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/2303.11648v1
Improving Content Retrievability in Search with Controllable Query Generation
Gustavo Penha, Enrico Palumbo, Maryam Aziz, Alice Wang, Hugues Bouchard
cs.IR, cs.CL, cs.LG
An important goal of online platforms is to enable content discovery, i.e. allow users to find a catalog entity they were not familiar with. A pre-requisite to discover an entity, e.g. a book, with a search engine is that the entity is retrievable, i.e. there are queries for which the system will surface such entity in...
http://arxiv.org/pdf/2303.11648v1
2023-03-21T07:46:57
2023-03-21T07:46:57
{ "title": [ 0.0011692047, 0.00010210276, -0.022338867, -0.04031372, 0.019012451, -0.007499695, -0.0069007874, -0.02999878, -0.001698494, 0.04119873, -0.0104904175, 0.012504578, -0.007587433, -0.005672455, -0.04699707, 0.0018129349, -0.021743774, ...
http://arxiv.org/abs/2303.13373v1
Fine-tuning ClimateBert transformer with ClimaText for the disclosure analysis of climate-related financial risks
Eduardo C. Garrido-Merchán, Cristina González-Barthe, María Coronado Vaca
cs.CL, cs.AI
In recent years there has been a growing demand from financial agents, especially from particular and institutional investors, for companies to report on climate-related financial risks. A vast amount of information, in text format, can be expected to be disclosed in the short term by firms in order to identify these t...
http://arxiv.org/pdf/2303.13373v1
2023-03-21T07:25:36
2023-03-21T07:25:36
{ "title": [ 0.012275696, -0.0025749207, -0.0390625, -0.038757324, -0.017623901, -0.0028629303, -0.026428223, 0.0036830902, -0.035858154, 0.06402588, -0.061798096, 0.042785645, 0.037475586, 0.008125305, -0.012084961, -0.035339355, 0.0023059845, 0...
http://arxiv.org/abs/2303.11621v1
Heterogeneous-Branch Collaborative Learning for Dialogue Generation
Yiwei Li, Shaoxiong Feng, Bin Sun, Kan Li
cs.CL, cs.AI
With the development of deep learning, advanced dialogue generation methods usually require a greater amount of computational resources. One promising approach to obtaining a high-performance and lightweight model is knowledge distillation, which relies heavily on the pre-trained powerful teacher. Collaborative learnin...
http://arxiv.org/pdf/2303.11621v1
2023-03-21T06:41:50
2023-03-21T06:41:50
{ "title": [ -0.021347046, -0.0021896362, -0.055358887, -0.06317139, 0.0012483597, 0.003545761, -0.056854248, 0.011672974, -0.0010528564, 0.026275635, -0.004119873, 0.00856781, -0.048309326, -0.017333984, -0.044006348, -0.024490356, -0.024597168, ...
http://arxiv.org/abs/2303.11607v1
Transformers in Speech Processing: A Survey
Siddique Latif, Aun Zaidi, Heriberto Cuayahuitl, Fahad Shamshad, Moazzam Shoukat, Junaid Qadir
cs.CL, cs.SD, eess.AS
The remarkable success of transformers in the field of natural language processing has sparked the interest of the speech-processing community, leading to an exploration of their potential for modeling long-range dependencies within speech sequences. Recently, transformers have gained prominence across various speech-r...
http://arxiv.org/pdf/2303.11607v1
2023-03-21T06:00:39
2023-03-21T06:00:39
{ "title": [ -0.0067443848, 0.0038795471, -0.024780273, -0.0692749, 0.027542114, 0.009254456, -0.033813477, -0.009223938, -0.009239197, 0.0501709, -0.0362854, -0.0041046143, 0.028579712, -0.011024475, -0.056793213, 0.0070381165, -0.01449585, 0.06...
http://arxiv.org/abs/2303.11593v3
How does Transformer model evolve to learn diverse chemical structures?
Yasuhiro Yoshikai, Tadahaya Mizuno, Shumpei Nemoto, Hiroyuki Kusuhara
cs.LG, cs.CL, physics.chem-ph, q-bio.BM, J.2; I.2.7
Recent years have seen rapid development of descriptor generation based on representation learning of extremely diverse molecules, especially those that apply natural language processing (NLP) models to SMILES, a literal representation of molecular structure. However, little research has been done on how these models u...
http://arxiv.org/pdf/2303.11593v3
2023-10-07T06:34:16
2023-03-21T04:47:45
{ "title": [ -0.021270752, 0.0023117065, -0.051727295, -0.0055236816, -0.018585205, -0.009666443, -0.058013916, -0.009719849, -0.029052734, 0.020599365, -0.024047852, 0.013221741, -0.023513794, -0.025802612, 0.027267456, -0.023483276, 0.028945923, ...
http://arxiv.org/abs/2303.11525v2
Sparse Iso-FLOP Transformations for Maximizing Training Efficiency
Shreyas Saxena, Vithursan Thangarasa, Abhay Gupta, Sean Lie
cs.LG, cs.CL, cs.CV
Recent works have explored the use of weight sparsity to improve the training efficiency (test accuracy w.r.t training FLOPs) of deep neural networks (DNNs). These works aim to reduce training FLOPs but training with sparse weights often leads to accuracy loss or requires longer training schedules, making the resulting...
http://arxiv.org/pdf/2303.11525v2
2023-03-25T15:35:03
2023-03-21T01:06:37
{ "title": [ -0.032409668, -0.0023288727, -0.048431396, -0.017532349, 0.0053215027, -0.0062446594, -0.0056381226, -0.013381958, 0.007007599, 0.04876709, 0.0017728806, -0.0032730103, 0.015014648, -0.02406311, -0.04736328, 0.008117676, 0.01374054, ...
http://arxiv.org/abs/2303.11504v2
Language Model Behavior: A Comprehensive Survey
Tyler A. Chang, Benjamin K. Bergen
cs.CL
Transformer language models have received widespread public attention, yet their generated text is often surprising even to NLP researchers. In this survey, we discuss over 250 recent studies of English language model behavior before task-specific fine-tuning. Language models possess basic capabilities in syntax, seman...
http://arxiv.org/pdf/2303.11504v2
2023-08-26T00:33:23
2023-03-20T23:54:26
{ "title": [ -0.03543091, -0.011627197, -0.028793335, -0.070129395, 0.007865906, 0.019851685, -0.059906006, 0.0087509155, 0.00072193146, 0.037750244, -0.00806427, 0.007156372, 0.012145996, -0.01713562, -0.056549072, 0.027160645, -0.02470398, 0.03...
http://arxiv.org/abs/2303.11455v1
Large Language Models and Simple, Stupid Bugs
Kevin Jesse, Toufique Ahmed, Premkumar T. Devanbu, Emily Morgan
cs.SE, cs.CL, cs.LG
With the advent of powerful neural language models, AI-based systems to assist developers in coding tasks are becoming widely available; Copilot is one such system. Copilot uses Codex, a large language model (LLM), to complete code conditioned on a preceding "prompt". Codex, however, is trained on public GitHub reposit...
http://arxiv.org/pdf/2303.11455v1
2023-03-20T21:14:06
2023-03-20T21:14:06
{ "title": [ 0.00068092346, -0.02268982, -0.026535034, -0.066223145, -0.027709961, -0.011314392, -0.03994751, 0.012863159, 0.008018494, 0.038238525, -0.03842163, -0.026855469, -0.014678955, -0.027282715, -0.027740479, 0.0016422272, 0.0046463013, ...
http://arxiv.org/abs/2303.11436v2
Mind meets machine: Unravelling GPT-4's cognitive psychology
Sifatkaur Dhingra, Manmeet Singh, Vaisakh SB, Neetiraj Malviya, Sukhpal Singh Gill
cs.CL, cs.AI
Cognitive psychology delves on understanding perception, attention, memory, language, problem-solving, decision-making, and reasoning. Large language models (LLMs) are emerging as potent tools increasingly capable of performing human-level tasks. The recent development in the form of GPT-4 and its demonstrated success ...
http://arxiv.org/pdf/2303.11436v2
2023-04-12T15:46:20
2023-03-20T20:28:26
{ "title": [ -0.0087890625, 0.004333496, -0.016082764, -0.019622803, 0.018127441, -0.022125244, 0.015609741, -0.049591064, -0.010726929, 0.009590149, -0.009780884, 0.005393982, -0.017745972, -0.036743164, -0.029266357, 0.0046081543, -0.010437012, ...
http://arxiv.org/abs/2303.11403v4
eP-ALM: Efficient Perceptual Augmentation of Language Models
Mustafa Shukor, Corentin Dancette, Matthieu Cord
cs.CV, cs.CL, cs.LG
Large Language Models (LLMs) have so far impressed the world, with unprecedented capabilities that emerge in models at large scales. On the vision side, transformer models (i.e., ViT) are following the same trend, achieving the best performance on challenging benchmarks. With the abundance of such unimodal models, a na...
http://arxiv.org/pdf/2303.11403v4
2023-10-27T16:38:40
2023-03-20T19:20:34
{ "title": [ -0.0022182465, -0.007003784, -0.022857666, -0.064208984, -0.0009727478, -0.0047073364, -0.048553467, -0.00052547455, 0.016326904, 0.04046631, -0.013389587, 0.0070266724, 0.02494812, -0.009490967, -0.042266846, 0.027648926, -0.0064888, ...
http://arxiv.org/abs/2303.11381v1
MM-REACT: Prompting ChatGPT for Multimodal Reasoning and Action
Zhengyuan Yang, Linjie Li, Jianfeng Wang, Kevin Lin, Ehsan Azarnasab, Faisal Ahmed, Zicheng Liu, Ce Liu, Michael Zeng, Lijuan Wang
cs.CV, cs.CL, cs.LG
We propose MM-REACT, a system paradigm that integrates ChatGPT with a pool of vision experts to achieve multimodal reasoning and action. In this paper, we define and explore a comprehensive list of advanced vision tasks that are intriguing to solve, but may exceed the capabilities of existing vision and vision-language...
http://arxiv.org/pdf/2303.11381v1
2023-03-20T18:31:47
2023-03-20T18:31:47
{ "title": [ -0.013549805, -0.016403198, -0.026748657, -0.04345703, -0.014968872, 0.008102417, -0.042999268, 0.0032367706, -0.0096206665, 0.0030002594, -0.028457642, 0.02444458, -0.0060157776, -0.020126343, -0.0017967224, -0.007232666, -0.009269714, ...
http://arxiv.org/abs/2303.11366v4
Reflexion: Language Agents with Verbal Reinforcement Learning
Noah Shinn, Federico Cassano, Edward Berman, Ashwin Gopinath, Karthik Narasimhan, Shunyu Yao
cs.AI, cs.CL, cs.LG
Large language models (LLMs) have been increasingly used to interact with external environments (e.g., games, compilers, APIs) as goal-driven agents. However, it remains challenging for these language agents to quickly and efficiently learn from trial-and-error as traditional reinforcement learning methods require exte...
http://arxiv.org/pdf/2303.11366v4
2023-10-10T05:21:45
2023-03-20T18:08:50
{ "title": [ -0.019348145, -0.008850098, -0.017730713, -0.051483154, -0.0035591125, 0.00059604645, -0.01020813, -0.0049476624, -0.008811951, 0.019546509, -0.027450562, 0.0069503784, -0.019378662, -0.009437561, -0.03640747, -0.012252808, -0.017791748,...
http://arxiv.org/abs/2303.11331v2
EVA-02: A Visual Representation for Neon Genesis
Yuxin Fang, Quan Sun, Xinggang Wang, Tiejun Huang, Xinlong Wang, Yue Cao
cs.CV, cs.CL
We launch EVA-02, a next-generation Transformer-based visual representation pre-trained to reconstruct strong and robust language-aligned vision features via masked image modeling. With an updated plain Transformer architecture as well as extensive pre-training from an open & accessible giant CLIP vision encoder, EVA-0...
http://arxiv.org/pdf/2303.11331v2
2023-03-22T14:10:37
2023-03-20T17:59:59
{ "title": [ -0.0032653809, 0.0069885254, -0.037994385, 0.021469116, -0.02067566, -0.017349243, 0.0109939575, -0.046569824, 0.008758545, 0.050842285, 0.025726318, -0.02861023, 0.055786133, -0.059692383, -0.02355957, 0.016082764, 0.011505127, 0.01...
http://arxiv.org/abs/2303.11315v2
Context-faithful Prompting for Large Language Models
Wenxuan Zhou, Sheng Zhang, Hoifung Poon, Muhao Chen
cs.CL
Large language models (LLMs) encode parametric knowledge about world facts and have shown remarkable performance in knowledge-driven NLP tasks. However, their reliance on parametric knowledge may cause them to overlook contextual cues, leading to incorrect predictions in context-sensitive NLP tasks (e.g., knowledge acq...
http://arxiv.org/pdf/2303.11315v2
2023-10-23T03:25:13
2023-03-20T17:54:58
{ "title": [ -0.0104904175, -0.008377075, -0.027648926, -0.08911133, 0.0025749207, 0.008453369, -0.0758667, 0.0074272156, 0.013076782, 0.031280518, -0.04067993, 0.0007367134, 0.018005371, -0.0046958923, -0.036010742, 0.011360168, 0.009277344, -0....
http://arxiv.org/abs/2303.13314v1
Leveraging Foundation Models for Clinical Text Analysis
Shaina Raza, Syed Raza Bashir
cs.CL, cs.AI
Infectious diseases are a significant public health concern globally, and extracting relevant information from scientific literature can facilitate the development of effective prevention and treatment strategies. However, the large amount of clinical data available presents a challenge for information extraction. To a...
http://arxiv.org/pdf/2303.13314v1
2023-03-20T17:05:13
2023-03-20T17:05:13
{ "title": [ -0.036254883, 0.0007634163, -0.05029297, -0.07739258, 0.00007921457, 0.021713257, -0.08282471, -0.0034046173, 0.0075416565, 0.033843994, -0.010772705, 0.016189575, -0.00856781, -0.010574341, -0.048187256, 0.010276794, -0.015838623, 0...
http://arxiv.org/abs/2303.13475v2
Learning Semantic Text Similarity to rank Hypernyms of Financial Terms
Sohom Ghosh, Ankush Chopra, Sudip Kumar Naskar
cs.CL
Over the years, there has been a paradigm shift in how users access financial services. With the advancement of digitalization more users have been preferring the online mode of performing financial activities. This has led to the generation of a huge volume of financial content. Most investors prefer to go through the...
http://arxiv.org/pdf/2303.13475v2
2023-08-12T23:51:53
2023-03-20T16:53:36
{ "title": [ -0.018218994, 0.011650085, -0.039978027, -0.039245605, 0.008216858, -0.003829956, -0.020874023, -0.01550293, 0.0070152283, 0.053894043, -0.035705566, 0.03353882, 0.02557373, 0.007095337, -0.052215576, 0.0012140274, -0.01737976, 0.019...
http://arxiv.org/abs/2303.13375v2
Capabilities of GPT-4 on Medical Challenge Problems
Harsha Nori, Nicholas King, Scott Mayer McKinney, Dean Carignan, Eric Horvitz
cs.CL, cs.AI
Large language models (LLMs) have demonstrated remarkable capabilities in natural language understanding and generation across various domains, including medicine. We present a comprehensive evaluation of GPT-4, a state-of-the-art LLM, on medical competency examinations and benchmark datasets. GPT-4 is a general-purpos...
http://arxiv.org/pdf/2303.13375v2
2023-04-12T16:48:39
2023-03-20T16:18:38
{ "title": [ 0.05609131, 0.015655518, 0.007369995, -0.039733887, 0.0001757145, -0.022979736, -0.009124756, -0.033569336, -0.00982666, 0.028366089, 0.030654907, -0.020904541, -0.008911133, -0.068115234, -0.012512207, 0.0149002075, 0.0116119385, 0....
http://arxiv.org/abs/2303.11192v1
Multimodal Shannon Game with Images
Vilém Zouhar, Sunit Bhattacharya, Ondřej Bojar
cs.CL
The Shannon game has long been used as a thought experiment in linguistics and NLP, asking participants to guess the next letter in a sentence based on its preceding context. We extend the game by introducing an optional extra modality in the form of image information. To investigate the impact of multimodal informatio...
http://arxiv.org/pdf/2303.11192v1
2023-03-20T15:22:11
2023-03-20T15:22:11
{ "title": [ 0.0045700073, 0.019058228, -0.010604858, 0.013671875, -0.015365601, -0.034179688, 0.011619568, -0.048034668, -0.011421204, 0.03338623, -0.02784729, 0.0023517609, -0.016998291, -0.021652222, -0.014785767, -0.019699097, 0.0066223145, 0...
http://arxiv.org/abs/2303.11184v1
Conversation Modeling to Predict Derailment
Jiaqing Yuan, Munindar P. Singh
cs.CL
Conversations among online users sometimes derail, i.e., break down into personal attacks. Such derailment has a negative impact on the healthy growth of cyberspace communities. The ability to predict whether ongoing conversations are likely to derail could provide valuable real-time insight to interlocutors and modera...
http://arxiv.org/pdf/2303.11184v1
2023-03-20T15:10:45
2023-03-20T15:10:45
{ "title": [ 0.008285522, 0.016723633, -0.04107666, -0.07122803, 0.0025939941, 0.0015850067, -0.041778564, 0.013427734, 0.01184845, 0.0126953125, -0.021331787, -0.014297485, 0.005519867, -0.039794922, 0.0042495728, -0.0004324913, -0.003929138, 0....
http://arxiv.org/abs/2303.11141v2
DocRED-FE: A Document-Level Fine-Grained Entity And Relation Extraction Dataset
Hongbo Wang, Weimin Xiong, Yifan Song, Dawei Zhu, Yu Xia, Sujian Li
cs.CL
Joint entity and relation extraction (JERE) is one of the most important tasks in information extraction. However, most existing works focus on sentence-level coarse-grained JERE, which have limitations in real-world scenarios. In this paper, we construct a large-scale document-level fine-grained JERE dataset DocRED-FE...
http://arxiv.org/pdf/2303.11141v2
2023-03-21T09:03:14
2023-03-20T14:19:58
{ "title": [ -0.010856628, -0.014045715, -0.07128906, -0.08135986, 0.030197144, -0.012672424, -0.033966064, 0.0029087067, 0.015144348, 0.028457642, -0.00831604, 0.009284973, 0.009468079, -0.007972717, -0.05432129, 0.013725281, -0.011672974, 0.047...
http://arxiv.org/abs/2303.11131v1
Cocktail HuBERT: Generalized Self-Supervised Pre-training for Mixture and Single-Source Speech
Maryam Fazel-Zarandi, Wei-Ning Hsu
cs.CL, cs.LG, cs.SD, eess.AS
Self-supervised learning leverages unlabeled data effectively, improving label efficiency and generalization to domains without labeled data. While recent work has studied generalization to more acoustic/linguistic domains, languages, and modalities, these investigations are limited to single-source speech with one pri...
http://arxiv.org/pdf/2303.11131v1
2023-03-20T14:07:13
2023-03-20T14:07:13
{ "title": [ 0.002861023, 0.0036506653, -0.058135986, -0.09820557, 0.008369446, 0.0005264282, -0.061676025, 0.022003174, -0.041259766, 0.047790527, 0.004283905, 0.03475952, 0.015274048, -0.035461426, -0.01020813, 0.017532349, -0.008468628, 0.0545...
http://arxiv.org/abs/2303.11117v3
EmotionIC: Emotional Inertia and Contagion-Driven Dependency Modeling for Emotion Recognition in Conversation
Yingjian Liu, Jiang Li, Xiaoping Wang, Zhigang Zeng
cs.CL
Emotion Recognition in Conversation (ERC) has attracted growing attention in recent years as a result of the advancement and implementation of human-computer interface technologies. In this paper, we propose a novel approach to dependency modeling driven by Emotional Inertia and Contagion (EmotionIC) for ERC task. Our ...
http://arxiv.org/pdf/2303.11117v3
2023-08-28T08:44:09
2023-03-20T13:58:35
{ "title": [ 0.0030288696, -0.0011892319, -0.052001953, -0.061920166, 0.0109939575, -0.0067100525, -0.037872314, 0.0075035095, 0.036132812, 0.034301758, -0.027954102, -0.0099105835, 0.020599365, -0.022338867, -0.06286621, 0.00036978722, -0.009185791,...
http://arxiv.org/abs/2303.11082v1
Evaluating Language Models for Knowledge Base Completion
Blerta Veseli, Sneha Singhania, Simon Razniewski, Gerhard Weikum
cs.CL, cs.AI
Structured knowledge bases (KBs) are a foundation of many intelligent applications, yet are notoriously incomplete. Language models (LMs) have recently been proposed for unsupervised knowledge base completion (KBC), yet, despite encouraging initial results, questions regarding their suitability remain open. Existing ev...
http://arxiv.org/pdf/2303.11082v1
2023-03-20T13:14:59
2023-03-20T13:14:59
{ "title": [ -0.03353882, -0.009132385, -0.04437256, -0.08947754, 0.004032135, 0.03010559, -0.08276367, 0.01247406, 0.012512207, 0.033447266, -0.044403076, 0.003112793, 0.002784729, 0.008796692, -0.06488037, 0.002910614, -0.016921997, 0.038391113...
http://arxiv.org/abs/2303.11032v1
DeID-GPT: Zero-shot Medical Text De-Identification by GPT-4
Zhengliang Liu, Xiaowei Yu, Lu Zhang, Zihao Wu, Chao Cao, Haixing Dai, Lin Zhao, Wei Liu, Dinggang Shen, Quanzheng Li, Tianming Liu, Dajiang Zhu, Xiang Li
cs.CL, cs.CY
The digitization of healthcare has facilitated the sharing and re-using of medical data but has also raised concerns about confidentiality and privacy. HIPAA (Health Insurance Portability and Accountability Act) mandates removing re-identifying information before the dissemination of medical records. Thus, effective an...
http://arxiv.org/pdf/2303.11032v1
2023-03-20T11:34:37
2023-03-20T11:34:37
{ "title": [ 0.02319336, 0.01285553, -0.028518677, -0.038513184, 0.011421204, -0.018157959, -0.012145996, -0.03475952, -0.0034370422, 0.021347046, -0.020645142, 0.0026454926, 0.018447876, -0.04638672, -0.0068244934, -0.0015544891, -0.0054244995, ...
http://arxiv.org/abs/2303.11005v1
Controllable Ancient Chinese Lyrics Generation Based on Phrase Prototype Retrieving
Li Yi
cs.CL, cs.AI, cs.IR, cs.LG
Generating lyrics and poems is one of the essential downstream tasks in the Natural Language Processing (NLP) field. Current methods have performed well in some lyrics generation scenarios but need further improvements in tasks requiring fine control. We propose a novel method for generating ancient Chinese lyrics (Son...
http://arxiv.org/pdf/2303.11005v1
2023-03-20T10:33:06
2023-03-20T10:33:06
{ "title": [ -0.0317688, 0.00459671, -0.023406982, -0.0028629303, -0.014884949, 0.006690979, -0.018844604, -0.013244629, 0.00699234, 0.04626465, -0.02809143, 0.016082764, 0.006713867, -0.005569458, -0.04663086, 0.007499695, -0.013435364, 0.002933...
http://arxiv.org/abs/2303.10974v2
Translate your gibberish: black-box adversarial attack on machine translation systems
Andrei Chertkov, Olga Tsymboi, Mikhail Pautov, Ivan Oseledets
cs.CL, cs.AI
Neural networks are deployed widely in natural language processing tasks on the industrial scale, and perhaps the most often they are used as compounds of automatic machine translation systems. In this work, we present a simple approach to fool state-of-the-art machine translation tools in the task of translation from ...
http://arxiv.org/pdf/2303.10974v2
2023-05-23T19:19:54
2023-03-20T09:52:52
{ "title": [ -0.023956299, -0.022598267, -0.014465332, -0.036193848, -0.0042686462, -0.002817154, -0.0010881424, -0.015083313, -0.043945312, 0.055633545, -0.003982544, 0.019332886, 0.02760315, -0.010864258, -0.041748047, -0.0061035156, -0.012840271, ...
http://arxiv.org/abs/2303.10966v2
Towards Reliable Neural Machine Translation with Consistency-Aware Meta-Learning
Rongxiang Weng, Qiang Wang, Wensen Cheng, Changfeng Zhu, Min Zhang
cs.CL
Neural machine translation (NMT) has achieved remarkable success in producing high-quality translations. However, current NMT systems suffer from a lack of reliability, as their outputs that are often affected by lexical or syntactic changes in inputs, resulting in large variations in quality. This limitation hinders t...
http://arxiv.org/pdf/2303.10966v2
2023-09-19T07:37:31
2023-03-20T09:41:28
{ "title": [ -0.029953003, -0.00082445145, -0.016220093, -0.0473938, -0.0041885376, 0.021057129, -0.050842285, -0.00843811, -0.0119018555, 0.06210327, 0.010818481, -0.005973816, 0.024658203, -0.018432617, -0.062194824, 0.02168274, -0.005504608, 0...
http://arxiv.org/abs/2303.10949v1
Code-Switching Text Generation and Injection in Mandarin-English ASR
Haibin Yu, Yuxuan Hu, Yao Qian, Ma Jin, Linquan Liu, Shujie Liu, Yu Shi, Yanmin Qian, Edward Lin, Michael Zeng
eess.AS, cs.CL, cs.SD
Code-switching speech refers to a means of expression by mixing two or more languages within a single utterance. Automatic Speech Recognition (ASR) with End-to-End (E2E) modeling for such speech can be a challenging task due to the lack of data. In this study, we investigate text generation and injection for improving ...
http://arxiv.org/pdf/2303.10949v1
2023-03-20T09:13:27
2023-03-20T09:13:27
{ "title": [ -0.013847351, -0.0026569366, -0.012023926, -0.0129776, -0.020004272, 0.001033783, -0.02104187, -0.0018024445, -0.05102539, 0.041168213, -0.004322052, 0.016189575, 0.013328552, -0.034606934, -0.012016296, -0.007949829, -0.025238037, 0...
http://arxiv.org/abs/2303.10942v1
On-the-fly Text Retrieval for End-to-End ASR Adaptation
Bolaji Yusuf, Aditya Gourav, Ankur Gandhe, Ivan Bulyko
cs.CL, cs.SD, eess.AS
End-to-end speech recognition models are improved by incorporating external text sources, typically by fusion with an external language model. Such language models have to be retrained whenever the corpus of interest changes. Furthermore, since they store the entire corpus in their parameters, rare words can be challen...
http://arxiv.org/pdf/2303.10942v1
2023-03-20T08:54:40
2023-03-20T08:54:40
{ "title": [ -0.0289917, -0.014373779, -0.03933716, -0.070129395, -0.010231018, 0.004611969, -0.034698486, 0.029922485, 0.004009247, 0.058746338, -0.01876831, -0.011558533, 0.015205383, -0.0021305084, -0.061035156, 0.011184692, 0.0008749962, 0.03...
http://arxiv.org/abs/2303.10912v1
Exploring Representation Learning for Small-Footprint Keyword Spotting
Fan Cui, Liyong Guo, Quandong Wang, Peng Gao, Yujun Wang
cs.SD, cs.CL, cs.LG, eess.AS
In this paper, we investigate representation learning for low-resource keyword spotting (KWS). The main challenges of KWS are limited labeled data and limited available device resources. To address those challenges, we explore representation learning for KWS by self-supervised contrastive learning and self-training wit...
http://arxiv.org/pdf/2303.10912v1
2023-03-20T07:09:26
2023-03-20T07:09:26
{ "title": [ 0.0024147034, 0.00598526, -0.023223877, -0.024658203, -0.008071899, -0.023544312, -0.0011253357, -0.024353027, 0.0007648468, 0.056610107, -0.041137695, -0.0064201355, 0.019515991, 0.002708435, -0.053466797, -0.018112183, 0.005897522, ...
http://arxiv.org/abs/2303.10897v1
Relate auditory speech to EEG by shallow-deep attention-based network
Fan Cui, Liyong Guo, Lang He, Jiyao Liu, ErCheng Pei, Yujun Wang, Dongmei Jiang
cs.SD, cs.CL, eess.AS, q-bio.NC
Electroencephalography (EEG) plays a vital role in detecting how brain responses to different stimulus. In this paper, we propose a novel Shallow-Deep Attention-based Network (SDANet) to classify the correct auditory stimulus evoking the EEG signal. It adopts the Attention-based Correlation Module (ACM) to discover the...
http://arxiv.org/pdf/2303.10897v1
2023-03-20T06:34:22
2023-03-20T06:34:22
{ "title": [ 0.010955811, 0.014823914, -0.025466919, -0.022903442, -0.012680054, -0.026245117, -0.0118637085, -0.021713257, -0.012290955, 0.06021118, -0.008728027, -0.010131836, 0.063964844, -0.04763794, -0.021392822, 0.0065574646, 0.005924225, 0...
http://arxiv.org/abs/2303.10893v2
Character, Word, or Both? Revisiting the Segmentation Granularity for Chinese Pre-trained Language Models
Xinnian Liang, Zefan Zhou, Hui Huang, Shuangzhi Wu, Tong Xiao, Muyun Yang, Zhoujun Li, Chao Bian
cs.CL
Pretrained language models (PLMs) have shown marvelous improvements across various NLP tasks. Most Chinese PLMs simply treat an input text as a sequence of characters, and completely ignore word information. Although Whole Word Masking can alleviate this, the semantics in words is still not well represented. In this pa...
http://arxiv.org/pdf/2303.10893v2
2023-03-22T03:20:27
2023-03-20T06:20:03
{ "title": [ -0.024353027, -0.0020961761, -0.008270264, -0.04232788, -0.022949219, 0.008720398, -0.054626465, -0.008560181, -0.019622803, 0.02619934, -0.0018167496, 0.006072998, 0.026000977, -0.035949707, -0.025375366, 0.0104522705, -0.011550903, ...
http://arxiv.org/abs/2303.11945v1
Unsupervised Cross-Domain Rumor Detection with Contrastive Learning and Cross-Attention
Hongyan Ran, Caiyan Jia
cs.SI, cs.AI, cs.CL, cs.LG
Massive rumors usually appear along with breaking news or trending topics, seriously hindering the truth. Existing rumor detection methods are mostly focused on the same domain, and thus have poor performance in cross-domain scenarios due to domain shift. In this work, we propose an end-to-end instance-wise and prototy...
http://arxiv.org/pdf/2303.11945v1
2023-03-20T06:19:49
2023-03-20T06:19:49
{ "title": [ 0.006893158, 0.015586853, -0.034179688, -0.025756836, 0.015640259, -0.030471802, 0.0023479462, -0.043182373, 0.014091492, 0.02897644, -0.052459717, 0.007507324, 0.017074585, -0.0018501282, -0.028793335, -0.01449585, -0.046173096, 0.0...
http://arxiv.org/abs/2303.10888v1
Self-Improving-Leaderboard(SIL): A Call for Real-World Centric Natural Language Processing Leaderboards
Chanjun Park, Hyeonseok Moon, Seolhwa Lee, Jaehyung Seo, Sugyeong Eo, Heuiseok Lim
cs.CL, cs.AI
Leaderboard systems allow researchers to objectively evaluate Natural Language Processing (NLP) models and are typically used to identify models that exhibit superior performance on a given task in a predetermined setting. However, we argue that evaluation on a given test dataset is just one of many performance indicat...
http://arxiv.org/pdf/2303.10888v1
2023-03-20T06:13:03
2023-03-20T06:13:03
{ "title": [ 0.01335144, 0.0009536743, -0.017318726, -0.046905518, -0.016967773, -0.010597229, -0.02532959, -0.0061798096, -0.035705566, 0.04598999, -0.031173706, 0.031219482, -0.008323669, -0.020523071, -0.0042037964, -0.010292053, -0.0134887695, ...
http://arxiv.org/abs/2303.10870v1
Multi-task Transformer with Relation-attention and Type-attention for Named Entity Recognition
Ying Mo, Hongyin Tang, Jiahao Liu, Qifan Wang, Zenglin Xu, Jingang Wang, Wei Wu, Zhoujun Li
cs.CL
Named entity recognition (NER) is an important research problem in natural language processing. There are three types of NER tasks, including flat, nested and discontinuous entity recognition. Most previous sequential labeling models are task-specific, while recent years have witnessed the rising of generative models d...
http://arxiv.org/pdf/2303.10870v1
2023-03-20T05:11:22
2023-03-20T05:11:22
{ "title": [ 0.0064582825, 0.013137817, -0.05871582, -0.08129883, 0.016952515, -0.005420685, -0.041992188, 0.017471313, 0.024017334, 0.036743164, -0.020462036, 0.011314392, 0.039520264, -0.0029830933, -0.028808594, 0.018249512, -0.011810303, 0.02...
http://arxiv.org/abs/2303.10868v2
Retrieving Multimodal Information for Augmented Generation: A Survey
Ruochen Zhao, Hailin Chen, Weishi Wang, Fangkai Jiao, Xuan Long Do, Chengwei Qin, Bosheng Ding, Xiaobao Guo, Minzhi Li, Xingxuan Li, Shafiq Joty
cs.CL
As Large Language Models (LLMs) become popular, there emerged an important trend of using multimodality to augment the LLMs' generation ability, which enables LLMs to better interact with the world. However, there lacks a unified perception of at which stage and how to incorporate different modalities. In this survey, ...
http://arxiv.org/pdf/2303.10868v2
2023-10-08T03:26:35
2023-03-20T05:07:41
{ "title": [ -0.024917603, 0.007194519, -0.049072266, -0.056793213, 0.0051002502, 0.0043563843, -0.041259766, -0.007335663, -0.004497528, 0.048217773, 0.008644104, 0.0039100647, 0.0032749176, -0.02973938, -0.050445557, 0.0050582886, -0.017822266, ...
http://arxiv.org/abs/2303.10845v1
PanGu-Σ: Towards Trillion Parameter Language Model with Sparse Heterogeneous Computing
Xiaozhe Ren, Pingyi Zhou, Xinfan Meng, Xinjing Huang, Yadao Wang, Weichao Wang, Pengfei Li, Xiaoda Zhang, Alexander Podolskiy, Grigory Arshinov, Andrey Bout, Irina Piontkovskaya, Jiansheng Wei, Xin Jiang, Teng Su, Qun Liu, Jun Yao
cs.CL
The scaling of large language models has greatly improved natural language understanding, generation, and reasoning. In this work, we develop a system that trained a trillion-parameter language model on a cluster of Ascend 910 AI processors and MindSpore framework, and present the language model with 1.085T parameters ...
http://arxiv.org/pdf/2303.10845v1
2023-03-20T03:39:27
2023-03-20T03:39:27
{ "title": [ -0.040374756, -0.013618469, -0.056549072, -0.07867432, 0.0011367798, 0.011795044, -0.0690918, 0.028579712, -0.00019454956, 0.041870117, 0.0024852753, -0.0032081604, 0.008972168, -0.025756836, -0.028442383, 0.021347046, -0.009346008, ...
http://arxiv.org/abs/2304.02016v1
The Multimodal And Modular Ai Chef: Complex Recipe Generation From Imagery
David Noever, Samantha Elizabeth Miller Noever
cs.CL, cs.CV, cs.LG
The AI community has embraced multi-sensory or multi-modal approaches to advance this generation of AI models to resemble expected intelligent understanding. Combining language and imagery represents a familiar method for specific tasks like image captioning or generation from descriptions. This paper compares these mo...
http://arxiv.org/pdf/2304.02016v1
2023-03-20T01:57:52
2023-03-20T01:57:52
{ "title": [ -0.013496399, 0.0022506714, -0.032562256, -0.0362854, 0.0029144287, -0.011932373, -0.012908936, -0.017623901, -0.0010318756, 0.038116455, -0.0016174316, 0.01864624, 0.013900757, -0.02104187, -0.031677246, 0.014572144, -0.0009288788, ...
http://arxiv.org/abs/2303.10794v2
PheME: A deep ensemble framework for improving phenotype prediction from multi-modal data
Shenghan Zhang, Haoxuan Li, Ruixiang Tang, Sirui Ding, Laila Rasmy, Degui Zhi, Na Zou, Xia Hu
cs.LG, cs.CL, cs.MM, q-bio.QM
Detailed phenotype information is fundamental to accurate diagnosis and risk estimation of diseases. As a rich source of phenotype information, electronic health records (EHRs) promise to empower diagnostic variant interpretation. However, how to accurately and efficiently extract phenotypes from the heterogeneous EHR ...
http://arxiv.org/pdf/2303.10794v2
2023-04-26T20:40:43
2023-03-19T23:41:04
{ "title": [ -0.012886047, -0.005844116, -0.04550171, -0.06762695, 0.009727478, -0.005252838, -0.055908203, -0.0025539398, 0.017120361, 0.018981934, 0.00041127205, 0.0051193237, 0.011413574, -0.04660034, -0.029144287, 0.020645142, -0.0074653625, ...
http://arxiv.org/abs/2303.10782v1
On the Importance of Signer Overlap for Sign Language Detection
Abhilash Pal, Stephan Huber, Cyrine Chaabani, Alessandro Manzotti, Oscar Koller
cs.CL, cs.CV
Sign language detection, identifying if someone is signing or not, is becoming crucially important for its applications in remote conferencing software and for selecting useful sign data for training sign language recognition or translation tasks. We argue that the current benchmark data sets for sign language detectio...
http://arxiv.org/pdf/2303.10782v1
2023-03-19T22:15:05
2023-03-19T22:15:05
{ "title": [ -0.019088745, 0.02268982, -0.051239014, -0.03048706, 0.007881165, -0.015327454, -0.033843994, -0.03366089, 0.0013189316, 0.015144348, -0.05166626, 0.0031089783, -0.017120361, -0.017944336, -0.033233643, -0.026153564, -0.017807007, 0....
http://arxiv.org/abs/2303.10699v1
FVQA 2.0: Introducing Adversarial Samples into Fact-based Visual Question Answering
Weizhe Lin, Zhilin Wang, Bill Byrne
cs.CL, cs.CV
The widely used Fact-based Visual Question Answering (FVQA) dataset contains visually-grounded questions that require information retrieval using common sense knowledge graphs to answer. It has been observed that the original dataset is highly imbalanced and concentrated on a small portion of its associated knowledge g...
http://arxiv.org/pdf/2303.10699v1
2023-03-19T16:07:42
2023-03-19T16:07:42
{ "title": [ -0.0013589859, -0.0037460327, -0.013664246, -0.03527832, -0.0020961761, 0.006980896, -0.003019333, -0.008087158, 0.0055503845, 0.018188477, -0.014770508, 0.011741638, -0.021972656, -0.015136719, -0.048553467, -0.017837524, 0.017669678, ...
http://arxiv.org/abs/2303.10659v2
COVID-19 event extraction from Twitter via extractive question answering with continuous prompts
Yuhang Jiang, Ramakanth Kavuluru
cs.CL
As COVID-19 ravages the world, social media analytics could augment traditional surveys in assessing how the pandemic evolves and capturing consumer chatter that could help healthcare agencies in addressing it. This typically involves mining disclosure events that mention testing positive for the disease or discussions...
http://arxiv.org/pdf/2303.10659v2
2023-03-22T15:27:08
2023-03-19T13:47:56
{ "title": [ 0.016021729, 0.0047416687, -0.057525635, -0.10449219, 0.021331787, 0.0010843277, -0.03845215, 0.022155762, -0.0005249977, 0.026901245, -0.028320312, 0.022583008, 0.008834839, -0.013656616, -0.03656006, 0.0026340485, -0.0007696152, 0....
http://arxiv.org/abs/2303.10612v1
Bangla Grammatical Error Detection Using T5 Transformer Model
H. A. Z. Sameen Shahgir, Khondker Salman Sayeed
cs.CL, cs.LG
This paper presents a method for detecting grammatical errors in Bangla using a Text-to-Text Transfer Transformer (T5) Language Model, using the small variant of BanglaT5, fine-tuned on a corpus of 9385 sentences where errors were bracketed by the dedicated demarcation symbol. The T5 model was primarily designed for tr...
http://arxiv.org/pdf/2303.10612v1
2023-03-19T09:24:48
2023-03-19T09:24:48
{ "title": [ 0.009208679, 0.017562866, -0.01625061, -0.039978027, -0.006465912, -0.0028076172, -0.035461426, -0.01776123, -0.0012111664, 0.029785156, -0.011802673, -0.02355957, 0.028457642, -0.036499023, -0.03152466, -0.009925842, -0.02319336, 0....
http://arxiv.org/abs/2303.10606v1
CTRAN: CNN-Transformer-based Network for Natural Language Understanding
Mehrdad Rafiepour, Javad Salimi Sartakhti
cs.CL
Intent-detection and slot-filling are the two main tasks in natural language understanding. In this study, we propose CTRAN, a novel encoder-decoder CNN-Transformer-based architecture for intent-detection and slot-filling. In the encoder, we use BERT, followed by several convolutional layers, and rearrange the output u...
http://arxiv.org/pdf/2303.10606v1
2023-03-19T08:57:39
2023-03-19T08:57:39
{ "title": [ -0.033111572, -0.0038661957, -0.042907715, -0.058563232, 0.0068855286, 0.0034160614, -0.042999268, 0.0053253174, 0.004447937, 0.024642944, -0.016693115, 0.0003578663, 0.0069503784, -0.01222229, -0.043945312, -0.00008165836, -0.012710571,...
http://arxiv.org/abs/2303.10583v2
Toward Artificial Empathy for Human-Centered Design: A Framework
Qihao Zhu, Jianxi Luo
cs.HC, cs.CL
In the early stages of the design process, designers explore opportunities by discovering unmet needs and developing innovative concepts as potential solutions. From a human-centered design perspective, designers must develop empathy with people to truly understand their needs. However, developing empathy is a complex ...
http://arxiv.org/pdf/2303.10583v2
2023-05-13T05:47:16
2023-03-19T06:09:52
{ "title": [ -0.025222778, -0.015014648, -0.007030487, -0.008926392, -0.008613586, -0.016159058, 0.0009880066, -0.026504517, 0.003528595, 0.01600647, -0.030212402, 0.00042200089, -0.01058197, -0.023803711, -0.0390625, -0.0050315857, -0.0101623535, ...
http://arxiv.org/abs/2303.10573v1
Extracting Incidents, Effects, and Requested Advice from MeToo Posts
Vaibhav Garg, Jiaqing Yuan, Rujie Xi, Munindar P. Singh
cs.CL
Survivors of sexual harassment frequently share their experiences on social media, revealing their feelings and emotions and seeking advice. We observed that on Reddit, survivors regularly share long posts that describe a combination of (i) a sexual harassment incident, (ii) its effect on the survivor, including their ...
http://arxiv.org/pdf/2303.10573v1
2023-03-19T05:22:12
2023-03-19T05:22:12
{ "title": [ 0.047851562, -0.017028809, -0.024765015, -0.033721924, -0.005756378, -0.027832031, 0.040740967, 0.009307861, -0.03616333, 0.034240723, -0.020599365, 0.04852295, 0.018569946, -0.009941101, 0.002603531, 0.01184082, 0.014335632, 0.02517...
http://arxiv.org/abs/2303.12808v1
PACO: Provocation Involving Action, Culture, and Oppression
Vaibhav Garg, Ganning Xu, Munindar P. Singh
cs.CL, cs.AI
In India, people identify with a particular group based on certain attributes such as religion. The same religious groups are often provoked against each other. Previous studies show the role of provocation in increasing tensions between India's two prominent religious groups: Hindus and Muslims. With the advent of the...
http://arxiv.org/pdf/2303.12808v1
2023-03-19T04:39:36
2023-03-19T04:39:36
{ "title": [ 0.06359863, 0.026397705, -0.002084732, -0.049468994, -0.0006365776, -0.02017212, -0.01739502, -0.027389526, -0.04144287, 0.0050621033, 0.027633667, -0.0032081604, 0.005584717, -0.062683105, 0.051483154, -0.024047852, -0.0056991577, 0...
http://arxiv.org/abs/2303.10560v1
How People Respond to the COVID-19 Pandemic on Twitter: A Comparative Analysis of Emotional Expressions from US and India
Brandon Siyuan Loh, Raj Kumar Gupta, Ajay Vishwanath, Andrew Ortony, Yinping Yang
cs.CL
The COVID-19 pandemic has claimed millions of lives worldwide and elicited heightened emotions. This study examines the expression of various emotions pertaining to COVID-19 in the United States and India as manifested in over 54 million tweets, covering the fifteen-month period from February 2020 through April 2021, a...
http://arxiv.org/pdf/2303.10560v1
2023-03-19T04:05:10
2023-03-19T04:05:10
{ "title": [ 0.004371643, -0.004261017, -0.054138184, -0.048217773, 0.01361084, -0.024597168, -0.0041389465, 0.013298035, -0.020462036, 0.038726807, -0.027023315, 0.030685425, 0.010543823, -0.01247406, -0.020477295, -0.015426636, 0.014152527, 0.0...
http://arxiv.org/abs/2303.10527v1
Two Kinds of Recall
Yoav Goldberg
cs.CL
It is an established assumption that pattern-based models are good at precision, while learning based models are better at recall. But is that really the case? I argue that there are two kinds of recall: d-recall, reflecting diversity, and e-recall, reflecting exhaustiveness. I demonstrate through experiments that whil...
http://arxiv.org/pdf/2303.10527v1
2023-03-19T01:40:31
2023-03-19T01:40:31
{ "title": [ 0.06335449, -0.0060768127, 0.011528015, -0.066833496, -0.015045166, -0.026916504, -0.021255493, 0.011161804, -0.012176514, 0.05105591, -0.021942139, -0.009918213, -0.025436401, -0.009216309, -0.0045166016, -0.038360596, 0.03427124, -...
http://arxiv.org/abs/2303.10512v1
Adaptive Budget Allocation for Parameter-Efficient Fine-Tuning
Qingru Zhang, Minshuo Chen, Alexander Bukharin, Pengcheng He, Yu Cheng, Weizhu Chen, Tuo Zhao
cs.CL, cs.LG
Fine-tuning large pre-trained language models on downstream tasks has become an important paradigm in NLP. However, common practice fine-tunes all of the parameters in a pre-trained model, which becomes prohibitive when a large number of downstream tasks are present. Therefore, many fine-tuning methods are proposed to ...
http://arxiv.org/pdf/2303.10512v1
2023-03-18T22:36:25
2023-03-18T22:36:25
{ "title": [ -0.039642334, -0.013938904, -0.029708862, -0.06866455, 0.005744934, 0.015792847, -0.029037476, 0.013175964, -0.028549194, 0.032989502, -0.0051116943, 0.02420044, -0.030685425, -0.01701355, -0.016784668, 0.008415222, 0.00217247, 0.010...
http://arxiv.org/abs/2303.10510v2
A Deep Learning System for Domain-specific Speech Recognition
Yanan Jia
cs.CL, cs.LG, cs.SD, eess.AS
As human-machine voice interfaces provide easy access to increasingly intelligent machines, many state-of-the-art automatic speech recognition (ASR) systems are proposed. However, commercial ASR systems usually have poor performance on domain-specific speech especially under low-resource settings. The author works with...
http://arxiv.org/pdf/2303.10510v2
2023-09-27T17:32:30
2023-03-18T22:19:09
{ "title": [ -0.0113220215, 0.0027236938, -0.059020996, -0.08428955, 0.02003479, -0.011238098, -0.05380249, 0.025253296, -0.0049552917, 0.038635254, -0.006515503, -0.000048696995, 0.004234314, -0.018875122, -0.044158936, 0.0071907043, -0.01537323, ...
http://arxiv.org/abs/2303.10475v5
Is Prompt All You Need? No. A Comprehensive and Broader View of Instruction Learning
Renze Lou, Kai Zhang, Wenpeng Yin
cs.CL
Task semantics can be expressed by a set of input-to-output examples or a piece of textual instruction. Conventional machine learning approaches for natural language processing (NLP) mainly rely on the availability of large-scale sets of task-specific examples. Two issues arise: first, collecting task-specific labeled ...
http://arxiv.org/pdf/2303.10475v5
2023-05-03T00:45:11
2023-03-18T19:17:47
{ "title": [ -0.004459381, -0.009590149, -0.02609253, -0.03173828, -0.013954163, -0.014556885, -0.011756897, 0.019729614, -0.019042969, 0.011070251, -0.07366943, 0.0060272217, -0.0058250427, -0.02330017, 0.029174805, -0.0418396, 0.013114929, -0.0...
http://arxiv.org/abs/2303.10464v2
SPDF: Sparse Pre-training and Dense Fine-tuning for Large Language Models
Vithursan Thangarasa, Abhay Gupta, William Marshall, Tianda Li, Kevin Leong, Dennis DeCoste, Sean Lie, Shreyas Saxena
cs.LG, cs.CL
The pre-training and fine-tuning paradigm has contributed to a number of breakthroughs in Natural Language Processing (NLP). Instead of directly training on a downstream task, language models are first pre-trained on large datasets with cross-domain knowledge (e.g., Pile, MassiveText, etc.) and then fine-tuned on task-...
http://arxiv.org/pdf/2303.10464v2
2023-07-29T19:56:50
2023-03-18T17:56:01
{ "title": [ -0.017990112, -0.028518677, -0.03253174, -0.091430664, -0.0020198822, 0.017501831, -0.06896973, 0.023269653, 0.008773804, 0.026245117, -0.014678955, 0.014839172, 0.012062073, -0.013832092, -0.03765869, 0.016296387, 0.0034599304, 0.02...
http://arxiv.org/abs/2303.13365v1
Requirement Formalisation using Natural Language Processing and Machine Learning: A Systematic Review
Shekoufeh Kolahdouz-Rahimi, Kevin Lano, Chenghua Lin
cs.CL, cs.AI, cs.LG, cs.SE
Improvement of software development methodologies attracts developers to automatic Requirement Formalisation (RF) in the Requirement Engineering (RE) field. The potential advantages by applying Natural Language Processing (NLP) and Machine Learning (ML) in reducing the ambiguity and incompleteness of requirement writte...
http://arxiv.org/pdf/2303.13365v1
2023-03-18T17:36:21
2023-03-18T17:36:21
{ "title": [ -0.0067443848, -0.025390625, -0.031280518, -0.09423828, 0.0065727234, 0.023742676, -0.07055664, 0.0038776398, -0.03149414, 0.0149002075, -0.0050849915, -0.009361267, -0.012489319, -0.027908325, -0.032409668, -0.023117065, -0.032104492, ...
http://arxiv.org/abs/2303.10443v1
GazeReader: Detecting Unknown Word Using Webcam for English as a Second Language (ESL) Learners
Jiexin Ding, Bowen Zhao, Yuqi Huang, Yuntao Wang, Yuanchun Shi
cs.HC, cs.CL
Automatic unknown word detection techniques can enable new applications for assisting English as a Second Language (ESL) learners, thus improving their reading experiences. However, most modern unknown word detection methods require dedicated eye-tracking devices with high precision that are not easily accessible to en...
http://arxiv.org/pdf/2303.10443v1
2023-03-18T15:55:49
2023-03-18T15:55:49
{ "title": [ -0.017700195, 0.0056762695, -0.02947998, 0.044769287, -0.033996582, -0.032196045, 0.015205383, -0.043029785, 0.004634857, 0.024047852, -0.03652954, -0.0010328293, 0.02482605, -0.030014038, -0.012008667, -0.026382446, 0.0045394897, 0....
http://arxiv.org/abs/2303.10439v2
Stop Words for Processing Software Engineering Documents: Do they Matter?
Yaohou Fan, Chetan Arora, Christoph Treude
cs.SE, cs.CL
Stop words, which are considered non-predictive, are often eliminated in natural language processing tasks. However, the definition of uninformative vocabulary is vague, so most algorithms use general knowledge-based stop lists to remove stop words. There is an ongoing debate among academics about the usefulness of sto...
http://arxiv.org/pdf/2303.10439v2
2023-06-12T12:36:28
2023-03-18T15:39:23
{ "title": [ -0.05810547, -0.017501831, -0.035491943, -0.0013456345, -0.030532837, -0.012374878, -0.014175415, 0.0030727386, 0.011672974, 0.03201294, -0.023269653, -0.011222839, 0.02758789, -0.01991272, -0.038513184, 0.016647339, -0.0069732666, 0...
http://arxiv.org/abs/2303.10430v1
NoisyHate: Benchmarking Content Moderation Machine Learning Models with Human-Written Perturbations Online
Yiran Ye, Thai Le, Dongwon Lee
cs.LG, cs.AI, cs.CL, cs.CY
Online texts with toxic content are a threat in social media that might cause cyber harassment. Although many platforms applied measures, such as machine learning-based hate-speech detection systems, to diminish their effect, those toxic content publishers can still evade the system by modifying the spelling of toxic w...
http://arxiv.org/pdf/2303.10430v1
2023-03-18T14:54:57
2023-03-18T14:54:57
{ "title": [ 0.024780273, 0.012756348, 0.011726379, -0.0063972473, -0.0074882507, -0.031082153, 0.014335632, -0.044036865, -0.011680603, 0.039764404, -0.03366089, 0.028030396, 0.040039062, -0.025344849, -0.008148193, -0.012168884, -0.011207581, 0...
http://arxiv.org/abs/2303.10420v1
A Comprehensive Capability Analysis of GPT-3 and GPT-3.5 Series Models
Junjie Ye, Xuanting Chen, Nuo Xu, Can Zu, Zekai Shao, Shichun Liu, Yuhan Cui, Zeyang Zhou, Chao Gong, Yang Shen, Jie Zhou, Siming Chen, Tao Gui, Qi Zhang, Xuanjing Huang
cs.CL, 68-06, I.2
GPT series models, such as GPT-3, CodeX, InstructGPT, ChatGPT, and so on, have gained considerable attention due to their exceptional natural language processing capabilities. However, despite the abundance of research on the difference in capabilities between GPT series models and fine-tuned models, there has been lim...
http://arxiv.org/pdf/2303.10420v1
2023-03-18T14:02:04
2023-03-18T14:02:04
{ "title": [ 0.03692627, 0.029571533, -0.024902344, -0.04425049, 0.007167816, -0.02357483, -0.05609131, -0.0496521, -0.009056091, 0.03302002, -0.010955811, -0.010269165, 0.018463135, -0.0340271, 0.0022068024, 0.015235901, 0.009292603, 0.029953003...
http://arxiv.org/abs/2303.10395v1
A Graph-Guided Reasoning Approach for Open-ended Commonsense Question Answering
Zhen Han, Yue Feng, Mingming Sun
cs.CL, cs.LG
Recently, end-to-end trained models for multiple-choice commonsense question answering (QA) have delivered promising results. However, such question-answering systems cannot be directly applied in real-world scenarios where answer candidates are not provided. Hence, a new benchmark challenge set for open-ended commonse...
http://arxiv.org/pdf/2303.10395v1
2023-03-18T11:15:33
2023-03-18T11:15:33
{ "title": [ -0.00075244904, -0.018188477, -0.016647339, -0.07556152, -0.01436615, 0.015052795, -0.046783447, 0.016479492, 0.015083313, 0.015045166, -0.009506226, -0.002298355, -0.022247314, -0.026321411, -0.032684326, -0.009147644, -0.0036792755, ...
http://arxiv.org/abs/2303.10384v1
Powerful and Extensible WFST Framework for RNN-Transducer Losses
Aleksandr Laptev, Vladimir Bataev, Igor Gitman, Boris Ginsburg
eess.AS, cs.AI, cs.CL, cs.LG, cs.SD
This paper presents a framework based on Weighted Finite-State Transducers (WFST) to simplify the development of modifications for RNN-Transducer (RNN-T) loss. Existing implementations of RNN-T use CUDA-related code, which is hard to extend and debug. WFSTs are easy to construct and extend, and allow debugging through ...
http://arxiv.org/pdf/2303.10384v1
2023-03-18T10:36:33
2023-03-18T10:36:33
{ "title": [ -0.02217102, -0.015716553, -0.031555176, -0.046417236, -0.010322571, -0.0035095215, -0.039398193, -0.013877869, -0.004760742, 0.04031372, 0.03289795, -0.017089844, 0.021331787, -0.049102783, -0.02192688, 0.027420044, 0.008796692, 0.0...
http://arxiv.org/abs/2303.10368v1
An Empirical Study of Pre-trained Language Models in Simple Knowledge Graph Question Answering
Nan Hu, Yike Wu, Guilin Qi, Dehai Min, Jiaoyan Chen, Jeff Z. Pan, Zafar Ali
cs.CL
Large-scale pre-trained language models (PLMs) such as BERT have recently achieved great success and become a milestone in natural language processing (NLP). It is now the consensus of the NLP community to adopt PLMs as the backbone for downstream tasks. In recent works on knowledge graph question answering (KGQA), BER...
http://arxiv.org/pdf/2303.10368v1
2023-03-18T08:57:09
2023-03-18T08:57:09
{ "title": [ -0.02407837, -0.018661499, -0.0362854, -0.064575195, -0.02645874, 0.0020160675, -0.06665039, 0.013809204, -0.007637024, 0.036132812, -0.007659912, 0.003944397, 0.0017194748, -0.007980347, -0.051513672, -0.0011034012, 0.004070282, 0.0...
http://arxiv.org/abs/2303.10330v3
Exploring Partial Knowledge Base Inference in Biomedical Entity Linking
Hongyi Yuan, Keming Lu, Zheng Yuan
cs.CL
Biomedical entity linking (EL) consists of named entity recognition (NER) and named entity disambiguation (NED). EL models are trained on corpora labeled by a predefined KB. However, it is a common scenario that only entities within a subset of the KB are precious to stakeholders. We name this scenario partial knowledg...
http://arxiv.org/pdf/2303.10330v3
2023-06-03T07:57:51
2023-03-18T04:31:07
{ "title": [ -0.014968872, -0.01360321, -0.043304443, -0.10406494, 0.009773254, 0.01020813, -0.052459717, 0.026367188, -0.003025055, 0.011932373, 0.0036144257, -0.012199402, 0.008728027, -0.040802002, -0.015945435, 0.0043945312, 0.007843018, 0.01...
http://arxiv.org/abs/2303.10328v1
Revisiting Automatic Question Summarization Evaluation in the Biomedical Domain
Hongyi Yuan, Yaoyun Zhang, Fei Huang, Songfang Huang
cs.CL
Automatic evaluation metrics have been facilitating the rapid development of automatic summarization methods by providing instant and fair assessments of the quality of summaries. Most metrics have been developed for the general domain, especially news and meeting notes, or other language-generation tasks. However, the...
http://arxiv.org/pdf/2303.10328v1
2023-03-18T04:28:01
2023-03-18T04:28:01
{ "title": [ -0.0061531067, 0.0076026917, -0.011787415, -0.057128906, -0.0042495728, 0.012077332, -0.032196045, -0.022323608, 0.01663208, 0.03213501, -0.0013380051, 0.004508972, -0.009986877, -0.013374329, -0.07342529, 0.014663696, -0.002565384, ...
http://arxiv.org/abs/2303.10311v1
On the rise of fear speech in online social media
Punyajoy Saha, Kiran Garimella, Narla Komal Kalyan, Saurabh Kumar Pandey, Pauras Mangesh Meher, Binny Mathew, Animesh Mukherjee
cs.SI, cs.CL, cs.CY
Recently, social media platforms are heavily moderated to prevent the spread of online hate speech, which is usually fertile in toxic words and is directed toward an individual or a community. Owing to such heavy moderation, newer and more subtle techniques are being deployed. One of the most striking among these is fe...
http://arxiv.org/pdf/2303.10311v1
2023-03-18T02:46:49
2023-03-18T02:46:49
{ "title": [ 0.08703613, -0.0011787415, -0.003572464, -0.024154663, -0.0044136047, -0.037322998, 0.040649414, 0.00060272217, -0.05279541, 0.042419434, -0.0044555664, 0.021438599, 0.019699097, -0.040039062, 0.023040771, -0.014450073, 0.000936985, ...
http://arxiv.org/abs/2303.10255v2
Feedback Effect in User Interaction with Intelligent Assistants: Delayed Engagement, Adaption and Drop-out
Zidi Xiu, Kai-Chen Cheng, David Q. Sun, Jiannan Lu, Hadas Kotek, Yuhan Zhang, Paul McCarthy, Christopher Klein, Stephen Pulman, Jason D. Williams
cs.HC, cs.CL
With the growing popularity of intelligent assistants (IAs), evaluating IA quality becomes an increasingly active field of research. This paper identifies and quantifies the feedback effect, a novel component in IA-user interactions: how the capabilities and limitations of the IA influence user behavior over time. Firs...
http://arxiv.org/pdf/2303.10255v2
2023-04-18T15:26:44
2023-03-17T21:39:33
{ "title": [ -0.021850586, -0.0050354004, -0.010673523, 0.0024700165, 0.00248909, -0.013130188, 0.008926392, -0.043395996, 0.003019333, 0.046051025, -0.037750244, 0.0084991455, 0.011039734, -0.0029087067, -0.06329346, -0.0070114136, -0.00856781, ...
http://arxiv.org/abs/2303.10227v1
Conversational Tree Search: A New Hybrid Dialog Task
Dirk Väth, Lindsey Vanderlyn, Ngoc Thang Vu
cs.CL
Conversational interfaces provide a flexible and easy way for users to seek information that may otherwise be difficult or inconvenient to obtain. However, existing interfaces generally fall into one of two categories: FAQs, where users must have a concrete question in order to retrieve a general answer, or dialogs, wh...
http://arxiv.org/pdf/2303.10227v1
2023-03-17T19:50:51
2023-03-17T19:50:51
{ "title": [ -0.017944336, -0.01826477, -0.032836914, -0.037872314, -0.01852417, -0.008674622, -0.043884277, -0.00025606155, 0.022781372, 0.028945923, -0.015106201, 0.0010757446, -0.019836426, -0.013664246, -0.060699463, -0.005554199, -0.023895264, ...
http://arxiv.org/abs/2303.13379v2
Practical and Ethical Challenges of Large Language Models in Education: A Systematic Scoping Review
Lixiang Yan, Lele Sha, Linxuan Zhao, Yuheng Li, Roberto Martinez-Maldonado, Guanliang Chen, Xinyu Li, Yueqiao Jin, Dragan Gašević
cs.CL, cs.AI, cs.CY
Educational technology innovations leveraging large language models (LLMs) have shown the potential to automate the laborious process of generating and analysing textual content. While various innovations have been developed to automate a range of educational tasks (e.g., question generation, feedback provision, and es...
http://arxiv.org/pdf/2303.13379v2
2023-07-22T15:26:28
2023-03-17T18:14:46
{ "title": [ -0.041778564, -0.014602661, -0.04269409, -0.03286743, -0.0105896, -0.0079956055, -0.026565552, 0.00039315224, -0.039215088, 0.0395813, 0.0012626648, 0.003458023, -0.009712219, -0.030319214, -0.01953125, -0.006515503, -0.028656006, 0....
http://arxiv.org/abs/2303.11156v2
Can AI-Generated Text be Reliably Detected?
Vinu Sankar Sadasivan, Aounon Kumar, Sriram Balasubramanian, Wenxiao Wang, Soheil Feizi
cs.CL, cs.AI, cs.LG
In this paper, both empirically and theoretically, we show that several AI-text detectors are not reliable in practical scenarios. Empirically, we show that paraphrasing attacks, where a light paraphraser is applied on top of a large language model (LLM), can break a whole range of detectors, including ones using water...
http://arxiv.org/pdf/2303.11156v2
2023-06-28T20:29:16
2023-03-17T17:53:19
{ "title": [ -0.015014648, 0.0052261353, 0.002614975, 0.026184082, -0.038879395, 0.0030231476, -0.0033493042, -0.06311035, -0.012245178, 0.034332275, -0.020339966, -0.0077438354, 0.020858765, -0.036987305, 0.013015747, -0.022232056, -0.008728027, ...
http://arxiv.org/abs/2303.10128v2
Direct and indirect evidence of compression of word lengths. Zipf's law of abbreviation revisited
Sonia Petrini, Antoni Casas-i-Muñoz, Jordi Cluet-i-Martinell, Mengxue Wang, Chris Bentz, Ramon Ferrer-i-Cancho
cs.CL, physics.soc-ph
Zipf's law of abbreviation, the tendency of more frequent words to be shorter, is one of the most solid candidates for a linguistic universal, in the sense that it has the potential for being exceptionless or with a number of exceptions that is vanishingly small compared to the number of languages on Earth. Since Zipf'...
http://arxiv.org/pdf/2303.10128v2
2023-05-27T08:36:57
2023-03-17T17:12:18
{ "title": [ -0.06225586, -0.01789856, -0.057769775, -0.015960693, -0.04360962, -0.018432617, -0.055114746, -0.0025577545, -0.035736084, 0.029144287, 0.001909256, 0.024765015, -0.0015258789, -0.04547119, -0.0030593872, 0.0045547485, -0.006072998, ...
http://arxiv.org/abs/2303.10093v2
Investigating the Role of Attribute Context in Vision-Language Models for Object Recognition and Detection
Kyle Buettner, Adriana Kovashka
cs.CV, cs.AI, cs.CL, cs.LG
Vision-language alignment learned from image-caption pairs has been shown to benefit tasks like object recognition and detection. Methods are mostly evaluated in terms of how well object class names are learned, but captions also contain rich attribute context that should be considered when learning object alignment. I...
http://arxiv.org/pdf/2303.10093v2
2023-11-06T20:58:11
2023-03-17T16:14:37
{ "title": [ -0.014907837, 0.0064582825, -0.051239014, -0.050109863, 0.008598328, -0.024887085, -0.029693604, -0.007083893, 0.00818634, 0.039154053, -0.034240723, -0.005050659, 0.011421204, -0.0057640076, -0.051696777, -0.0025196075, 0.009132385, ...
http://arxiv.org/abs/2303.09905v1
More Robust Schema-Guided Dialogue State Tracking via Tree-Based Paraphrase Ranking
A. Coca, B. H. Tseng, W. Lin, B. Byrne
cs.CL
The schema-guided paradigm overcomes scalability issues inherent in building task-oriented dialogue (TOD) agents with static ontologies. Instead of operating on dialogue context alone, agents have access to hierarchical schemas containing task-relevant natural language descriptions. Fine-tuned language models excel at ...
http://arxiv.org/pdf/2303.09905v1
2023-03-17T11:43:08
2023-03-17T11:43:08
{ "title": [ 0.015029907, -0.026000977, -0.0085372925, -0.0791626, -0.030517578, 0.0046653748, -0.05392456, 0.002380371, 0.0047569275, 0.019119263, -0.01134491, 0.008407593, -0.001791954, -0.02949524, -0.026489258, -0.008354187, -0.021865845, 0.0...
http://arxiv.org/abs/2303.09901v3
mCPT at SemEval-2023 Task 3: Multilingual Label-Aware Contrastive Pre-Training of Transformers for Few- and Zero-shot Framing Detection
Markus Reiter-Haas, Alexander Ertl, Kevin Innerebner, Elisabeth Lex
cs.CL, cs.AI, cs.LG, I.2.7; I.7; J.7
This paper presents the winning system for the zero-shot Spanish framing detection task, which also achieves competitive places in eight additional languages. The challenge of the framing detection task lies in identifying a set of 14 frames when only a few or zero samples are available, i.e., a multilingual multi-labe...
http://arxiv.org/pdf/2303.09901v3
2023-08-01T15:16:52
2023-03-17T11:33:06
{ "title": [ 0.053466797, -0.007572174, 0.004962921, -0.05709839, -0.008178711, -0.011886597, -0.004486084, -0.016174316, 0.000033974648, 0.032440186, -0.017196655, 0.014625549, 0.009353638, -0.021499634, -0.015792847, -0.01260376, -0.007434845, ...
http://arxiv.org/abs/2303.09892v3
Memotion 3: Dataset on Sentiment and Emotion Analysis of Codemixed Hindi-English Memes
Shreyash Mishra, S Suryavardan, Parth Patwa, Megha Chakraborty, Anku Rani, Aishwarya Reganti, Aman Chadha, Amitava Das, Amit Sheth, Manoj Chinnakotla, Asif Ekbal, Srijan Kumar
cs.CL
Memes are the new-age conveyance mechanism for humor on social media sites. Memes often include an image and some text. Memes can be used to promote disinformation or hatred, thus it is crucial to investigate in details. We introduce Memotion 3, a new dataset with 10,000 annotated memes. Unlike other prevalent datasets...
http://arxiv.org/pdf/2303.09892v3
2023-10-02T14:28:03
2023-03-17T11:13:30
{ "title": [ 0.008110046, 0.0011119843, -0.021591187, -0.013999939, -0.0154953, -0.0116119385, -0.030563354, -0.010848999, 0.03704834, 0.049072266, -0.017959595, 0.0030879974, 0.02279663, -0.024749756, -0.036895752, 0.007221222, 0.00459671, 0.016...
http://arxiv.org/abs/2303.09859v3
Trained on 100 million words and still in shape: BERT meets British National Corpus
David Samuel, Andrey Kutuzov, Lilja Øvrelid, Erik Velldal
cs.CL
While modern masked language models (LMs) are trained on ever larger corpora, we here explore the effects of down-scaling training to a modestly-sized but representative, well-balanced, and publicly available English text source -- the British National Corpus. We show that pre-training on this carefully curated corpus ...
http://arxiv.org/pdf/2303.09859v3
2023-05-05T23:59:35
2023-03-17T09:53:33
{ "title": [ 0.023742676, -0.02947998, -0.042938232, -0.049743652, -0.026931763, -0.02166748, -0.034820557, 0.014198303, -0.019042969, 0.020462036, -0.03756714, 0.024368286, 0.014190674, -0.030807495, 0.006793976, -0.02456665, -0.015342712, 0.017...
http://arxiv.org/abs/2303.09827v1
DORIC : Domain Robust Fine-Tuning for Open Intent Clustering through Dependency Parsing
Jihyun Lee, Seungyeon Seo, Yunsu Kim, Gary Geunbae Lee
cs.CL, cs.AI
We present our work on Track 2 in the Dialog System Technology Challenges 11 (DSTC11). DSTC11-Track2 aims to provide a benchmark for zero-shot, cross-domain, intent-set induction. In the absence of in-domain training dataset, robust utterance representation that can be used across domains is necessary to induce users' ...
http://arxiv.org/pdf/2303.09827v1
2023-03-17T08:12:36
2023-03-17T08:12:36
{ "title": [ 0.0011930466, 0.015136719, -0.060150146, -0.092041016, 0.033111572, -0.006465912, -0.0592041, -0.017486572, -0.041900635, 0.031173706, -0.0152282715, 0.042510986, -0.019180298, -0.011314392, -0.018722534, -0.010475159, -0.01625061, 0...
http://arxiv.org/abs/2303.09823v1
Transformers and Ensemble methods: A solution for Hate Speech Detection in Arabic languages
Angel Felipe Magnossão de Paula, Imene Bensalem, Paolo Rosso, Wajdi Zaghouani
cs.CL, cs.AI, cs.LG
This paper describes our participation in the shared task of hate speech detection, which is one of the subtasks of the CERIST NLP Challenge 2022. Our experiments evaluate the performance of six transformer models and their combination using 2 ensemble approaches. The best results on the training set, in a five-fold cr...
http://arxiv.org/pdf/2303.09823v1
2023-03-17T08:02:54
2023-03-17T08:02:54
{ "title": [ 0.010116577, 0.0033569336, -0.02218628, -0.043884277, 0.02029419, -0.01576233, 0.0036373138, -0.006401062, -0.009262085, 0.022827148, -0.026901245, -0.0049476624, 0.0073623657, -0.026260376, -0.051879883, -0.015510559, -0.0105896, 0....
http://arxiv.org/abs/2303.09752v3
CoLT5: Faster Long-Range Transformers with Conditional Computation
Joshua Ainslie, Tao Lei, Michiel de Jong, Santiago Ontañón, Siddhartha Brahma, Yury Zemlyanskiy, David Uthus, Mandy Guo, James Lee-Thorp, Yi Tay, Yun-Hsuan Sung, Sumit Sanghai
cs.CL, cs.LG
Many natural language processing tasks benefit from long inputs, but processing long documents with Transformers is expensive -- not only due to quadratic attention complexity but also from applying feedforward and projection layers to every token. However, not all tokens are equally important, especially for longer do...
http://arxiv.org/pdf/2303.09752v3
2023-10-24T00:51:49
2023-03-17T03:28:17
{ "title": [ -0.020111084, 0.02029419, -0.041900635, -0.09649658, 0.03475952, 0.019638062, -0.04815674, -0.017089844, -0.0007367134, 0.030151367, -0.03475952, 0.010223389, 0.023880005, -0.018753052, -0.020019531, 0.003835678, -0.0059509277, 0.032...
http://arxiv.org/abs/2303.09719v1
Learning towards Selective Data Augmentation for Dialogue Generation
Xiuying Chen, Mingzhe Li, Jiayi Zhang, Xiaoqiang Xia, Chen Wei, Jianwei Cui, Xin Gao, Xiangliang Zhang, Rui Yan
cs.CL
As it is cumbersome and expensive to acquire a huge amount of data for training neural dialog models, data augmentation is proposed to effectively utilize existing training samples. However, current data augmentation techniques on the dialog generation task mostly augment all cases in the training dataset without consi...
http://arxiv.org/pdf/2303.09719v1
2023-03-17T01:26:39
2023-03-17T01:26:39
{ "title": [ -0.0011224747, -0.010353088, -0.044403076, -0.07824707, 0.00856781, -0.001332283, -0.040924072, 0.020950317, 0.010749817, 0.0074920654, -0.024353027, 0.0059547424, -0.025772095, -0.012901306, -0.045898438, -0.009727478, -0.024108887, ...
http://arxiv.org/abs/2303.09713v2
CHAMPAGNE: Learning Real-world Conversation from Large-Scale Web Videos
Seungju Han, Jack Hessel, Nouha Dziri, Yejin Choi, Youngjae Yu
cs.CL, cs.AI
Visual information is central to conversation: body gestures and physical behaviour, for example, contribute to meaning that transcends words alone. To date, however, most neural conversational models are limited to just text. We introduce CHAMPAGNE, a generative model of conversations that can account for visual conte...
http://arxiv.org/pdf/2303.09713v2
2023-08-16T08:17:02
2023-03-17T01:10:33
{ "title": [ -0.009246826, 0.005088806, -0.056610107, -0.05911255, 0.01689148, -0.021133423, -0.032714844, 0.00919342, 0.0048828125, 0.008811951, -0.02432251, 0.00062799454, -0.015136719, -0.017044067, -0.037017822, -0.021484375, -0.008384705, 0....
http://arxiv.org/abs/2303.09639v2
Neural Architecture Search for Effective Teacher-Student Knowledge Transfer in Language Models
Aashka Trivedi, Takuma Udagawa, Michele Merler, Rameswar Panda, Yousef El-Kurdi, Bishwaranjan Bhattacharjee
cs.CL
Large pretrained language models have achieved state-of-the-art results on a variety of downstream tasks. Knowledge Distillation (KD) into a smaller student model addresses their inefficiency, allowing for deployment in resource-constrained environments. However, KD can be ineffective when the student is manually selec...
http://arxiv.org/pdf/2303.09639v2
2023-10-13T21:34:39
2023-03-16T20:39:44
{ "title": [ -0.023864746, -0.00198555, -0.009902954, -0.0031833649, -0.015838623, -0.016799927, 0.0007009506, -0.020889282, -0.021331787, 0.04534912, -0.005836487, 0.005065918, 0.0048217773, -0.0068740845, -0.054473877, -0.010505676, -0.014091492, ...
http://arxiv.org/abs/2303.13355v1
Revealing Weaknesses of Vietnamese Language Models Through Unanswerable Questions in Machine Reading Comprehension
Son Quoc Tran, Phong Nguyen-Thuan Do, Kiet Van Nguyen, Ngan Luu-Thuy Nguyen
cs.CL, cs.AI
Although the curse of multilinguality significantly restricts the language abilities of multilingual models in monolingual settings, researchers now still have to rely on multilingual models to develop state-of-the-art systems in Vietnamese Machine Reading Comprehension. This difficulty in researching is because of the...
http://arxiv.org/pdf/2303.13355v1
2023-03-16T20:32:58
2023-03-16T20:32:58
{ "title": [ -0.045928955, -0.018234253, 0.0019989014, -0.004737854, -0.0491333, 0.009773254, -0.038970947, -0.036956787, 0.009880066, 0.035064697, -0.05154419, -0.0093307495, 0.00843811, -0.0074691772, -0.064819336, -0.018569946, -0.016708374, 0...
http://arxiv.org/abs/2303.09618v1
HIVE: Harnessing Human Feedback for Instructional Visual Editing
Shu Zhang, Xinyi Yang, Yihao Feng, Can Qin, Chia-Chih Chen, Ning Yu, Zeyuan Chen, Huan Wang, Silvio Savarese, Stefano Ermon, Caiming Xiong, Ran Xu
cs.CV, cs.AI, cs.CL, cs.HC, cs.LG
Incorporating human feedback has been shown to be crucial to align text generated by large language models to human preferences. We hypothesize that state-of-the-art instructional image editing models, where outputs are generated based on an input image and an editing instruction, could similarly benefit from human fee...
http://arxiv.org/pdf/2303.09618v1
2023-03-16T19:47:41
2023-03-16T19:47:41
{ "title": [ -0.032714844, -0.0062408447, -0.016937256, -0.00843811, -0.020065308, -0.0064849854, -0.016860962, -0.039001465, 0.0071792603, 0.026046753, -0.03768921, -0.004749298, -0.0019226074, -0.0047836304, -0.05545044, -0.020812988, -0.0070991516...
http://arxiv.org/abs/2303.09601v1
Psychotherapy AI Companion with Reinforcement Learning Recommendations and Interpretable Policy Dynamics
Baihan Lin, Guillermo Cecchi, Djallel Bouneffouf
cs.LG, cs.AI, cs.CL, cs.HC, q-bio.NC
We introduce a Reinforcement Learning Psychotherapy AI Companion that generates topic recommendations for therapists based on patient responses. The system uses Deep Reinforcement Learning (DRL) to generate multi-objective policies for four different psychiatric conditions: anxiety, depression, schizophrenia, and suici...
http://arxiv.org/pdf/2303.09601v1
2023-03-16T19:01:29
2023-03-16T19:01:29
{ "title": [ -0.0043640137, 0.0007839203, -0.011558533, -0.016159058, -0.024124146, -0.018951416, -0.007888794, -0.01084137, -0.039001465, 0.0118255615, -0.02619934, 0.01828003, 0.010726929, -0.050964355, 0.05053711, -0.030914307, 0.01979065, 0.0...
http://arxiv.org/abs/2303.09522v3
P+: Extended Textual Conditioning in Text-to-Image Generation
Andrey Voynov, Qinghao Chu, Daniel Cohen-Or, Kfir Aberman
cs.CV, cs.CL, cs.GR, cs.LG
We introduce an Extended Textual Conditioning space in text-to-image models, referred to as $P+$. This space consists of multiple textual conditions, derived from per-layer prompts, each corresponding to a layer of the denoising U-net of the diffusion model. We show that the extended space provides greater disentangl...
http://arxiv.org/pdf/2303.09522v3
2023-07-15T15:48:48
2023-03-16T17:38:15
{ "title": [ 0.001704216, 0.0005168915, -0.010925293, -0.047729492, 0.005622864, 0.0033130646, -0.011154175, -0.03225708, 0.02507019, 0.04168701, -0.023834229, 0.027877808, 0.0044136047, -0.009880066, -0.036865234, 0.008926392, -0.0067329407, 0.0...
http://arxiv.org/abs/2303.12734v1
MultiModal Bias: Introducing a Framework for Stereotypical Bias Assessment beyond Gender and Race in Vision Language Models
Sepehr Janghorbani, Gerard de Melo
cs.CV, cs.CL, cs.LG
Recent breakthroughs in self supervised training have led to a new class of pretrained vision language models. While there have been investigations of bias in multimodal models, they have mostly focused on gender and racial bias, giving much less attention to other relevant groups, such as minorities with regard to rel...
http://arxiv.org/pdf/2303.12734v1
2023-03-16T17:36:37
2023-03-16T17:36:37
{ "title": [ -0.020523071, -0.004486084, -0.05456543, -0.03677368, -0.007423401, -0.029052734, -0.048187256, -0.015304565, 0.030654907, 0.022720337, -0.026672363, -0.019241333, 0.020568848, -0.022705078, -0.0491333, -0.014778137, -0.02027893, 0.0...
http://arxiv.org/abs/2303.09438v1
Trustera: A Live Conversation Redaction System
Evandro Gouvêa, Ali Dadgar, Shahab Jalalvand, Rathi Chengalvarayan, Badrinath Jayakumar, Ryan Price, Nicholas Ruiz, Jennifer McGovern, Srinivas Bangalore, Ben Stern
cs.CL, cs.SD, eess.AS
Trustera, the first functional system that redacts personally identifiable information (PII) in real-time spoken conversations to remove agents' need to hear sensitive information while preserving the naturalness of live customer-agent conversations. As opposed to post-call redaction, audio masking starts as soon as th...
http://arxiv.org/pdf/2303.09438v1
2023-03-16T16:13:36
2023-03-16T16:13:36
{ "title": [ -0.048095703, -0.005264282, -0.08068848, -0.0847168, 0.00068330765, 0.0006761551, -0.070007324, 0.021591187, -0.004547119, 0.04736328, 0.011787415, -0.01474762, 0.02909851, -0.03756714, -0.041381836, 0.01663208, -0.007785797, 0.02236...
http://arxiv.org/abs/2303.09435v1
Jump to Conclusions: Short-Cutting Transformers With Linear Transformations
Alexander Yom Din, Taelin Karidi, Leshem Choshen, Mor Geva
cs.CL
Transformer-based language models (LMs) create hidden representations of their inputs at every layer, but only use final-layer representations for prediction. This obscures the internal decision-making process of the model and the utility of its intermediate representations. One way to elucidate this is to cast the hid...
http://arxiv.org/pdf/2303.09435v1
2023-03-16T16:10:16
2023-03-16T16:10:16
{ "title": [ -0.009643555, 0.0055503845, -0.022201538, -0.016113281, -0.023101807, -0.02104187, -0.023391724, -0.029205322, -0.044891357, 0.021224976, -0.010940552, 0.021942139, -0.016540527, -0.035583496, 0.0057411194, -0.012001038, 0.020492554, ...
http://arxiv.org/abs/2303.09421v2
SheffieldVeraAI at SemEval-2023 Task 3: Mono and multilingual approaches for news genre, topic and persuasion technique classification
Ben Wu, Olesya Razuvayevskaya, Freddy Heppell, João A. Leite, Carolina Scarton, Kalina Bontcheva, Xingyi Song
cs.CL
This paper describes our approach for SemEval-2023 Task 3: Detecting the category, the framing, and the persuasion techniques in online news in a multi-lingual setup. For Subtask 1 (News Genre), we propose an ensemble of fully trained and adapter mBERT models which was ranked joint-first for German, and had the highest...
http://arxiv.org/pdf/2303.09421v2
2023-05-09T09:33:33
2023-03-16T15:54:23
{ "title": [ 0.0028743744, 0.014343262, -0.05795288, -0.007633209, 0.012084961, -0.035064697, -0.0040779114, -0.014190674, -0.03363037, 0.031829834, 0.0020122528, 0.03869629, -0.018966675, -0.032073975, 0.008781433, -0.012893677, -0.028167725, 0....
http://arxiv.org/abs/2303.13360v1
Towards the Scalable Evaluation of Cooperativeness in Language Models
Alan Chan, Maxime Riché, Jesse Clifton
cs.CL, cs.LG
It is likely that AI systems driven by pre-trained language models (PLMs) will increasingly be used to assist humans in high-stakes interactions with other agents, such as negotiation or conflict resolution. Consistent with the goals of Cooperative AI \citep{dafoe_open_2020}, we wish to understand and shape the multi-a...
http://arxiv.org/pdf/2303.13360v1
2023-03-16T15:34:23
2023-03-16T15:34:23
{ "title": [ -0.021148682, -0.000746727, -0.032989502, -0.07434082, 0.0016126633, -0.002084732, -0.060516357, 0.0063552856, 0.013336182, 0.04824829, -0.020965576, -0.0064430237, 0.022094727, -0.00018298626, -0.07421875, 0.007865906, -0.043029785, ...
http://arxiv.org/abs/2303.09366v2
The Scope of In-Context Learning for the Extraction of Medical Temporal Constraints
Parker Seegmiller, Joseph Gatto, Madhusudan Basak, Diane Cook, Hassan Ghasemzadeh, John Stankovic, Sarah Preum
cs.CL, cs.LG
Medications often impose temporal constraints on everyday patient activity. Violations of such medical temporal constraints (MTCs) lead to a lack of treatment adherence, in addition to poor health outcomes and increased healthcare expenses. These MTCs are found in drug usage guidelines (DUGs) in both patient education ...
http://arxiv.org/pdf/2303.09366v2
2023-10-16T14:22:41
2023-03-16T14:51:44
{ "title": [ 0.017990112, -0.0012340546, -0.05569458, -0.0859375, 0.028686523, -0.020080566, -0.02607727, 0.004886627, -0.00944519, 0.021942139, -0.015029907, 0.00074863434, -0.00944519, -0.03491211, -0.026626587, -0.020004272, 0.013313293, 0.009...
http://arxiv.org/abs/2303.09364v1
Tollywood Emotions: Annotation of Valence-Arousal in Telugu Song Lyrics
R Guru Ravi Shanker, B Manikanta Gupta, BV Koushik, Vinoo Alluri
cs.CL
Emotion recognition from a given music track has heavily relied on acoustic features, social tags, and metadata but is seldom focused on lyrics. There are no datasets of Indian language songs that contain both valence and arousal manual ratings of lyrics. We present a new manually annotated dataset of Telugu songs' lyr...
http://arxiv.org/pdf/2303.09364v1
2023-03-16T14:47:52
2023-03-16T14:47:52
{ "title": [ 0.03692627, -0.014579773, -0.033599854, -0.028335571, -0.016998291, -0.026855469, 0.018157959, 0.022842407, -0.01966858, 0.061767578, -0.00041389465, -0.00013911724, 0.044281006, -0.04135132, -0.040130615, -0.01965332, -0.00061941147, ...
http://arxiv.org/abs/2303.09325v1
Can Generative Pre-trained Transformers (GPT) Pass Assessments in Higher Education Programming Courses?
Jaromir Savelka, Arav Agarwal, Christopher Bogart, Yifan Song, Majd Sakr
cs.AI, cs.CL
We evaluated the capability of generative pre-trained transformers (GPT), to pass assessments in introductory and intermediate Python programming courses at the postsecondary level. Discussions of potential uses (e.g., exercise generation, code explanation) and misuses (e.g., cheating) of this emerging technology in pr...
http://arxiv.org/pdf/2303.09325v1
2023-03-16T13:58:45
2023-03-16T13:58:45
{ "title": [ 0.0446167, 0.0015220642, 0.003080368, -0.008422852, -0.0019350052, -0.032165527, 0.030532837, -0.035095215, 0.012031555, 0.015823364, -0.04748535, -0.009895325, 0.014831543, -0.042022705, -0.014450073, -0.030914307, 0.001250267, 0.00...
http://arxiv.org/abs/2303.09306v2
BanglaCoNER: Towards Robust Bangla Complex Named Entity Recognition
HAZ Sameen Shahgir, Ramisa Alam, Md. Zarif Ul Alam
cs.CL, cs.AI
Named Entity Recognition (NER) is a fundamental task in natural language processing that involves identifying and classifying named entities in text. But much work hasn't been done for complex named entity recognition in Bangla, despite being the seventh most spoken language globally. CNER is a more challenging task th...
http://arxiv.org/pdf/2303.09306v2
2023-03-17T15:13:01
2023-03-16T13:31:31
{ "title": [ -0.011184692, -0.0104599, -0.0345459, -0.082214355, 0.019104004, 0.0075683594, -0.039367676, 0.017593384, 0.007171631, 0.030761719, -0.012870789, 0.0027866364, 0.0036525726, -0.008255005, -0.056121826, -0.006313324, -0.021224976, 0.0...
http://arxiv.org/abs/2303.09266v2
SmartBERT: A Promotion of Dynamic Early Exiting Mechanism for Accelerating BERT Inference
Boren Hu, Yun Zhu, Jiacheng Li, Siliang Tang
cs.CL, cs.AI, cs.LG
Dynamic early exiting has been proven to improve the inference speed of the pre-trained language model like BERT. However, all samples must go through all consecutive layers before early exiting and more complex samples usually go through more layers, which still exists redundant computation. In this paper, we propose ...
http://arxiv.org/pdf/2303.09266v2
2023-05-08T13:05:00
2023-03-16T12:44:16
{ "title": [ -0.01612854, -0.0362854, -0.018692017, -0.07421875, -0.027618408, 0.014961243, -0.05078125, 0.004432678, 0.0115737915, 0.036712646, 0.013786316, -0.004699707, 0.028015137, -0.0335083, -0.035339355, 0.031951904, 0.01612854, 0.02600097...
http://arxiv.org/abs/2303.09184v2
Block-wise Bit-Compression of Transformer-based Models
Gaochen Dong, Wei Chen
cs.CL, cs.LG
With the popularity of the recent Transformer-based models represented by BERT, GPT-3 and ChatGPT, there has been state-of-the-art performance in a range of natural language processing tasks. However, the massive computations, huge memory footprint, and thus high latency of Transformer-based models is an inevitable cha...
http://arxiv.org/pdf/2303.09184v2
2023-04-01T12:50:29
2023-03-16T09:53:57
{ "title": [ -0.04248047, 0.0118637085, -0.03933716, -0.034820557, -0.006187439, 0.01423645, -0.06262207, -0.0013465881, 0.009460449, 0.062042236, -0.0036144257, -0.017730713, 0.03756714, -0.037994385, -0.038024902, 0.042114258, 0.043548584, -0.0...
http://arxiv.org/abs/2304.02015v1
How well do Large Language Models perform in Arithmetic tasks?
Zheng Yuan, Hongyi Yuan, Chuanqi Tan, Wei Wang, Songfang Huang
cs.CL, cs.AI
Large language models have emerged abilities including chain-of-thought to answer math word problems step by step. Solving math word problems not only requires abilities to disassemble problems via chain-of-thought but also needs to calculate arithmetic expressions correctly for each step. To the best of our knowledge,...
http://arxiv.org/pdf/2304.02015v1
2023-03-16T09:28:15
2023-03-16T09:28:15
{ "title": [ -0.020568848, -0.009490967, -0.035339355, -0.017425537, -0.043060303, -0.023956299, -0.040252686, 0.0038890839, 0.016448975, 0.030960083, -0.026412964, -0.035339355, 0.038757324, -0.030807495, -0.019500732, -0.0036811829, 0.0016231537, ...
http://arxiv.org/abs/2303.09136v1
A Short Survey of Viewing Large Language Models in Legal Aspect
Zhongxiang Sun
cs.CL
Large language models (LLMs) have transformed many fields, including natural language processing, computer vision, and reinforcement learning. These models have also made a significant impact in the field of law, where they are being increasingly utilized to automate various legal tasks, such as legal judgement predict...
http://arxiv.org/pdf/2303.09136v1
2023-03-16T08:01:22
2023-03-16T08:01:22
{ "title": [ -0.02508545, -0.019485474, -0.035583496, -0.056671143, -0.022842407, -0.014129639, -0.052886963, 0.015327454, 0.005302429, 0.019119263, -0.047973633, -0.01121521, 0.016311646, -0.022247314, -0.013923645, -0.014930725, -0.013061523, 0...
http://arxiv.org/abs/2303.09128v1
Exploring Distributional Shifts in Large Language Models for Code Analysis
Shushan Arakelyan, Rocktim Jyoti Das, Yi Mao, Xiang Ren
cs.CL, cs.LG, cs.SE
We systematically study the capacity of two large language models for code - CodeT5 and Codex - to generalize to out-of-domain data. In this study, we consider two fundamental applications - code summarization, and code generation. We split data into domains following its natural boundaries - by an organization, by a p...
http://arxiv.org/pdf/2303.09128v1
2023-03-16T07:45:46
2023-03-16T07:45:46
{ "title": [ -0.04550171, -0.01448822, -0.05508423, -0.045715332, 0.00022542477, -0.009986877, -0.054626465, 0.00944519, 0.018218994, 0.033081055, -0.01725769, 0.0022449493, 0.011428833, -0.017852783, -0.033447266, 0.010467529, -0.013137817, 0.02...
http://arxiv.org/abs/2303.09100v1
Patch-Token Aligned Bayesian Prompt Learning for Vision-Language Models
Xinyang Liu, Dongsheng Wang, Miaoge Li, Zhibin Duan, Yishi Xu, Bo Chen, Mingyuan Zhou
cs.CV, cs.CL, cs.LG
For downstream applications of vision-language pre-trained models, there has been significant interest in constructing effective prompts. Existing works on prompt engineering, which either require laborious manual designs or optimize the prompt tuning as a point estimation problem, may fail to describe diverse characte...
http://arxiv.org/pdf/2303.09100v1
2023-03-16T06:09:15
2023-03-16T06:09:15
{ "title": [ -0.020019531, -0.008590698, -0.05090332, -0.044647217, -0.025390625, -0.020370483, -0.058288574, 0.012756348, 0.01209259, 0.004760742, -0.060058594, -0.0132751465, 0.0016698837, -0.021621704, -0.026031494, -0.02722168, 0.022323608, 0...
http://arxiv.org/abs/2303.09093v3
GLEN: General-Purpose Event Detection for Thousands of Types
Qiusi Zhan, Sha Li, Kathryn Conger, Martha Palmer, Heng Ji, Jiawei Han
cs.CL
The progress of event extraction research has been hindered by the absence of wide-coverage, large-scale datasets. To make event extraction systems more accessible, we build a general-purpose event detection dataset GLEN, which covers 205K event mentions with 3,465 different types, making it more than 20x larger in ont...
http://arxiv.org/pdf/2303.09093v3
2023-10-31T17:21:45
2023-03-16T05:36:38
{ "title": [ 0.011886597, 0.003829956, -0.044799805, -0.08630371, 0.04034424, -0.011108398, -0.034118652, -0.022460938, -0.008049011, 0.018600464, 0.0020828247, 0.014442444, -0.001364708, -0.036987305, -0.018310547, 0.01914978, -0.006259918, 0.02...