagoudi, Muhammad Abdul-Mageed
cs.CL
Understanding Arabic text and generating human-like responses is a challenging endeavor. While many researchers have proposed models and solutions for individual problems, there is an acute shortage of a comprehensive Arabic natural language generation toolkit that is capable of handling a wide range of tasks. In this ...
http://arxiv.org/pdf/2310.16127v1
2023-10-24T19:06:55
2023-10-24T19:06:55
{ "title": [ -0.040283203, -0.038146973, -0.03201294, -0.06829834, -0.022094727, 0.02734375, -0.064453125, 0.03265381, -0.004600525, 0.022079468, 0.008140564, 0.0082473755, -0.019714355, -0.03665161, -0.03274536, -0.000045359135, -0.0061912537, 0...
http://arxiv.org/abs/2310.16117v1
NADI 2023: The Fourth Nuanced Arabic Dialect Identification Shared Task
Muhammad Abdul-Mageed, AbdelRahim Elmadany, Chiyu Zhang, El Moatez Billah Nagoudi, Houda Bouamor, Nizar Habash
cs.CL
We describe the findings of the fourth Nuanced Arabic Dialect Identification Shared Task (NADI 2023). The objective of NADI is to help advance state-of-the-art Arabic NLP by creating opportunities for teams of researchers to collaboratively compete under standardized conditions. It does so with a focus on Arabic dialec...
http://arxiv.org/pdf/2310.16117v1
2023-10-24T18:41:24
2023-10-24T18:41:24
{ "title": [ -0.019897461, -0.00484848, -0.048217773, -0.048950195, 0.006904602, -0.0029773712, -0.037475586, 0.0028457642, -0.017501831, 0.032562256, 0.019592285, -0.0047340393, -0.0052871704, -0.055541992, -0.020629883, 0.017089844, 0.019424438, ...
http://arxiv.org/abs/2310.16111v1
Locally Differentially Private Document Generation Using Zero Shot Prompting
Saiteja Utpala, Sara Hooker, Pin Yu Chen
cs.CL, cs.CR, cs.LG
Numerous studies have highlighted the privacy risks associated with pretrained large language models. In contrast, our research offers a unique perspective by demonstrating that pretrained large language models can effectively contribute to privacy preservation. We propose a locally differentially private mechanism cal...
http://arxiv.org/pdf/2310.16111v1
2023-10-24T18:25:13
2023-10-24T18:25:13
{ "title": [ -0.025177002, -0.02986145, -0.029830933, -0.070129395, -0.028457642, 0.01687622, -0.066711426, 0.028823853, -0.013954163, 0.006126404, -0.015899658, 0.00390625, 0.0048217773, -0.029556274, 0.002866745, -0.01424408, 0.009513855, 0.033...
http://arxiv.org/abs/2310.16095v1
CR-COPEC: Causal Rationale of Corporate Performance Changes to Learn from Financial Reports
Ye Eun Chun, Sunjae Kwon, Kyunghwan Sohn, Nakwon Sung, Junyoup Lee, Byungki Seo, Kevin Compher, Seung-won Hwang, Jaesik Choi
cs.CL, cs.CE
In this paper, we introduce CR-COPEC called Causal Rationale of Corporate Performance Changes from financial reports. This is a comprehensive large-scale domain-adaptation causal sentence dataset to detect financial performance changes of corporate. CR-COPEC contributes to two major achievements. First, it detects caus...
http://arxiv.org/pdf/2310.16095v1
2023-10-24T18:00:40
2023-10-24T18:00:40
{ "title": [ -0.039733887, -0.025970459, -0.066223145, -0.045898438, -0.029464722, -0.015991211, -0.042266846, 0.024276733, -0.027770996, 0.00019598007, -0.064086914, 0.019866943, -0.017242432, -0.01789856, 0.011787415, -0.037353516, -0.0107421875, ...
http://arxiv.org/abs/2310.16049v1
MuSR: Testing the Limits of Chain-of-thought with Multistep Soft Reasoning
Zayne Sprague, Xi Ye, Kaj Bostrom, Swarat Chaudhuri, Greg Durrett
cs.CL
While large language models (LLMs) equipped with techniques like chain-of-thought prompting have demonstrated impressive capabilities, they still fall short in their ability to reason robustly in complex settings. However, evaluating LLM reasoning is challenging because system capabilities continue to grow while benchm...
http://arxiv.org/pdf/2310.16049v1
2023-10-24T17:59:20
2023-10-24T17:59:20
{ "title": [ -0.028259277, -0.027511597, -0.037322998, -0.047576904, -0.0078086853, -0.00006467104, -0.013061523, 0.0135269165, 0.0066604614, -0.00044751167, -0.02619934, -0.004055023, -0.0051002502, -0.029937744, -0.0146484375, -0.009239197, 0.01040...
http://arxiv.org/abs/2310.16048v1
AI Alignment and Social Choice: Fundamental Limitations and Policy Implications
Abhilash Mishra
cs.AI, cs.CL, cs.CY, cs.HC, cs.LG
Aligning AI agents to human intentions and values is a key bottleneck in building safe and deployable AI applications. But whose values should AI agents be aligned with? Reinforcement learning with human feedback (RLHF) has emerged as the key framework for AI alignment. RLHF uses feedback from human reinforcers to fine...
http://arxiv.org/pdf/2310.16048v1
2023-10-24T17:59:04
2023-10-24T17:59:04
{ "title": [ 0.00057697296, 0.012893677, -0.01890564, -0.012611389, 0.018737793, -0.02418518, 0.04360962, -0.040222168, -0.016677856, 0.030151367, -0.034606934, 0.01473999, -0.027175903, 0.00039243698, -0.054992676, -0.015571594, -0.0070343018, 0...
http://arxiv.org/abs/2310.16045v1
Woodpecker: Hallucination Correction for Multimodal Large Language Models
Shukang Yin, Chaoyou Fu, Sirui Zhao, Tong Xu, Hao Wang, Dianbo Sui, Yunhang Shen, Ke Li, Xing Sun, Enhong Chen
cs.CV, cs.AI, cs.CL, cs.LG
Hallucination is a big shadow hanging over the rapidly evolving Multimodal Large Language Models (MLLMs), referring to the phenomenon that the generated text is inconsistent with the image content. In order to mitigate hallucinations, existing studies mainly resort to an instruction-tuning manner that requires retraini...
http://arxiv.org/pdf/2310.16045v1
2023-10-24T17:58:07
2023-10-24T17:58:07
{ "title": [ -0.03375244, -0.002313614, -0.04107666, -0.042022705, -0.012817383, -0.0011491776, -0.050109863, -0.0038166046, 0.0102005005, 0.03829956, -0.008056641, 0.004135132, 0.012336731, -0.008140564, -0.05227661, 0.012069702, -0.020568848, 0...
http://arxiv.org/abs/2310.16042v2
WebWISE: Web Interface Control and Sequential Exploration with Large Language Models
Heyi Tao, Sethuraman T V, Michal Shlapentokh-Rothman, Derek Hoiem
cs.CL, cs.AI
The paper investigates using a Large Language Model (LLM) to automatically perform web software tasks using click, scroll, and text input operations. Previous approaches, such as reinforcement learning (RL) or imitation learning, are inefficient to train and task-specific. Our method uses filtered Document Object Model...
http://arxiv.org/pdf/2310.16042v2
2023-10-25T03:54:11
2023-10-24T17:57:03
{ "title": [ -0.020111084, -0.025482178, -0.0289917, -0.05343628, 0.01448822, -0.0070610046, -0.026672363, -0.013824463, -0.010231018, 0.032043457, -0.018951416, -0.00944519, -0.004322052, -0.024642944, -0.046173096, -0.009124756, -0.008529663, 0...
http://arxiv.org/abs/2310.16040v1
Instruct and Extract: Instruction Tuning for On-Demand Information Extraction
Yizhu Jiao, Ming Zhong, Sha Li, Ruining Zhao, Siru Ouyang, Heng Ji, Jiawei Han
cs.CL, cs.AI
Large language models with instruction-following capabilities open the door to a wider group of users. However, when it comes to information extraction - a classic task in natural language processing - most task-specific systems cannot align well with long-tail ad hoc extraction use cases for non-expert users. To addre...
http://arxiv.org/pdf/2310.16040v1
2023-10-24T17:54:25
2023-10-24T17:54:25
{ "title": [ -0.05783081, -0.030319214, -0.041381836, -0.064819336, -0.015548706, 0.0061302185, -0.039978027, 0.010505676, -0.006225586, 0.012260437, -0.051574707, 0.026046753, -0.03302002, 0.0041160583, -0.036102295, -0.010810852, -0.0067596436, ...
http://arxiv.org/abs/2310.16035v1
What's Left? Concept Grounding with Logic-Enhanced Foundation Models
Joy Hsu, Jiayuan Mao, Joshua B. Tenenbaum, Jiajun Wu
cs.CV, cs.AI, cs.CL, cs.LG, stat.ML
Recent works such as VisProg and ViperGPT have smartly composed foundation models for visual reasoning-using large language models (LLMs) to produce programs that can be executed by pre-trained vision-language models. However, they operate in limited domains, such as 2D images, not fully exploiting the generalization o...
http://arxiv.org/pdf/2310.16035v1
2023-10-24T17:50:20
2023-10-24T17:50:20
{ "title": [ -0.0017347336, 0.0018854141, -0.042877197, -0.08093262, 0.014556885, -0.010559082, -0.04647827, -0.0042533875, 0.020874023, 0.012207031, -0.02305603, 0.011993408, -0.029693604, -0.021575928, -0.04574585, -0.00091362, -0.018615723, 0....
http://arxiv.org/abs/2310.16033v1
Visual Cropping Improves Zero-Shot Question Answering of Multimodal Large Language Models
Jiarui Zhang, Mahyar Khayatkhoei, Prateek Chhikara, Filip Ilievski
cs.CV, cs.CL
Multimodal Large Language Models (LLMs) have recently achieved promising zero-shot accuracy on visual question answering (VQA) -- a fundamental task affecting various downstream applications and domains. Given the great potential for the broad use of these models, it is important to investigate their limitations in dea...
http://arxiv.org/pdf/2310.16033v1
2023-10-24T17:48:04
2023-10-24T17:48:04
{ "title": [ -0.0054779053, -0.007896423, -0.015258789, -0.050933838, -0.040618896, -0.0032901764, -0.06341553, -0.007358551, 0.058044434, 0.022521973, -0.021240234, -0.007511139, -0.0013847351, -0.008140564, -0.07757568, 0.015075684, 0.0036792755, ...
http://arxiv.org/abs/2310.16028v1
What Algorithms can Transformers Learn? A Study in Length Generalization
Hattie Zhou, Arwen Bradley, Etai Littwin, Noam Razin, Omid Saremi, Josh Susskind, Samy Bengio, Preetum Nakkiran
cs.LG, cs.AI, cs.CL, stat.ML
Large language models exhibit surprising emergent generalization properties, yet also struggle on many simple reasoning tasks such as arithmetic and parity. This raises the question of if and when Transformer models can learn the true algorithm for solving a task. We study the scope of Transformers' abilities in the sp...
http://arxiv.org/pdf/2310.16028v1
2023-10-24T17:43:29
2023-10-24T17:43:29
{ "title": [ -0.03173828, 0.0035324097, -0.023529053, -0.013336182, -0.01222229, -0.006713867, -0.023284912, -0.029647827, -0.02947998, 0.03125, -0.043060303, 0.042755127, -0.027542114, -0.0025901794, -0.023956299, -0.012290955, -0.0037059784, 0....
http://arxiv.org/abs/2310.15987v1
Dissecting In-Context Learning of Translations in GPTs
Vikas Raunak, Hany Hassan Awadalla, Arul Menezes
cs.CL, cs.AI
Most of the recent work in leveraging Large Language Models (LLMs) such as GPT-3 for Machine Translation (MT) has focused on selecting the few-shot samples for prompting. In this work, we try to better understand the role of demonstration attributes for the in-context learning of translations through perturbations of h...
http://arxiv.org/pdf/2310.15987v1
2023-10-24T16:37:18
2023-10-24T16:37:18
{ "title": [ -0.016235352, -0.0055732727, -0.012123108, -0.035369873, -0.015068054, -0.010910034, -0.02798462, -0.011451721, -0.030319214, 0.029708862, -0.034240723, 0.030685425, -0.0065231323, -0.01727295, -0.018997192, 0.00019299984, 0.004962921, ...
http://arxiv.org/abs/2310.15970v3
Accented Speech Recognition With Accent-specific Codebooks
Darshan Prabhu, Preethi Jyothi, Sriram Ganapathy, Vinit Unni
cs.CL, cs.AI, cs.LG
Speech accents pose a significant challenge to state-of-the-art automatic speech recognition (ASR) systems. Degradation in performance across underrepresented accents is a severe deterrent to the inclusive adoption of ASR. In this work, we propose a novel accent adaptation approach for end-to-end ASR systems using cros...
http://arxiv.org/pdf/2310.15970v3
2023-10-27T02:54:29
2023-10-24T16:10:58
{ "title": [ -0.003446579, 0.009513855, -0.021362305, -0.02758789, -0.018035889, -0.012306213, -0.029129028, -0.010185242, -0.009811401, 0.04788208, -0.036071777, 0.01121521, 0.033691406, -0.03253174, -0.014213562, 0.017654419, 0.0066108704, 0.01...
http://arxiv.org/abs/2310.15961v1
Mixture of Tokens: Efficient LLMs through Cross-Example Aggregation
Szymon Antoniak, Sebastian Jaszczur, Michał Krutul, Maciej Pióro, Jakub Krajewski, Jan Ludziejewski, Tomasz Odrzygóźdź, Marek Cygan
cs.CL, cs.LG
Despite the promise of Mixture of Experts (MoE) models in increasing parameter counts of Transformer models while maintaining training and inference costs, their application carries notable drawbacks. The key strategy of these models is to, for each processed token, activate at most a few experts - subsets of an extens...
http://arxiv.org/pdf/2310.15961v1
2023-10-24T16:03:57
2023-10-24T16:03:57
{ "title": [ -0.020584106, -0.02267456, -0.059814453, -0.087768555, -0.0024757385, 0.00026893616, -0.05645752, 0.013626099, 0.0019111633, 0.020187378, -0.047424316, 0.03717041, 0.011894226, -0.01689148, 0.0019626617, -0.0023269653, -0.016464233, ...
http://arxiv.org/abs/2310.15959v1
NoteChat: A Dataset of Synthetic Doctor-Patient Conversations Conditioned on Clinical Notes
Junda Wang, Zonghai Yao, Zhichao Yang, Huixue Zhou, Rumeng Li, Xun Wang, Yucheng Xu, Hong Yu
cs.CL
The detailed clinical records drafted by doctors after each patient's visit are crucial for medical practitioners and researchers. Automating the creation of these notes with language models can reduce the workload of doctors. However, training such models can be difficult due to the limited public availability of conv...
http://arxiv.org/pdf/2310.15959v1
2023-10-24T15:59:43
2023-10-24T15:59:43
{ "title": [ 0.0062561035, -0.008735657, -0.022979736, -0.0692749, 0.0017652512, 0.0019426346, -0.04348755, 0.0030517578, 0.0010232925, 0.031463623, -0.04348755, 0.016693115, -0.005065918, -0.0012083054, -0.038848877, -0.023361206, -0.0034713745, ...