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/2311.07767v1
GreekT5: A Series of Greek Sequence-to-Sequence Models for News Summarization
Nikolaos Giarelis, Charalampos Mastrokostas, Nikos Karacapilidis
cs.CL, cs.AI, 68T07, 68T50, I.2.7
Text summarization (TS) is a natural language processing (NLP) subtask pertaining to the automatic formulation of a concise and coherent summary that covers the major concepts and topics from one or multiple documents. Recent advancements in deep learning have led to the development of abstractive summarization transfo...
http://arxiv.org/pdf/2311.07767v1
2023-11-13T21:33:12
2023-11-13T21:33:12
{ "title": [ -0.016342163, 0.0023994446, -0.021255493, -0.0703125, 0.00054836273, 0.01890564, -0.04837036, -0.006111145, 0.0024871826, 0.048706055, -0.018585205, 0.026031494, 0.022460938, -0.004322052, -0.042541504, 0.026428223, 0.0037441254, 0.0...
http://arxiv.org/abs/2311.07766v1
Vision-Language Integration in Multimodal Video Transformers (Partially) Aligns with the Brain
Dota Tianai Dong, Mariya Toneva
cs.CV, cs.AI, cs.CL, cs.LG
Integrating information from multiple modalities is arguably one of the essential prerequisites for grounding artificial intelligence systems with an understanding of the real world. Recent advances in video transformers that jointly learn from vision, text, and sound over time have made some progress toward this goal,...
http://arxiv.org/pdf/2311.07766v1
2023-11-13T21:32:37
2023-11-13T21:32:37
{ "title": [ 0.0104904175, 0.010925293, -0.047729492, -0.013458252, -0.012908936, -0.046844482, -0.0079956055, -0.020523071, -0.0075569153, 0.03604126, -0.030532837, -0.015853882, 0.047912598, -0.041870117, -0.007194519, -0.014022827, 0.024108887, ...
http://arxiv.org/abs/2311.07723v1
Generalization Analogies (GENIES): A Testbed for Generalizing AI Oversight to Hard-To-Measure Domains
Joshua Clymer, Garrett Baker, Rohan Subramani, Sam Wang
cs.AI, cs.CL, cs.LG
As AI systems become more intelligent and their behavior becomes more challenging to assess, they may learn to game the flaws of human feedback instead of genuinely striving to follow instructions; however, this risk can be mitigated by controlling how LLMs generalize human feedback to situations where it is unreliable...
http://arxiv.org/pdf/2311.07723v1
2023-11-13T20:07:36
2023-11-13T20:07:36
{ "title": [ -0.027709961, 0.009864807, -0.0076675415, -0.011428833, -0.014633179, -0.010520935, 0.000005066395, -0.045166016, 0.01474762, 0.017303467, -0.042938232, 0.02671814, -0.0048599243, -0.009292603, -0.027862549, -0.007537842, -0.017669678, ...
http://arxiv.org/abs/2311.07715v1
PolyIE: A Dataset of Information Extraction from Polymer Material Scientific Literature
Jerry Junyang Cheung, Yuchen Zhuang, Yinghao Li, Pranav Shetty, Wantian Zhao, Sanjeev Grampurohit, Rampi Ramprasad, Chao Zhang
cs.CL, cs.AI, cs.IR, cs.LG
Scientific information extraction (SciIE), which aims to automatically extract information from scientific literature, is becoming more important than ever. However, there are no existing SciIE datasets for polymer materials, which is an important class of materials used ubiquitously in our daily lives. To bridge this ...
http://arxiv.org/pdf/2311.07715v1
2023-11-13T19:56:18
2023-11-13T19:56:18
{ "title": [ -0.06829834, -0.017959595, -0.09899902, -0.07006836, -0.0010728836, -0.018478394, -0.068237305, 0.03213501, -0.010688782, 0.0154800415, -0.021942139, 0.018325806, -0.008155823, -0.030090332, 0.002248764, 0.01133728, -0.0046920776, 0....
http://arxiv.org/abs/2311.07703v1
Measuring Entrainment in Spontaneous Code-switched Speech
Debasmita Bhattacharya, Siying Ding, Alayna Nguyen, Julia Hirschberg
cs.CL, cs.SD, eess.AS
It is well-known that interlocutors who entrain to one another have more successful conversations than those who do not. Previous research has shown that interlocutors entrain on linguistic features in both written and spoken monolingual domains. More recent work on code-switched communication has also shown preliminar...
http://arxiv.org/pdf/2311.07703v1
2023-11-13T19:41:34
2023-11-13T19:41:34
{ "title": [ -0.02748108, -0.0055885315, -0.033416748, 0.00021028519, -0.010353088, -0.01184845, 0.0027809143, -0.0055770874, 0.0020313263, 0.039123535, -0.030151367, 0.014144897, 0.015258789, -0.02558899, -0.020080566, 0.005962372, -0.011703491, ...
http://arxiv.org/abs/2311.07700v1
AuthentiGPT: Detecting Machine-Generated Text via Black-Box Language Models Denoising
Zhen Guo, Shangdi Yu
cs.CL, cs.AI, cs.LG
Large language models (LLMs) have opened up enormous opportunities while simultaneously posing ethical dilemmas. One of the major concerns is their ability to create text that closely mimics human writing, which can lead to potential misuse, such as academic misconduct, disinformation, and fraud. To address this proble...
http://arxiv.org/pdf/2311.07700v1
2023-11-13T19:36:54
2023-11-13T19:36:54
{ "title": [ -0.018005371, 0.008277893, -0.036865234, -0.033111572, 0.012802124, -0.011878967, -0.03286743, -0.03274536, -0.01751709, 0.034484863, -0.0073890686, 0.033996582, 0.0035438538, -0.010482788, -0.041107178, 0.006412506, -0.03250122, 0.0...
http://arxiv.org/abs/2311.07692v1
On The Truthfulness of 'Surprisingly Likely' Responses of Large Language Models
Naman Goel
cs.LG, cs.AI, cs.CL, cs.GT
The surprisingly likely criterion in the seminal work of Prelec (the Bayesian Truth Serum) guarantees truthfulness in a game-theoretic multi-agent setting, by rewarding rational agents to maximise the expected information gain with their answers w.r.t. their probabilistic beliefs. We investigate the relevance of a simi...
http://arxiv.org/pdf/2311.07692v1
2023-11-13T19:21:25
2023-11-13T19:21:25
{ "title": [ -0.028396606, -0.007457733, -0.039154053, -0.05050659, -0.0027065277, -0.012931824, -0.046081543, -0.015113831, -0.00031971931, 0.015571594, -0.017074585, -0.0014257431, -0.013084412, -0.021011353, -0.04144287, 0.0023021698, -0.01852417,...
http://arxiv.org/abs/2311.07689v1
MART: Improving LLM Safety with Multi-round Automatic Red-Teaming
Suyu Ge, Chunting Zhou, Rui Hou, Madian Khabsa, Yi-Chia Wang, Qifan Wang, Jiawei Han, Yuning Mao
cs.CL
Red-teaming is a common practice for mitigating unsafe behaviors in Large Language Models (LLMs), which involves thoroughly assessing LLMs to identify potential flaws and addressing them with responsible and accurate responses. While effective, manual red-teaming is costly, and existing automatic red-teaming typically ...
http://arxiv.org/pdf/2311.07689v1
2023-11-13T19:13:29
2023-11-13T19:13:29
{ "title": [ -0.041778564, 0.0023097992, -0.05731201, -0.05984497, -0.019851685, 0.00079345703, -0.061920166, 0.00806427, -0.03643799, 0.015266418, -0.024261475, 0.047210693, -0.030639648, -0.037841797, 0.019851685, -0.014808655, -0.01071167, -0....
http://arxiv.org/abs/2311.07687v1
Language Model-In-The-Loop: Data Optimal Approach to Learn-To-Recommend Actions in Text Games
Arjun Vaithilingam Sudhakar, Prasanna Parthasarathi, Janarthanan Rajendran, Sarath Chandar
cs.CL, cs.AI, cs.LG
Large Language Models (LLMs) have demonstrated superior performance in language understanding benchmarks. CALM, a popular approach, leverages linguistic priors of LLMs -- GPT-2 -- for action candidate recommendations to improve the performance in text games in Jericho without environment-provided actions. However, CALM...
http://arxiv.org/pdf/2311.07687v1
2023-11-13T19:12:49
2023-11-13T19:12:49
{ "title": [ 0.0044059753, -0.008056641, 0.0049705505, -0.018661499, -0.02178955, -0.007965088, 0.0023975372, -0.0045204163, 0.013885498, 0.026382446, -0.045135498, 0.023010254, -0.020019531, -0.007980347, -0.027450562, -0.015319824, -0.018630981, ...
http://arxiv.org/abs/2311.07682v1
Fuse to Forget: Bias Reduction and Selective Memorization through Model Fusion
Kerem Zaman, Leshem Choshen, Shashank Srivastava
cs.CL, cs.AI, cs.LG
Model fusion research aims to aggregate the knowledge of multiple models to enhance performance by combining their weights. In this work, we study the inverse, investigating whether and how can model fusion interfere and reduce unwanted knowledge. We delve into the effects of model fusion on the evolution of learned sh...
http://arxiv.org/pdf/2311.07682v1
2023-11-13T19:02:56
2023-11-13T19:02:56
{ "title": [ 0.009750366, 0.0022449493, -0.031463623, -0.053710938, 0.008049011, -0.0115356445, -0.026977539, -0.0262146, 0.004184723, 0.030471802, -0.021362305, 0.002439499, 0.01713562, -0.020965576, -0.037475586, -0.0016021729, -0.007232666, 0....
http://arxiv.org/abs/2311.07575v1
SPHINX: The Joint Mixing of Weights, Tasks, and Visual Embeddings for Multi-modal Large Language Models
Ziyi Lin, Chris Liu, Renrui Zhang, Peng Gao, Longtian Qiu, Han Xiao, Han Qiu, Chen Lin, Wenqi Shao, Keqin Chen, Jiaming Han, Siyuan Huang, Yichi Zhang, Xuming He, Hongsheng Li, Yu Qiao
cs.CV, cs.AI, cs.CL, cs.LG
We present SPHINX, a versatile multi-modal large language model (MLLM) with a joint mixing of model weights, tuning tasks, and visual embeddings. First, for stronger vision-language alignment, we unfreeze the large language model (LLM) during pre-training, and introduce a weight mix strategy between LLMs trained by rea...
http://arxiv.org/pdf/2311.07575v1
2023-11-13T18:59:47
2023-11-13T18:59:47
{ "title": [ -0.0066833496, 0.000078856945, -0.052581787, -0.06567383, -0.002632141, -0.009437561, -0.068481445, -0.007820129, 0.020370483, 0.02658081, -0.017578125, 0.016525269, 0.0071792603, -0.022735596, -0.034118652, 0.008476257, -0.00315094, ...
http://arxiv.org/abs/2311.07564v1
Can Authorship Attribution Models Distinguish Speakers in Speech Transcripts?
Cristina Aggazzotti, Nicholas Andrews, Elizabeth Allyn Smith
cs.CL, cs.LG
Authorship verification is the problem of determining if two distinct writing samples share the same author and is typically concerned with the attribution of written text. In this paper, we explore the attribution of transcribed speech, which poses novel challenges. The main challenge is that many stylistic features, ...
http://arxiv.org/pdf/2311.07564v1
2023-11-13T18:54:17
2023-11-13T18:54:17
{ "title": [ -0.020507812, 0.030334473, -0.055664062, -0.039031982, -0.017440796, -0.009773254, -0.07019043, -0.02330017, -0.038269043, 0.03857422, 0.0031852722, 0.012107849, 0.0002553463, -0.026229858, -0.01033783, -0.007511139, -0.039367676, 0....
http://arxiv.org/abs/2311.07556v1
Using Natural Language Explanations to Improve Robustness of In-context Learning for Natural Language Inference
Xuanli He, Yuxiang Wu, Oana-Maria Camburu, Pasquale Minervini, Pontus Stenetorp
cs.CL
Recent studies have demonstrated that large language models (LLMs) excel in diverse tasks through in-context learning (ICL) facilitated by task-specific prompts and examples. However, the existing literature shows that ICL encounters performance deterioration when exposed to adversarial inputs. Enhanced performance has...
http://arxiv.org/pdf/2311.07556v1
2023-11-13T18:49:13
2023-11-13T18:49:13
{ "title": [ 0.004283905, -0.019561768, -0.025604248, -0.07159424, -0.0013799667, -0.0029277802, -0.035583496, -0.004360199, 0.0015630722, 0.0046463013, -0.044799805, 0.020339966, -0.004261017, -0.009094238, -0.013023376, -0.011154175, -0.010017395, ...
http://arxiv.org/abs/2311.07547v1
GPT-4V(ision) as A Social Media Analysis Engine
Hanjia Lyu, Jinfa Huang, Daoan Zhang, Yongsheng Yu, Xinyi Mou, Jinsheng Pan, Zhengyuan Yang, Zhongyu Wei, Jiebo Luo
cs.CV, cs.AI, cs.CL, cs.MM
Recent research has offered insights into the extraordinary capabilities of Large Multimodal Models (LMMs) in various general vision and language tasks. There is growing interest in how LMMs perform in more specialized domains. Social media content, inherently multimodal, blends text, images, videos, and sometimes audi...
http://arxiv.org/pdf/2311.07547v1
2023-11-13T18:36:50
2023-11-13T18:36:50
{ "title": [ 0.024383545, 0.0039520264, -0.023101807, -0.03543091, -0.00092840195, -0.02116394, -0.01234436, -0.029037476, -0.04815674, 0.052947998, -0.013206482, 0.015808105, 0.015686035, -0.029373169, -0.016647339, 0.0037956238, 0.0017194748, 0...
http://arxiv.org/abs/2311.07538v1
Leveraging Multiple Teachers for Test-Time Adaptation of Language-Guided Classifiers
Kangda Wei, Sayan Ghosh, Rakesh R. Menon, Shashank Srivastava
cs.CL, cs.LG
Recent approaches have explored language-guided classifiers capable of classifying examples from novel tasks when provided with task-specific natural language explanations, instructions or prompts (Sanh et al., 2022; R. Menon et al., 2022). While these classifiers can generalize in zero-shot settings, their task perfor...
http://arxiv.org/pdf/2311.07538v1
2023-11-13T18:28:25
2023-11-13T18:28:25
{ "title": [ -0.029785156, -0.034851074, -0.022262573, -0.03668213, -0.020263672, -0.014595032, -0.019943237, 0.00844574, -0.0065345764, 0.005214691, -0.04083252, 0.034423828, -0.05834961, 0.0025081635, -0.0335083, -0.04336548, -0.00932312, 0.048...
http://arxiv.org/abs/2311.07536v1
A Comprehensive Evaluation of GPT-4V on Knowledge-Intensive Visual Question Answering
Yunxin Li, Longyue Wang, Baotian Hu, Xinyu Chen, Wanqi Zhong, Chenyang Lyu, Min Zhang
cs.CL
The emergence of multimodal large models (MLMs) has significantly advanced the field of visual understanding, offering remarkable capabilities in the realm of visual question answering (VQA). Yet, the true challenge lies in the domain of knowledge-intensive VQA tasks, which necessitate not just recognition of visual el...
http://arxiv.org/pdf/2311.07536v1
2023-11-13T18:22:32
2023-11-13T18:22:32
{ "title": [ 0.006378174, 0.007896423, -0.00894928, -0.026153564, -0.010902405, -0.010169983, -0.0018949509, -0.014030457, 0.03378296, 0.023162842, -0.0309906, -0.0059394836, -0.017028809, -0.0110321045, -0.07373047, 0.0016021729, 0.0074806213, 0...
http://arxiv.org/abs/2311.07532v1
It's Not Easy Being Wrong: Evaluating Process of Elimination Reasoning in Large Language Models
Nishant Balepur, Shramay Palta, Rachel Rudinger
cs.CL
Chain-of-thought (COT) prompting can help large language models (LLMs) reason toward correct answers, but its efficacy in reasoning toward incorrect answers is unexplored. This strategy of process of elimination (PoE), when used with COT, has the potential to enhance interpretability in tasks like medical diagnoses of ...
http://arxiv.org/pdf/2311.07532v1
2023-11-13T18:18:22
2023-11-13T18:18:22
{ "title": [ -0.020385742, -0.013885498, -0.029663086, -0.06060791, -0.0026512146, -0.0006160736, -0.040985107, 0.0020332336, 0.01977539, 0.023757935, -0.015029907, -0.015365601, 0.0129776, -0.015083313, -0.055236816, 0.00085639954, -0.027297974, ...
http://arxiv.org/abs/2311.07509v1
A Benchmark to Understand the Role of Knowledge Graphs on Large Language Model's Accuracy for Question Answering on Enterprise SQL Databases
Juan Sequeda, Dean Allemang, Bryon Jacob
cs.AI, cs.CL, cs.DB
Enterprise applications of Large Language Models (LLMs) hold promise for question answering on enterprise SQL databases. However, the extent to which LLMs can accurately respond to enterprise questions in such databases remains unclear, given the absence of suitable Text-to-SQL benchmarks tailored to enterprise setting...
http://arxiv.org/pdf/2311.07509v1
2023-11-13T17:54:50
2023-11-13T17:54:50
{ "title": [ -0.01737976, -0.027069092, -0.044952393, -0.048461914, -0.043060303, -0.014373779, -0.06085205, 0.030944824, 0.025619507, 0.052093506, -0.034698486, 0.019302368, 0.012718201, -0.0062065125, -0.050048828, -0.014389038, 0.022018433, 0....
http://arxiv.org/abs/2311.07497v1
Multilingual Nonce Dependency Treebanks: Understanding how LLMs represent and process syntactic structure
David Arps, Laura Kallmeyer, Younes Samih, Hassan Sajjad
cs.CL
We introduce SPUD (Semantically Perturbed Universal Dependencies), a framework for creating nonce treebanks for the multilingual Universal Dependencies (UD) corpora. SPUD data satisfies syntactic argument structure, provides syntactic annotations, and ensures grammaticality via language-specific rules. We create nonce ...
http://arxiv.org/pdf/2311.07497v1
2023-11-13T17:36:58
2023-11-13T17:36:58
{ "title": [ -0.035461426, -0.024673462, -0.030883789, -0.08166504, -0.038238525, 0.02355957, -0.10308838, 0.030014038, -0.009887695, 0.03164673, -0.025970459, 0.031204224, -0.011268616, -0.021514893, -0.017486572, 0.0047798157, -0.014968872, 0.0...
http://arxiv.org/abs/2311.07491v1
A Step Closer to Comprehensive Answers: Constrained Multi-Stage Question Decomposition with Large Language Models
Hejing Cao, Zhenwei An, Jiazhan Feng, Kun Xu, Liwei Chen, Dongyan Zhao
cs.CL
While large language models exhibit remarkable performance in the Question Answering task, they are susceptible to hallucinations. Challenges arise when these models grapple with understanding multi-hop relations in complex questions or lack the necessary knowledge for a comprehensive response. To address this issue, w...
http://arxiv.org/pdf/2311.07491v1
2023-11-13T17:28:03
2023-11-13T17:28:03
{ "title": [ -0.01763916, -0.001876831, -0.026428223, -0.06750488, 0.00093221664, 0.011405945, -0.042541504, 0.0053710938, 0.0053100586, 0.008575439, -0.0078048706, 0.018249512, -0.018234253, -0.021224976, -0.026931763, 0.013664246, 0.006614685, ...
http://arxiv.org/abs/2311.07484v1
Psychometric Predictive Power of Large Language Models
Tatsuki Kuribayashi, Yohei Oseki, Timothy Baldwin
cs.CL, cs.AI
Next-word probabilities from language models have been shown to successfully simulate human reading behavior. Building on this, we show that, interestingly, instruction-tuned large language models (LLMs) yield worse psychometric predictive power (PPP) for human reading behavior than base LLMs with equivalent perplexiti...
http://arxiv.org/pdf/2311.07484v1
2023-11-13T17:19:14
2023-11-13T17:19:14
{ "title": [ -0.028442383, -0.01966858, -0.034698486, -0.050598145, -0.023345947, -0.0034675598, -0.054748535, -0.00592041, -0.0061416626, 0.02796936, -0.008544922, -0.0077209473, 0.0012073517, -0.032714844, -0.033477783, 0.008102417, -0.0069885254, ...
http://arxiv.org/abs/2311.07470v1
Finding and Editing Multi-Modal Neurons in Pre-Trained Transformer
Haowen Pan, Yixin Cao, Xiaozhi Wang, Xun Yang
cs.CL
Multi-modal large language models (LLM) have achieved powerful capabilities for visual semantic understanding in recent years. However, little is known about how LLMs comprehend visual information and interpret different modalities of features. In this paper, we propose a new method for identifying multi-modal neurons ...
http://arxiv.org/pdf/2311.07470v1
2023-11-13T17:03:02
2023-11-13T17:03:02
{ "title": [ 0.03439331, 0.027877808, -0.011138916, 0.0023288727, -0.0053482056, -0.028213501, -0.0019626617, -0.07006836, -0.022033691, 0.032196045, 0.0013370514, -0.014602661, 0.036712646, -0.058410645, 0.012001038, 0.0031852722, 0.0019388199, ...
http://arxiv.org/abs/2311.07469v1
InCA: Rethinking In-Car Conversational System Assessment Leveraging Large Language Models
Ken E. Friedl, Abbas Goher Khan, Soumya Ranjan Sahoo, Md Rashad Al Hasan Rony, Jana Germies, Christian Süß
cs.CL, cs.AI
The assessment of advanced generative large language models (LLMs) poses a significant challenge, given their heightened complexity in recent developments. Furthermore, evaluating the performance of LLM-based applications in various industries, as indicated by Key Performance Indicators (KPIs), is a complex undertaking...
http://arxiv.org/pdf/2311.07469v1
2023-11-13T17:02:06
2023-11-13T17:02:06
{ "title": [ -0.036590576, -0.021850586, -0.037597656, -0.009689331, -0.024612427, -0.020431519, -0.04220581, -0.003408432, -0.005756378, 0.024719238, -0.03463745, 0.00843811, -0.0033740997, -0.020751953, -0.041931152, -0.015792847, -0.025970459, ...
http://arxiv.org/abs/2311.07468v1
Are We Falling in a Middle-Intelligence Trap? An Analysis and Mitigation of the Reversal Curse
Ang Lv, Kaiyi Zhang, Shufang Xie, Quan Tu, Yuhan Chen, Ji-Rong Wen, Rui Yan
cs.CL, cs.AI, cs.LG
Recent studies have highlighted a phenomenon in large language models (LLMs) known as "the reversal curse," in which the order of knowledge entities in the training data biases the models' comprehension. For example, if a model is trained on sentences where entity A consistently appears before entity B, it can respond ...
http://arxiv.org/pdf/2311.07468v1
2023-11-13T17:01:12
2023-11-13T17:01:12
{ "title": [ -0.001964569, -0.006755829, -0.014305115, 0.00598526, 0.015090942, -0.027709961, 0.052490234, -0.052978516, -0.04562378, 0.045135498, -0.013717651, 0.007797241, -0.0077552795, -0.00137043, -0.06011963, -0.01713562, -0.0053138733, 0.0...
http://arxiv.org/abs/2311.07466v1
On Measuring Faithfulness of Natural Language Explanations
Letitia Parcalabescu, Anette Frank
cs.CL, cs.AI, cs.LG, 68Txx, I.2.7; I.2.10
Large language models (LLMs) can explain their own predictions, through post-hoc or Chain-of-Thought (CoT) explanations. However the LLM could make up reasonably sounding explanations that are unfaithful to its underlying reasoning. Recent work has designed tests that aim to judge the faithfulness of either post-hoc or...
http://arxiv.org/pdf/2311.07466v1
2023-11-13T16:53:51
2023-11-13T16:53:51
{ "title": [ -0.024002075, -0.004989624, -0.021636963, -0.07720947, -0.005672455, 0.013000488, -0.064697266, -0.022460938, -0.017486572, 0.019226074, -0.01574707, 0.012641907, -0.0138549805, -0.015823364, -0.028442383, 0.0006275177, -0.010772705, ...
http://arxiv.org/abs/2311.07463v1
MEGAVERSE: Benchmarking Large Language Models Across Languages, Modalities, Models and Tasks
Sanchit Ahuja, Divyanshu Aggarwal, Varun Gumma, Ishaan Watts, Ashutosh Sathe, Millicent Ochieng, Rishav Hada, Prachi Jain, Maxamed Axmed, Kalika Bali, Sunayana Sitaram
cs.CL
Recently, there has been a rapid advancement in research on Large Language Models (LLMs), resulting in significant progress in several Natural Language Processing (NLP) tasks. Consequently, there has been a surge in LLM evaluation research to comprehend the models' capabilities and limitations. However, much of this re...
http://arxiv.org/pdf/2311.07463v1
2023-11-13T16:45:37
2023-11-13T16:45:37
{ "title": [ -0.036865234, -0.020812988, -0.041900635, -0.06726074, -0.000017106533, 0.00712204, -0.072143555, 0.010116577, 0.0067749023, 0.030456543, -0.00010848045, 0.009399414, 0.013763428, -0.029800415, -0.034118652, 0.026031494, -0.008865356, ...
http://arxiv.org/abs/2311.07453v1
ChartCheck: An Evidence-Based Fact-Checking Dataset over Real-World Chart Images
Mubashara Akhtar, Nikesh Subedi, Vivek Gupta, Sahar Tahmasebi, Oana Cocarascu, Elena Simperl
cs.CL, cs.CV
Data visualizations are common in the real-world. We often use them in data sources such as scientific documents, news articles, textbooks, and social media to summarize key information in a visual form. Charts can also mislead its audience by communicating false information or biasing them towards a specific agenda. V...
http://arxiv.org/pdf/2311.07453v1
2023-11-13T16:35:29
2023-11-13T16:35:29
{ "title": [ 0.016571045, -0.0018224716, -0.035614014, -0.05831909, 0.005138397, -0.021408081, -0.026367188, -0.02331543, -0.00022161007, 0.028747559, -0.017562866, 0.0029182434, -0.0059661865, -0.028579712, -0.03817749, 0.00050115585, 0.0035877228, ...
http://arxiv.org/abs/2311.07445v1
Think Before You Speak: Cultivating Communication Skills of Large Language Models via Inner Monologue
Junkai Zhou, Liang Pang, Huawei Shen, Xueqi Cheng
cs.CL, cs.AI
The emergence of large language models (LLMs) further improves the capabilities of open-domain dialogue systems and can generate fluent, coherent, and diverse responses. However, LLMs still lack an important ability: communication skills, which makes them more like information seeking tools than anthropomorphic chatbot...
http://arxiv.org/pdf/2311.07445v1
2023-11-13T16:19:42
2023-11-13T16:19:42
{ "title": [ -0.03552246, -0.020095825, -0.039794922, -0.018661499, -0.020874023, -0.015289307, -0.025100708, -0.0055770874, -0.03744507, 0.011016846, -0.009246826, 0.015716553, 0.004600525, -0.032073975, -0.0070228577, -0.0107803345, -0.012886047, ...
http://arxiv.org/abs/2311.07439v2
Investigating Multi-Pivot Ensembling with Massively Multilingual Machine Translation Models
Alireza Mohammadshahi, Jannis Vamvas, Rico Sennrich
cs.CL, cs.AI, cs.LG
Massively multilingual machine translation models allow for the translation of a large number of languages with a single model, but have limited performance on low- and very-low-resource translation directions. Pivoting via high-resource languages remains a strong strategy for low-resource directions, and in this paper...
http://arxiv.org/pdf/2311.07439v2
2023-11-14T14:01:46
2023-11-13T16:15:20
{ "title": [ -0.04144287, -0.026657104, -0.013092041, -0.056243896, -0.036071777, 0.035369873, -0.064819336, 0.007785797, 0.015853882, 0.05682373, -0.017623901, 0.032836914, 0.027008057, -0.009757996, -0.024749756, 0.030578613, -0.0048217773, 0.0...
http://arxiv.org/abs/2311.07430v1
Controlled Text Generation for Black-box Language Models via Score-based Progressive Editor
Sangwon Yu, Changmin Lee, Hojin Lee, Sungroh Yoon
cs.CL
Despite recent progress in language models, generating constrained text for specific domains remains a challenge, particularly when utilizing black-box models that lack domain-specific knowledge. In this paper, we introduce ScoPE (Score-based Progressive Editor) generation, a novel approach for controlled text generati...
http://arxiv.org/pdf/2311.07430v1
2023-11-13T16:03:23
2023-11-13T16:03:23
{ "title": [ -0.02949524, -0.023727417, -0.03515625, -0.06488037, -0.027557373, 0.008514404, -0.07409668, 0.008171082, 0.0040016174, 0.03111267, -0.015609741, 0.00969696, -0.0063667297, -0.008430481, -0.056884766, -0.0044555664, -0.015945435, 0.0...
http://arxiv.org/abs/2311.07424v1
Hallucination Augmented Recitations for Language Models
Abdullatif Köksal, Renat Aksitov, Chung-Ching Chang
cs.CL, cs.AI
Attribution is a key concept in large language models (LLMs) as it enables control over information sources and enhances the factuality of LLMs. While existing approaches utilize open book question answering to improve attribution, factual datasets may reward language models to recall facts that they already know from ...
http://arxiv.org/pdf/2311.07424v1
2023-11-13T15:58:18
2023-11-13T15:58:18
{ "title": [ -0.011192322, 0.0016565323, -0.027801514, -0.05999756, -0.026565552, -0.000053465366, -0.058380127, -0.006385803, 0.010620117, 0.051116943, -0.010360718, 0.0013523102, 0.019592285, -0.010063171, -0.057434082, 0.026657104, 0.00017213821, ...
http://arxiv.org/abs/2311.07418v1
Speech-based Slot Filling using Large Language Models
Guangzhi Sun, Shutong Feng, Dongcheng Jiang, Chao Zhang, Milica Gašić, Philip C. Woodland
cs.CL, cs.SD, eess.AS
Recently, advancements in large language models (LLMs) have shown an unprecedented ability across various language tasks. This paper investigates the potential application of LLMs to slot filling with noisy ASR transcriptions, via both in-context learning and task-specific fine-tuning. Dedicated prompt designs and fine...
http://arxiv.org/pdf/2311.07418v1
2023-11-13T15:54:30
2023-11-13T15:54:30
{ "title": [ -0.016738892, -0.004722595, -0.023864746, -0.057128906, 0.006668091, 0.0146865845, -0.060394287, -0.003818512, -0.0043678284, 0.047424316, -0.04269409, -0.018066406, 0.0066375732, -0.002565384, -0.061767578, -0.016220093, -0.022720337, ...
http://arxiv.org/abs/2311.07397v1
An LLM-free Multi-dimensional Benchmark for MLLMs Hallucination Evaluation
Junyang Wang, Yuhang Wang, Guohai Xu, Jing Zhang, Yukai Gu, Haitao Jia, Ming Yan, Ji Zhang, Jitao Sang
cs.CL, cs.CV
Despite making significant progress in multi-modal tasks, current Multi-modal Large Language Models (MLLMs) encounter the significant challenge of hallucination, which may lead to harmful consequences. Therefore, evaluating MLLMs' hallucinations is becoming increasingly important in model improvement and practical appl...
http://arxiv.org/pdf/2311.07397v1
2023-11-13T15:25:42
2023-11-13T15:25:42
{ "title": [ -0.0045928955, -0.0075912476, -0.01574707, -0.04248047, -0.03829956, 0.00095033646, -0.021133423, -0.010971069, -0.0068893433, 0.045166016, 0.02268982, 0.031066895, 0.007789612, -0.021408081, -0.01751709, 0.022415161, -0.009674072, 0...
http://arxiv.org/abs/2311.07387v1
Assessing Logical Puzzle Solving in Large Language Models: Insights from a Minesweeper Case Study
Yinghao Li, Haorui Wang, Chao Zhang
cs.CL
Large Language Models (LLMs) have shown remarkable proficiency in language understanding and have been successfully applied to a variety of real-world tasks through task-specific fine-tuning or prompt engineering. Despite these advancements, it remains an open question whether LLMs are fundamentally capable of reasonin...
http://arxiv.org/pdf/2311.07387v1
2023-11-13T15:11:26
2023-11-13T15:11:26
{ "title": [ -0.039123535, -0.022735596, -0.015975952, 0.0015745163, -0.025634766, -0.0128479, -0.030090332, -0.014167786, -0.010292053, 0.026275635, -0.014297485, 0.0022068024, 0.0010814667, -0.018920898, -0.033813477, -0.01272583, -0.03479004, ...
http://arxiv.org/abs/2311.07383v1
LM-Polygraph: Uncertainty Estimation for Language Models
Ekaterina Fadeeva, Roman Vashurin, Akim Tsvigun, Artem Vazhentsev, Sergey Petrakov, Kirill Fedyanin, Daniil Vasilev, Elizaveta Goncharova, Alexander Panchenko, Maxim Panov, Timothy Baldwin, Artem Shelmanov
cs.CL, cs.LG
Recent advancements in the capabilities of large language models (LLMs) have paved the way for a myriad of groundbreaking applications in various fields. However, a significant challenge arises as these models often "hallucinate", i.e., fabricate facts without providing users an apparent means to discern the veracity o...
http://arxiv.org/pdf/2311.07383v1
2023-11-13T15:08:59
2023-11-13T15:08:59
{ "title": [ -0.06567383, -0.016143799, -0.049224854, -0.028366089, -0.020141602, -0.0016613007, -0.052520752, -0.009063721, -0.0030498505, 0.0073661804, -0.010894775, 0.0031814575, 0.006275177, -0.033172607, -0.02142334, 0.012176514, -0.013900757, ...
http://arxiv.org/abs/2311.07635v1
Past as a Guide: Leveraging Retrospective Learning for Python Code Completion
Seunggyoon Shin, Seunggyu Chang, Sungjoon Choi
cs.SE, cs.AI, cs.CL
This work presents Past as a Guide (PaG), a simple approach for Large Language Models (LLMs) to improve the coding capabilities by integrating the past history with interactive and iterative code refinements. To be specific, inspired by human cognitive processes, the proposed method enables LLMs to utilize previous pro...
http://arxiv.org/pdf/2311.07635v1
2023-11-13T14:40:33
2023-11-13T14:40:33
{ "title": [ -0.0036201477, -0.022598267, -0.057434082, -0.09277344, 0.0021438599, -0.0107803345, -0.035614014, 0.04736328, 0.037261963, 0.03543091, -0.031555176, 0.003686905, -0.009857178, -0.0069084167, -0.07366943, -0.0104599, -0.0034217834, 0...
http://arxiv.org/abs/2311.07362v2
Volcano: Mitigating Multimodal Hallucination through Self-Feedback Guided Revision
Seongyun Lee, Sue Hyun Park, Yongrae Jo, Minjoon Seo
cs.CL, cs.CV
Large multimodal models (LMMs) suffer from multimodal hallucination, where they provide incorrect responses misaligned with the given visual information. Recent works have conjectured that one of the reasons behind multimodal hallucination might be due to the vision encoder failing to ground on the image properly. To m...
http://arxiv.org/pdf/2311.07362v2
2023-11-14T06:04:31
2023-11-13T14:26:24
{ "title": [ -0.016174316, 0.009681702, -0.031280518, -0.034088135, -0.019119263, -0.007980347, -0.03152466, -0.030410767, -0.03186035, 0.031402588, -0.0049209595, 0.028137207, 0.0095825195, -0.02607727, -0.020996094, 0.008041382, 0.013824463, 0....
http://arxiv.org/abs/2311.07361v1
The Impact of Large Language Models on Scientific Discovery: a Preliminary Study using GPT-4
Microsoft Research AI4Science, Microsoft Azure Quantum
cs.CL, cs.AI
In recent years, groundbreaking advancements in natural language processing have culminated in the emergence of powerful large language models (LLMs), which have showcased remarkable capabilities across a vast array of domains, including the understanding, generation, and translation of natural language, and even tasks...
http://arxiv.org/pdf/2311.07361v1
2023-11-13T14:26:12
2023-11-13T14:26:12
{ "title": [ -0.037200928, -0.014450073, -0.045806885, -0.02760315, -0.007396698, -0.0064430237, -0.033294678, -0.008605957, -0.00050354004, 0.022628784, 0.010475159, -0.0057525635, -0.012428284, -0.029464722, -0.03970337, 0.013160706, -0.019134521, ...
http://arxiv.org/abs/2311.07314v1
Semi-automatic Data Enhancement for Document-Level Relation Extraction with Distant Supervision from Large Language Models
Junpeng Li, Zixia Jia, Zilong Zheng
cs.CL, cs.AI
Document-level Relation Extraction (DocRE), which aims to extract relations from a long context, is a critical challenge in achieving fine-grained structural comprehension and generating interpretable document representations. Inspired by recent advances in in-context learning capabilities emergent from large language ...
http://arxiv.org/pdf/2311.07314v1
2023-11-13T13:10:44
2023-11-13T13:10:44
{ "title": [ -0.0033226013, -0.0048332214, -0.045440674, -0.10534668, 0.007068634, 0.003545761, -0.06604004, 0.012573242, 0.030593872, 0.008361816, -0.035217285, -0.009605408, 0.011802673, -0.011222839, -0.035858154, -0.010955811, -0.018936157, 0...
http://arxiv.org/abs/2311.07311v1
Do large language models and humans have similar behaviors in causal inference with script knowledge?
Xudong Hong, Margarita Ryzhova, Daniel Adrian Biondi, Vera Demberg
cs.CL, cs.AI, I.2.7; I.2.0
Recently, large pre-trained language models (LLMs) have demonstrated superior language understanding abilities, including zero-shot causal reasoning. However, it is unclear to what extent their capabilities are similar to human ones. We here study the processing of an event $B$ in a script-based story, which causally d...
http://arxiv.org/pdf/2311.07311v1
2023-11-13T13:05:15
2023-11-13T13:05:15
{ "title": [ -0.022842407, -0.018478394, -0.029129028, -0.042663574, -0.058044434, -0.016098022, -0.0491333, 0.025436401, -0.010147095, -0.00566864, -0.052825928, 0.0065460205, 0.008399963, -0.022827148, 0.021240234, -0.028717041, 0.009895325, 0....
http://arxiv.org/abs/2311.07296v1
BIDRN: A Method of Bidirectional Recurrent Neural Network for Sentiment Analysis
Dr. D Muthusankar, Dr. P Kaladevi, Dr. V R Sadasivam, R Praveen
cs.CL, cs.IR
Text mining research has grown in importance in recent years due to the tremendous increase in the volume of unstructured textual data. This has resulted in immense potential as well as obstacles in the sector, which may be efficiently addressed with adequate analytical and study methods. Deep Bidirectional Recurrent N...
http://arxiv.org/pdf/2311.07296v1
2023-11-13T12:36:53
2023-11-13T12:36:53
{ "title": [ 0.0025520325, -0.0039863586, -0.031982422, -0.06173706, 0.01977539, -0.0028495789, -0.015434265, -0.015380859, 0.0206604, 0.0491333, -0.009300232, 0.0044784546, -0.014137268, -0.0074882507, -0.07574463, 0.0067634583, -0.022644043, 0....
http://arxiv.org/abs/2311.07277v1
AdaCCD: Adaptive Semantic Contrasts Discovery based Cross Lingual Adaptation for Code Clone Detection
Yangkai Du, Tengfei Ma, Lingfei Wu, Xuhong Zhang, Shouling Ji
cs.SE, cs.CL
Code Clone Detection, which aims to retrieve functionally similar programs from large code bases, has been attracting increasing attention. Modern software often involves a diverse range of programming languages. However, current code clone detection methods are generally limited to only a few popular programming langu...
http://arxiv.org/pdf/2311.07277v1
2023-11-13T12:20:48
2023-11-13T12:20:48
{ "title": [ -0.031555176, -0.017440796, -0.038970947, -0.034851074, -0.0138168335, -0.008735657, -0.0446167, -0.009063721, 0.0037784576, 0.055541992, -0.02507019, 0.021606445, 0.030380249, -0.009140015, -0.019332886, 0.010375977, -0.010360718, 0...
http://arxiv.org/abs/2311.07264v1
Danish Foundation Models
Kenneth Enevoldsen, Lasse Hansen, Dan S. Nielsen, Rasmus A. F. Egebæk, Søren V. Holm, Martin C. Nielsen, Martin Bernstorff, Rasmus Larsen, Peter B. Jørgensen, Malte Højmark-Bertelsen, Peter B. Vahlstrup, Per Møldrup-Dalum, Kristoffer Nielbo
cs.CL
Large language models, sometimes referred to as foundation models, have transformed multiple fields of research. However, smaller languages risk falling behind due to high training costs and small incentives for large companies to train these models. To combat this, the Danish Foundation Models project seeks to provide...
http://arxiv.org/pdf/2311.07264v1
2023-11-13T12:03:52
2023-11-13T12:03:52
{ "title": [ -0.022247314, 0.0036201477, -0.052368164, -0.011680603, -0.048461914, -0.018997192, -0.061645508, -0.0050964355, 0.01247406, 0.045074463, 0.016723633, -0.034820557, 0.019546509, -0.032104492, 0.014549255, 0.0058937073, 0.026763916, -...
http://arxiv.org/abs/2311.07237v1
In Search of the Long-Tail: Systematic Generation of Long-Tail Knowledge via Logical Rule Guided Search
Huihan Li, Yuting Ning, Zeyi Liao, Siyuan Wang, Xiang Lorraine Li, Ximing Lu, Faeze Brahman, Wenting Zhao, Yejin Choi, Xiang Ren
cs.CL, cs.AI
Since large language models have approached human-level performance on many tasks, it has become increasingly harder for researchers to find tasks that are still challenging to the models. Failure cases usually come from the long-tail distribution - data that an oracle language model could assign a probability on the l...
http://arxiv.org/pdf/2311.07237v1
2023-11-13T10:56:59
2023-11-13T10:56:59
{ "title": [ -0.01878357, -0.019317627, -0.055847168, -0.06088257, -0.003917694, -0.0061569214, -0.025802612, 0.0024433136, -0.014717102, 0.021331787, -0.020690918, 0.015213013, -0.03338623, -0.02999878, -0.04675293, -0.031082153, 0.0062713623, 0...
http://arxiv.org/abs/2311.07230v1
How are Prompts Different in Terms of Sensitivity?
Sheng Lu, Hendrik Schuff, Iryna Gurevych
cs.CL
In-context learning (ICL) has become one of the most popular learning paradigms. While there is a growing body of literature focusing on prompt engineering, there is a lack of systematic analysis comparing the effects of prompts across different models and tasks. To address this gap, we present a comprehensive prompt a...
http://arxiv.org/pdf/2311.07230v1
2023-11-13T10:52:01
2023-11-13T10:52:01
{ "title": [ 0.02809143, 0.0118255615, 0.016555786, 0.033569336, -0.018325806, -0.017196655, 0.009422302, -0.06567383, 0.01096344, 0.030181885, -0.0017795563, -0.0143585205, -0.016342163, -0.0256958, -0.01335144, -0.01991272, 0.022399902, -0.0131...
http://arxiv.org/abs/2311.07217v1
Troubles and Failures in Interactional Language. Towards a Linguistically Informed Taxonomy
Martina Wiltschko
cs.CL
The goal of this talk is to introduce a systematic research agenda which aims to understand the nature of interaction between humans and artificial conversational agents (CA) (henceforth humanmachine interaction, HMI). Specifically, we shall take an explicit linguistic perspective focusing on linguistically defined var...
http://arxiv.org/pdf/2311.07217v1
2023-11-13T10:24:51
2023-11-13T10:24:51
{ "title": [ -0.006099701, -0.010559082, -0.010238647, -0.04562378, 0.00090026855, 0.004562378, -0.029083252, -0.025131226, -0.0076293945, 0.02897644, -0.0082473755, 0.010765076, -0.011230469, -0.023071289, -0.04574585, 0.00012791157, -0.015716553, ...
http://arxiv.org/abs/2311.07215v1
Coffee: Boost Your Code LLMs by Fixing Bugs with Feedback
Seungjun Moon, Yongho Song, Hyungjoo Chae, Dongjin Kang, Taeyoon Kwon, Kai Tzu-iunn Ong, Seung-won Hwang, Jinyoung Yeo
cs.CL, cs.SE
Code editing is an essential step towards reliable program synthesis to automatically correct critical errors generated from code LLMs. Recent studies have demonstrated that closed-source LLMs (i.e., ChatGPT and GPT-4) are capable of generating corrective feedback to edit erroneous inputs. However, it remains challengi...
http://arxiv.org/pdf/2311.07215v1
2023-11-13T10:15:19
2023-11-13T10:15:19
{ "title": [ -0.03756714, -0.044647217, -0.043701172, -0.04232788, -0.034332275, -0.0024795532, -0.032836914, 0.028442383, -0.027679443, -0.004207611, -0.038391113, 0.021621704, -0.0019779205, -0.037750244, 0.018539429, -0.051513672, -0.0003607273, ...
http://arxiv.org/abs/2311.07194v1
Exploring the Dialogue Comprehension Ability of Large Language Models
Shuaijie She, Shujian Huang, Xingyun Wang, Yanke Zhou, Jiajun Chen
cs.CL
The recent emergence of large language models (LLMs) have attracted considerable attention. LLMs may interact with users in the form of dialogue and generate responses following their instructions, which naturally require dialogue comprehension abilities. Without correct comprehension of the dialogue, the model may ine...
http://arxiv.org/pdf/2311.07194v1
2023-11-13T09:32:12
2023-11-13T09:32:12
{ "title": [ -0.040893555, -0.007221222, -0.027542114, -0.064575195, -0.0141067505, 0.011207581, -0.07879639, 0.0049934387, 0.02243042, 0.02078247, -0.017700195, -0.025054932, -0.0027446747, -0.018249512, -0.059936523, 0.004901886, -0.025634766, ...
http://arxiv.org/abs/2311.07172v1
VerityMath: Advancing Mathematical Reasoning by Self-Verification Through Unit Consistency
Vernon Toh, Ratish Puduppully, Nancy F. Chen
cs.CL, cs.PL
Large Language Models (LLMs) combined with program-based solving techniques are increasingly demonstrating proficiency in mathematical reasoning. However, such progress is mostly demonstrated in closed-source models such as OpenAI-GPT4 and Claude. In this paper, we seek to study the performance of strong open-source LL...
http://arxiv.org/pdf/2311.07172v1
2023-11-13T09:06:58
2023-11-13T09:06:58
{ "title": [ -0.018554688, -0.000035583973, -0.060272217, -0.0690918, 0.02658081, -0.019470215, -0.01448822, -0.006462097, -0.00819397, 0.02279663, -0.009689331, 0.009002686, 0.012870789, -0.03741455, -0.023788452, 0.0054626465, -0.0073623657, 0....
http://arxiv.org/abs/2311.07171v1
calamanCy: A Tagalog Natural Language Processing Toolkit
Lester James V. Miranda
cs.CL
We introduce calamanCy, an open-source toolkit for constructing natural language processing (NLP) pipelines for Tagalog. It is built on top of spaCy, enabling easy experimentation and integration with other frameworks. calamanCy addresses the development gap by providing a consistent API for building NLP applications a...
http://arxiv.org/pdf/2311.07171v1
2023-11-13T09:06:43
2023-11-13T09:06:43
{ "title": [ 0.0019130707, -0.036315918, -0.017654419, -0.05718994, -0.020523071, 0.008880615, -0.044311523, 0.009536743, -0.008880615, 0.04421997, -0.031555176, 0.021942139, 0.014122009, -0.003540039, -0.032440186, -0.001490593, -0.0056991577, 0...
http://arxiv.org/abs/2311.07167v1
STEER: Unified Style Transfer with Expert Reinforcement
Skyler Hallinan, Faeze Brahman, Ximing Lu, Jaehun Jung, Sean Welleck, Yejin Choi
cs.CL, cs.AI
While text style transfer has many applications across natural language processing, the core premise of transferring from a single source style is unrealistic in a real-world setting. In this work, we focus on arbitrary style transfer: rewriting a text from an arbitrary, unknown style to a target style. We propose ST...
http://arxiv.org/pdf/2311.07167v1
2023-11-13T09:02:30
2023-11-13T09:02:30
{ "title": [ -0.022232056, -0.0054397583, -0.041534424, -0.06088257, 0.010070801, 0.008201599, -0.03564453, -0.010612488, -0.01234436, 0.0362854, 0.012519836, 0.01687622, 0.02027893, -0.032165527, -0.00025057793, 0.025177002, 0.00459671, 0.023147...
http://arxiv.org/abs/2311.07161v1
Developing a Named Entity Recognition Dataset for Tagalog
Lester James V. Miranda
cs.CL
We present the development of a Named Entity Recognition (NER) dataset for Tagalog. This corpus helps fill the resource gap present in Philippine languages today, where NER resources are scarce. The texts were obtained from a pretraining corpora containing news reports, and were labeled by native speakers in an iterati...
http://arxiv.org/pdf/2311.07161v1
2023-11-13T08:56:47
2023-11-13T08:56:47
{ "title": [ 0.0011940002, -0.007987976, -0.040252686, -0.06365967, -0.0031700134, -0.007659912, -0.054016113, 0.0030727386, -0.01007843, 0.043548584, -0.05001831, 0.023345947, -0.004173279, -0.0004863739, -0.047546387, -0.016281128, 0.0009860992, ...
http://arxiv.org/abs/2311.07150v1
Interaction is all You Need? A Study of Robots Ability to Understand and Execute
Kushal Koshti, Nidhir Bhavsar
cs.RO, cs.AI, cs.CL, cs.CV
This paper aims to address a critical challenge in robotics, which is enabling them to operate seamlessly in human environments through natural language interactions. Our primary focus is to equip robots with the ability to understand and execute complex instructions in coherent dialogs to facilitate intricate task-sol...
http://arxiv.org/pdf/2311.07150v1
2023-11-13T08:39:06
2023-11-13T08:39:06
{ "title": [ -0.0066452026, 0.0022678375, -0.031097412, -0.028182983, 0.019958496, -0.029891968, -0.0052719116, -0.024017334, -0.015808105, 0.024734497, -0.029022217, -0.00050497055, -0.0034980774, -0.015205383, -0.040740967, -0.020324707, 0.01169586...
http://arxiv.org/abs/2311.07138v1
WaterBench: Towards Holistic Evaluation of Watermarks for Large Language Models
Shangqing Tu, Yuliang Sun, Yushi Bai, Jifan Yu, Lei Hou, Juanzi Li
cs.CL, cs.AI
To mitigate the potential misuse of large language models (LLMs), recent research has developed watermarking algorithms, which restrict the generation process to leave an invisible trace for watermark detection. Due to the two-stage nature of the task, most studies evaluate the generation and detection separately, ther...
http://arxiv.org/pdf/2311.07138v1
2023-11-13T08:09:01
2023-11-13T08:09:01
{ "title": [ -0.019012451, 0.0032978058, -0.036102295, -0.06774902, 0.020736694, 0.0021629333, -0.058166504, -0.008384705, 0.0031147003, 0.036621094, -0.009109497, 0.008544922, 0.009284973, -0.016067505, -0.04272461, 0.01512146, -0.021713257, 0.0...
http://arxiv.org/abs/2311.07115v1
Gen-Z: Generative Zero-Shot Text Classification with Contextualized Label Descriptions
Sachin Kumar, Chan Young Park, Yulia Tsvetkov
cs.CL
Language model (LM) prompting--a popular paradigm for solving NLP tasks--has been shown to be susceptible to miscalibration and brittleness to slight prompt variations, caused by its discriminative prompting approach, i.e., predicting the label given the input. To address these issues, we propose Gen-Z--a generative pr...
http://arxiv.org/pdf/2311.07115v1
2023-11-13T07:12:57
2023-11-13T07:12:57
{ "title": [ -0.010887146, 0.000726223, -0.035095215, -0.055664062, 0.0011844635, -0.0052833557, -0.026321411, -0.008377075, 0.0141067505, 0.03793335, -0.028259277, 0.030960083, -0.007549286, 0.0027503967, -0.050964355, -0.00016629696, -0.025482178, ...
http://arxiv.org/abs/2311.07102v1
Fovea Transformer: Efficient Long-Context Modeling with Structured Fine-to-Coarse Attention
Ziwei He, Jian Yuan, Le Zhou, Jingwen Leng, Bo Jiang
cs.CL
The quadratic complexity of self-attention in Transformers has hindered the processing of long text. To alleviate this problem, previous works have proposed to sparsify the attention matrix, taking advantage of the observation that crucial information about a token can be derived from its neighbors. These methods typic...
http://arxiv.org/pdf/2311.07102v1
2023-11-13T06:24:27
2023-11-13T06:24:27
{ "title": [ 0.009544373, 0.01448822, -0.042236328, -0.076660156, 0.026992798, -0.007572174, -0.02722168, -0.009086609, 0.018814087, 0.031188965, -0.028198242, -0.008224487, 0.015167236, -0.016174316, -0.034851074, 0.0010070801, 0.0005326271, 0.0...
http://arxiv.org/abs/2311.07099v1
Explanation-aware Soft Ensemble Empowers Large Language Model In-context Learning
Yue Yu, Jiaming Shen, Tianqi Liu, Zhen Qin, Jing Nathan Yan, Jialu Liu, Chao Zhang, Michael Bendersky
cs.CL, cs.AI
Large language models (LLMs) have shown remarkable capabilities in various natural language understanding tasks. With only a few demonstration examples, these LLMs can quickly adapt to target tasks without expensive gradient updates. Common strategies to boost such 'in-context' learning ability are to ensemble multiple...
http://arxiv.org/pdf/2311.07099v1
2023-11-13T06:13:38
2023-11-13T06:13:38
{ "title": [ -0.008255005, -0.018035889, -0.037506104, -0.083984375, -0.0058517456, -0.0038051605, -0.035217285, 0.04510498, 0.00016248226, 0.016067505, -0.033416748, -0.0017366409, 0.023895264, -0.02229309, -0.0036811829, -0.0025291443, 0.0000905394...
http://arxiv.org/abs/2311.07093v2
On the Effectiveness of ASR Representations in Real-world Noisy Speech Emotion Recognition
Xiaohan Shi, Jiajun He, Xingfeng Li, Tomoki Toda
cs.SD, cs.CL, eess.AS
This paper proposes an efficient attempt to noisy speech emotion recognition (NSER). Conventional NSER approaches have proven effective in mitigating the impact of artificial noise sources, such as white Gaussian noise, but are limited to non-stationary noises in real-world environments due to their complexity and unce...
http://arxiv.org/pdf/2311.07093v2
2023-11-14T13:09:51
2023-11-13T05:45:55
{ "title": [ 0.0062713623, 0.001156807, -0.021957397, -0.06262207, 0.0060653687, -0.0112838745, -0.02067566, 0.011077881, 0.011657715, 0.04663086, -0.022384644, -0.014205933, 0.042755127, -0.020843506, -0.04751587, 0.008377075, 0.008895874, 0.041...
http://arxiv.org/abs/2311.07092v1
To Tell The Truth: Language of Deception and Language Models
Bodhisattwa Prasad Majumder, Sanchaita Hazra
cs.CL, cs.AI
Text-based misinformation permeates online discourses, yet evidence of people's ability to discern truth from such deceptive textual content is scarce. We analyze a novel TV game show data where conversations in a high-stake environment between individuals with conflicting objectives result in lies. We investigate the ...
http://arxiv.org/pdf/2311.07092v1
2023-11-13T05:40:11
2023-11-13T05:40:11
{ "title": [ -0.039001465, -0.014831543, -0.039886475, -0.024414062, -0.009147644, -0.0019798279, -0.015556335, -0.014678955, -0.03164673, 0.032165527, 0.003704071, 0.012680054, 0.009986877, -0.0256958, -0.020721436, 0.009017944, -0.02381897, 0.0...
http://arxiv.org/abs/2311.07076v1
On the Discussion of Large Language Models: Symmetry of Agents and Interplay with Prompts
Qineng Wang, Zihao Wang, Ying Su, Yangqiu Song
cs.CL
Two ways has been discussed to unlock the reasoning capability of a large language model. The first one is prompt engineering and the second one is to combine the multiple inferences of large language models, or the multi-agent discussion. Theoretically, this paper justifies the multi-agent discussion mechanisms from t...
http://arxiv.org/pdf/2311.07076v1
2023-11-13T04:56:48
2023-11-13T04:56:48
{ "title": [ -0.022888184, -0.0063438416, -0.035949707, -0.035003662, -0.016601562, -0.022750854, -0.0368042, 0.005657196, -0.02444458, 0.031921387, -0.041931152, 0.0033988953, 0.0206604, -0.018203735, -0.008766174, -0.009422302, -0.0055389404, 0...
http://arxiv.org/abs/2311.07070v1
Explain-then-Translate: An Analysis on Improving Program Translation with Self-generated Explanations
Zilu Tang, Mayank Agarwal, Alex Shypula, Bailin Wang, Derry Wijaya, Jie Chen, Yoon Kim
cs.CL
This work explores the use of self-generated natural language explanations as an intermediate step for code-to-code translation with language models. Across three types of explanations and 19 programming languages constructed from the MultiPL-E dataset, we find the explanations to be particularly effective in the zero-...
http://arxiv.org/pdf/2311.07070v1
2023-11-13T04:28:49
2023-11-13T04:28:49
{ "title": [ -0.044952393, -0.016189575, 0.0014858246, -0.050445557, -0.014007568, 0.01852417, -0.054718018, -0.003414154, -0.022827148, 0.024673462, 0.016036987, 0.017654419, -0.028289795, -0.026016235, -0.044677734, 0.0018157959, 0.001083374, -...
http://arxiv.org/abs/2311.07066v1
Context Consistency between Training and Testing in Simultaneous Machine Translation
Meizhi Zhong, Lemao Liu, Kehai Chen, Mingming Yang, Min Zhang
cs.CL
Simultaneous Machine Translation (SiMT) aims to yield a real-time partial translation with a monotonically growing the source-side context. However, there is a counterintuitive phenomenon about the context usage between training and testing: e.g., the wait-k testing model consistently trained with wait-k is much worse ...
http://arxiv.org/pdf/2311.07066v1
2023-11-13T04:11:32
2023-11-13T04:11:32
{ "title": [ -0.03265381, -0.0066184998, -0.02607727, -0.072021484, -0.011543274, 0.006340027, -0.04083252, 0.008338928, -0.033569336, 0.05847168, -0.008979797, 0.022109985, 0.018051147, -0.012252808, -0.029129028, 0.0016832352, -0.014129639, 0.0...
http://arxiv.org/abs/2311.07064v1
PROPANE: Prompt design as an inverse problem
Rimon Melamed, Lucas H. McCabe, Tanay Wakhare, Yejin Kim, H. Howie Huang, Enric Boix-Adsera
cs.CL
Carefully-designed prompts are key to inducing desired behavior in Large Language Models (LLMs). As a result, great effort has been dedicated to engineering prompts that guide LLMs toward particular behaviors. In this work, we propose an automatic prompt optimization framework, PROPANE, which aims to find a prompt that...
http://arxiv.org/pdf/2311.07064v1
2023-11-13T04:08:49
2023-11-13T04:08:49
{ "title": [ -0.0009121895, 0.0017604828, -0.04171753, -0.0031471252, -0.015594482, -0.031341553, -0.0028743744, -0.03186035, -0.031082153, -0.02067566, -0.040924072, -0.008766174, -0.06072998, -0.04534912, 0.01675415, -0.04574585, -0.021255493, ...
http://arxiv.org/abs/2311.07052v1
Towards the Law of Capacity Gap in Distilling Language Models
Chen Zhang, Dawei Song, Zheyu Ye, Yan Gao
cs.CL, cs.AI, cs.LG
Language model (LM) distillation is a trending area that aims to distil the knowledge resided in a large teacher LM to a small student one. While various methods have been proposed to push the distillation to its limits, it is still a pain distilling LMs when a large capacity gap is exhibited between the teacher and th...
http://arxiv.org/pdf/2311.07052v1
2023-11-13T03:36:18
2023-11-13T03:36:18
{ "title": [ -0.033966064, -0.0022144318, -0.05419922, -0.07952881, -0.004180908, 0.01398468, -0.09106445, 0.0042915344, 0.023147583, 0.020401001, -0.04559326, -0.020141602, 0.002948761, -0.008705139, -0.06329346, -0.01537323, -0.01737976, 0.0517...
http://arxiv.org/abs/2311.07037v1
Phonological Level wav2vec2-based Mispronunciation Detection and Diagnosis Method
Mostafa Shahin, Julien Epps, Beena Ahmed
eess.AS, cs.AI, cs.CL
The automatic identification and analysis of pronunciation errors, known as Mispronunciation Detection and Diagnosis (MDD) plays a crucial role in Computer Aided Pronunciation Learning (CAPL) tools such as Second-Language (L2) learning or speech therapy applications. Existing MDD methods relying on analysing phonemes c...
http://arxiv.org/pdf/2311.07037v1
2023-11-13T02:41:41
2023-11-13T02:41:41
{ "title": [ 0.012542725, -0.006313324, -0.06085205, -0.046722412, -0.0075950623, -0.02519226, -0.011276245, 0.021194458, -0.018814087, 0.055999756, -0.01965332, -0.011558533, 0.047180176, -0.05154419, -0.004119873, -0.0013513565, -0.014404297, 0...
http://arxiv.org/abs/2311.07032v1
ExpNote: Black-box Large Language Models are Better Task Solvers with Experience Notebook
Wangtao Sun, Xuanqing Yu, Shizhu He, Jun Zhao, Kang Liu
cs.CL, cs.AI
Black-box Large Language Models (LLMs) have shown great power in solving various tasks and are considered general problem solvers. However, LLMs still fail in many specific tasks although understand the task instruction. In this paper, we focus on the problem of boosting the ability of black-box LLMs to solve downstrea...
http://arxiv.org/pdf/2311.07032v1
2023-11-13T02:31:16
2023-11-13T02:31:16
{ "title": [ -0.007896423, -0.035003662, -0.01701355, -0.02670288, -0.04800415, -0.015930176, -0.03488159, -0.0069503784, -0.0012149811, 0.032409668, -0.026885986, 0.001499176, 0.003753662, -0.028686523, -0.0031204224, -0.028442383, 0.0027370453, ...
http://arxiv.org/abs/2311.07022v1
ViLMA: A Zero-Shot Benchmark for Linguistic and Temporal Grounding in Video-Language Models
Ilker Kesen, Andrea Pedrotti, Mustafa Dogan, Michele Cafagna, Emre Can Acikgoz, Letitia Parcalabescu, Iacer Calixto, Anette Frank, Albert Gatt, Aykut Erdem, Erkut Erdem
cs.CL, cs.AI, cs.CV
With the ever-increasing popularity of pretrained Video-Language Models (VidLMs), there is a pressing need to develop robust evaluation methodologies that delve deeper into their visio-linguistic capabilities. To address this challenge, we present ViLMA (Video Language Model Assessment), a task-agnostic benchmark that ...
http://arxiv.org/pdf/2311.07022v1
2023-11-13T02:13:13
2023-11-13T02:13:13
{ "title": [ -0.014549255, -0.004825592, -0.043395996, -0.07849121, 0.0018520355, -0.00983429, -0.05517578, 0.009277344, 0.020690918, 0.037231445, -0.008300781, 0.0030174255, 0.009605408, -0.0073013306, -0.062072754, 0.014793396, -0.011581421, 0....
http://arxiv.org/abs/2311.07014v1
Teach me with a Whisper: Enhancing Large Language Models for Analyzing Spoken Transcripts using Speech Embeddings
Fatema Hasan, Yulong Li, James Foulds, Shimei Pan, Bishwaranjan Bhattacharjee
cs.CL, cs.SD, eess.AS
Speech data has rich acoustic and paralinguistic information with important cues for understanding a speaker's tone, emotion, and intent, yet traditional large language models such as BERT do not incorporate this information. There has been an increased interest in multi-modal language models leveraging audio and/or vi...
http://arxiv.org/pdf/2311.07014v1
2023-11-13T01:53:12
2023-11-13T01:53:12
{ "title": [ -0.016479492, -0.009307861, -0.04473877, -0.07574463, -0.0046806335, -0.00579834, -0.06854248, 0.020355225, 0.010917664, 0.026870728, -0.027008057, 0.010513306, 0.012138367, -0.003276825, -0.029663086, 0.0062026978, -0.013877869, 0.0...
http://arxiv.org/abs/2311.07006v1
Context-dependent Instruction Tuning for Dialogue Response Generation
Jin Myung Kwak, Minseon Kim, Sung Ju Hwang
cs.CL, cs.AI
Recent language models have achieved impressive performance in natural language tasks by incorporating instructions with task input during fine-tuning. Since all samples in the same natural language task can be explained with the same task instructions, many instruction datasets only provide a few instructions for the ...
http://arxiv.org/pdf/2311.07006v1
2023-11-13T01:25:30
2023-11-13T01:25:30
{ "title": [ -0.03302002, -0.026397705, -0.027786255, -0.054534912, -0.020339966, 0.0036144257, -0.05014038, 0.016937256, -0.014099121, 0.025421143, -0.005874634, 0.011245728, -0.03692627, -0.018066406, -0.042022705, -0.015365601, -0.007507324, 0...
http://arxiv.org/abs/2311.06985v1
SELF-EXPLAIN: Teaching Large Language Models to Reason Complex Questions by Themselves
Jiachen Zhao, Zonghai Yao, Zhichao Yang, Hong Yu
cs.CL
Large language models (LLMs) can generate intermediate reasoning steps. To elicit the reliable reasoning, the common practice is to employ few-shot chain-of-thought prompting, where several in-context demonstrations for reasoning are prepended to the question. However, such chain-of-thought examples are expensive to cr...
http://arxiv.org/pdf/2311.06985v1
2023-11-12T23:14:43
2023-11-12T23:14:43
{ "title": [ -0.028945923, -0.024139404, -0.01977539, -0.06695557, -0.017211914, 0.010597229, -0.04537964, 0.004470825, -0.0028247833, 0.00008517504, -0.0095825195, 0.006969452, -0.032104492, -0.022949219, -0.028625488, -0.0074005127, 0.0010967255, ...
http://arxiv.org/abs/2311.06899v1
Flames: Benchmarking Value Alignment of Chinese Large Language Models
Kexin Huang, Xiangyang Liu, Qianyu Guo, Tianxiang Sun, Jiawei Sun, Yaru Wang, Zeyang Zhou, Yixu Wang, Yan Teng, Xipeng Qiu, Yingchun Wang, Dahua Lin
cs.CL, cs.AI
The widespread adoption of large language models (LLMs) across various regions underscores the urgent need to evaluate their alignment with human values. Current benchmarks, however, fall short of effectively uncovering safety vulnerabilities in LLMs. Despite numerous models achieving high scores and 'topping the chart...
http://arxiv.org/pdf/2311.06899v1
2023-11-12T17:18:21
2023-11-12T17:18:21
{ "title": [ -0.04031372, -0.0050582886, -0.04321289, -0.04788208, -0.0049743652, -0.0029392242, -0.054748535, -0.0015878677, 0.010009766, 0.0368042, -0.0075683594, -0.0003798008, 0.013389587, -0.027740479, -0.048583984, 0.0077171326, -0.022033691, ...
http://arxiv.org/abs/2311.06898v1
Retrieval and Generative Approaches for a Pregnancy Chatbot in Nepali with Stemmed and Non-Stemmed Data : A Comparative Study
Sujan Poudel, Nabin Ghimire, Bipesh Subedi, Saugat Singh
cs.CL, cs.LG
The field of Natural Language Processing which involves the use of artificial intelligence to support human languages has seen tremendous growth due to its high-quality features. Its applications such as language translation, chatbots, virtual assistants, search autocomplete, and autocorrect are widely used in various ...
http://arxiv.org/pdf/2311.06898v1
2023-11-12T17:16:46
2023-11-12T17:16:46
{ "title": [ -0.008476257, -0.02909851, 0.0017175674, -0.031921387, -0.03845215, 0.00793457, -0.045166016, -0.0025367737, 0.03149414, 0.062347412, -0.020584106, -0.011154175, 0.046661377, -0.020767212, -0.064697266, 0.0069732666, 0.009277344, 0.0...
http://arxiv.org/abs/2311.06858v1
Can Large Language Models Augment a Biomedical Ontology with missing Concepts and Relations?
Antonio Zaitoun, Tomer Sagi, Szymon Wilk, Mor Peleg
cs.CL, cs.AI
Ontologies play a crucial role in organizing and representing knowledge. However, even current ontologies do not encompass all relevant concepts and relationships. Here, we explore the potential of large language models (LLM) to expand an existing ontology in a semi-automated fashion. We demonstrate our approach on the...
http://arxiv.org/pdf/2311.06858v1
2023-11-12T14:20:55
2023-11-12T14:20:55
{ "title": [ -0.014175415, -0.009750366, -0.038269043, -0.061035156, -0.024536133, 0.011924744, -0.071777344, 0.0023422241, 0.020965576, 0.008926392, -0.03164673, -0.017974854, 0.016479492, -0.025619507, -0.015670776, -0.008460999, 0.0036029816, ...
http://arxiv.org/abs/2311.06855v1
DialMAT: Dialogue-Enabled Transformer with Moment-Based Adversarial Training
Kanta Kaneda, Ryosuke Korekata, Yuiga Wada, Shunya Nagashima, Motonari Kambara, Yui Iioka, Haruka Matsuo, Yuto Imai, Takayuki Nishimura, Komei Sugiura
cs.CV, cs.CL, cs.RO
This paper focuses on the DialFRED task, which is the task of embodied instruction following in a setting where an agent can actively ask questions about the task. To address this task, we propose DialMAT. DialMAT introduces Moment-based Adversarial Training, which incorporates adversarial perturbations into the latent...
http://arxiv.org/pdf/2311.06855v1
2023-11-12T14:12:19
2023-11-12T14:12:19
{ "title": [ -0.0077056885, 0.00051784515, -0.042236328, -0.056884766, 0.027053833, -0.0022583008, -0.01940918, 0.0014314651, -0.030014038, 0.019500732, 0.010238647, 0.00831604, -0.0015382767, -0.03366089, -0.012374878, -0.0062294006, -0.021865845, ...
http://arxiv.org/abs/2311.06851v2
Automatic Textual Normalization for Hate Speech Detection
Anh Thi-Hoang Nguyen, Dung Ha Nguyen, Nguyet Thi Nguyen, Khanh Thanh-Duy Ho, Kiet Van Nguyen
cs.CL
Social media data is a valuable resource for research, yet it contains a wide range of non-standard words (NSW). These irregularities hinder the effective operation of NLP tools. Current state-of-the-art methods for the Vietnamese language address this issue as a problem of lexical normalization, involving the creation...
http://arxiv.org/pdf/2311.06851v2
2023-11-15T17:33:39
2023-11-12T14:01:38
{ "title": [ 0.0008497238, 0.010894775, -0.034454346, -0.0473938, 0.0262146, -0.0076522827, -0.002878189, -0.01374054, -0.010902405, 0.042510986, -0.020401001, 0.023239136, 0.012908936, 0.0020198822, -0.05633545, 0.0053710938, -0.009140015, 0.053...
http://arxiv.org/abs/2311.06838v1
GIELLM: Japanese General Information Extraction Large Language Model Utilizing Mutual Reinforcement Effect
Chengguang Gan, Qinghao Zhang, Tatsunori Mori
cs.CL
Information Extraction (IE) stands as a cornerstone in natural language processing, traditionally segmented into distinct sub-tasks. The advent of Large Language Models (LLMs) heralds a paradigm shift, suggesting the feasibility of a singular model addressing multiple IE subtasks. In this vein, we introduce the General...
http://arxiv.org/pdf/2311.06838v1
2023-11-12T13:30:38
2023-11-12T13:30:38
{ "title": [ -0.039215088, -0.009635925, -0.044525146, -0.029251099, -0.013114929, -0.007007599, -0.03640747, -0.00093078613, -0.010650635, 0.031433105, -0.033599854, -0.009552002, 0.02545166, -0.015022278, -0.02746582, -0.016799927, -0.029754639, ...
http://arxiv.org/abs/2311.06818v1
Cricket Player Profiling: Unraveling Strengths and Weaknesses Using Text Commentary Data
Swarup Ranjan Behera, Vijaya V. Saradhi
cs.LG, cs.CL, I.2.7
Devising player-specific strategies in cricket necessitates a meticulous understanding of each player's unique strengths and weaknesses. Nevertheless, the absence of a definitive computational approach to extract such insights from cricket players poses a significant challenge. This paper seeks to address this gap by e...
http://arxiv.org/pdf/2311.06818v1
2023-11-12T11:51:05
2023-11-12T11:51:05
{ "title": [ -0.01939392, -0.001953125, -0.040618896, -0.014724731, 0.0049858093, -0.019485474, -0.0017614365, -0.0138168335, -0.0067329407, 0.036834717, -0.036010742, 0.027893066, 0.004108429, 0.004562378, -0.04714966, -0.024627686, -0.005290985, ...
http://arxiv.org/abs/2311.06815v1
Evaluation of GPT-4 for chest X-ray impression generation: A reader study on performance and perception
Sebastian Ziegelmayer, Alexander W. Marka, Nicolas Lenhart, Nadja Nehls, Stefan Reischl, Felix Harder, Andreas Sauter, Marcus Makowski, Markus Graf, Joshua Gawlitza
cs.CL
The remarkable generative capabilities of multimodal foundation models are currently being explored for a variety of applications. Generating radiological impressions is a challenging task that could significantly reduce the workload of radiologists. In our study we explored and analyzed the generative abilities of GPT...
http://arxiv.org/pdf/2311.06815v1
2023-11-12T11:40:57
2023-11-12T11:40:57
{ "title": [ -0.002614975, 0.015686035, -0.026290894, -0.01953125, 0.0005812645, -0.026550293, -0.008460999, -0.028701782, -0.012161255, 0.009468079, -0.018722534, -0.002035141, -0.03540039, -0.047210693, -0.044921875, -0.009475708, -0.007209778, ...
http://arxiv.org/abs/2311.06807v1
On the Robustness of Question Rewriting Systems to Questions of Varying Hardness
Hai Ye, Hwee Tou Ng, Wenjuan Han
cs.CL
In conversational question answering (CQA), the task of question rewriting~(QR) in context aims to rewrite a context-dependent question into an equivalent self-contained question that gives the same answer. In this paper, we are interested in the robustness of a QR system to questions varying in rewriting hardness or d...
http://arxiv.org/pdf/2311.06807v1
2023-11-12T11:09:30
2023-11-12T11:09:30
{ "title": [ -0.005844116, -0.00970459, -0.008102417, -0.08685303, -0.0021018982, 0.023605347, -0.05996704, -0.010818481, 0.016571045, 0.014053345, -0.02029419, 0.014213562, -0.042785645, 0.0004427433, -0.062927246, -0.007881165, 0.00011372566, -...
http://arxiv.org/abs/2311.06805v1
Tunable Soft Prompts are Messengers in Federated Learning
Chenhe Dong, Yuexiang Xie, Bolin Ding, Ying Shen, Yaliang Li
cs.CL
Federated learning (FL) enables multiple participants to collaboratively train machine learning models using decentralized data sources, alleviating privacy concerns that arise from directly sharing local data. However, the lack of model privacy protection in FL becomes an unneglectable challenge, especially when peopl...
http://arxiv.org/pdf/2311.06805v1
2023-11-12T11:01:10
2023-11-12T11:01:10
{ "title": [ 0.01687622, 0.0068626404, -0.030517578, -0.03616333, -0.024108887, -0.028396606, -0.010566711, -0.0066871643, -0.054534912, 0.021728516, -0.045959473, 0.012931824, 0.0038909912, -0.023406982, 0.035583496, -0.03869629, 0.025360107, 0....
http://arxiv.org/abs/2311.06771v1
Learning Globally Optimized Language Structure via Adversarial Training
Xuwang Yin
cs.CL, cs.AI
Recent work has explored integrating autoregressive language models with energy-based models (EBMs) to enhance text generation capabilities. However, learning effective EBMs for text is challenged by the discrete nature of language. This work proposes an adversarial training strategy to address limitations in prior eff...
http://arxiv.org/pdf/2311.06771v1
2023-11-12T08:21:43
2023-11-12T08:21:43
{ "title": [ -0.033294678, -0.013557434, -0.030700684, -0.06689453, -0.008872986, 0.019104004, -0.06365967, 0.01802063, -0.033477783, 0.022903442, 0.010620117, 0.019348145, -0.02192688, -0.026168823, -0.013763428, 0.00013697147, -0.001001358, 0.0...
http://arxiv.org/abs/2311.07618v1
Large Language Models' Understanding of Math: Source Criticism and Extrapolation
Roozbeh Yousefzadeh, Xuenan Cao
cs.LG, cs.AI, cs.CL, math.HO
It has been suggested that large language models such as GPT-4 have acquired some form of understanding beyond the correlations among the words in text including some understanding of mathematics as well. Here, we perform a critical inquiry into this claim by evaluating the mathematical understanding of the GPT-4 model...
http://arxiv.org/pdf/2311.07618v1
2023-11-12T07:52:32
2023-11-12T07:52:32
{ "title": [ -0.04434204, -0.013420105, -0.012954712, -0.024520874, -0.028137207, -0.009292603, -0.02848816, -0.00868988, -0.015335083, 0.019104004, -0.04284668, 0.0010480881, -0.0025939941, -0.01499176, -0.025680542, -0.015335083, -0.0146865845, ...
http://arxiv.org/abs/2311.07617v1
CLAMP: A Contrastive Language And Molecule Pre-training Network
Neel Redkar
cs.CL, cs.IR, 8.2.D.2.5
This paper highlights a shift in how to approach material generation. Instead of material-to-material, we propose a language-to-material generation architecture that utilizes millions of untapped data points. Using a web scraper to collect crystal text pairs from open-source research papers, a contrastive model can be ...
http://arxiv.org/pdf/2311.07617v1
2023-11-12T07:45:35
2023-11-12T07:45:35
{ "title": [ -0.020584106, -0.02330017, -0.03491211, -0.04940796, -0.017227173, -0.0028209686, -0.03857422, -0.0019330978, 0.01675415, 0.015296936, -0.0021362305, 0.0045433044, -0.015914917, -0.03515625, -0.028289795, 0.016723633, -0.009651184, 0...
http://arxiv.org/abs/2311.06761v1
Learning Knowledge-Enhanced Contextual Language Representations for Domain Natural Language Understanding
Ruyao Xu, Taolin Zhang, Chengyu Wang, Zhongjie Duan, Cen Chen, Minghui Qiu, Dawei Cheng, Xiaofeng He, Weining Qian
cs.CL
Knowledge-Enhanced Pre-trained Language Models (KEPLMs) improve the performance of various downstream NLP tasks by injecting knowledge facts from large-scale Knowledge Graphs (KGs). However, existing methods for pre-training KEPLMs with relational triples are difficult to be adapted to close domains due to the lack of ...
http://arxiv.org/pdf/2311.06761v1
2023-11-12T07:37:24
2023-11-12T07:37:24
{ "title": [ -0.008346558, -0.0073280334, -0.03756714, -0.08239746, 0.0098724365, 0.0000641942, -0.052093506, 0.011642456, 0.011962891, 0.023422241, -0.02407837, -0.003929138, 0.0040512085, -0.008888245, -0.044799805, 0.0022945404, -0.008598328, ...
http://arxiv.org/abs/2311.06758v1
Sharing, Teaching and Aligning: Knowledgeable Transfer Learning for Cross-Lingual Machine Reading Comprehension
Tingfeng Cao, Chengyu Wang, Chuanqi Tan, Jun Huang, Jinhui Zhu
cs.CL
In cross-lingual language understanding, machine translation is often utilized to enhance the transferability of models across languages, either by translating the training data from the source language to the target, or from the target to the source to aid inference. However, in cross-lingual machine reading comprehen...
http://arxiv.org/pdf/2311.06758v1
2023-11-12T07:20:37
2023-11-12T07:20:37
{ "title": [ -0.051452637, -0.014503479, -0.019989014, -0.03955078, -0.027648926, 0.011985779, -0.048736572, -0.012168884, 0.0037117004, 0.045806885, -0.030303955, 0.016357422, 0.017242432, 0.0077781677, -0.04864502, 0.0047416687, -0.016189575, 0...
http://arxiv.org/abs/2311.06754v1
From Complex to Simple: Unraveling the Cognitive Tree for Reasoning with Small Language Models
Junbing Yan, Chengyu Wang, Taolin Zhang, Xiaofeng He, Jun Huang, Wei Zhang
cs.CL
Reasoning is a distinctive human capacity, enabling us to address complex problems by breaking them down into a series of manageable cognitive steps. Yet, complex logical reasoning is still cumbersome for language models. Based on the dual process theory in cognitive science, we are the first to unravel the cognitive r...
http://arxiv.org/pdf/2311.06754v1
2023-11-12T06:56:21
2023-11-12T06:56:21
{ "title": [ -0.046813965, -0.030899048, -0.03189087, -0.06567383, -0.015090942, 0.025970459, -0.075927734, 0.002981186, 0.004512787, 0.01939392, -0.010368347, -0.014854431, -0.0146484375, -0.032989502, -0.053131104, -0.009567261, -0.0027275085, ...
http://arxiv.org/abs/2311.06753v1
Towards General-Purpose Speech Abilities for Large Language Models Using Unpaired Data
Yassir Fathullah, Chunyang Wu, Egor Lakomkin, Junteng Jia, Yuan Shangguan, Jay Mahadeokar, Ozlem Kalinli, Christian Fuegen, Mike Seltzer
cs.CL, cs.AI
In this work, we extend the instruction-tuned Llama-2 model with end-to-end general-purpose speech processing and reasoning abilities while maintaining the wide range of LLM capabilities, without using any carefully curated paired data. The proposed model can utilize audio prompts as a replacement for text and sustain ...
http://arxiv.org/pdf/2311.06753v1
2023-11-12T06:56:14
2023-11-12T06:56:14
{ "title": [ -0.021270752, 0.0022602081, -0.06414795, -0.07720947, -0.00340271, -0.013183594, -0.058624268, 0.031143188, 0.020462036, 0.03414917, -0.018676758, -0.02003479, 0.01777649, -0.020904541, -0.056121826, 0.016677856, -0.017852783, 0.0539...
http://arxiv.org/abs/2311.06752v1
BeautifulPrompt: Towards Automatic Prompt Engineering for Text-to-Image Synthesis
Tingfeng Cao, Chengyu Wang, Bingyan Liu, Ziheng Wu, Jinhui Zhu, Jun Huang
cs.CL
Recently, diffusion-based deep generative models (e.g., Stable Diffusion) have shown impressive results in text-to-image synthesis. However, current text-to-image models often require multiple passes of prompt engineering by humans in order to produce satisfactory results for real-world applications. We propose Beautif...
http://arxiv.org/pdf/2311.06752v1
2023-11-12T06:39:00
2023-11-12T06:39:00
{ "title": [ -0.032714844, -0.011154175, -0.01763916, -0.03567505, -0.012390137, 0.0124435425, -0.03567505, -0.019958496, 0.026046753, 0.036376953, -0.050689697, 0.030578613, 0.0030517578, -0.0076446533, -0.039367676, 0.0004799366, -0.0071983337, ...
http://arxiv.org/abs/2311.06737v1
Detecting and Correcting Hate Speech in Multimodal Memes with Large Visual Language Model
Minh-Hao Van, Xintao Wu
cs.CL, cs.AI
Recently, large language models (LLMs) have taken the spotlight in natural language processing. Further, integrating LLMs with vision enables the users to explore more emergent abilities in multimodality. Visual language models (VLMs), such as LLaVA, Flamingo, or GPT-4, have demonstrated impressive performance on vario...
http://arxiv.org/pdf/2311.06737v1
2023-11-12T05:20:20
2023-11-12T05:20:20
{ "title": [ 0.028305054, 0.019165039, -0.03189087, -0.024993896, 0.008407593, -0.029891968, -0.018127441, -0.032043457, 0.017089844, 0.0395813, -0.011367798, 0.008171082, 0.019348145, -0.016357422, -0.04638672, 0.012741089, -0.02458191, 0.055419...
http://arxiv.org/abs/2311.06736v1
Are LLMs Rigorous Logical Reasoner? Empowering Natural Language Proof Generation with Contrastive Stepwise Decoding
Ying Su, Xiaojin Fu, Mingwen Liu, Zhijiang Guo
cs.CL
Logical reasoning remains a pivotal component within the realm of artificial intelligence. The recent evolution of large language models (LLMs) has marked significant progress in this domain. The adoption of strategies like chain-of-thought (CoT) has enhanced the performance of LLMs across diverse reasoning tasks. None...
http://arxiv.org/pdf/2311.06736v1
2023-11-12T05:12:49
2023-11-12T05:12:49
{ "title": [ -0.039611816, -0.02053833, -0.052703857, -0.06427002, -0.029846191, -0.008560181, -0.0803833, 0.0067863464, -0.018341064, 0.01675415, -0.038848877, 0.02583313, -0.013198853, -0.016815186, -0.013504028, -0.019851685, -0.02809143, 0.05...
http://arxiv.org/abs/2311.06729v1
Comprehending Lexical and Affective Ontologies in the Demographically Diverse Spatial Social Media Discourse
Salim Sazzed
cs.CL
This study aims to comprehend linguistic and socio-demographic features, encompassing English language styles, conveyed sentiments, and lexical diversity within spatial online social media review data. To this end, we undertake a case study that scrutinizes reviews composed by two distinct and demographically diverse g...
http://arxiv.org/pdf/2311.06729v1
2023-11-12T04:23:33
2023-11-12T04:23:33
{ "title": [ 0.016342163, -0.0012359619, -0.014228821, -0.057006836, 0.0020427704, -0.0031738281, -0.013824463, -0.011993408, 0.018722534, 0.02520752, -0.036346436, 0.0022277832, 0.007801056, 0.0017299652, -0.043884277, -0.0129470825, -0.0068244934, ...
http://arxiv.org/abs/2311.06724v1
Controllable Topic-Focused Abstractive Summarization
Seyed Ali Bahrainian, Martin Jaggi, Carsten Eickhoff
cs.CL, cs.LG
Controlled abstractive summarization focuses on producing condensed versions of a source article to cover specific aspects by shifting the distribution of generated text towards a desired style, e.g., a set of topics. Subsequently, the resulting summaries may be tailored to user-defined requirements. This paper present...
http://arxiv.org/pdf/2311.06724v1
2023-11-12T03:51:38
2023-11-12T03:51:38
{ "title": [ 0.0005302429, 0.010429382, -0.0335083, -0.07965088, -0.008110046, 0.0005145073, -0.058898926, -0.018463135, 0.018829346, 0.046142578, -0.02268982, 0.035247803, -0.01121521, -0.00071287155, -0.03955078, 0.015235901, -0.0017251968, 0.0...
http://arxiv.org/abs/2311.06720v1
Cappy: Outperforming and Boosting Large Multi-Task LMs with a Small Scorer
Bowen Tan, Yun Zhu, Lijuan Liu, Eric Xing, Zhiting Hu, Jindong Chen
cs.LG, cs.CL
Large language models (LLMs) such as T0, FLAN, and OPT-IML, excel in multi-tasking under a unified instruction-following paradigm, where they also exhibit remarkable generalization abilities to unseen tasks. Despite their impressive performance, these LLMs, with sizes ranging from several billion to hundreds of billion...
http://arxiv.org/pdf/2311.06720v1
2023-11-12T03:25:34
2023-11-12T03:25:34
{ "title": [ -0.021392822, -0.0025234222, -0.051757812, -0.028060913, 0.012191772, -0.025939941, -0.0074005127, -0.0119018555, -0.014572144, 0.032714844, -0.021881104, 0.011398315, -0.008232117, -0.0025691986, -0.03817749, -0.0012540817, -0.010627747...
http://arxiv.org/abs/2311.06714v1
What factors influence the popularity of user-generated text in the creative domain? A case study of book reviews
Salim Sazzed
cs.CL, cs.IR
This study investigates a range of psychological, lexical, semantic, and readability features of book reviews to elucidate the factors underlying their perceived popularity. To this end, we conduct statistical analyses of various features, including the types and frequency of opinion and emotion-conveying terms, connec...
http://arxiv.org/pdf/2311.06714v1
2023-11-12T02:54:11
2023-11-12T02:54:11
{ "title": [ -0.0027198792, 0.008811951, -0.018493652, 0.017105103, 0.0015869141, -0.039367676, 0.024276733, -0.043121338, -0.05029297, 0.0473938, -0.008796692, 0.0423584, -0.012924194, -0.0051841736, -0.047424316, -0.011154175, -0.035614014, 0.0...
http://arxiv.org/abs/2311.06697v1
Trusted Source Alignment in Large Language Models
Vasilisa Bashlovkina, Zhaobin Kuang, Riley Matthews, Edward Clifford, Yennie Jun, William W. Cohen, Simon Baumgartner
cs.CL
Large language models (LLMs) are trained on web-scale corpora that inevitably include contradictory factual information from sources of varying reliability. In this paper, we propose measuring an LLM property called trusted source alignment (TSA): the model's propensity to align with content produced by trusted publish...
http://arxiv.org/pdf/2311.06697v1
2023-11-12T00:25:25
2023-11-12T00:25:25
{ "title": [ -0.04083252, -0.0057296753, -0.068359375, -0.08050537, 0.0149002075, 0.0035171509, -0.0713501, 0.021575928, -0.0073165894, 0.038269043, -0.002105713, -0.0022411346, 0.020629883, -0.023468018, -0.03012085, 0.01448822, -0.015357971, 0....
http://arxiv.org/abs/2311.06696v1
Simple and Effective Input Reformulations for Translation
Brian Yu, Hansen Lillemark, Kurt Keutzer
cs.CL, cs.LG
Foundation language models learn from their finetuning input context in different ways. In this paper, we reformulate inputs during finetuning for challenging translation tasks, leveraging model strengths from pretraining in novel ways to improve downstream performance. These reformulations are simple data level modifi...
http://arxiv.org/pdf/2311.06696v1
2023-11-12T00:23:37
2023-11-12T00:23:37
{ "title": [ -0.012191772, -0.0024547577, 0.006729126, -0.05392456, -0.031982422, 0.004306793, -0.06793213, -0.022476196, -0.011657715, 0.05529785, 0.017547607, -0.006477356, 0.0138168335, -0.044555664, -0.031311035, 0.023452759, 0.017044067, -0....
http://arxiv.org/abs/2311.06694v1
Comparative Multi-View Language Grounding
Chancharik Mitra, Abrar Anwar, Rodolfo Corona, Dan Klein, Jesse Thomason
cs.CL, cs.AI, cs.CV, cs.RO
In this work, we consider the task of resolving object referents when given a comparative language description. We present a Multi-view Approach to Grounding in Context (MAGiC) that leverages transformers to pragmatically reason over both objects given multiple image views and a language description. In contrast to pas...
http://arxiv.org/pdf/2311.06694v1
2023-11-12T00:21:58
2023-11-12T00:21:58
{ "title": [ -0.0072288513, -0.0023269653, -0.03555298, -0.08166504, -0.016235352, -0.0032367706, -0.067993164, 0.015899658, -0.0038280487, 0.045776367, -0.028121948, 0.0102005005, 0.014884949, -0.025497437, -0.026794434, 0.0028915405, 0.0050735474, ...
http://arxiv.org/abs/2311.06668v1
In-context Vectors: Making In Context Learning More Effective and Controllable Through Latent Space Steering
Sheng Liu, Lei Xing, James Zou
cs.LG, cs.AI, cs.CL
Large language models (LLMs) demonstrate emergent in-context learning capabilities, where they adapt to new tasks based on example demonstrations. However, in-context learning has seen limited effectiveness in many settings, is difficult to quantitatively control and takes up context window space. To overcome these lim...
http://arxiv.org/pdf/2311.06668v1
2023-11-11T21:19:44
2023-11-11T21:19:44
{ "title": [ 0.021865845, -0.0038452148, -0.027374268, -0.018127441, 0.011878967, -0.03488159, 0.033050537, -0.038604736, -0.027435303, 0.03427124, -0.042877197, 0.04751587, -0.012184143, -0.013252258, -0.005065918, -0.013412476, 0.006954193, 0.0...
http://arxiv.org/abs/2311.07611v1
Intentional Biases in LLM Responses
Nicklaus Badyal, Derek Jacoby, Yvonne Coady
cs.CL
In this study we intentionally introduce biases into large language model responses in an attempt to create specific personas for interactive media purposes. We explore the differences between open source models such as Falcon-7b and the GPT-4 model from Open AI, and we quantify some differences in responses afforded b...
http://arxiv.org/pdf/2311.07611v1
2023-11-11T19:59:24
2023-11-11T19:59:24
{ "title": [ -0.018081665, 0.0028572083, -0.03463745, 0.0007529259, -0.05014038, -0.018081665, -0.034118652, -0.029922485, -0.015380859, 0.014350891, -0.02734375, -0.009796143, -0.004917145, -0.056793213, 0.0107040405, -0.019210815, -0.015594482, ...
http://arxiv.org/abs/2311.06649v1
A Template Is All You Meme
Luke Bates, Peter Ebert Christensen, Preslav Nakov, Iryna Gurevych
cs.CL
Memes are a modern form of communication and meme templates possess a base semantics that is customizable by whomever posts it on social media. Machine learning systems struggle with memes, which is likely due to such systems having insufficient context to understand memes, as there is more to memes than the obvious im...
http://arxiv.org/pdf/2311.06649v1
2023-11-11T19:38:14
2023-11-11T19:38:14
{ "title": [ 0.047729492, -0.00034308434, 0.011756897, -0.03866577, -0.040374756, -0.01651001, -0.018310547, -0.0030765533, 0.006881714, 0.090148926, 0.0016794205, 0.020858765, 0.019821167, -0.0126571655, -0.026367188, 0.03918457, 0.026245117, -0...