id
stringlengths
33
34
title
stringlengths
6
218
authors
stringlengths
5
6.98k
categories
stringlengths
5
188
summary
stringlengths
34
3.26k
link_pdf
stringlengths
33
34
updated
timestamp[s]
published
timestamp[s]
embeddings
dict
http://arxiv.org/abs/2305.18200v1
Contextual Knowledge Learning For Dialogue Generation
Wen Zheng, Natasa Milic-Frayling, Ke Zhou
cs.CL
Incorporating conversational context and knowledge into dialogue generation models has been essential for improving the quality of the generated responses. The context, comprising utterances from previous dialogue exchanges, is used as a source of content for response generation and as a means of selecting external kno...
http://arxiv.org/pdf/2305.18200v1
2023-05-29T16:54:10
2023-05-29T16:54:10
{ "title": [ -0.013885498, -0.006942749, -0.051086426, -0.079711914, 0.010467529, 0.0015239716, -0.055999756, 0.014205933, 0.00074338913, 0.037902832, -0.009849548, -0.0041618347, -0.02281189, -0.02078247, -0.06286621, -0.010360718, -0.005996704, ...
http://arxiv.org/abs/2305.18189v1
Marked Personas: Using Natural Language Prompts to Measure Stereotypes in Language Models
Myra Cheng, Esin Durmus, Dan Jurafsky
cs.CL, cs.AI, cs.CY
To recognize and mitigate harms from large language models (LLMs), we need to understand the prevalence and nuances of stereotypes in LLM outputs. Toward this end, we present Marked Personas, a prompt-based method to measure stereotypes in LLMs for intersectional demographic groups without any lexicon or data labeling....
http://arxiv.org/pdf/2305.18189v1
2023-05-29T16:29:22
2023-05-29T16:29:22
{ "title": [ -0.020935059, -0.004825592, -0.028213501, -0.03186035, -0.006668091, -0.005153656, -0.027404785, -0.019760132, 0.00077581406, 0.020767212, -0.03616333, 0.008071899, 0.002117157, -0.01751709, -0.041870117, -0.009643555, -0.035614014, ...
http://arxiv.org/abs/2305.18185v2
Syntax and Semantics Meet in the "Middle": Probing the Syntax-Semantics Interface of LMs Through Agentivity
Lindia Tjuatja, Emmy Liu, Lori Levin, Graham Neubig
cs.CL
Recent advances in large language models have prompted researchers to examine their abilities across a variety of linguistic tasks, but little has been done to investigate how models handle the interactions in meaning across words and larger syntactic forms -- i.e. phenomena at the intersection of syntax and semantics....
http://arxiv.org/pdf/2305.18185v2
2023-07-10T13:10:40
2023-05-29T16:24:01
{ "title": [ -0.016326904, -0.017044067, -0.021606445, -0.03540039, -0.015670776, -0.013793945, -0.009025574, -0.011436462, -0.017562866, 0.014183044, -0.0390625, 0.0063819885, 0.005756378, -0.031082153, -0.016937256, -0.023147583, -0.011810303, ...
http://arxiv.org/abs/2305.18176v1
Perceived Trustworthiness of Natural Language Generators
Beatriz Cabrero-Daniel, Andrea Sanagustín Cabrero
cs.HC, cs.AI, cs.CL
Natural Language Generation tools, such as chatbots that can generate human-like conversational text, are becoming more common both for personal and professional use. However, there are concerns about their trustworthiness and ethical implications. The paper addresses the problem of understanding how different users (e...
http://arxiv.org/pdf/2305.18176v1
2023-05-29T16:09:58
2023-05-29T16:09:58
{ "title": [ -0.011680603, 0.000957489, -0.07104492, -0.044921875, 0.012283325, -0.025650024, -0.003271103, 0.012275696, -0.02798462, 0.031036377, -0.00970459, 0.008598328, -0.0013217926, -0.02116394, -0.034423828, -0.009811401, -0.025405884, 0.0...
http://arxiv.org/abs/2305.18170v2
Leveraging Training Data in Few-Shot Prompting for Numerical Reasoning
Zhanming Jie, Wei Lu
cs.CL
Chain-of-thought (CoT) prompting with large language models has proven effective in numerous natural language processing tasks, but designing prompts that generalize well to diverse problem types can be challenging, especially in the context of math word problem (MWP) solving. Additionally, it is common to have a large...
http://arxiv.org/pdf/2305.18170v2
2023-06-09T05:00:30
2023-05-29T16:01:40
{ "title": [ 0.020126343, -0.034423828, -0.022827148, -0.042785645, -0.011894226, -0.02470398, 0.0036888123, 0.016677856, -0.0074157715, 0.007583618, -0.04260254, 0.00073862076, -0.02017212, -0.020996094, -0.030288696, -0.043395996, 0.012802124, ...
http://arxiv.org/abs/2305.18169v3
LM-CPPF: Paraphrasing-Guided Data Augmentation for Contrastive Prompt-Based Few-Shot Fine-Tuning
Amirhossein Abaskohi, Sascha Rothe, Yadollah Yaghoobzadeh
cs.CL
In recent years, there has been significant progress in developing pre-trained language models for NLP. However, these models often struggle when fine-tuned on small datasets. To address this issue, researchers have proposed various adaptation approaches. Prompt-based tuning is arguably the most common way, especially ...
http://arxiv.org/pdf/2305.18169v3
2023-07-05T09:15:55
2023-05-29T15:59:51
{ "title": [ 0.028274536, -0.025558472, -0.017044067, -0.062072754, -0.021575928, 0.0014896393, -0.017593384, -0.0034828186, 0.0015745163, 0.024475098, -0.039001465, 0.016403198, 0.00038814545, -0.018249512, -0.011688232, -0.023391724, -0.007724762, ...
http://arxiv.org/abs/2305.18156v1
Exploring Effectiveness of GPT-3 in Grammatical Error Correction: A Study on Performance and Controllability in Prompt-Based Methods
Mengsay Loem, Masahiro Kaneko, Sho Takase, Naoaki Okazaki
cs.CL, cs.AI
Large-scale pre-trained language models such as GPT-3 have shown remarkable performance across various natural language processing tasks. However, applying prompt-based methods with GPT-3 for Grammatical Error Correction (GEC) tasks and their controllability remains underexplored. Controllability in GEC is crucial for ...
http://arxiv.org/pdf/2305.18156v1
2023-05-29T15:31:29
2023-05-29T15:31:29
{ "title": [ 0.020935059, 0.022659302, -0.016494751, -0.023330688, 0.0061035156, -0.014770508, -0.021408081, -0.04284668, -0.0059165955, 0.01940918, -0.034576416, 0.016738892, 0.0019292831, -0.013153076, -0.034301758, 0.00066804886, -0.03488159, ...
http://arxiv.org/abs/2305.18153v2
Do Large Language Models Know What They Don't Know?
Zhangyue Yin, Qiushi Sun, Qipeng Guo, Jiawen Wu, Xipeng Qiu, Xuanjing Huang
cs.CL
Large language models (LLMs) have a wealth of knowledge that allows them to excel in various Natural Language Processing (NLP) tasks. Current research focuses on enhancing their performance within their existing knowledge. Despite their vast knowledge, LLMs are still limited by the amount of information they can accomm...
http://arxiv.org/pdf/2305.18153v2
2023-05-30T15:14:06
2023-05-29T15:30:13
{ "title": [ -0.052886963, 0.006580353, -0.03692627, -0.02809143, -0.04360962, 0.00178051, -0.059509277, -0.020355225, 0.023162842, 0.009231567, -0.06359863, -0.0015239716, -0.0028514862, -0.0027008057, -0.036956787, -0.0063095093, 0.00052165985, ...
http://arxiv.org/abs/2305.18152v1
Extrinsic Factors Affecting the Accuracy of Biomedical NER
Zhiyi Li, Shengjie Zhang, Yujie Song, Jungyeul Park
cs.CL
Biomedical named entity recognition (NER) is a critial task that aims to identify structured information in clinical text, which is often replete with complex, technical terms and a high degree of variability. Accurate and reliable NER can facilitate the extraction and analysis of important biomedical information, whic...
http://arxiv.org/pdf/2305.18152v1
2023-05-29T15:29:49
2023-05-29T15:29:49
{ "title": [ 0.010787964, -0.0018587112, 0.011985779, -0.0113220215, -0.032714844, 0.002878189, -0.023208618, -0.04119873, -0.03768921, 0.030151367, 0.00077962875, -0.006362915, 0.018600464, -0.05392456, 0.0026187897, 0.0029945374, -0.0050582886, ...
http://arxiv.org/abs/2305.18149v3
Multiscale Positive-Unlabeled Detection of AI-Generated Texts
Yuchuan Tian, Hanting Chen, Xutao Wang, Zheyuan Bai, Qinghua Zhang, Ruifeng Li, Chao Xu, Yunhe Wang
cs.CL, cs.AI
Recent releases of Large Language Models (LLMs), e.g. ChatGPT, are astonishing at generating human-like texts, but they may impact the authenticity of texts. Previous works proposed methods to detect these AI-generated texts, including simple ML classifiers, pretrained-model-based zero-shot methods, and finetuned langu...
http://arxiv.org/pdf/2305.18149v3
2023-09-29T14:23:22
2023-05-29T15:25:00
{ "title": [ -0.01448822, 0.02708435, -0.03805542, -0.036376953, 0.024337769, -0.0060043335, -0.0128479, -0.02331543, 0.009529114, 0.03656006, -0.02217102, 0.023849487, 0.004486084, -0.0110321045, -0.04269409, 0.007888794, -0.018997192, 0.0477600...
http://arxiv.org/abs/2305.18144v2
GripRank: Bridging the Gap between Retrieval and Generation via the Generative Knowledge Improved Passage Ranking
Jiaqi Bai, Hongcheng Guo, Jiaheng Liu, Jian Yang, Xinnian Liang, Zhao Yan, Zhoujun Li
cs.CL, cs.AI
Retrieval-enhanced text generation has shown remarkable progress on knowledge-intensive language tasks, such as open-domain question answering and knowledge-enhanced dialogue generation, by leveraging passages retrieved from a large passage corpus for delivering a proper answer given the input query. However, the retri...
http://arxiv.org/pdf/2305.18144v2
2023-08-15T17:10:57
2023-05-29T15:15:53
{ "title": [ -0.010765076, -0.0042533875, -0.05001831, -0.057678223, -0.008338928, -0.013343811, -0.037628174, -0.024108887, 0.026123047, 0.027435303, -0.029769897, 0.038116455, 0.0025482178, -0.015808105, -0.023742676, 0.014015198, 0.0035991669, ...
http://arxiv.org/abs/2305.18507v2
Code Prompting: a Neural Symbolic Method for Complex Reasoning in Large Language Models
Yi Hu, Haotong Yang, Zhouchen Lin, Muhan Zhang
cs.CL, cs.AI
Large language models (LLMs) have scaled up to unlock a wide range of complex reasoning tasks with the aid of various prompting methods. However, current prompting methods generate natural language intermediate steps to help reasoning, which can cause imperfect task reduction and confusion. To mitigate such limitations...
http://arxiv.org/pdf/2305.18507v2
2023-10-07T08:07:46
2023-05-29T15:14:09
{ "title": [ -0.022003174, -0.03137207, -0.0044898987, -0.04296875, -0.031707764, 0.004299164, -0.03643799, 0.004776001, 0.00018250942, 0.019897461, -0.014030457, -0.0104522705, 0.018920898, -0.037719727, -0.019515991, 0.0049934387, -0.0012874603, ...
http://arxiv.org/abs/2305.18503v1
From Adversarial Arms Race to Model-centric Evaluation: Motivating a Unified Automatic Robustness Evaluation Framework
Yangyi Chen, Hongcheng Gao, Ganqu Cui, Lifan Yuan, Dehan Kong, Hanlu Wu, Ning Shi, Bo Yuan, Longtao Huang, Hui Xue, Zhiyuan Liu, Maosong Sun, Heng Ji
cs.CL, cs.CR, cs.LG
Textual adversarial attacks can discover models' weaknesses by adding semantic-preserved but misleading perturbations to the inputs. The long-lasting adversarial attack-and-defense arms race in Natural Language Processing (NLP) is algorithm-centric, providing valuable techniques for automatic robustness evaluation. How...
http://arxiv.org/pdf/2305.18503v1
2023-05-29T14:55:20
2023-05-29T14:55:20
{ "title": [ 0.0039634705, 0.024536133, -0.013801575, -0.047332764, 0.020492554, 0.0021648407, -0.027740479, -0.04437256, -0.027832031, 0.041046143, -0.003660202, 0.030700684, -0.028411865, -0.005493164, -0.044128418, -0.005504608, -0.022277832, ...
http://arxiv.org/abs/2305.18500v2
VAST: A Vision-Audio-Subtitle-Text Omni-Modality Foundation Model and Dataset
Sihan Chen, Handong Li, Qunbo Wang, Zijia Zhao, Mingzhen Sun, Xinxin Zhu, Jing Liu
cs.CV, cs.AI, cs.CL, cs.LG, eess.AS
Vision and text have been fully explored in contemporary video-text foundational models, while other modalities such as audio and subtitles in videos have not received sufficient attention. In this paper, we resort to establish connections between multi-modality video tracks, including Vision, Audio, and Subtitle, and ...
http://arxiv.org/pdf/2305.18500v2
2023-10-07T12:58:26
2023-05-29T14:34:50
{ "title": [ -0.010070801, -0.007335663, -0.054138184, -0.046875, 0.005012512, -0.028793335, -0.024261475, -0.009056091, 0.010986328, 0.047912598, -0.017959595, -0.0052337646, 0.036712646, -0.01889038, -0.039245605, 0.008354187, -0.009735107, 0.0...
http://arxiv.org/abs/2305.18109v1
Medical Dialogue Generation via Dual Flow Modeling
Kaishuai Xu, Wenjun Hou, Yi Cheng, Jian Wang, Wenjie Li
cs.CL, cs.AI
Medical dialogue systems (MDS) aim to provide patients with medical services, such as diagnosis and prescription. Since most patients cannot precisely describe their symptoms, dialogue understanding is challenging for MDS. Previous studies mainly addressed this by extracting the mentioned medical entities as critical d...
http://arxiv.org/pdf/2305.18109v1
2023-05-29T14:23:34
2023-05-29T14:23:34
{ "title": [ 0.003353119, -0.014289856, -0.022064209, -0.08703613, -0.009811401, 0.012268066, -0.072143555, 0.007385254, -0.014251709, 0.048858643, 0.026641846, -0.00097608566, -0.008277893, -0.03656006, -0.050689697, 0.009269714, -0.006828308, 0...
http://arxiv.org/abs/2305.18498v1
ANPL: Compiling Natural Programs with Interactive Decomposition
Di Huang, Ziyuan Nan, Xing Hu, Pengwei Jin, Shaohui Peng, Yuanbo Wen, Rui Zhang, Zidong Du, Qi Guo, Yewen Pu, Yunji Chen
cs.PL, cs.AI, cs.CL, cs.LG
The advents of Large Language Models (LLMs) have shown promise in augmenting programming using natural interactions. However, while LLMs are proficient in compiling common usage patterns into a programming language, e.g., Python, it remains a challenge how to edit and debug an LLM-generated program. We introduce ANPL, ...
http://arxiv.org/pdf/2305.18498v1
2023-05-29T14:19:40
2023-05-29T14:19:40
{ "title": [ -0.062683105, -0.011619568, -0.040405273, -0.07501221, -0.0024280548, 0.02558899, -0.06744385, 0.01713562, 0.008171082, 0.009796143, -0.016586304, 0.026565552, -0.037597656, -0.02357483, -0.026428223, 0.0058937073, -0.0056991577, 0.0...
http://arxiv.org/abs/2305.18099v1
Writing user personas with Large Language Models: Testing phase 6 of a Thematic Analysis of semi-structured interviews
Stefano De Paoli
cs.CL, cs.CY
The goal of this paper is establishing if we can satisfactorily perform a Thematic Analysis (TA) of semi-structured interviews using a Large Language Model (more precisely GPT3.5-Turbo). Building on previous work by the author, which established an embryonal process for conducting a TA with the model, this paper will p...
http://arxiv.org/pdf/2305.18099v1
2023-05-29T14:09:14
2023-05-29T14:09:14
{ "title": [ -0.017974854, -0.004711151, -0.026367188, -0.022018433, -0.016952515, -0.013160706, -0.04171753, -0.021026611, -0.013305664, 0.039154053, 0.009613037, 0.010169983, 0.016693115, -0.03894043, -0.028762817, 0.008529663, -0.0340271, 0.01...
http://arxiv.org/abs/2305.18098v2
BigTranslate: Augmenting Large Language Models with Multilingual Translation Capability over 100 Languages
Wen Yang, Chong Li, Jiajun Zhang, Chengqing Zong
cs.CL
Large language models (LLMs) demonstrate promising translation performance among various natural languages. However, many LLMs especially the open-sourced ones, such as BLOOM and LLaMA, are English-dominant and support only dozens of natural languages, making the potential of LLMs on language translation less explored....
http://arxiv.org/pdf/2305.18098v2
2023-07-07T08:45:42
2023-05-29T14:07:52
{ "title": [ -0.03543091, -0.021621704, -0.048614502, -0.074279785, 0.006843567, 0.0039634705, -0.057434082, 0.0079956055, 0.0018167496, 0.0385437, -0.0016727448, -0.0010681152, 0.020202637, -0.025726318, -0.037506104, 0.0151901245, -0.0030097961, ...
http://arxiv.org/abs/2305.18096v2
Improving Textless Spoken Language Understanding with Discrete Units as Intermediate Target
Guan-Wei Wu, Guan-Ting Lin, Shang-Wen Li, Hung-yi Lee
cs.CL, eess.AS
Spoken Language Understanding (SLU) is a task that aims to extract semantic information from spoken utterances. Previous research has made progress in end-to-end SLU by using paired speech-text data, such as pre-trained Automatic Speech Recognition (ASR) models or paired text as intermediate targets. However, acquiring...
http://arxiv.org/pdf/2305.18096v2
2023-07-08T07:25:14
2023-05-29T14:00:24
{ "title": [ -0.004257202, -0.013076782, -0.048828125, -0.07745361, -0.005760193, -0.009162903, -0.045410156, 0.03326416, -0.017562866, 0.033233643, -0.051086426, -0.00035238266, 0.01600647, -0.006046295, -0.04751587, -0.028198242, -0.017547607, ...
http://arxiv.org/abs/2305.18486v4
A Systematic Study and Comprehensive Evaluation of ChatGPT on Benchmark Datasets
Md Tahmid Rahman Laskar, M Saiful Bari, Mizanur Rahman, Md Amran Hossen Bhuiyan, Shafiq Joty, Jimmy Xiangji Huang
cs.CL, cs.AI, cs.LG
The development of large language models (LLMs) such as ChatGPT has brought a lot of attention recently. However, their evaluation in the benchmark academic datasets remains under-explored due to the difficulty of evaluating the generative outputs produced by this model against the ground truth. In this paper, we aim t...
http://arxiv.org/pdf/2305.18486v4
2023-07-05T16:19:38
2023-05-29T12:37:21
{ "title": [ -0.014190674, -0.015464783, -0.021011353, -0.043762207, 0.009185791, 0.0105896, -0.027694702, -0.012336731, 0.004535675, 0.01953125, -0.0027656555, 0.034851074, -0.031829834, -0.0138168335, -0.031829834, 0.020385742, -0.024612427, 0....
http://arxiv.org/abs/2306.04480v1
Exploring the Compositional Generalization in Context Dependent Text-to-SQL Parsing
Aiwei Liu, Wei Liu, Xuming Hu, Shuang Li, Fukun Ma, Yawen Yang, Lijie Wen
cs.CL, 68T50, I.2.7
In the context-dependent Text-to-SQL task, the generated SQL statements are refined iteratively based on the user input utterance from each interaction. The input text from each interaction can be viewed as component modifications to the previous SQL statements, which could be further extracted as the modification patt...
http://arxiv.org/pdf/2306.04480v1
2023-05-29T12:36:56
2023-05-29T12:36:56
{ "title": [ -0.0029640198, 0.016220093, -0.03768921, -0.10772705, 0.03652954, 0.021118164, -0.062469482, 0.022384644, 0.0037441254, 0.064819336, -0.04043579, 0.017211914, 0.0037670135, 0.008049011, -0.0736084, -0.020095825, -0.016082764, 0.04724...
http://arxiv.org/abs/2305.18034v1
A Corpus for Sentence-level Subjectivity Detection on English News Articles
Francesco Antici, Andrea Galassi, Federico Ruggeri, Katerina Korre, Arianna Muti, Alessandra Bardi, Alice Fedotova, Alberto Barrón-Cedeño
cs.CL
We present a novel corpus for subjectivity detection at the sentence level. We develop new annotation guidelines for the task, which are not limited to language-specific cues, and apply them to produce a new corpus in English. The corpus consists of 411 subjective and 638 objective sentences extracted from ongoing cove...
http://arxiv.org/pdf/2305.18034v1
2023-05-29T11:54:50
2023-05-29T11:54:50
{ "title": [ 0.003156662, 0.0026950836, -0.015434265, -0.032562256, 0.010307312, -0.0031356812, 0.010292053, -0.029388428, 0.03668213, 0.026657104, -0.0231781, 0.0036392212, -0.0070114136, 0.002210617, -0.07800293, 0.007797241, -0.032714844, 0.02...
http://arxiv.org/abs/2305.18029v2
Faithfulness Tests for Natural Language Explanations
Pepa Atanasova, Oana-Maria Camburu, Christina Lioma, Thomas Lukasiewicz, Jakob Grue Simonsen, Isabelle Augenstein
cs.CL, cs.AI, 68T50, I.2.7
Explanations of neural models aim to reveal a model's decision-making process for its predictions. However, recent work shows that current methods giving explanations such as saliency maps or counterfactuals can be misleading, as they are prone to present reasons that are unfaithful to the model's inner workings. This ...
http://arxiv.org/pdf/2305.18029v2
2023-06-30T15:41:47
2023-05-29T11:40:37
{ "title": [ -0.019119263, -0.0077552795, -0.020095825, -0.087402344, -0.0034008026, 0.018188477, -0.060272217, -0.014282227, -0.005344391, 0.018325806, -0.0132369995, 0.0028209686, -0.007904053, -0.024154663, -0.029541016, 0.0059280396, -0.013076782...
http://arxiv.org/abs/2305.18028v1
ADAPTERMIX: Exploring the Efficacy of Mixture of Adapters for Low-Resource TTS Adaptation
Ambuj Mehrish, Abhinav Ramesh Kashyap, Li Yingting, Navonil Majumder, Soujanya Poria
cs.SD, cs.AI, cs.CL, eess.AS
There are significant challenges for speaker adaptation in text-to-speech for languages that are not widely spoken or for speakers with accents or dialects that are not well-represented in the training data. To address this issue, we propose the use of the "mixture of adapters" method. This approach involves adding mul...
http://arxiv.org/pdf/2305.18028v1
2023-05-29T11:39:01
2023-05-29T11:39:01
{ "title": [ -0.010818481, -0.0014686584, -0.055786133, -0.066833496, 0.003967285, -0.010276794, -0.045654297, 0.0032787323, -0.011947632, 0.071777344, -0.008552551, 0.007835388, 0.030593872, -0.017990112, -0.05230713, 0.01979065, -0.012252808, 0...
http://arxiv.org/abs/2305.18026v1
Semantic Role Labeling Guided Out-of-distribution Detection
Jinan Zou, Maihao Guo, Yu Tian, Yuhao Lin, Haiyao Cao, Lingqiao Liu, Ehsan Abbasnejad, Javen Qinfeng Shi
cs.CL
Identifying unexpected domain-shifted instances in natural language processing is crucial in real-world applications. Previous works identify the OOD instance by leveraging a single global feature embedding to represent the sentence, which cannot characterize subtle OOD patterns well. Another major challenge current OO...
http://arxiv.org/pdf/2305.18026v1
2023-05-29T11:34:14
2023-05-29T11:34:14
{ "title": [ 0.023864746, 0.00041794777, -0.02456665, -0.052215576, 0.00970459, -0.002532959, -0.019348145, -0.015731812, 0.0049324036, 0.0051994324, -0.029708862, 0.023223877, -0.014694214, -0.036895752, 0.004146576, -0.017730713, -0.024520874, ...
http://arxiv.org/abs/2305.18023v1
Abstractive Summarization as Augmentation for Document-Level Event Detection
Janko Vidaković, Filip Karlo Došilović, Domagoj Pluščec
cs.CL
Transformer-based models have consistently produced substantial performance gains across a variety of NLP tasks, compared to shallow models. However, deep models are orders of magnitude more computationally expensive than shallow models, especially on tasks with large sequence lengths, such as document-level event dete...
http://arxiv.org/pdf/2305.18023v1
2023-05-29T11:28:26
2023-05-29T11:28:26
{ "title": [ 0.029708862, 0.006752014, -0.023071289, -0.10021973, 0.01889038, 0.0058135986, -0.0552063, -0.023712158, 0.032409668, 0.034179688, -0.020111084, 0.020858765, -0.008453369, -0.0070114136, -0.0418396, 0.020568848, -0.0061531067, 0.0324...
http://arxiv.org/abs/2305.18011v1
Can We Trust Explainable AI Methods on ASR? An Evaluation on Phoneme Recognition
Xiaoliang Wu, Peter Bell, Ajitha Rajan
cs.CL, cs.SD, eess.AS
Explainable AI (XAI) techniques have been widely used to help explain and understand the output of deep learning models in fields such as image classification and Natural Language Processing. Interest in using XAI techniques to explain deep learning-based automatic speech recognition (ASR) is emerging. but there is not...
http://arxiv.org/pdf/2305.18011v1
2023-05-29T11:04:13
2023-05-29T11:04:13
{ "title": [ -0.020202637, -0.0103302, -0.03543091, -0.05505371, -0.014076233, -0.0059661865, -0.016052246, 0.019744873, -0.028244019, 0.032104492, -0.025161743, 0.0063323975, 0.00030064583, -0.016662598, -0.035980225, -0.021347046, -0.022506714, ...
http://arxiv.org/abs/2305.18004v2
The Effects of Political Martyrdom on Election Results: The Assassination of Abe
Miu Nicole Takagi
cs.CL
In developed nations assassinations are rare and thus the impact of such acts on the electoral and political landscape is understudied. In this paper, we focus on Twitter data to examine the effects of Japan's former Primer Minister Abe's assassination on the Japanese House of Councillors elections in 2022. We utilize ...
http://arxiv.org/pdf/2305.18004v2
2023-08-28T09:34:39
2023-05-29T10:33:08
{ "title": [ 0.07086182, 0.004119873, -0.016281128, -0.031066895, 0.012702942, -0.046203613, 0.021713257, -0.035980225, -0.02772522, 0.05029297, 0.028381348, -0.0059928894, 0.047210693, -0.05456543, -0.010406494, 0.012191772, -0.0011053085, -0.01...
http://arxiv.org/abs/2305.17984v1
minOffense: Inter-Agreement Hate Terms for Stable Rules, Concepts, Transitivities, and Lattices
Animesh Chaturvedi, Rajesh Sharma
cs.CL, cs.AI, cs.SI, https://www.youtube.com/watch?v=iRGXiJGp3Cc&list=PLtvWi5o3JBnF3yxcjGdT4KCDLxRBIpsyR
Hate speech classification has become an important problem due to the spread of hate speech on social media platforms. For a given set of Hate Terms lists (HTs-lists) and Hate Speech data (HS-data), it is challenging to understand which hate term contributes the most for hate speech classification. This paper contribut...
http://arxiv.org/pdf/2305.17984v1
2023-05-29T09:47:36
2023-05-29T09:47:36
{ "title": [ 0.000084877014, -0.014556885, -0.052520752, -0.052459717, -0.019577026, -0.033813477, -0.001077652, 0.0037593842, -0.0014467239, 0.060028076, -0.04537964, 0.03741455, 0.04348755, -0.007411957, -0.021621704, 0.0012712479, -0.021224976, ...
http://arxiv.org/abs/2305.17968v1
Data Augmentation for Low-Resource Keyphrase Generation
Krishna Garg, Jishnu Ray Chowdhury, Cornelia Caragea
cs.CL
Keyphrase generation is the task of summarizing the contents of any given article into a few salient phrases (or keyphrases). Existing works for the task mostly rely on large-scale annotated datasets, which are not easy to acquire. Very few works address the problem of keyphrase generation in low-resource settings, but...
http://arxiv.org/pdf/2305.17968v1
2023-05-29T09:20:34
2023-05-29T09:20:34
{ "title": [ -0.034942627, -0.025894165, -0.056762695, -0.05239868, -0.012039185, -0.0033855438, -0.042541504, 0.03010559, 0.0067749023, 0.05230713, -0.015716553, 0.0076904297, 0.023086548, -0.008285522, -0.046081543, 0.010955811, -0.023147583, 0...
http://arxiv.org/abs/2305.17951v1
ContrastNER: Contrastive-based Prompt Tuning for Few-shot NER
Amirhossein Layegh, Amir H. Payberah, Ahmet Soylu, Dumitru Roman, Mihhail Matskin
cs.CL, cs.AI
Prompt-based language models have produced encouraging results in numerous applications, including Named Entity Recognition (NER) tasks. NER aims to identify entities in a sentence and provide their types. However, the strong performance of most available NER approaches is heavily dependent on the design of discrete pr...
http://arxiv.org/pdf/2305.17951v1
2023-05-29T08:24:42
2023-05-29T08:24:42
{ "title": [ 0.014656067, -0.033813477, 0.009765625, -0.048828125, -0.035888672, 0.009010315, -0.021835327, -0.014274597, -0.018051147, 0.018341064, -0.016937256, 0.012916565, 0.0051994324, -0.045196533, 0.00067043304, -0.014129639, -0.002313614, ...
http://arxiv.org/abs/2305.18466v2
Test-Time Training on Nearest Neighbors for Large Language Models
Moritz Hardt, Yu Sun
cs.CL, cs.LG
Many recent efforts aim to augment language models with relevant information retrieved from a database at test time. We avoid the need for prompt engineering by directly fine-tuning the model on data retrieved at test time using its standard training setup. For this purpose, we build a large-scale distributed nearest n...
http://arxiv.org/pdf/2305.18466v2
2023-06-07T06:21:30
2023-05-29T08:03:28
{ "title": [ -0.037719727, -0.01940918, -0.024932861, -0.058258057, -0.011459351, 0.0040740967, -0.048950195, 0.0053482056, 0.0018825531, 0.028442383, -0.026565552, 0.011116028, -0.004524231, -0.005748749, -0.034454346, -0.000916481, -0.009033203, ...
http://arxiv.org/abs/2305.17926v2
Large Language Models are not Fair Evaluators
Peiyi Wang, Lei Li, Liang Chen, Zefan Cai, Dawei Zhu, Binghuai Lin, Yunbo Cao, Qi Liu, Tianyu Liu, Zhifang Sui
cs.CL, cs.AI, cs.IR
In this paper, we uncover a systematic bias in the evaluation paradigm of adopting large language models~(LLMs), e.g., GPT-4, as a referee to score and compare the quality of responses generated by candidate models. We find that the quality ranking of candidate responses can be easily hacked by simply altering their or...
http://arxiv.org/pdf/2305.17926v2
2023-08-30T13:22:35
2023-05-29T07:41:03
{ "title": [ -0.015167236, -0.009765625, -0.016052246, -0.028884888, -0.044799805, -0.009666443, -0.023635864, -0.006351471, -0.012962341, 0.03213501, -0.04510498, -0.045776367, -0.01360321, -0.030441284, -0.02722168, -0.033416748, -0.010696411, ...
http://arxiv.org/abs/2305.18462v2
Membership Inference Attacks against Language Models via Neighbourhood Comparison
Justus Mattern, Fatemehsadat Mireshghallah, Zhijing Jin, Bernhard Schölkopf, Mrinmaya Sachan, Taylor Berg-Kirkpatrick
cs.CL, cs.CR, cs.LG
Membership Inference attacks (MIAs) aim to predict whether a data sample was present in the training data of a machine learning model or not, and are widely used for assessing the privacy risks of language models. Most existing attacks rely on the observation that models tend to assign higher probabilities to their tra...
http://arxiv.org/pdf/2305.18462v2
2023-08-07T06:32:56
2023-05-29T07:06:03
{ "title": [ -0.027618408, -0.006511688, -0.03604126, -0.05697632, 0.013557434, 0.0012044907, -0.031341553, 0.005596161, -0.032470703, 0.028869629, -0.0025997162, 0.022583008, 0.0017747879, -0.013122559, -0.014251709, -0.0066070557, -0.011062622, ...
http://arxiv.org/abs/2305.17911v1
TotalDefMeme: A Multi-Attribute Meme dataset on Total Defence in Singapore
Nirmalendu Prakash, Ming Shan Hee, Roy Ka-Wei Lee
cs.SI, cs.AI, cs.CL, cs.CV, I.2.7
Total Defence is a defence policy combining and extending the concept of military defence and civil defence. While several countries have adopted total defence as their defence policy, very few studies have investigated its effectiveness. With the rapid proliferation of social media and digitalisation, many social stud...
http://arxiv.org/pdf/2305.17911v1
2023-05-29T06:43:37
2023-05-29T06:43:37
{ "title": [ 0.07055664, -0.015930176, -0.051818848, -0.04385376, -0.010177612, -0.06237793, 0.008880615, 0.013412476, -0.020629883, 0.03970337, 0.011169434, 0.05709839, -0.038360596, -0.035247803, -0.0051231384, 0.0030174255, 0.0062217712, 0.014...
http://arxiv.org/abs/2305.17906v1
Byte-Level Grammatical Error Correction Using Synthetic and Curated Corpora
Svanhvít Lilja Ingólfsdóttir, Pétur Orri Ragnarsson, Haukur Páll Jónsson, Haukur Barri Símonarson, Vilhjálmur Þorsteinsson, Vésteinn Snæbjarnarson
cs.CL
Grammatical error correction (GEC) is the task of correcting typos, spelling, punctuation and grammatical issues in text. Approaching the problem as a sequence-to-sequence task, we compare the use of a common subword unit vocabulary and byte-level encoding. Initial synthetic training data is created using an error-gene...
http://arxiv.org/pdf/2305.17906v1
2023-05-29T06:35:40
2023-05-29T06:35:40
{ "title": [ 0.0052719116, -0.000323534, -0.013916016, -0.07751465, -0.000078082085, 0.014724731, -0.04385376, 0.009628296, 0.0053634644, 0.035614014, -0.03842163, -0.000087976456, 0.017547607, -0.00066661835, -0.050567627, -0.008956909, -0.03552246,...
http://arxiv.org/abs/2305.17888v1
LLM-QAT: Data-Free Quantization Aware Training for Large Language Models
Zechun Liu, Barlas Oguz, Changsheng Zhao, Ernie Chang, Pierre Stock, Yashar Mehdad, Yangyang Shi, Raghuraman Krishnamoorthi, Vikas Chandra
cs.CL
Several post-training quantization methods have been applied to large language models (LLMs), and have been shown to perform well down to 8-bits. We find that these methods break down at lower bit precision, and investigate quantization aware training for LLMs (LLM-QAT) to push quantization levels even further. We prop...
http://arxiv.org/pdf/2305.17888v1
2023-05-29T05:22:11
2023-05-29T05:22:11
{ "title": [ -0.02947998, -0.007003784, -0.047546387, -0.052642822, -0.005306244, -0.012214661, -0.045318604, 0.009666443, -0.0047721863, 0.05166626, 0.005382538, -0.007007599, 0.018463135, -0.023788452, -0.0541687, 0.0118255615, -0.012451172, 0....
http://arxiv.org/abs/2305.17878v1
Ask an Expert: Leveraging Language Models to Improve Strategic Reasoning in Goal-Oriented Dialogue Models
Qiang Zhang, Jason Naradowsky, Yusuke Miyao
cs.CL, cs.AI, cs.HC
Existing dialogue models may encounter scenarios which are not well-represented in the training data, and as a result generate responses that are unnatural, inappropriate, or unhelpful. We propose the "Ask an Expert" framework in which the model is trained with access to an "expert" which it can consult at each turn. A...
http://arxiv.org/pdf/2305.17878v1
2023-05-29T04:19:35
2023-05-29T04:19:35
{ "title": [ -0.023666382, -0.01966858, -0.038085938, -0.057739258, -0.015838623, -0.006000519, -0.072143555, 0.009429932, -0.024261475, 0.023864746, 0.011329651, -0.0026454926, 0.00034236908, -0.034698486, -0.028305054, -0.00223732, -0.007865906, ...
http://arxiv.org/abs/2305.18449v1
Taming AI Bots: Controllability of Neural States in Large Language Models
Stefano Soatto, Paulo Tabuada, Pratik Chaudhari, Tian Yu Liu
cs.AI, cs.CL, cs.LG, cs.SY, eess.SY
We tackle the question of whether an agent can, by suitable choice of prompts, control an AI bot to any state. To that end, we first introduce a formal definition of ``meaning'' that is amenable to analysis. Then, we characterize ``meaningful data'' on which large language models (LLMs) are ostensibly trained, and ``we...
http://arxiv.org/pdf/2305.18449v1
2023-05-29T03:58:33
2023-05-29T03:58:33
{ "title": [ 0.000087082386, 0.0035209656, 0.0009455681, -0.025665283, 0.008796692, -0.009361267, 0.0042419434, -0.045013428, -0.015037537, 0.020614624, -0.023757935, 0.0071372986, -0.009460449, -0.006164551, -0.03866577, -0.012138367, -0.013435364, ...
http://arxiv.org/abs/2305.17855v1
Vec2Gloss: definition modeling leveraging contextualized vectors with Wordnet gloss
Yu-Hsiang Tseng, Mao-Chang Ku, Wei-Ling Chen, Yu-Lin Chang, Shu-Kai Hsieh
cs.CL
Contextualized embeddings are proven to be powerful tools in multiple NLP tasks. Nonetheless, challenges regarding their interpretability and capability to represent lexical semantics still remain. In this paper, we propose that the task of definition modeling, which aims to generate the human-readable definition of th...
http://arxiv.org/pdf/2305.17855v1
2023-05-29T02:37:37
2023-05-29T02:37:37
{ "title": [ -0.013923645, -0.0058059692, -0.05697632, -0.021255493, -0.019577026, -0.021224976, -0.028945923, -0.015388489, -0.015144348, 0.029846191, -0.043121338, 0.0038986206, 0.019058228, -0.021499634, -0.018112183, -0.010047913, -0.0049438477, ...
http://arxiv.org/abs/2305.17854v1
E-NER: Evidential Deep Learning for Trustworthy Named Entity Recognition
Zhen Zhang, Mengting Hu, Shiwan Zhao, Minlie Huang, Haotian Wang, Lemao Liu, Zhirui Zhang, Zhe Liu, Bingzhe Wu
cs.CL
Most named entity recognition (NER) systems focus on improving model performance, ignoring the need to quantify model uncertainty, which is critical to the reliability of NER systems in open environments. Evidential deep learning (EDL) has recently been proposed as a promising solution to explicitly model predictive un...
http://arxiv.org/pdf/2305.17854v1
2023-05-29T02:36:16
2023-05-29T02:36:16
{ "title": [ -0.01184082, 0.016052246, -0.07196045, -0.06945801, 0.029693604, -0.012123108, -0.03253174, -0.003944397, -0.003293991, 0.01701355, -0.037384033, 0.028625488, 0.019760132, -0.01725769, -0.010948181, 0.009666443, -0.026870728, 0.06063...
http://arxiv.org/abs/2305.17846v1
Retraining-free Customized ASR for Enharmonic Words Based on a Named-Entity-Aware Model and Phoneme Similarity Estimation
Yui Sudo, Kazuya Hata, Kazuhiro Nakadai
cs.SD, cs.CL, eess.AS
End-to-end automatic speech recognition (E2E-ASR) has the potential to improve performance, but a specific issue that needs to be addressed is the difficulty it has in handling enharmonic words: named entities (NEs) with the same pronunciation and part of speech that are spelled differently. This often occurs with Japa...
http://arxiv.org/pdf/2305.17846v1
2023-05-29T02:10:13
2023-05-29T02:10:13
{ "title": [ 0.006416321, 0.006137848, -0.042541504, -0.083740234, -0.0042381287, -0.0028095245, -0.0440979, 0.031982422, -0.0057678223, 0.03387451, 0.00086688995, 0.0028419495, 0.02796936, -0.028045654, -0.022323608, 0.011413574, 0.0025901794, 0...
http://arxiv.org/abs/2305.17826v1
NOTABLE: Transferable Backdoor Attacks Against Prompt-based NLP Models
Kai Mei, Zheng Li, Zhenting Wang, Yang Zhang, Shiqing Ma
cs.CL, cs.CR
Prompt-based learning is vulnerable to backdoor attacks. Existing backdoor attacks against prompt-based models consider injecting backdoors into the entire embedding layers or word embedding vectors. Such attacks can be easily affected by retraining on downstream tasks and with different prompting strategies, limiting ...
http://arxiv.org/pdf/2305.17826v1
2023-05-28T23:35:17
2023-05-28T23:35:17
{ "title": [ -0.010040283, -0.010795593, -0.019927979, -0.01889038, -0.017852783, -0.01524353, -0.032836914, -0.01777649, -0.05328369, 0.0045700073, -0.0058250427, 0.0071144104, -0.0041389465, -0.052947998, 0.014282227, -0.028884888, -0.0062675476, ...
http://arxiv.org/abs/2305.17819v1
Large Language Models, scientific knowledge and factuality: A systematic analysis in antibiotic discovery
Magdalena Wysocka, Oskar Wysocki, Maxime Delmas, Vincent Mutel, Andre Freitas
cs.CL, cs.AI
Inferring over and extracting information from Large Language Models (LLMs) trained on a large corpus of scientific literature can potentially drive a new era in biomedical research, reducing the barriers for accessing existing medical evidence. This work examines the potential of LLMs for dialoguing with biomedical ba...
http://arxiv.org/pdf/2305.17819v1
2023-05-28T22:46:21
2023-05-28T22:46:21
{ "title": [ -0.04534912, -0.023910522, -0.060638428, -0.041259766, -0.020965576, -0.0141067505, -0.05480957, 0.008468628, 0.010147095, 0.022857666, -0.028564453, 0.0062675476, -0.012001038, -0.0036296844, -0.052490234, -0.012809753, 0.004085541, ...
http://arxiv.org/abs/2305.17817v1
Transfer Learning for Power Outage Detection Task with Limited Training Data
Olukunle Owolabi
cs.CL, stat.AP
Early detection of power outages is crucial for maintaining a reliable power distribution system. This research investigates the use of transfer learning and language models in detecting outages with limited labeled data. By leveraging pretraining and transfer learning, models can generalize to unseen classes. Using ...
http://arxiv.org/pdf/2305.17817v1
2023-05-28T22:36:35
2023-05-28T22:36:35
{ "title": [ 0.0073928833, 0.025604248, -0.013435364, -0.012779236, 0.015090942, -0.012069702, 0.009284973, -0.051635742, -0.012619019, 0.016464233, -0.016189575, -0.0001090765, -0.0038833618, -0.033447266, -0.012260437, -0.0113220215, -0.011192322, ...
http://arxiv.org/abs/2306.03763v4
ChatGPT Informed Graph Neural Network for Stock Movement Prediction
Zihan Chen, Lei Nico Zheng, Cheng Lu, Jialu Yuan, Di Zhu
q-fin.ST, cs.AI, cs.CL, cs.LG, q-fin.CP, I.2.7; J.1
ChatGPT has demonstrated remarkable capabilities across various natural language processing (NLP) tasks. However, its potential for inferring dynamic network structures from temporal textual data, specifically financial news, remains an unexplored frontier. In this research, we introduce a novel framework that leverage...
http://arxiv.org/pdf/2306.03763v4
2023-09-18T20:26:04
2023-05-28T21:11:59
{ "title": [ -0.007797241, 0.0034866333, -0.03982544, -0.029037476, -0.0035076141, -0.015510559, -0.0075798035, -0.016082764, -0.016342163, 0.05596924, 0.003900528, -0.0008687973, -0.000006020069, -0.024032593, -0.049621582, 0.004627228, -0.005336761...
http://arxiv.org/abs/2305.17812v1
Tab-CoT: Zero-shot Tabular Chain of Thought
Ziqi Jin, Wei Lu
cs.CL
The chain-of-though (CoT) prompting methods were successful in various natural language processing (NLP) tasks thanks to their ability to unveil the underlying complex reasoning processes. Such reasoning processes typically exhibit implicitly structured steps. Recent efforts also started investigating methods to encour...
http://arxiv.org/pdf/2305.17812v1
2023-05-28T20:49:52
2023-05-28T20:49:52
{ "title": [ 0.0038471222, -0.051513672, -0.029693604, -0.036102295, -0.07751465, -0.007068634, -0.032928467, 0.013282776, 0.009277344, 0.03652954, -0.021728516, -0.029144287, 0.042938232, -0.04208374, -0.013671875, -0.008888245, 0.013557434, -0....
http://arxiv.org/abs/2306.01768v1
A Quantitative Review on Language Model Efficiency Research
Meng Jiang, Hy Dang, Lingbo Tong
cs.LG, cs.CL
Language models (LMs) are being scaled and becoming powerful. Improving their efficiency is one of the core research topics in neural information processing systems. Tay et al. (2022) provided a comprehensive overview of efficient Transformers that have become an indispensable staple in the field of NLP. However, in th...
http://arxiv.org/pdf/2306.01768v1
2023-05-28T20:25:20
2023-05-28T20:25:20
{ "title": [ -0.049835205, -0.009841919, -0.05291748, -0.05114746, -0.0036945343, -0.0048942566, -0.061950684, 0.00096702576, 0.015037537, 0.042297363, -0.011878967, 0.0025978088, 0.018692017, -0.020706177, -0.050811768, 0.029953003, -0.031143188, ...
http://arxiv.org/abs/2305.17804v1
Targeted Data Generation: Finding and Fixing Model Weaknesses
Zexue He, Marco Tulio Ribeiro, Fereshte Khani
cs.CL
Even when aggregate accuracy is high, state-of-the-art NLP models often fail systematically on specific subgroups of data, resulting in unfair outcomes and eroding user trust. Additional data collection may not help in addressing these weaknesses, as such challenging subgroups may be unknown to users, and underrepresen...
http://arxiv.org/pdf/2305.17804v1
2023-05-28T19:36:50
2023-05-28T19:36:50
{ "title": [ -0.006439209, 0.004917145, -0.033569336, -0.033813477, -0.008201599, -0.0016613007, -0.052459717, -0.045776367, -0.018127441, 0.005962372, -0.021499634, 0.0015010834, -0.03086853, -0.039031982, -0.01737976, -0.012817383, -0.0017843246, ...
http://arxiv.org/abs/2305.18419v1
Semantic Segmentation with Bidirectional Language Models Improves Long-form ASR
W. Ronny Huang, Hao Zhang, Shankar Kumar, Shuo-yiin Chang, Tara N. Sainath
cs.CL, cs.LG, cs.SD, eess.AS
We propose a method of segmenting long-form speech by separating semantically complete sentences within the utterance. This prevents the ASR decoder from needlessly processing faraway context while also preventing it from missing relevant context within the current sentence. Semantically complete sentence boundaries ar...
http://arxiv.org/pdf/2305.18419v1
2023-05-28T19:31:45
2023-05-28T19:31:45
{ "title": [ 0.007522583, 0.00617218, -0.041229248, -0.08758545, -0.009757996, -0.0030670166, -0.076416016, 0.012779236, -0.032806396, 0.04321289, -0.0019760132, 0.0012102127, 0.0048446655, -0.03012085, -0.021057129, 0.0004553795, 0.0026283264, 0...
http://arxiv.org/abs/2306.13089v3
GIMLET: A Unified Graph-Text Model for Instruction-Based Molecule Zero-Shot Learning
Haiteng Zhao, Shengchao Liu, Chang Ma, Hannan Xu, Jie Fu, Zhi-Hong Deng, Lingpeng Kong, Qi Liu
cs.LG, cs.CL, q-bio.BM
Molecule property prediction has gained significant attention in recent years. The main bottleneck is the label insufficiency caused by expensive lab experiments. In order to alleviate this issue and to better leverage textual knowledge for tasks, this study investigates the feasibility of employing natural language in...
http://arxiv.org/pdf/2306.13089v3
2023-10-22T18:13:40
2023-05-28T18:27:59
{ "title": [ -0.060058594, -0.046844482, -0.03918457, -0.036346436, -0.050048828, 0.0028915405, -0.052093506, 0.03414917, 0.011245728, 0.019897461, -0.03366089, 0.028442383, -0.017059326, -0.018295288, -0.019226074, 0.0036849976, -0.00084352493, ...
http://arxiv.org/abs/2305.17782v1
RASR2: The RWTH ASR Toolkit for Generic Sequence-to-sequence Speech Recognition
Wei Zhou, Eugen Beck, Simon Berger, Ralf Schlüter, Hermann Ney
cs.CL, cs.SD, eess.AS
Modern public ASR tools usually provide rich support for training various sequence-to-sequence (S2S) models, but rather simple support for decoding open-vocabulary scenarios only. For closed-vocabulary scenarios, public tools supporting lexical-constrained decoding are usually only for classical ASR, or do not support ...
http://arxiv.org/pdf/2305.17782v1
2023-05-28T17:48:48
2023-05-28T17:48:48
{ "title": [ -0.02897644, -0.0037708282, -0.05102539, -0.08050537, 0.023727417, 0.0051879883, -0.05429077, 0.013931274, 0.0014810562, 0.031677246, 0.0010929108, 0.0022716522, 0.0010356903, -0.032226562, -0.038879395, 0.015342712, -0.0042304993, 0...
http://arxiv.org/abs/2305.17779v1
Generating EDU Extracts for Plan-Guided Summary Re-Ranking
Griffin Adams, Alexander R. Fabbri, Faisal Ladhak, Kathleen McKeown, Noémie Elhadad
cs.CL
Two-step approaches, in which summary candidates are generated-then-reranked to return a single summary, can improve ROUGE scores over the standard single-step approach. Yet, standard decoding methods (i.e., beam search, nucleus sampling, and diverse beam search) produce candidates with redundant, and often low quality...
http://arxiv.org/pdf/2305.17779v1
2023-05-28T17:22:04
2023-05-28T17:22:04
{ "title": [ 0.0015821457, 0.0015211105, -0.021835327, -0.042510986, -0.03955078, -0.0007762909, -0.03765869, -0.023117065, -0.028060913, 0.050872803, -0.049926758, 0.044891357, -0.006225586, 0.013397217, -0.017852783, -0.01007843, -0.0028381348, ...
http://arxiv.org/abs/2305.18410v1
Understanding Breast Cancer Survival: Using Causality and Language Models on Multi-omics Data
Mugariya Farooq, Shahad Hardan, Aigerim Zhumbhayeva, Yujia Zheng, Preslav Nakov, Kun Zhang
cs.LG, cs.CL, q-bio.GN, stat.ME
The need for more usable and explainable machine learning models in healthcare increases the importance of developing and utilizing causal discovery algorithms, which aim to discover causal relations by analyzing observational data. Explainable approaches aid clinicians and biologists in predicting the prognosis of dis...
http://arxiv.org/pdf/2305.18410v1
2023-05-28T17:07:46
2023-05-28T17:07:46
{ "title": [ -0.016052246, -0.0037879944, -0.04284668, -0.064453125, -0.011245728, 0.0005350113, -0.05606079, 0.009597778, -0.0032024384, 0.015930176, 0.0029754639, -0.0023021698, 0.019760132, -0.040924072, -0.002538681, 0.010063171, -0.0045928955, ...
http://arxiv.org/abs/2305.17760v5
Language Models are Bounded Pragmatic Speakers: Understanding RLHF from a Bayesian Cognitive Modeling Perspective
Khanh Nguyen
cs.CL, cs.LG
How do language models "think"? This paper formulates a probabilistic cognitive model called the bounded pragmatic speaker, which can characterize the operation of different variations of language models. Specifically, we demonstrate that large language models fine-tuned with reinforcement learning from human feedback ...
http://arxiv.org/pdf/2305.17760v5
2023-07-29T13:43:50
2023-05-28T16:04:48
{ "title": [ -0.013130188, -0.0062789917, -0.014312744, -0.06695557, -0.03567505, 0.0031471252, -0.072265625, 0.002319336, -0.01852417, 0.006011963, -0.004875183, -0.013313293, -0.01965332, -0.042175293, -0.032470703, -0.02230835, -0.025161743, 0...
http://arxiv.org/abs/2305.19148v3
Mitigating Label Biases for In-context Learning
Yu Fei, Yifan Hou, Zeming Chen, Antoine Bosselut
cs.CL, cs.AI, cs.LG
Various design settings for in-context learning (ICL), such as the choice and order of the in-context examples, can bias a model toward a particular prediction without being reflective of an understanding of the task. While many studies discuss these design choices, there have been few systematic investigations into ca...
http://arxiv.org/pdf/2305.19148v3
2023-08-04T15:43:19
2023-05-28T15:37:39
{ "title": [ 0.005630493, 0.0033493042, -0.045196533, -0.03050232, 0.016723633, -0.036468506, 0.015136719, -0.02507019, -0.0035705566, 0.011878967, -0.029190063, 0.017196655, -0.01876831, -0.022598267, -0.008232117, -0.011398315, -0.010765076, 0....
http://arxiv.org/abs/2305.18404v3
Conformal Prediction with Large Language Models for Multi-Choice Question Answering
Bhawesh Kumar, Charlie Lu, Gauri Gupta, Anil Palepu, David Bellamy, Ramesh Raskar, Andrew Beam
cs.CL, cs.LG, stat.ML
As large language models continue to be widely developed, robust uncertainty quantification techniques will become crucial for their safe deployment in high-stakes scenarios. In this work, we explore how conformal prediction can be used to provide uncertainty quantification in language models for the specific task of m...
http://arxiv.org/pdf/2305.18404v3
2023-07-08T02:20:29
2023-05-28T15:26:10
{ "title": [ -0.020141602, -0.0018415451, -0.02861023, -0.064208984, -0.011062622, 0.008865356, -0.050567627, 0.0069084167, 0.020217896, 0.0335083, -0.019729614, -0.009269714, 0.006462097, -0.006931305, -0.06628418, 0.0041923523, -0.006752014, 0....
http://arxiv.org/abs/2305.17750v1
Reliable and Interpretable Drift Detection in Streams of Short Texts
Ella Rabinovich, Matan Vetzler, Samuel Ackerman, Ateret Anaby-Tavor
cs.CL
Data drift is the change in model input data that is one of the key factors leading to machine learning models performance degradation over time. Monitoring drift helps detecting these issues and preventing their harmful consequences. Meaningful drift interpretation is a fundamental step towards effective re-training o...
http://arxiv.org/pdf/2305.17750v1
2023-05-28T15:14:54
2023-05-28T15:14:54
{ "title": [ 0.010345459, 0.0116119385, -0.008911133, -0.042053223, 0.01687622, -0.009841919, -0.017684937, -0.038604736, 0.025787354, 0.028381348, -0.04107666, -0.002948761, -0.005508423, 0.002773285, -0.07104492, -0.014724731, -0.02104187, 0.03...
http://arxiv.org/abs/2305.17746v2
Whitening-based Contrastive Learning of Sentence Embeddings
Wenjie Zhuo, Yifan Sun, Xiaohan Wang, Linchao Zhu, Yi Yang
cs.CL
This paper presents a whitening-based contrastive learning method for sentence embedding learning (WhitenedCSE), which combines contrastive learning with a novel shuffled group whitening. Generally, contrastive learning pulls distortions of a single sample (i.e., positive samples) close and push negative samples far aw...
http://arxiv.org/pdf/2305.17746v2
2023-06-08T05:33:55
2023-05-28T14:58:10
{ "title": [ 0.0066375732, 0.016952515, -0.020828247, -0.052215576, 0.019577026, -0.009391785, -0.01826477, -0.017669678, 0.027404785, 0.02734375, -0.035827637, 0.036254883, -0.012542725, -0.0022945404, -0.04147339, 0.0053253174, -0.006603241, 0....
http://arxiv.org/abs/2305.17740v1
Breaking Language Barriers with a LEAP: Learning Strategies for Polyglot LLMs
Akshay Nambi, Vaibhav Balloli, Mercy Ranjit, Tanuja Ganu, Kabir Ahuja, Sunayana Sitaram, Kalika Bali
cs.CL, cs.AI
Large language models (LLMs) are at the forefront of transforming numerous domains globally. However, their inclusivity and effectiveness remain limited for non-Latin scripts and low-resource languages. This paper tackles the imperative challenge of enhancing the multilingual performance of LLMs, specifically focusing ...
http://arxiv.org/pdf/2305.17740v1
2023-05-28T14:48:38
2023-05-28T14:48:38
{ "title": [ -0.052337646, -0.01725769, -0.051727295, -0.024551392, -0.020370483, -0.008354187, -0.036712646, 0.0115737915, -0.026229858, 0.025039673, -0.016525269, 0.0357666, -0.014350891, -0.031097412, 0.004020691, -0.010719299, -0.009674072, 0...
http://arxiv.org/abs/2305.17739v1
Range-Based Equal Error Rate for Spoof Localization
Lin Zhang, Xin Wang, Erica Cooper, Nicholas Evans, Junichi Yamagishi
cs.SD, cs.CL, eess.AS
Spoof localization, also called segment-level detection, is a crucial task that aims to locate spoofs in partially spoofed audio. The equal error rate (EER) is widely used to measure performance for such biometric scenarios. Although EER is the only threshold-free metric, it is usually calculated in a point-based way t...
http://arxiv.org/pdf/2305.17739v1
2023-05-28T14:46:54
2023-05-28T14:46:54
{ "title": [ -0.02015686, 0.013046265, -0.013366699, 0.026046753, 0.008415222, -0.013702393, 0.044311523, -0.072021484, -0.022918701, 0.033325195, -0.0001963377, 0.0069465637, 0.002828598, -0.029373169, -0.0056419373, 0.023010254, -0.023132324, 0...
http://arxiv.org/abs/2305.17733v1
Investigating Pre-trained Audio Encoders in the Low-Resource Condition
Hao Yang, Jinming Zhao, Gholamreza Haffari, Ehsan Shareghi
cs.CL, cs.SD, eess.AS
Pre-trained speech encoders have been central to pushing state-of-the-art results across various speech understanding and generation tasks. Nonetheless, the capabilities of these encoders in low-resource settings are yet to be thoroughly explored. To address this, we conduct a comprehensive set of experiments using a r...
http://arxiv.org/pdf/2305.17733v1
2023-05-28T14:15:19
2023-05-28T14:15:19
{ "title": [ 0.019500732, 0.0026931763, -0.021133423, -0.025772095, -0.030029297, -0.03024292, -0.003604889, -0.018798828, -0.009208679, 0.046936035, -0.0065345764, -0.004600525, 0.0211792, -0.04727173, -0.003194809, 0.014472961, -0.00028276443, ...
http://arxiv.org/abs/2305.17729v1
Tri-level Joint Natural Language Understanding for Multi-turn Conversational Datasets
Henry Weld, Sijia Hu, Siqu Long, Josiah Poon, Soyeon Caren Han
cs.CL
Natural language understanding typically maps single utterances to a dual level semantic frame, sentence level intent and slot labels at the word level. The best performing models force explicit interaction between intent detection and slot filling. We present a novel tri-level joint natural language understanding appr...
http://arxiv.org/pdf/2305.17729v1
2023-05-28T13:59:58
2023-05-28T13:59:58
{ "title": [ -0.0019874573, -0.0065612793, -0.038757324, -0.09820557, 0.0029830933, 0.008102417, -0.06640625, 0.026412964, 0.010597229, 0.013671875, -0.035339355, -0.0037631989, -0.023208618, -0.0053520203, -0.050750732, -0.022201538, -0.022476196, ...
http://arxiv.org/abs/2305.17727v1
Learning a Structural Causal Model for Intuition Reasoning in Conversation
Hang Chen, Bingyu Liao, Jing Luo, Wenjing Zhu, Xinyu Yang
cs.CL
Reasoning, a crucial aspect of NLP research, has not been adequately addressed by prevailing models including Large Language Model. Conversation reasoning, as a critical component of it, remains largely unexplored due to the absence of a well-designed cognitive model. In this paper, inspired by intuition theory on conv...
http://arxiv.org/pdf/2305.17727v1
2023-05-28T13:54:09
2023-05-28T13:54:09
{ "title": [ -0.018432617, -0.004333496, -0.040222168, -0.06365967, -0.007095337, 0.0032348633, -0.06323242, 0.0030937195, -0.00944519, 0.013145447, -0.021621704, -0.0006556511, -0.01108551, -0.013290405, -0.03186035, -0.022979736, -0.018249512, ...
http://arxiv.org/abs/2305.17721v1
Rethinking Masked Language Modeling for Chinese Spelling Correction
Hongqiu Wu, Shaohua Zhang, Yuchen Zhang, Hai Zhao
cs.CL
In this paper, we study Chinese Spelling Correction (CSC) as a joint decision made by two separate models: a language model and an error model. Through empirical analysis, we find that fine-tuning BERT tends to over-fit the error model while under-fit the language model, resulting in poor generalization to out-of-distr...
http://arxiv.org/pdf/2305.17721v1
2023-05-28T13:19:12
2023-05-28T13:19:12
{ "title": [ -0.045440674, 0.01335144, -0.035858154, 0.014793396, -0.017669678, -0.014091492, -0.022750854, -0.024734497, 0.0071983337, 0.04232788, -0.027236938, -0.02394104, 0.05581665, -0.032104492, -0.046691895, 0.0076942444, -0.042541504, 0.0...
http://arxiv.org/abs/2305.17718v1
FuseCap: Leveraging Large Language Models to Fuse Visual Data into Enriched Image Captions
Noam Rotstein, David Bensaid, Shaked Brody, Roy Ganz, Ron Kimmel
cs.CV, cs.AI, cs.CL
Image captioning is a central task in computer vision which has experienced substantial progress following the advent of vision-language pre-training techniques. In this paper, we highlight a frequently overlooked limitation of captioning models that often fail to capture semantically significant elements. This drawbac...
http://arxiv.org/pdf/2305.17718v1
2023-05-28T13:16:03
2023-05-28T13:16:03
{ "title": [ -0.009536743, -0.0027198792, -0.036315918, -0.048858643, -0.0017633438, -0.0137786865, -0.029693604, -0.014755249, 0.03012085, 0.045440674, -0.038024902, 0.0146102905, 0.01486969, -0.0006160736, -0.05355835, 0.007698059, -0.0062675476, ...
http://arxiv.org/abs/2305.18396v2
LLMs Can Understand Encrypted Prompt: Towards Privacy-Computing Friendly Transformers
Xuanqi Liu, Zhuotao Liu
cs.LG, cs.CL, cs.CR
The community explored to build private inference frameworks for transformer-based large language models (LLMs) in a server-client setting, where the server holds the model parameters and the client inputs its private data (or prompt) for inference. However, these frameworks impose significant overhead when the private...
http://arxiv.org/pdf/2305.18396v2
2023-10-23T10:06:52
2023-05-28T13:08:13
{ "title": [ -0.009048462, -0.012069702, -0.020141602, -0.03186035, -0.027664185, -0.019485474, -0.023742676, 0.0074386597, -0.042266846, -0.011405945, -0.02999878, -0.009666443, 0.0078125, -0.064208984, 0.039154053, -0.050842285, 0.0064086914, 0...
http://arxiv.org/abs/2305.18395v2
Knowledge-Augmented Reasoning Distillation for Small Language Models in Knowledge-Intensive Tasks
Minki Kang, Seanie Lee, Jinheon Baek, Kenji Kawaguchi, Sung Ju Hwang
cs.CL, cs.AI, cs.LG
Large Language Models (LLMs) have shown promising performance in knowledge-intensive reasoning tasks that require a compound understanding of knowledge. However, deployment of the LLMs in real-world applications can be challenging due to their high computational requirements and concerns on data privacy. Previous studi...
http://arxiv.org/pdf/2305.18395v2
2023-10-30T08:20:14
2023-05-28T13:00:00
{ "title": [ -0.040039062, -0.030426025, -0.06829834, -0.072021484, -0.016616821, -0.0004892349, -0.09112549, 0.021865845, 0.02645874, 0.025466919, -0.019683838, -0.025619507, 0.001748085, -0.025482178, -0.051116943, -0.006931305, 0.005996704, -0...
http://arxiv.org/abs/2305.17714v1
An Open-Source Gloss-Based Baseline for Spoken to Signed Language Translation
Amit Moryossef, Mathias Müller, Anne Göhring, Zifan Jiang, Yoav Goldberg, Sarah Ebling
cs.CL, cs.CV
Sign language translation systems are complex and require many components. As a result, it is very hard to compare methods across publications. We present an open-source implementation of a text-to-gloss-to-pose-to-video pipeline approach, demonstrating conversion from German to Swiss German Sign Language, French to Fr...
http://arxiv.org/pdf/2305.17714v1
2023-05-28T12:57:20
2023-05-28T12:57:20
{ "title": [ -0.060028076, -0.018814087, -0.039123535, -0.06109619, -0.022354126, 0.0256958, -0.084106445, 0.018127441, -0.0028209686, 0.042419434, -0.029418945, 0.012123108, 0.0022602081, -0.009117126, -0.04257202, -0.006542206, -0.008026123, 0....
http://arxiv.org/abs/2305.17709v1
Parallel Data Helps Neural Entity Coreference Resolution
Gongbo Tang, Christian Hardmeier
cs.CL
Coreference resolution is the task of finding expressions that refer to the same entity in a text. Coreference models are generally trained on monolingual annotated data but annotating coreference is expensive and challenging. Hardmeier et al.(2013) have shown that parallel data contains latent anaphoric knowledge, but...
http://arxiv.org/pdf/2305.17709v1
2023-05-28T12:30:23
2023-05-28T12:30:23
{ "title": [ 0.02670288, 0.009674072, -0.043182373, -0.06933594, -0.02583313, -0.036346436, -0.033325195, 0.0027770996, -0.026245117, 0.08380127, -0.0060920715, 0.0035190582, 0.076293945, -0.032409668, -0.0036144257, 0.03564453, 0.0049552917, 0.0...
http://arxiv.org/abs/2305.17701v2
KoSBi: A Dataset for Mitigating Social Bias Risks Towards Safer Large Language Model Application
Hwaran Lee, Seokhee Hong, Joonsuk Park, Takyoung Kim, Gunhee Kim, Jung-Woo Ha
cs.CL
Large language models (LLMs) learn not only natural text generation abilities but also social biases against different demographic groups from real-world data. This poses a critical risk when deploying LLM-based applications. Existing research and resources are not readily applicable in South Korea due to the differenc...
http://arxiv.org/pdf/2305.17701v2
2023-05-30T01:42:07
2023-05-28T12:07:16
{ "title": [ -0.015655518, -0.009284973, -0.050079346, -0.06842041, 0.00013697147, -0.008003235, -0.052856445, 0.0018119812, -0.014533997, 0.038208008, -0.021560669, 0.0062446594, 0.040771484, -0.014541626, -0.02168274, 0.0059776306, -0.022750854, ...
http://arxiv.org/abs/2305.17699v1
Decoupling Pseudo Label Disambiguation and Representation Learning for Generalized Intent Discovery
Yutao Mou, Xiaoshuai Song, Keqing He, Chen Zeng, Pei Wang, Jingang Wang, Yunsen Xian, Weiran Xu
cs.CL
Generalized intent discovery aims to extend a closed-set in-domain intent classifier to an open-world intent set including in-domain and out-of-domain intents. The key challenges lie in pseudo label disambiguation and representation learning. Previous methods suffer from a coupling of pseudo label disambiguation and re...
http://arxiv.org/pdf/2305.17699v1
2023-05-28T12:01:34
2023-05-28T12:01:34
{ "title": [ -0.0035934448, 0.0071487427, -0.05114746, -0.036590576, 0.011779785, -0.022018433, 0.0016717911, -0.011856079, -0.008735657, 0.023864746, -0.017471313, 0.0062942505, -0.0013780594, -0.031799316, -0.011634827, -0.015609741, -0.01121521, ...
http://arxiv.org/abs/2305.17698v1
Neural Machine Translation with Dynamic Graph Convolutional Decoder
Lei Li, Kai Fan, Lingyu Yang, Hongjia Li, Chun Yuan
cs.CL
Existing wisdom demonstrates the significance of syntactic knowledge for the improvement of neural machine translation models. However, most previous works merely focus on leveraging the source syntax in the well-known encoder-decoder framework. In sharp contrast, this paper proposes an end-to-end translation architect...
http://arxiv.org/pdf/2305.17698v1
2023-05-28T11:58:07
2023-05-28T11:58:07
{ "title": [ -0.040802002, -0.014862061, -0.028656006, -0.05731201, -0.014778137, 0.0116119385, -0.0501709, 0.008125305, -0.012649536, 0.04498291, 0.011299133, -0.0043144226, 0.024505615, -0.02848816, -0.032714844, 0.022964478, 0.003917694, 0.035...
http://arxiv.org/abs/2305.17696v1
SQuARe: A Large-Scale Dataset of Sensitive Questions and Acceptable Responses Created Through Human-Machine Collaboration
Hwaran Lee, Seokhee Hong, Joonsuk Park, Takyoung Kim, Meeyoung Cha, Yejin Choi, Byoung Pil Kim, Gunhee Kim, Eun-Ju Lee, Yong Lim, Alice Oh, Sangchul Park, Jung-Woo Ha
cs.CL
The potential social harms that large language models pose, such as generating offensive content and reinforcing biases, are steeply rising. Existing works focus on coping with this concern while interacting with ill-intentioned users, such as those who explicitly make hate speech or elicit harmful responses. However, ...
http://arxiv.org/pdf/2305.17696v1
2023-05-28T11:51:20
2023-05-28T11:51:20
{ "title": [ 0.011566162, -0.018341064, -0.07312012, -0.013633728, -0.019729614, -0.06185913, 0.011985779, 0.01209259, -0.005405426, 0.020309448, -0.023101807, -0.007980347, 0.0016593933, -0.036468506, -0.021057129, -0.037506104, -0.0009241104, -...
http://arxiv.org/abs/2305.18391v2
MemeGraphs: Linking Memes to Knowledge Graphs
Vasiliki Kougia, Simon Fetzel, Thomas Kirchmair, Erion Çano, Sina Moayed Baharlou, Sahand Sharifzadeh, Benjamin Roth
cs.LG, cs.CL, cs.CV
Memes are a popular form of communicating trends and ideas in social media and on the internet in general, combining the modalities of images and text. They can express humor and sarcasm but can also have offensive content. Analyzing and classifying memes automatically is challenging since their interpretation relies o...
http://arxiv.org/pdf/2305.18391v2
2023-06-26T16:15:48
2023-05-28T11:17:30
{ "title": [ -0.0029125214, -0.010513306, -0.052764893, -0.05291748, 0.0015487671, -0.0072135925, -0.02458191, -0.008773804, 0.029129028, 0.032928467, -0.015045166, 0.0149383545, -0.011161804, -0.014785767, -0.04248047, 0.01828003, -0.0040779114, ...
http://arxiv.org/abs/2305.18390v2
Emergent Modularity in Pre-trained Transformers
Zhengyan Zhang, Zhiyuan Zeng, Yankai Lin, Chaojun Xiao, Xiaozhi Wang, Xu Han, Zhiyuan Liu, Ruobing Xie, Maosong Sun, Jie Zhou
cs.CL, cs.LG
This work examines the presence of modularity in pre-trained Transformers, a feature commonly found in human brains and thought to be vital for general intelligence. In analogy to human brains, we consider two main characteristics of modularity: (1) functional specialization of neurons: we evaluate whether each neuron ...
http://arxiv.org/pdf/2305.18390v2
2023-10-30T07:40:35
2023-05-28T11:02:32
{ "title": [ -0.0082473755, 0.0126571655, -0.033569336, -0.08673096, 0.017486572, 0.0007390976, -0.036590576, 0.00088500977, -0.034057617, 0.03930664, 0.0016975403, 0.022445679, -0.008079529, -0.03149414, -0.023925781, 0.007499695, 0.002861023, 0...
http://arxiv.org/abs/2305.17691v1
Plug-and-Play Knowledge Injection for Pre-trained Language Models
Zhengyan Zhang, Zhiyuan Zeng, Yankai Lin, Huadong Wang, Deming Ye, Chaojun Xiao, Xu Han, Zhiyuan Liu, Peng Li, Maosong Sun, Jie Zhou
cs.CL
Injecting external knowledge can improve the performance of pre-trained language models (PLMs) on various downstream NLP tasks. However, massive retraining is required to deploy new knowledge injection methods or knowledge bases for downstream tasks. In this work, we are the first to study how to improve the flexibilit...
http://arxiv.org/pdf/2305.17691v1
2023-05-28T10:58:00
2023-05-28T10:58:00
{ "title": [ -0.015838623, -0.025802612, -0.034973145, -0.07720947, -0.011428833, 0.001376152, -0.0513916, 0.010383606, -0.016525269, 0.018127441, -0.038360596, 0.016998291, -0.009750366, -0.012084961, -0.024658203, -0.007583618, -0.002008438, 0....
http://arxiv.org/abs/2305.17690v1
HaVQA: A Dataset for Visual Question Answering and Multimodal Research in Hausa Language
Shantipriya Parida, Idris Abdulmumin, Shamsuddeen Hassan Muhammad, Aneesh Bose, Guneet Singh Kohli, Ibrahim Said Ahmad, Ketan Kotwal, Sayan Deb Sarkar, Ondřej Bojar, Habeebah Adamu Kakudi
cs.CL
This paper presents HaVQA, the first multimodal dataset for visual question-answering (VQA) tasks in the Hausa language. The dataset was created by manually translating 6,022 English question-answer pairs, which are associated with 1,555 unique images from the Visual Genome dataset. As a result, the dataset provides 12...
http://arxiv.org/pdf/2305.17690v1
2023-05-28T10:55:31
2023-05-28T10:55:31
{ "title": [ -0.028244019, -0.027297974, -0.008056641, -0.039245605, -0.039093018, 0.012039185, -0.057037354, 0.0023021698, 0.027709961, 0.022033691, -0.03274536, -0.0045547485, 0.0027675629, -0.028900146, -0.04736328, -0.0042419434, 0.021606445, ...
http://arxiv.org/abs/2305.17682v2
One Network, Many Masks: Towards More Parameter-Efficient Transfer Learning
Guangtao Zeng, Peiyuan Zhang, Wei Lu
cs.CL, cs.AI
Fine-tuning pre-trained language models for multiple tasks tends to be expensive in terms of storage. To mitigate this, parameter-efficient transfer learning (PETL) methods have been proposed to address this issue, but they still require a significant number of parameters and storage when being applied to broader range...
http://arxiv.org/pdf/2305.17682v2
2023-06-12T02:44:26
2023-05-28T10:27:14
{ "title": [ -0.020645142, 0.008964539, -0.055389404, -0.04446411, 0.02633667, -0.01084137, -0.018981934, -0.011482239, 0.00039553642, 0.0129470825, -0.0023345947, 0.023345947, -0.016540527, -0.025497437, -0.0109939575, 0.016967773, -0.0032749176, ...
http://arxiv.org/abs/2305.17680v4
Evaluating GPT-3 Generated Explanations for Hateful Content Moderation
Han Wang, Ming Shan Hee, Md Rabiul Awal, Kenny Tsu Wei Choo, Roy Ka-Wei Lee
cs.CL, cs.AI, I.2.7
Recent research has focused on using large language models (LLMs) to generate explanations for hate speech through fine-tuning or prompting. Despite the growing interest in this area, these generated explanations' effectiveness and potential limitations remain poorly understood. A key concern is that these explanations...
http://arxiv.org/pdf/2305.17680v4
2023-08-30T16:17:27
2023-05-28T10:05:13
{ "title": [ 0.034179688, 0.015060425, -0.008163452, -0.027450562, 0.016967773, -0.024627686, 0.018692017, -0.027740479, -0.026901245, 0.030059814, -0.010749817, 0.018035889, 0.01852417, -0.030197144, -0.016525269, 0.006198883, -0.02859497, 0.052...
http://arxiv.org/abs/2305.17679v1
RuSentNE-2023: Evaluating Entity-Oriented Sentiment Analysis on Russian News Texts
Anton Golubev, Nicolay Rusnachenko, Natalia Loukachevitch
cs.CL, I.2.7
The paper describes the RuSentNE-2023 evaluation devoted to targeted sentiment analysis in Russian news texts. The task is to predict sentiment towards a named entity in a single sentence. The dataset for RuSentNE-2023 evaluation is based on the Russian news corpus RuSentNE having rich sentiment-related annotation. The...
http://arxiv.org/pdf/2305.17679v1
2023-05-28T10:04:15
2023-05-28T10:04:15
{ "title": [ -0.00032901764, -0.012809753, -0.016113281, -0.029418945, 0.010734558, -0.007080078, 0.015670776, -0.0059280396, 0.011955261, 0.035980225, -0.03149414, 0.020004272, 0.0061798096, 0.0038108826, -0.05670166, 0.002216339, -0.024978638, ...
http://arxiv.org/abs/2305.17678v2
Decoding the Underlying Meaning of Multimodal Hateful Memes
Ming Shan Hee, Wen-Haw Chong, Roy Ka-Wei Lee
cs.CL, cs.AI, cs.CV, I.2.7; I.2.10
Recent studies have proposed models that yielded promising performance for the hateful meme classification task. Nevertheless, these proposed models do not generate interpretable explanations that uncover the underlying meaning and support the classification output. A major reason for the lack of explainable hateful me...
http://arxiv.org/pdf/2305.17678v2
2023-06-19T17:07:33
2023-05-28T10:02:59
{ "title": [ 0.02961731, 0.0063209534, -0.029907227, -0.021362305, 0.0046806335, -0.030670166, 0.0050239563, -0.026031494, -0.010650635, 0.04196167, 0.011192322, 0.020629883, 0.018157959, -0.024612427, -0.02722168, 0.017578125, 0.000014066696, 0....
http://arxiv.org/abs/2305.17670v1
Stochastic Bridges as Effective Regularizers for Parameter-Efficient Tuning
Weize Chen, Xu Han, Yankai Lin, Zhiyuan Liu, Maosong Sun, Jie Zhou
cs.CL
Parameter-efficient tuning methods (PETs) have achieved promising results in tuning large pre-trained language models (PLMs). By formalizing frozen PLMs and additional tunable parameters as systems and controls respectively, PETs can be theoretically grounded to optimal control and further viewed as optimizing the term...
http://arxiv.org/pdf/2305.17670v1
2023-05-28T09:22:44
2023-05-28T09:22:44
{ "title": [ -0.04348755, 0.0058898926, -0.027328491, -0.072265625, 0.016616821, 0.01676941, -0.060760498, 0.006565094, -0.034454346, 0.020019531, 0.037231445, 0.022109985, -0.019897461, -0.04449463, -0.015541077, 0.0016860962, -0.0054626465, 0.0...
http://arxiv.org/abs/2305.17663v1
Lexical Retrieval Hypothesis in Multimodal Context
Po-Ya Angela Wang, Pin-Er Chen, Hsin-Yu Chou, Yu-Hsiang Tseng, Shu-Kai Hsieh
cs.CL
Multimodal corpora have become an essential language resource for language science and grounded natural language processing (NLP) systems due to the growing need to understand and interpret human communication across various channels. In this paper, we first present our efforts in building the first Multimodal Corpus f...
http://arxiv.org/pdf/2305.17663v1
2023-05-28T08:17:07
2023-05-28T08:17:07
{ "title": [ -0.015777588, 0.0077819824, -0.04095459, -0.028457642, -0.011451721, -0.022262573, -0.020523071, -0.009918213, 0.0079574585, 0.054260254, -0.026748657, 0.0070877075, 0.026992798, -0.0093307495, -0.04107666, 0.010688782, -0.015022278, ...
http://arxiv.org/abs/2305.17660v1
Plug-and-Play Document Modules for Pre-trained Models
Chaojun Xiao, Zhengyan Zhang, Xu Han, Chi-Min Chan, Yankai Lin, Zhiyuan Liu, Xiangyang Li, Zhonghua Li, Zhao Cao, Maosong Sun
cs.CL
Large-scale pre-trained models (PTMs) have been widely used in document-oriented NLP tasks, such as question answering. However, the encoding-task coupling requirement results in the repeated encoding of the same documents for different tasks and queries, which is highly computationally inefficient. To this end, we tar...
http://arxiv.org/pdf/2305.17660v1
2023-05-28T08:01:40
2023-05-28T08:01:40
{ "title": [ 0.0118637085, -0.01197052, -0.038208008, -0.09240723, -0.03491211, -0.0067977905, -0.089904785, -0.007713318, -0.039215088, 0.015914917, -0.04257202, 0.009986877, -0.01373291, -0.027862549, 0.024169922, -0.027114868, 0.033203125, 0.0...
http://arxiv.org/abs/2305.17653v1
Prompt-Guided Retrieval Augmentation for Non-Knowledge-Intensive Tasks
Zhicheng Guo, Sijie Cheng, Yile Wang, Peng Li, Yang Liu
cs.CL
Retrieval-augmented methods have received increasing attention to support downstream tasks by leveraging useful information from external resources. Recent studies mainly focus on exploring retrieval to solve knowledge-intensive (KI) tasks. However, the potential of retrieval for most non-knowledge-intensive (NKI) task...
http://arxiv.org/pdf/2305.17653v1
2023-05-28T07:27:12
2023-05-28T07:27:12
{ "title": [ 0.027755737, -0.0054397583, -0.038146973, -0.036315918, 0.002407074, -0.024856567, -0.012550354, -0.01424408, 0.016159058, 0.03515625, -0.018157959, 0.0027637482, 0.018005371, -0.01689148, -0.04168701, 0.017700195, 0.0061149597, 0.00...
http://arxiv.org/abs/2305.17652v1
ConaCLIP: Exploring Distillation of Fully-Connected Knowledge Interaction Graph for Lightweight Text-Image Retrieval
Jiapeng Wang, Chengyu Wang, Xiaodan Wang, Jun Huang, Lianwen Jin
cs.CV, cs.CL
Large-scale pre-trained text-image models with dual-encoder architectures (such as CLIP) are typically adopted for various vision-language applications, including text-image retrieval. However,these models are still less practical on edge devices or for real-time situations, due to the substantial indexing and inferenc...
http://arxiv.org/pdf/2305.17652v1
2023-05-28T07:16:44
2023-05-28T07:16:44
{ "title": [ -0.0061836243, -0.014533997, -0.030792236, -0.04650879, -0.010757446, -0.0029029846, -0.020690918, -0.0032520294, 0.03656006, 0.049346924, -0.048339844, 0.019927979, 0.0019683838, 0.005748749, -0.05407715, -0.009399414, 0.0008010864, ...
http://arxiv.org/abs/2305.17651v1
DPHuBERT: Joint Distillation and Pruning of Self-Supervised Speech Models
Yifan Peng, Yui Sudo, Shakeel Muhammad, Shinji Watanabe
cs.CL, cs.SD, eess.AS
Self-supervised learning (SSL) has achieved notable success in many speech processing tasks, but the large model size and heavy computational cost hinder the deployment. Knowledge distillation trains a small student model to mimic the behavior of a large teacher model. However, the student architecture usually needs to...
http://arxiv.org/pdf/2305.17651v1
2023-05-28T07:09:33
2023-05-28T07:09:33
{ "title": [ -0.0054626465, 0.004753113, -0.051605225, -0.09838867, 0.015335083, 0.002735138, -0.08337402, 0.009468079, -0.0023727417, 0.035003662, -0.0029067993, -0.0056610107, 0.006690979, -0.030471802, -0.03768921, 0.007133484, -0.010505676, 0...
http://arxiv.org/abs/2305.18373v1
KAFA: Rethinking Image Ad Understanding with Knowledge-Augmented Feature Adaptation of Vision-Language Models
Zhiwei Jia, Pradyumna Narayana, Arjun R. Akula, Garima Pruthi, Hao Su, Sugato Basu, Varun Jampani
cs.CV, cs.CL
Image ad understanding is a crucial task with wide real-world applications. Although highly challenging with the involvement of diverse atypical scenes, real-world entities, and reasoning over scene-texts, how to interpret image ads is relatively under-explored, especially in the era of foundational vision-language mod...
http://arxiv.org/pdf/2305.18373v1
2023-05-28T04:49:01
2023-05-28T04:49:01
{ "title": [ -0.015319824, -0.016235352, -0.026626587, -0.015823364, -0.001288414, -0.022781372, -0.019683838, -0.013038635, 0.017288208, 0.010673523, -0.027786255, -0.0066986084, -0.009086609, -0.01687622, -0.035583496, -0.012001038, 0.024749756, ...
http://arxiv.org/abs/2305.17627v1
Robust Natural Language Understanding with Residual Attention Debiasing
Fei Wang, James Y. Huang, Tianyi Yan, Wenxuan Zhou, Muhao Chen
cs.CL, cs.AI, cs.LG
Natural language understanding (NLU) models often suffer from unintended dataset biases. Among bias mitigation methods, ensemble-based debiasing methods, especially product-of-experts (PoE), have stood out for their impressive empirical success. However, previous ensemble-based debiasing methods typically apply debiasi...
http://arxiv.org/pdf/2305.17627v1
2023-05-28T04:25:04
2023-05-28T04:25:04
{ "title": [ 0.0009918213, 0.019485474, -0.035858154, -0.08465576, 0.008880615, 0.0047302246, -0.076171875, -0.014602661, 0.02722168, 0.015167236, -0.03375244, -0.009635925, 0.0075798035, 0.0032196045, -0.050720215, -0.00881958, -0.010444641, 0.0...
http://arxiv.org/abs/2305.17626v2
In-Context Analogical Reasoning with Pre-Trained Language Models
Xiaoyang Hu, Shane Storks, Richard L. Lewis, Joyce Chai
cs.AI, cs.CL, cs.LG
Analogical reasoning is a fundamental capacity of human cognition that allows us to reason abstractly about novel situations by relating them to past experiences. While it is thought to be essential for robust reasoning in AI systems, conventional approaches require significant training and/or hard-coding of domain kno...
http://arxiv.org/pdf/2305.17626v2
2023-06-05T06:57:29
2023-05-28T04:22:26
{ "title": [ 0.010749817, -0.018493652, 0.0022850037, -0.0602417, -0.0132369995, 0.0047836304, -0.039611816, -0.019317627, 0.019332886, 0.038635254, -0.0049057007, -0.009056091, 0.02142334, -0.02357483, -0.038146973, 0.01776123, 0.0010585785, 0.0...
http://arxiv.org/abs/2305.17619v1
AI Coach Assist: An Automated Approach for Call Recommendation in Contact Centers for Agent Coaching
Md Tahmid Rahman Laskar, Cheng Chen, Xue-Yong Fu, Mahsa Azizi, Shashi Bhushan, Simon Corston-Oliver
cs.CL, cs.AI, cs.LG
In recent years, the utilization of Artificial Intelligence (AI) in the contact center industry is on the rise. One area where AI can have a significant impact is in the coaching of contact center agents. By analyzing call transcripts using Natural Language Processing (NLP) techniques, it would be possible to quickly d...
http://arxiv.org/pdf/2305.17619v1
2023-05-28T03:29:59
2023-05-28T03:29:59
{ "title": [ 0.003129959, -0.021606445, -0.013771057, -0.032684326, -0.023727417, -0.0096206665, -0.004398346, -0.018066406, -0.000099897385, 0.038391113, -0.07122803, 0.02619934, 0.02458191, -0.0023841858, -0.019989014, -0.023025513, -0.012138367, ...
http://arxiv.org/abs/2305.17608v1
Reward Collapse in Aligning Large Language Models
Ziang Song, Tianle Cai, Jason D. Lee, Weijie J. Su
cs.LG, cs.AI, cs.CL, math.OC, stat.ML
The extraordinary capabilities of large language models (LLMs) such as ChatGPT and GPT-4 are in part unleashed by aligning them with reward models that are trained on human preferences, which are often represented as rankings of responses to prompts. In this paper, we document the phenomenon of \textit{reward collapse}...
http://arxiv.org/pdf/2305.17608v1
2023-05-28T02:12:00
2023-05-28T02:12:00
{ "title": [ -0.033721924, 0.0017061234, -0.037719727, -0.031555176, 0.0053634644, 0.002691269, -0.046844482, -0.015594482, 0.012260437, 0.03869629, -0.01977539, -0.010681152, 0.0034694672, -0.0071029663, -0.05355835, 0.008201599, -0.005302429, 0...
http://arxiv.org/abs/2305.17607v1
More than Classification: A Unified Framework for Event Temporal Relation Extraction
Quzhe Huang, Yutong Hu, Shengqi Zhu, Yansong Feng, Chang Liu, Dongyan Zhao
cs.CL
Event temporal relation extraction~(ETRE) is usually formulated as a multi-label classification task, where each type of relation is simply treated as a one-hot label. This formulation ignores the meaning of relations and wipes out their intrinsic dependency. After examining the relation definitions in various ETRE tas...
http://arxiv.org/pdf/2305.17607v1
2023-05-28T02:09:08
2023-05-28T02:09:08
{ "title": [ 0.017532349, 0.0023994446, -0.05682373, -0.12695312, 0.046142578, -0.0008068085, -0.05984497, 0.008140564, 0.032409668, 0.021011353, 0.007888794, -0.00035977364, -0.012245178, -0.019332886, -0.0574646, 0.0082473755, -0.015914917, 0.0...
http://arxiv.org/abs/2305.17588v2
Diagnosing Transformers: Illuminating Feature Spaces for Clinical Decision-Making
Aliyah R. Hsu, Yeshwanth Cherapanamjeri, Briton Park, Tristan Naumann, Anobel Y. Odisho, Bin Yu
cs.CL, cs.AI, cs.LG
Pre-trained transformers are often fine-tuned to aid clinical decision-making using limited clinical notes. Model interpretability is crucial, especially in high-stakes domains like medicine, to establish trust and ensure safety, which requires human engagement. We introduce SUFO, a systematic framework that enhances i...
http://arxiv.org/pdf/2305.17588v2
2023-10-28T17:07:45
2023-05-27T22:15:48
{ "title": [ 0.0138549805, -0.0034980774, -0.0046577454, -0.059326172, 0.0042800903, 0.0050086975, -0.024459839, -0.03741455, -0.011688232, 0.006587982, -0.043182373, -0.004470825, -0.015411377, -0.032592773, -0.02557373, -0.022354126, 0.008003235, ...
http://arxiv.org/abs/2305.17580v1
ArPanEmo: An Open-Source Dataset for Fine-Grained Emotion Recognition in Arabic Online Content during COVID-19 Pandemic
Maha Jarallah Althobaiti
cs.CL
Emotion recognition is a crucial task in Natural Language Processing (NLP) that enables machines to comprehend the feelings conveyed in the text. The applications of emotion recognition are diverse, including mental health diagnosis, student support, and the detection of online suspicious behavior. Despite the substant...
http://arxiv.org/pdf/2305.17580v1
2023-05-27T21:04:26
2023-05-27T21:04:26
{ "title": [ -0.008995056, -0.018569946, -0.051086426, -0.038391113, -0.014411926, -0.02130127, -0.01637268, 0.040924072, 0.037109375, 0.040527344, -0.0064697266, -0.012557983, 0.045715332, -0.052368164, -0.02961731, -0.00082588196, 0.0071029663, ...
http://arxiv.org/abs/2305.17561v1
Grounding Characters and Places in Narrative Texts
Sandeep Soni, Amanpreet Sihra, Elizabeth F. Evans, Matthew Wilkens, David Bamman
cs.CL
Tracking characters and locations throughout a story can help improve the understanding of its plot structure. Prior research has analyzed characters and locations from text independently without grounding characters to their locations in narrative time. Here, we address this gap by proposing a new spatial relationship...
http://arxiv.org/pdf/2305.17561v1
2023-05-27T19:31:41
2023-05-27T19:31:41
{ "title": [ 0.024093628, 0.00027394295, -0.015350342, -0.0546875, -0.007183075, -0.016098022, -0.0016317368, -0.0012893677, -0.011161804, 0.052612305, -0.032104492, 0.00762558, 0.028503418, -0.022003174, -0.0146484375, -0.004749298, -0.007987976, ...
http://arxiv.org/abs/2305.17553v2
Detecting Edit Failures In Large Language Models: An Improved Specificity Benchmark
Jason Hoelscher-Obermaier, Julia Persson, Esben Kran, Ioannis Konstas, Fazl Barez
cs.CL, cs.AI, cs.LG, I.2.7
Recent model editing techniques promise to mitigate the problem of memorizing false or outdated associations during LLM training. However, we show that these techniques can introduce large unwanted side effects which are not detected by existing specificity benchmarks. We extend the existing CounterFact benchmark to in...
http://arxiv.org/pdf/2305.17553v2
2023-06-03T08:01:11
2023-05-27T19:08:04
{ "title": [ -0.025238037, -0.009178162, -0.024597168, -0.08428955, -0.008132935, 0.013015747, -0.096069336, -0.008155823, 0.010551453, 0.029754639, -0.0025806427, 0.008514404, 0.026504517, -0.017562866, -0.03717041, 0.019927979, -0.034301758, 0....
http://arxiv.org/abs/2305.17547v2
Translatotron 3: Speech to Speech Translation with Monolingual Data
Eliya Nachmani, Alon Levkovitch, Yifan Ding, Chulayuth Asawaroengchai, Heiga Zen, Michelle Tadmor Ramanovich
cs.CL, cs.LG, cs.SD, eess.AS
This paper presents Translatotron 3, a novel approach to train a direct speech-to-speech translation model from monolingual speech-text datasets only in a fully unsupervised manner. Translatotron 3 combines masked autoencoder, unsupervised embedding mapping, and back-translation to achieve this goal. Experimental resul...
http://arxiv.org/pdf/2305.17547v2
2023-06-01T08:01:16
2023-05-27T18:30:54
{ "title": [ -0.029785156, -0.0014944077, -0.041015625, -0.06561279, 0.013145447, 0.0075531006, -0.05718994, -0.008636475, -0.007835388, 0.056427002, -0.001613617, 0.011779785, 0.027023315, -0.024673462, -0.038970947, 0.02671814, -0.0032444, 0.03...
http://arxiv.org/abs/2305.17542v1
Non-Sequential Graph Script Induction via Multimedia Grounding
Yu Zhou, Sha Li, Manling Li, Xudong Lin, Shih-Fu Chang, Mohit Bansal, Heng Ji
cs.CL, cs.MM
Online resources such as WikiHow compile a wide range of scripts for performing everyday tasks, which can assist models in learning to reason about procedures. However, the scripts are always presented in a linear manner, which does not reflect the flexibility displayed by people executing tasks in real life. For examp...
http://arxiv.org/pdf/2305.17542v1
2023-05-27T18:13:17
2023-05-27T18:13:17
{ "title": [ -0.007297516, -0.011787415, -0.01864624, -0.061920166, -0.018829346, 0.0031833649, -0.048065186, 0.0088272095, 0.0066375732, -0.00844574, -0.034942627, 0.025360107, -0.031677246, -0.02961731, -0.009880066, -0.028045654, -0.00096702576, ...