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/2303.01410v1
NLP Workbench: Efficient and Extensible Integration of State-of-the-art Text Mining Tools
Peiran Yao, Matej Kosmajac, Abeer Waheed, Kostyantyn Guzhva, Natalie Hervieux, Denilson Barbosa
cs.CL
NLP Workbench is a web-based platform for text mining that allows non-expert users to obtain semantic understanding of large-scale corpora using state-of-the-art text mining models. The platform is built upon latest pre-trained models and open source systems from academia that provide semantic analysis functionalities,...
http://arxiv.org/pdf/2303.01410v1
2023-03-02T16:59:31
2023-03-02T16:59:31
{ "title": [ -0.027053833, -0.00006043911, -0.046295166, -0.05706787, 0.028289795, 0.0011663437, -0.028671265, -0.010353088, 0.018829346, 0.019699097, -0.037109375, 0.020904541, -0.010734558, 0.00031590462, -0.048583984, -0.0026111603, -0.025512695, ...
http://arxiv.org/abs/2303.01396v1
MLANet: Multi-Level Attention Network with Sub-instruction for Continuous Vision-and-Language Navigation
Zongtao He, Liuyi Wang, Shu Li, Qingqing Yan, Chengju Liu, Qijun Chen
cs.CV, cs.CL, cs.MM
Vision-and-Language Navigation (VLN) aims to develop intelligent agents to navigate in unseen environments only through language and vision supervision. In the recently proposed continuous settings (continuous VLN), the agent must act in a free 3D space and faces tougher challenges like real-time execution, complex ins...
http://arxiv.org/pdf/2303.01396v1
2023-03-02T16:26:14
2023-03-02T16:26:14
{ "title": [ -0.010925293, -0.013748169, -0.021347046, -0.035827637, -0.0042419434, -0.010818481, -0.016067505, -0.0013570786, 0.015838623, 0.030929565, -0.013053894, -0.0099487305, 0.00579834, -0.021057129, -0.04537964, -0.0014762878, -0.0034942627,...
http://arxiv.org/abs/2303.01347v1
Letz Translate: Low-Resource Machine Translation for Luxembourgish
Yewei Song, Saad Ezzini, Jacques Klein, Tegawende Bissyande, Clément Lefebvre, Anne Goujon
cs.CL, cs.SE
Natural language processing of Low-Resource Languages (LRL) is often challenged by the lack of data. Therefore, achieving accurate machine translation (MT) in a low-resource environment is a real problem that requires practical solutions. Research in multilingual models have shown that some LRLs can be handled with suc...
http://arxiv.org/pdf/2303.01347v1
2023-03-02T15:26:46
2023-03-02T15:26:46
{ "title": [ -0.06726074, -0.042144775, -0.048339844, 0.012611389, -0.06365967, -0.005695343, -0.04147339, 0.0057411194, -0.037841797, 0.071777344, -0.01398468, -0.005680084, 0.0826416, -0.02748108, -0.023269653, -0.0014448166, 0.003107071, 0.038...
http://arxiv.org/abs/2303.01341v1
Matching-based Term Semantics Pre-training for Spoken Patient Query Understanding
Zefa Hu, Xiuyi Chen, Haoran Wu, Minglun Han, Ziyi Ni, Jing Shi, Shuang Xu, Bo Xu
cs.CL
Medical Slot Filling (MSF) task aims to convert medical queries into structured information, playing an essential role in diagnosis dialogue systems. However, the lack of sufficient term semantics learning makes existing approaches hard to capture semantically identical but colloquial expressions of terms in medical co...
http://arxiv.org/pdf/2303.01341v1
2023-03-02T15:17:38
2023-03-02T15:17:38
{ "title": [ 0.005012512, 0.0033168793, -0.034851074, -0.10144043, 0.019378662, -0.007156372, -0.0496521, 0.02810669, -0.017730713, 0.041503906, -0.023254395, 0.0017766953, -0.0011425018, -0.00957489, -0.039215088, -0.02279663, 0.009536743, -0.00...
http://arxiv.org/abs/2303.01217v1
Synthetic Misinformers: Generating and Combating Multimodal Misinformation
Stefanos-Iordanis Papadopoulos, Christos Koutlis, Symeon Papadopoulos, Panagiotis C. Petrantonakis
cs.MM, cs.AI, cs.CL
With the expansion of social media and the increasing dissemination of multimedia content, the spread of misinformation has become a major concern. This necessitates effective strategies for multimodal misinformation detection (MMD) that detect whether the combination of an image and its accompanying text could mislead...
http://arxiv.org/pdf/2303.01217v1
2023-03-02T12:59:01
2023-03-02T12:59:01
{ "title": [ 0.015686035, 0.012908936, -0.03955078, -0.06567383, 0.023757935, -0.0048103333, -0.02067566, -0.023788452, -0.055999756, 0.030639648, 0.015823364, 0.0143966675, -0.018920898, -0.031402588, -0.024154663, -0.006450653, -0.027557373, 0....
http://arxiv.org/abs/2303.01197v1
Document Provenance and Authentication through Authorship Classification
Muhammad Tayyab Zamir, Muhammad Asif Ayub, Jebran Khan, Muhammad Jawad Ikram, Nasir Ahmad, Kashif Ahmad
cs.CL, cs.DL
Style analysis, which is relatively a less explored topic, enables several interesting applications. For instance, it allows authors to adjust their writing style to produce a more coherent document in collaboration. Similarly, style analysis can also be used for document provenance and authentication as a primary step...
http://arxiv.org/pdf/2303.01197v1
2023-03-02T12:26:03
2023-03-02T12:26:03
{ "title": [ -0.033081055, 0.011741638, -0.051757812, -0.045776367, 0.013191223, -0.014884949, -0.03778076, -0.010520935, -0.025390625, 0.034973145, 0.0022640228, 0.011383057, 0.0038871765, -0.031951904, -0.029373169, -0.0076675415, -0.014404297, ...
http://arxiv.org/abs/2303.01194v2
UZH_CLyp at SemEval-2023 Task 9: Head-First Fine-Tuning and ChatGPT Data Generation for Cross-Lingual Learning in Tweet Intimacy Prediction
Andrianos Michail, Stefanos Konstantinou, Simon Clematide
cs.CL, cs.AI, 68T50
This paper describes the submission of UZH_CLyp for the SemEval 2023 Task 9 "Multilingual Tweet Intimacy Analysis". We achieved second-best results in all 10 languages according to the official Pearson's correlation regression evaluation measure. Our cross-lingual transfer learning approach explores the benefits of usi...
http://arxiv.org/pdf/2303.01194v2
2023-04-24T12:19:56
2023-03-02T12:18:53
{ "title": [ -0.0027160645, -0.040771484, 0.011993408, -0.02168274, -0.039489746, 0.012176514, -0.028701782, -0.004722595, 0.012870789, 0.057678223, -0.035827637, 0.044311523, 0.019439697, -0.00048089027, -0.0084991455, 0.021224976, -0.012931824, ...
http://arxiv.org/abs/2303.01191v1
Denoising-based UNMT is more robust to word-order divergence than MASS-based UNMT
Tamali Banerjee, Rudra Murthy V, Pushpak Bhattacharyya
cs.CL
We aim to investigate whether UNMT approaches with self-supervised pre-training are robust to word-order divergence between language pairs. We achieve this by comparing two models pre-trained with the same self-supervised pre-training objective. The first model is trained on language pairs with different word-orders, a...
http://arxiv.org/pdf/2303.01191v1
2023-03-02T12:11:58
2023-03-02T12:11:58
{ "title": [ 0.015838623, 0.025360107, -0.03152466, -0.05340576, -0.022644043, -0.014266968, -0.06774902, -0.029800415, -0.0043029785, 0.07330322, 0.009666443, -0.015541077, 0.07513428, -0.046203613, -0.007926941, 0.029953003, -0.012535095, 0.034...
http://arxiv.org/abs/2303.01510v1
INO at Factify 2: Structure Coherence based Multi-Modal Fact Verification
Yinuo Zhang, Zhulin Tao, Xi Wang, Tongyue Wang
cs.LG, cs.AI, cs.CL
This paper describes our approach to the multi-modal fact verification (FACTIFY) challenge at AAAI2023. In recent years, with the widespread use of social media, fake news can spread rapidly and negatively impact social security. Automatic claim verification becomes more and more crucial to combat fake news. In fact ve...
http://arxiv.org/pdf/2303.01510v1
2023-03-02T11:18:56
2023-03-02T11:18:56
{ "title": [ -0.017211914, 0.0042304993, -0.06121826, -0.114990234, 0.027862549, 0.012908936, -0.08929443, 0.0027389526, -0.01247406, 0.052490234, 0.016479492, 0.011528015, 0.03604126, -0.029205322, -0.02720642, 0.014549255, -0.043395996, 0.04922...
http://arxiv.org/abs/2303.01094v1
CTRLStruct: Dialogue Structure Learning for Open-Domain Response Generation
Congchi Yin, Piji Li, Zhaochun Ren
cs.CL
Dialogue structure discovery is essential in dialogue generation. Well-structured topic flow can leverage background information and predict future topics to help generate controllable and explainable responses. However, most previous work focused on dialogue structure learning in task-oriented dialogue other than open...
http://arxiv.org/pdf/2303.01094v1
2023-03-02T09:27:11
2023-03-02T09:27:11
{ "title": [ -0.008148193, -0.014579773, -0.031173706, -0.06304932, -0.011054993, 0.007507324, -0.054473877, 0.0012283325, -0.029693604, 0.0143966675, 0.015129089, 0.009346008, -0.05444336, -0.032165527, -0.033966064, -0.019134521, -0.0046539307, ...
http://arxiv.org/abs/2303.01086v1
LiteG2P: A fast, light and high accuracy model for grapheme-to-phoneme conversion
Chunfeng Wang, Peisong Huang, Yuxiang Zou, Haoyu Zhang, Shichao Liu, Xiang Yin, Zejun Ma
cs.CL, cs.SD, eess.AS
As a key component of automated speech recognition (ASR) and the front-end in text-to-speech (TTS), grapheme-to-phoneme (G2P) plays the role of converting letters to their corresponding pronunciations. Existing methods are either slow or poor in performance, and are limited in application scenarios, particularly in the...
http://arxiv.org/pdf/2303.01086v1
2023-03-02T09:16:21
2023-03-02T09:16:21
{ "title": [ -0.04244995, -0.021987915, -0.024932861, -0.027572632, -0.009025574, 0.013420105, -0.03366089, 0.0031471252, 0.0052108765, 0.046020508, 0.00491333, -0.0026893616, 0.031555176, -0.025115967, -0.037902832, 0.029663086, -0.016601562, 0....
http://arxiv.org/abs/2303.01081v1
Can BERT Refrain from Forgetting on Sequential Tasks? A Probing Study
Mingxu Tao, Yansong Feng, Dongyan Zhao
cs.CL, cs.AI
Large pre-trained language models help to achieve state of the art on a variety of natural language processing (NLP) tasks, nevertheless, they still suffer from forgetting when incrementally learning a sequence of tasks. To alleviate this problem, recent works enhance existing models by sparse experience replay and loc...
http://arxiv.org/pdf/2303.01081v1
2023-03-02T09:03:43
2023-03-02T09:03:43
{ "title": [ 0.006111145, -0.011268616, -0.034942627, -0.06793213, -0.012519836, -0.025726318, -0.022155762, 0.01335144, -0.016220093, 0.042510986, -0.0023479462, -0.0013084412, 0.026809692, -0.022232056, -0.041900635, 0.0012454987, -0.011940002, ...
http://arxiv.org/abs/2303.01080v1
LANDMARK: Language-guided Representation Enhancement Framework for Scene Graph Generation
Xiaoguang Chang, Teng Wang, Shaowei Cai, Changyin Sun
cs.CV, cs.AI, cs.CL
Scene graph generation (SGG) is a sophisticated task that suffers from both complex visual features and dataset long-tail problem. Recently, various unbiased strategies have been proposed by designing novel loss functions and data balancing strategies. Unfortunately, these unbiased methods fail to emphasize language pr...
http://arxiv.org/pdf/2303.01080v1
2023-03-02T09:03:11
2023-03-02T09:03:11
{ "title": [ -0.014190674, -0.026672363, -0.04498291, -0.060546875, -0.01626587, 0.00038576126, -0.051452637, 0.009757996, 0.007972717, 0.01789856, -0.016952515, 0.00705719, -0.023422241, -0.027160645, -0.012512207, -0.007457733, 0.0025901794, 0....
http://arxiv.org/abs/2303.03267v1
Evaluating Parameter-Efficient Transfer Learning Approaches on SURE Benchmark for Speech Understanding
Yingting Li, Ambuj Mehrish, Shuai Zhao, Rishabh Bhardwaj, Amir Zadeh, Navonil Majumder, Rada Mihalcea, Soujanya Poria
cs.CL, cs.SD, eess.AS
Fine-tuning is widely used as the default algorithm for transfer learning from pre-trained models. Parameter inefficiency can however arise when, during transfer learning, all the parameters of a large pre-trained model need to be updated for individual downstream tasks. As the number of parameters grows, fine-tuning i...
http://arxiv.org/pdf/2303.03267v1
2023-03-02T08:57:33
2023-03-02T08:57:33
{ "title": [ -0.008377075, 0.0019330978, -0.03451538, -0.06201172, 0.0027656555, -0.005012512, -0.046447754, 0.0027484894, 0.0003633499, 0.02923584, -0.011169434, 0.007675171, -0.0052452087, -0.022750854, -0.03640747, -0.0000320673, -0.013000488, ...
http://arxiv.org/abs/2303.01068v1
Targeted Adversarial Attacks against Neural Machine Translation
Sahar Sadrizadeh, AmirHossein Dabiri Aghdam, Ljiljana Dolamic, Pascal Frossard
cs.CL, cs.CR, cs.LG
Neural Machine Translation (NMT) systems are used in various applications. However, it has been shown that they are vulnerable to very small perturbations of their inputs, known as adversarial attacks. In this paper, we propose a new targeted adversarial attack against NMT models. In particular, our goal is to insert a...
http://arxiv.org/pdf/2303.01068v1
2023-03-02T08:43:30
2023-03-02T08:43:30
{ "title": [ -0.020080566, -0.01033783, -0.011772156, -0.03704834, 0.007282257, 0.0021324158, -0.0043907166, -0.006958008, -0.047027588, 0.043182373, 0.02078247, 0.013710022, -0.0041236877, -0.024032593, -0.03527832, 0.0034866333, -0.008201599, 0...
http://arxiv.org/abs/2303.01064v1
Adopting the Multi-answer Questioning Task with an Auxiliary Metric for Extreme Multi-label Text Classification Utilizing the Label Hierarchy
Li Wang, Ying Wah Teh, Mohammed Ali Al-Garadi
cs.CL
Extreme multi-label text classification utilizes the label hierarchy to partition extreme labels into multiple label groups, turning the task into simple multi-group multi-label classification tasks. Current research encodes labels as a vector with fixed length which needs establish multiple classifiers for different l...
http://arxiv.org/pdf/2303.01064v1
2023-03-02T08:40:31
2023-03-02T08:40:31
{ "title": [ -0.017211914, 0.00017416477, -0.025482178, -0.019378662, -0.00022149086, -0.010887146, 0.008857727, 0.0037956238, 0.005794525, 0.012039185, -0.014465332, 0.0368042, -0.01675415, -0.032196045, -0.01928711, 0.0052337646, -0.015510559, ...
http://arxiv.org/abs/2303.01037v3
Google USM: Scaling Automatic Speech Recognition Beyond 100 Languages
Yu Zhang, Wei Han, James Qin, Yongqiang Wang, Ankur Bapna, Zhehuai Chen, Nanxin Chen, Bo Li, Vera Axelrod, Gary Wang, Zhong Meng, Ke Hu, Andrew Rosenberg, Rohit Prabhavalkar, Daniel S. Park, Parisa Haghani, Jason Riesa, Ginger Perng, Hagen Soltau, Trevor Strohman, Bhuvana Ramabhadran, Tara Sainath, Pedro Moreno, Chung-...
cs.CL, cs.SD, eess.AS
We introduce the Universal Speech Model (USM), a single large model that performs automatic speech recognition (ASR) across 100+ languages. This is achieved by pre-training the encoder of the model on a large unlabeled multilingual dataset of 12 million (M) hours spanning over 300 languages, and fine-tuning on a smalle...
http://arxiv.org/pdf/2303.01037v3
2023-09-25T01:20:23
2023-03-02T07:47:18
{ "title": [ -0.012825012, -0.017745972, -0.067993164, -0.04916382, -0.005844116, -0.023849487, -0.04827881, 0.014533997, -0.024246216, 0.030563354, -0.022735596, -0.009963989, 0.015586853, -0.040985107, -0.011619568, -0.017196655, -0.00083112717, ...
http://arxiv.org/abs/2303.01962v1
Less is More: Mitigate Spurious Correlations for Open-Domain Dialogue Response Generation Models by Causal Discovery
Tao Feng, Lizhen Qu, Gholamreza Haffari
cs.CL, cs.AI
In this paper, we conduct the first study on spurious correlations for open-domain response generation models based on a corpus CGDIALOG curated in our work. The cur rent models indeed suffer from spurious correlations and have a tendency of generating irrelevant and generic responses. Inspired by causal discovery algo...
http://arxiv.org/pdf/2303.01962v1
2023-03-02T06:33:48
2023-03-02T06:33:48
{ "title": [ 0.0053367615, 0.000044882298, -0.04385376, -0.06567383, 0.003566742, -0.010498047, -0.041748047, -0.000053822994, -0.0340271, 0.020736694, 0.020431519, 0.029510498, -0.028198242, -0.032562256, -0.01525116, -0.013420105, -0.01687622, ...
http://arxiv.org/abs/2303.00978v1
Leveraging Large Text Corpora for End-to-End Speech Summarization
Kohei Matsuura, Takanori Ashihara, Takafumi Moriya, Tomohiro Tanaka, Atsunori Ogawa, Marc Delcroix, Ryo Masumura
cs.CL, eess.AS
End-to-end speech summarization (E2E SSum) is a technique to directly generate summary sentences from speech. Compared with the cascade approach, which combines automatic speech recognition (ASR) and text summarization models, the E2E approach is more promising because it mitigates ASR errors, incorporates nonverbal in...
http://arxiv.org/pdf/2303.00978v1
2023-03-02T05:19:49
2023-03-02T05:19:49
{ "title": [ 0.0063819885, 0.0061035156, -0.030349731, -0.09851074, 0.01576233, 0.009254456, -0.052886963, 0.016311646, -0.006374359, 0.054718018, -0.012161255, 0.016189575, 0.0060920715, 0.0010757446, -0.055236816, 0.0077323914, -0.0096588135, 0...
http://arxiv.org/abs/2303.00969v2
Rethinking the Reasonability of the Test Set for Simultaneous Machine Translation
Mengge Liu, Wen Zhang, Xiang Li, Jian Luan, Bin Wang, Yuhang Guo, Shuoying Chen
cs.CL
Simultaneous machine translation (SimulMT) models start translation before the end of the source sentence, making the translation monotonically aligned with the source sentence. However, the general full-sentence translation test set is acquired by offline translation of the entire source sentence, which is not designe...
http://arxiv.org/pdf/2303.00969v2
2023-03-13T12:14:30
2023-03-02T05:06:44
{ "title": [ -0.03894043, -0.015823364, -0.0151901245, -0.057617188, -0.0135269165, 0.015853882, -0.036956787, 0.01612854, -0.027297974, 0.040039062, -0.005279541, 0.0093307495, 0.0062446594, -0.013702393, -0.04977417, -0.00081443787, -0.012451172, ...
http://arxiv.org/abs/2303.00915v1
Large-Scale Domain-Specific Pretraining for Biomedical Vision-Language Processing
Sheng Zhang, Yanbo Xu, Naoto Usuyama, Jaspreet Bagga, Robert Tinn, Sam Preston, Rajesh Rao, Mu Wei, Naveen Valluri, Cliff Wong, Matthew P. Lungren, Tristan Naumann, Hoifung Poon
cs.CV, cs.CL
Contrastive pretraining on parallel image-text data has attained great success in vision-language processing (VLP), as exemplified by CLIP and related methods. However, prior explorations tend to focus on general domains in the web. Biomedical images and text are rather different, but publicly available datasets are sm...
http://arxiv.org/pdf/2303.00915v1
2023-03-02T02:20:04
2023-03-02T02:20:04
{ "title": [ 0.0019741058, -0.0029029846, -0.046142578, -0.06536865, 0.008171082, -0.021408081, -0.04269409, 0.006679535, 0.008041382, 0.034088135, -0.006603241, -0.0028305054, 0.0126571655, -0.035125732, -0.03265381, 0.008262634, 0.018737793, 0....
http://arxiv.org/abs/2303.00908v3
Interactive Text Generation
Felix Faltings, Michel Galley, Baolin Peng, Kianté Brantley, Weixin Cai, Yizhe Zhang, Jianfeng Gao, Bill Dolan
cs.CL
Users interact with text, image, code, or other editors on a daily basis. However, machine learning models are rarely trained in the settings that reflect the interactivity between users and their editor. This is understandable as training AI models with real users is not only slow and costly, but what these models lea...
http://arxiv.org/pdf/2303.00908v3
2023-11-11T20:43:13
2023-03-02T01:57:17
{ "title": [ -0.021362305, -0.0061531067, -0.02722168, -0.04119873, -0.04324341, 0.0018033981, -0.050720215, -0.0058937073, 0.0022296906, 0.06210327, -0.030410767, -0.0020771027, 0.002046585, -0.0146865845, -0.039855957, 0.003774643, 0.0073890686, ...
http://arxiv.org/abs/2303.00897v1
Stochastic Clustered Federated Learning
Dun Zeng, Xiangjing Hu, Shiyu Liu, Yue Yu, Qifan Wang, Zenglin Xu
cs.LG, cs.CL
Federated learning is a distributed learning framework that takes full advantage of private data samples kept on edge devices. In real-world federated learning systems, these data samples are often decentralized and Non-Independently Identically Distributed (Non-IID), causing divergence and performance degradation in t...
http://arxiv.org/pdf/2303.00897v1
2023-03-02T01:39:16
2023-03-02T01:39:16
{ "title": [ -0.012611389, 0.019577026, -0.066101074, -0.08239746, 0.04257202, 0.0025844574, -0.024719238, 0.0068359375, -0.031188965, 0.040740967, -0.03756714, 0.030578613, -0.024017334, 0.0076065063, -0.016799927, -0.0071144104, 0.0057029724, -...
http://arxiv.org/abs/2303.00855v1
Grounded Decoding: Guiding Text Generation with Grounded Models for Robot Control
Wenlong Huang, Fei Xia, Dhruv Shah, Danny Driess, Andy Zeng, Yao Lu, Pete Florence, Igor Mordatch, Sergey Levine, Karol Hausman, Brian Ichter
cs.RO, cs.AI, cs.CL, cs.CV, cs.LG
Recent progress in large language models (LLMs) has demonstrated the ability to learn and leverage Internet-scale knowledge through pre-training with autoregressive models. Unfortunately, applying such models to settings with embodied agents, such as robots, is challenging due to their lack of experience with the physi...
http://arxiv.org/pdf/2303.00855v1
2023-03-01T22:58:50
2023-03-01T22:58:50
{ "title": [ -0.056243896, -0.0075302124, -0.021865845, -0.008758545, -0.028625488, 0.00283432, -0.045898438, -0.029846191, -0.0049552917, 0.051635742, -0.031951904, 0.022109985, 0.026428223, -0.011245728, -0.04083252, -0.0012083054, -0.010566711, ...
http://arxiv.org/abs/2303.00815v1
Soft Prompt Guided Joint Learning for Cross-Domain Sentiment Analysis
Jingli Shi, Weihua Li, Quan Bai, Yi Yang, Jianhua Jiang
cs.CL, cs.AI
Aspect term extraction is a fundamental task in fine-grained sentiment analysis, which aims at detecting customer's opinion targets from reviews on product or service. The traditional supervised models can achieve promising results with annotated datasets, however, the performance dramatically decreases when they are a...
http://arxiv.org/pdf/2303.00815v1
2023-03-01T20:33:37
2023-03-01T20:33:37
{ "title": [ -0.009765625, 0.0010948181, -0.030715942, -0.052246094, 0.015808105, -0.012420654, -0.014968872, 0.000667572, -0.0031604767, 0.026184082, -0.037506104, 0.007320404, 0.010070801, -0.010864258, -0.026535034, -0.0023937225, -0.021453857, ...
http://arxiv.org/abs/2303.00807v3
UDAPDR: Unsupervised Domain Adaptation via LLM Prompting and Distillation of Rerankers
Jon Saad-Falcon, Omar Khattab, Keshav Santhanam, Radu Florian, Martin Franz, Salim Roukos, Avirup Sil, Md Arafat Sultan, Christopher Potts
cs.IR, cs.CL
Many information retrieval tasks require large labeled datasets for fine-tuning. However, such datasets are often unavailable, and their utility for real-world applications can diminish quickly due to domain shifts. To address this challenge, we develop and motivate a method for using large language models (LLMs) to ge...
http://arxiv.org/pdf/2303.00807v3
2023-10-13T17:23:04
2023-03-01T20:21:23
{ "title": [ -0.03387451, 0.008331299, -0.05682373, -0.048187256, 0.0027675629, -0.005092621, -0.049713135, -0.029769897, 0.0155181885, 0.016616821, -0.0016202927, 0.011749268, -0.0008635521, -0.031036377, -0.00072813034, 0.021591187, -0.0032157898, ...
http://arxiv.org/abs/2303.13622v1
Authorship attribution for Differences between Literary Texts by Bilingual Russian-French and Non-Bilingual French Authors
Margarita Makarova
cs.CL
Do bilingual Russian-French authors of the end of the twentieth century such as Andre\"i Makine, Val\'ery Afanassiev, Vladimir F\'edorovski, Iegor Gran, Luba Jurgenson have common stylistic traits in the novels they wrote in French? Can we distinguish between them and non-bilingual French writers' texts? Is the phenome...
http://arxiv.org/pdf/2303.13622v1
2023-03-01T20:20:05
2023-03-01T20:20:05
{ "title": [ -0.022506714, -0.000056147575, -0.05456543, -0.02532959, -0.014770508, -0.016403198, -0.027008057, -0.011184692, -0.058135986, 0.038604736, -0.0046539307, 0.025863647, 0.014549255, -0.03186035, 0.030044556, -0.021133423, -0.018615723, ...
http://arxiv.org/abs/2303.00802v1
Synthetic Cross-accent Data Augmentation for Automatic Speech Recognition
Philipp Klumpp, Pooja Chitkara, Leda Sarı, Prashant Serai, Jilong Wu, Irina-Elena Veliche, Rongqing Huang, Qing He
cs.CL, cs.SD, eess.AS
The awareness for biased ASR datasets or models has increased notably in recent years. Even for English, despite a vast amount of available training data, systems perform worse for non-native speakers. In this work, we improve an accent-conversion model (ACM) which transforms native US-English speech into accented pron...
http://arxiv.org/pdf/2303.00802v1
2023-03-01T20:05:19
2023-03-01T20:05:19
{ "title": [ 0.004299164, 0.00274086, -0.036621094, -0.051757812, 0.0077934265, -0.0063285828, -0.026733398, -0.0028572083, 0.0007596016, 0.045898438, -0.03237915, 0.0017175674, 0.01928711, -0.017700195, -0.042785645, 0.010498047, 0.00043821335, ...
http://arxiv.org/abs/2303.00786v2
Building High-accuracy Multilingual ASR with Gated Language Experts and Curriculum Training
Eric Sun, Jinyu Li, Yuxuan Hu, Yimeng Zhu, Long Zhou, Jian Xue, Peidong Wang, Linquan Liu, Shujie Liu, Edward Lin, Yifan Gong
cs.CL, eess.AS
We propose gated language experts and curriculum training to enhance multilingual transformer transducer models without requiring language identification (LID) input from users during inference. Our method incorporates a gating mechanism and LID loss, enabling transformer experts to learn language-specific information....
http://arxiv.org/pdf/2303.00786v2
2023-07-07T23:48:04
2023-03-01T19:20:01
{ "title": [ -0.010551453, -0.008201599, -0.057434082, -0.08227539, -0.024780273, -0.019546509, -0.06359863, 0.033050537, -0.045410156, 0.029342651, 0.0096206665, 0.009635925, 0.02229309, -0.056030273, 0.025497437, 0.006008148, 0.0055732727, 0.04...
http://arxiv.org/abs/2303.00733v1
SpeechPrompt v2: Prompt Tuning for Speech Classification Tasks
Kai-Wei Chang, Yu-Kai Wang, Hua Shen, Iu-thing Kang, Wei-Cheng Tseng, Shang-Wen Li, Hung-yi Lee
eess.AS, cs.AI, cs.CL, cs.LG, cs.SD
Prompt tuning is a technology that tunes a small set of parameters to steer a pre-trained language model (LM) to directly generate the output for downstream tasks. Recently, prompt tuning has demonstrated its storage and computation efficiency in both natural language processing (NLP) and speech processing fields. Thes...
http://arxiv.org/pdf/2303.00733v1
2023-03-01T18:47:41
2023-03-01T18:47:41
{ "title": [ -0.010871887, -0.01084137, -0.05532837, -0.07537842, 0.009094238, -0.020980835, -0.052001953, 0.010734558, -0.011428833, 0.015731812, -0.029663086, 0.0070648193, 0.004573822, -0.03488159, -0.01499176, -0.009605408, -0.018356323, 0.05...
http://arxiv.org/abs/2303.00722v1
A Systematic Analysis of Vocabulary and BPE Settings for Optimal Fine-tuning of NMT: A Case Study of In-domain Translation
J. Pourmostafa Roshan Sharami, D. Shterionov, P. Spronck
cs.CL, cs.AI
The effectiveness of Neural Machine Translation (NMT) models largely depends on the vocabulary used at training; small vocabularies can lead to out-of-vocabulary problems -- large ones, to memory issues. Subword (SW) tokenization has been successfully employed to mitigate these issues. The choice of vocabulary and SW t...
http://arxiv.org/pdf/2303.00722v1
2023-03-01T18:26:47
2023-03-01T18:26:47
{ "title": [ 0.003479004, -0.015403748, 0.008605957, -0.05255127, -0.0211792, -0.00011026859, -0.030700684, 0.0047454834, -0.028656006, 0.06213379, 0.011413574, 0.04458618, 0.013931274, -0.014541626, -0.015899658, 0.014083862, -0.027282715, 0.032...
http://arxiv.org/abs/2303.09402v1
ToxVis: Enabling Interpretability of Implicit vs. Explicit Toxicity Detection Models with Interactive Visualization
Uma Gunturi, Xiaohan Ding, Eugenia H. Rho
cs.CL, cs.HC
The rise of hate speech on online platforms has led to an urgent need for effective content moderation. However, the subjective and multi-faceted nature of hateful online content, including implicit hate speech, poses significant challenges to human moderators and content moderation systems. To address this issue, we d...
http://arxiv.org/pdf/2303.09402v1
2023-03-01T17:24:15
2023-03-01T17:24:15
{ "title": [ 0.025024414, 0.0054473877, 0.0031147003, -0.07006836, 0.017150879, 0.0048599243, -0.027633667, -0.0435791, -0.008087158, 0.003376007, -0.03463745, 0.016098022, -0.048675537, -0.01889038, -0.04559326, -0.025741577, -0.005935669, 0.035...
http://arxiv.org/abs/2303.01261v1
ParrotTTS: Text-to-Speech synthesis by exploiting self-supervised representations
Saiteja Kosgi, Neil Kumar Shah, Vishal Tambrahalli, Neha Sherin, Vineet Gandhi
cs.CL, cs.SD, eess.AS
Text-to-speech (TTS) systems are modelled as mel-synthesizers followed by speech-vocoders since the era of statistical TTS that is carried forward into neural designs. We propose an alternative approach to TTS modelling referred to as ParrotTTS borrowing from self-supervised learning (SSL) methods. ParrotTTS takes a tw...
http://arxiv.org/pdf/2303.01261v1
2023-03-01T17:23:12
2023-03-01T17:23:12
{ "title": [ -0.0041236877, 0.00042581558, -0.028549194, -0.074645996, 0.002670288, 0.011329651, -0.05307007, 0.018234253, -0.025314331, 0.055603027, -0.030319214, 0.029464722, 0.014480591, -0.0022888184, -0.04119873, -0.001086235, -0.015975952, ...
http://arxiv.org/abs/2303.00628v2
MuAViC: A Multilingual Audio-Visual Corpus for Robust Speech Recognition and Robust Speech-to-Text Translation
Mohamed Anwar, Bowen Shi, Vedanuj Goswami, Wei-Ning Hsu, Juan Pino, Changhan Wang
cs.CL, eess.AS
We introduce MuAViC, a multilingual audio-visual corpus for robust speech recognition and robust speech-to-text translation providing 1200 hours of audio-visual speech in 9 languages. It is fully transcribed and covers 6 English-to-X translation as well as 6 X-to-English translation directions. To the best of our knowl...
http://arxiv.org/pdf/2303.00628v2
2023-03-07T16:41:01
2023-03-01T16:31:01
{ "title": [ -0.016174316, -0.013702393, -0.029754639, -0.05795288, -0.0028133392, 0.0029773712, -0.04397583, 0.008102417, -0.020263672, 0.060272217, -0.033294678, 0.01247406, 0.041137695, -0.0064926147, -0.041656494, -0.0107040405, 0.0044517517, ...
http://arxiv.org/abs/2303.00595v2
A Universal Question-Answering Platform for Knowledge Graphs
Reham Omar, Ishika Dhall, Panos Kalnis, Essam Mansour
cs.AI, cs.CL, cs.DB
Knowledge from diverse application domains is organized as knowledge graphs (KGs) that are stored in RDF engines accessible in the web via SPARQL endpoints. Expressing a well-formed SPARQL query requires information about the graph structure and the exact URIs of its components, which is impractical for the average use...
http://arxiv.org/pdf/2303.00595v2
2023-08-08T17:16:03
2023-03-01T15:35:32
{ "title": [ -0.035614014, -0.015319824, -0.05114746, -0.08270264, -0.007247925, 0.020080566, -0.06585693, 0.02029419, 0.030563354, 0.030975342, -0.021835327, -0.0010213852, -0.0018987656, -0.0030059814, -0.059326172, 0.0057754517, 0.0009303093, ...
http://arxiv.org/abs/2303.00534v1
RAMM: Retrieval-augmented Biomedical Visual Question Answering with Multi-modal Pre-training
Zheng Yuan, Qiao Jin, Chuanqi Tan, Zhengyun Zhao, Hongyi Yuan, Fei Huang, Songfang Huang
cs.CV, cs.CL
Vision-and-language multi-modal pretraining and fine-tuning have shown great success in visual question answering (VQA). Compared to general domain VQA, the performance of biomedical VQA suffers from limited data. In this paper, we propose a retrieval-augmented pretrain-and-finetune paradigm named RAMM for biomedical V...
http://arxiv.org/pdf/2303.00534v1
2023-03-01T14:21:19
2023-03-01T14:21:19
{ "title": [ 0.0031147003, -0.0019950867, -0.018478394, -0.04711914, -0.008010864, 0.010360718, -0.035186768, -0.011299133, 0.028884888, 0.028518677, -0.0028438568, -0.0024642944, 0.0008177757, -0.025985718, -0.060913086, 0.0015134811, 0.019317627, ...
http://arxiv.org/abs/2303.00465v2
Uzbek text's correspondence with the educational potential of pupils: a case study of the School corpus
Khabibulla Madatov, Sanatbek Matlatipov, Mersaid Aripov
cs.CL
One of the major challenges of an educational system is choosing appropriate content considering pupils' age and intellectual potential. In this article the experiment of primary school grades (from 1st to 4th grades) is considered for automatically determining the correspondence of an educational materials recommended...
http://arxiv.org/pdf/2303.00465v2
2023-03-18T06:54:57
2023-03-01T12:46:00
{ "title": [ 0.043182373, 0.0058555603, 0.0005569458, 0.0146102905, -0.0022068024, -0.023666382, 0.035064697, -0.04623413, -0.01876831, 0.049438477, -0.0011119843, -0.010261536, 0.010467529, -0.039733887, -0.030792236, -0.032836914, -0.016937256, ...
http://arxiv.org/abs/2303.00461v1
Uzbek text summarization based on TF-IDF
Khabibulla Madatov, Shukurla Bekchanov, Jernej Vičič
cs.CL
The volume of information is increasing at an incredible rate with the rapid development of the Internet and electronic information services. Due to time constraints, we don't have the opportunity to read all this information. Even the task of analyzing textual data related to one field requires a lot of work. The text...
http://arxiv.org/pdf/2303.00461v1
2023-03-01T12:39:46
2023-03-01T12:39:46
{ "title": [ 0.018356323, 0.012756348, -0.005493164, -0.0501709, -0.0056533813, -0.007850647, -0.041168213, -0.030380249, -0.002960205, 0.08123779, 0.01890564, 0.00869751, 0.059661865, -0.020690918, -0.043518066, 0.030654907, 0.0038890839, -0.000...
http://arxiv.org/abs/2303.00456v3
N-best T5: Robust ASR Error Correction using Multiple Input Hypotheses and Constrained Decoding Space
Rao Ma, Mark J. F. Gales, Kate M. Knill, Mengjie Qian
cs.CL, cs.SD, eess.AS
Error correction models form an important part of Automatic Speech Recognition (ASR) post-processing to improve the readability and quality of transcriptions. Most prior works use the 1-best ASR hypothesis as input and therefore can only perform correction by leveraging the context within one sentence. In this work, we...
http://arxiv.org/pdf/2303.00456v3
2023-10-10T09:45:13
2023-03-01T12:32:34
{ "title": [ -0.023269653, -0.0129852295, -0.05517578, -0.064941406, -0.009986877, -0.012413025, -0.046875, 0.010177612, -0.009689331, 0.033111572, -0.0015964508, -0.015205383, 0.031921387, -0.040527344, -0.010223389, 0.001613617, 0.00008678436, ...
http://arxiv.org/abs/2303.00408v1
A Persian Benchmark for Joint Intent Detection and Slot Filling
Masoud Akbari, Amir Hossein Karimi, Tayyebeh Saeedi, Zeinab Saeidi, Kiana Ghezelbash, Fatemeh Shamsezat, Mohammad Akbari, Ali Mohades
cs.CL
Natural Language Understanding (NLU) is important in today's technology as it enables machines to comprehend and process human language, leading to improved human-computer interactions and advancements in fields such as virtual assistants, chatbots, and language-based AI systems. This paper highlights the significance ...
http://arxiv.org/pdf/2303.00408v1
2023-03-01T10:57:21
2023-03-01T10:57:21
{ "title": [ -0.012969971, 0.008773804, -0.02507019, -0.02796936, 0.012382507, -0.007499695, -0.010879517, -0.04034424, 0.008056641, 0.0143585205, -0.039764404, 0.016540527, -0.02670288, -0.020629883, -0.02960205, -0.015556335, -0.024749756, 0.02...
http://arxiv.org/abs/2303.01258v1
Domain-adapted large language models for classifying nuclear medicine reports
Zachary Huemann, Changhee Lee, Junjie Hu, Steve Y. Cho, Tyler Bradshaw
cs.CL, cs.AI, cs.LG
With the growing use of transformer-based language models in medicine, it is unclear how well these models generalize to nuclear medicine which has domain-specific vocabulary and unique reporting styles. In this study, we evaluated the value of domain adaptation in nuclear medicine by adapting language models for the p...
http://arxiv.org/pdf/2303.01258v1
2023-03-01T09:48:39
2023-03-01T09:48:39
{ "title": [ -0.02909851, 0.011642456, -0.043182373, -0.039642334, 0.0010232925, -0.0018863678, -0.040740967, -0.0137786865, -0.011909485, 0.023605347, 0.010887146, -0.026916504, -0.0021781921, -0.035186768, -0.03427124, 0.012046814, -0.016220093, ...
http://arxiv.org/abs/2303.00344v1
Inline Citation Classification using Peripheral Context and Time-evolving Augmentation
Priyanshi Gupta, Yash Kumar Atri, Apurva Nagvenkar, Sourish Dasgupta, Tanmoy Chakraborty
cs.CL
Citation plays a pivotal role in determining the associations among research articles. It portrays essential information in indicative, supportive, or contrastive studies. The task of inline citation classification aids in extrapolating these relationships; However, existing studies are still immature and demand furthe...
http://arxiv.org/pdf/2303.00344v1
2023-03-01T09:11:07
2023-03-01T09:11:07
{ "title": [ 0.01663208, 0.018249512, -0.048950195, -0.040527344, 0.019943237, -0.015991211, 0.00049734116, -0.018203735, -0.007972717, 0.035980225, -0.020568848, 0.011749268, -0.021011353, -0.015281677, -0.036071777, -0.023544312, -0.017486572, ...
http://arxiv.org/abs/2303.00333v3
Competence-Based Analysis of Language Models
Adam Davies, Jize Jiang, ChengXiang Zhai
cs.CL
Despite the recent success of large, pretrained neural language models (LLMs) on a variety of prompting tasks, these models can be alarmingly brittle to small changes in inputs or application contexts. To better understand such behavior and motivate the design of more robust LLMs, we provide a causal formulation of lin...
http://arxiv.org/pdf/2303.00333v3
2023-11-07T02:33:27
2023-03-01T08:53:36
{ "title": [ -0.01966858, 0.0035953522, -0.043823242, -0.064331055, 0.008201599, -0.002418518, -0.059265137, 0.0011472702, -0.0051612854, 0.05050659, -0.00031113625, -0.008972168, 0.020812988, -0.019317627, -0.05709839, 0.022994995, -0.008399963, ...
http://arxiv.org/abs/2303.04229v1
Understanding Natural Language Understanding Systems. A Critical Analysis
Alessandro Lenci
cs.AI, cs.CL
The development of machines that {\guillemotleft}talk like us{\guillemotright}, also known as Natural Language Understanding (NLU) systems, is the Holy Grail of Artificial Intelligence (AI), since language is the quintessence of human intelligence. The brief but intense life of NLU research in AI and Natural Language P...
http://arxiv.org/pdf/2303.04229v1
2023-03-01T08:32:55
2023-03-01T08:32:55
{ "title": [ -0.03152466, -0.010971069, -0.020645142, -0.062561035, 0.008735657, 0.018173218, -0.03262329, -0.0036125183, 0.0044174194, 0.011756897, -0.042877197, -0.008224487, -0.011657715, -0.0074005127, -0.055603027, -0.016036987, -0.022338867, ...
http://arxiv.org/abs/2303.00311v1
Modeling Multiple User Interests using Hierarchical Knowledge for Conversational Recommender System
Yuka Okuda, Katsuhito Sudoh, Seitaro Shinagawa, Satoshi Nakamura
cs.CL, cs.AI, cs.IR
A conversational recommender system (CRS) is a practical application for item recommendation through natural language conversation. Such a system estimates user interests for appropriate personalized recommendations. Users sometimes have various interests in different categories or genres, but existing studies assume a...
http://arxiv.org/pdf/2303.00311v1
2023-03-01T08:15:48
2023-03-01T08:15:48
{ "title": [ 0.02078247, 0.007259369, -0.06378174, -0.07672119, 0.022720337, -0.024032593, -0.024169922, 0.012191772, 0.008033752, 0.046539307, -0.014335632, 0.0184021, 0.0053977966, -0.010406494, -0.059906006, 0.006755829, -0.016159058, 0.019577...
http://arxiv.org/abs/2303.00293v1
How Robust is GPT-3.5 to Predecessors? A Comprehensive Study on Language Understanding Tasks
Xuanting Chen, Junjie Ye, Can Zu, Nuo Xu, Rui Zheng, Minlong Peng, Jie Zhou, Tao Gui, Qi Zhang, Xuanjing Huang
cs.CL, 68-06, I.2
The GPT-3.5 models have demonstrated impressive performance in various Natural Language Processing (NLP) tasks, showcasing their strong understanding and reasoning capabilities. However, their robustness and abilities to handle various complexities of the open world have yet to be explored, which is especially crucial ...
http://arxiv.org/pdf/2303.00293v1
2023-03-01T07:39:01
2023-03-01T07:39:01
{ "title": [ 0.006286621, 0.012107849, -0.019348145, -0.05215454, 0.010398865, -0.010421753, -0.04953003, -0.03363037, -0.0006957054, 0.007801056, -0.02720642, 0.009674072, -0.006462097, -0.014595032, -0.031036377, 0.006587982, -0.017807007, 0.03...
http://arxiv.org/abs/2303.00279v1
Coarse-to-Fine Covid-19 Segmentation via Vision-Language Alignment
Dandan Shan, Zihan Li, Wentao Chen, Qingde Li, Jie Tian, Qingqi Hong
eess.IV, cs.CL, cs.CV, cs.IR
Segmentation of COVID-19 lesions can assist physicians in better diagnosis and treatment of COVID-19. However, there are few relevant studies due to the lack of detailed information and high-quality annotation in the COVID-19 dataset. To solve the above problem, we propose C2FVL, a Coarse-to-Fine segmentation framework...
http://arxiv.org/pdf/2303.00279v1
2023-03-01T07:01:29
2023-03-01T07:01:29
{ "title": [ -0.000875473, -0.0040130615, -0.056640625, -0.07299805, 0.00315094, -0.014007568, -0.036712646, 0.0075683594, -0.029785156, 0.03250122, -0.03781128, 0.0011539459, 0.014953613, -0.022644043, -0.010269165, -0.018005371, 0.015731812, 0....
http://arxiv.org/abs/2303.00257v1
Hidden Markov Transformer for Simultaneous Machine Translation
Shaolei Zhang, Yang Feng
cs.CL
Simultaneous machine translation (SiMT) outputs the target sequence while receiving the source sequence, and hence learning when to start translating each target token is the core challenge for SiMT task. However, it is non-trivial to learn the optimal moment among many possible moments of starting translating, as the ...
http://arxiv.org/pdf/2303.00257v1
2023-03-01T06:28:33
2023-03-01T06:28:33
{ "title": [ -0.036590576, -0.007385254, -0.018341064, -0.059173584, 0.005203247, 0.01374054, -0.036376953, -0.0070610046, -0.01058197, 0.052459717, 0.0039634705, 0.00844574, 0.021438599, -0.019348145, -0.039886475, 0.025802612, -0.00233078, 0.03...
http://arxiv.org/abs/2303.01248v3
Can ChatGPT Assess Human Personalities? A General Evaluation Framework
Haocong Rao, Cyril Leung, Chunyan Miao
cs.CL, cs.AI
Large Language Models (LLMs) especially ChatGPT have produced impressive results in various areas, but their potential human-like psychology is still largely unexplored. Existing works study the virtual personalities of LLMs but rarely explore the possibility of analyzing human personalities via LLMs. This paper presen...
http://arxiv.org/pdf/2303.01248v3
2023-10-13T15:53:00
2023-03-01T06:16:14
{ "title": [ -0.013771057, -0.004951477, -0.01676941, -0.009223938, 0.0039405823, -0.0063476562, -0.00039100647, -0.0151901245, -0.020629883, 0.030883789, -0.0028362274, 0.014343262, -0.0045928955, -0.022079468, -0.030654907, 0.01108551, -0.026870728...
http://arxiv.org/abs/2303.01234v1
Frauds Bargain Attack: Generating Adversarial Text Samples via Word Manipulation Process
Mingze Ni, Zhensu Sun, Wei Liu
cs.CL, cs.AI
Recent studies on adversarial examples expose vulnerabilities of natural language processing (NLP) models. Existing techniques for generating adversarial examples are typically driven by deterministic heuristic rules that are agnostic to the optimal adversarial examples, a strategy that often results in attack failures...
http://arxiv.org/pdf/2303.01234v1
2023-03-01T06:04:25
2023-03-01T06:04:25
{ "title": [ -0.014915466, -0.008743286, -0.0128479, -0.02017212, -0.011543274, -0.003200531, -0.0069236755, -0.019851685, -0.06137085, 0.03643799, -0.0044403076, 0.046905518, 0.0058021545, -0.018356323, -0.014045715, -0.0113220215, -0.035247803, ...
http://arxiv.org/abs/2303.00242v1
DIFFQG: Generating Questions to Summarize Factual Changes
Jeremy R. Cole, Palak Jain, Julian Martin Eisenschlos, Michael J. Q. Zhang, Eunsol Choi, Bhuwan Dhingra
cs.CL
Identifying the difference between two versions of the same article is useful to update knowledge bases and to understand how articles evolve. Paired texts occur naturally in diverse situations: reporters write similar news stories and maintainers of authoritative websites must keep their information up to date. We pro...
http://arxiv.org/pdf/2303.00242v1
2023-03-01T05:45:48
2023-03-01T05:45:48
{ "title": [ -0.0071487427, -0.010772705, -0.054107666, -0.07495117, -0.011177063, -0.0076141357, -0.039794922, -0.0034389496, 0.019454956, 0.030181885, -0.017166138, 0.013420105, 0.003036499, -0.01374054, -0.04638672, 0.015625, 0.014541626, 0.02...
http://arxiv.org/abs/2303.01229v2
Almanac: Retrieval-Augmented Language Models for Clinical Medicine
Cyril Zakka, Akash Chaurasia, Rohan Shad, Alex R. Dalal, Jennifer L. Kim, Michael Moor, Kevin Alexander, Euan Ashley, Jack Boyd, Kathleen Boyd, Karen Hirsch, Curt Langlotz, Joanna Nelson, William Hiesinger
cs.CL, cs.AI
Large-language models have recently demonstrated impressive zero-shot capabilities in a variety of natural language tasks such as summarization, dialogue generation, and question-answering. Despite many promising applications in clinical medicine, adoption of these models in real-world settings has been largely limited...
http://arxiv.org/pdf/2303.01229v2
2023-05-31T21:17:13
2023-03-01T02:30:11
{ "title": [ -0.015052795, -0.012069702, -0.031280518, -0.066101074, 0.0004427433, 0.017105103, -0.054626465, -0.0044517517, 0.00031352043, 0.060394287, 0.0014076233, -0.008018494, 0.030303955, -0.023361206, -0.06011963, 0.023117065, 0.0049095154, ...
http://arxiv.org/abs/2303.00135v1
Deep learning for COVID-19 topic modelling via Twitter: Alpha, Delta and Omicron
Janhavi Lande, Arti Pillay, Rohitash Chandra
cs.LG, cs.CL, cs.IR
Topic modelling with innovative deep learning methods has gained interest for a wide range of applications that includes COVID-19. Topic modelling can provide, psychological, social and cultural insights for understanding human behaviour in extreme events such as the COVID-19 pandemic. In this paper, we use prominent d...
http://arxiv.org/pdf/2303.00135v1
2023-02-28T23:40:41
2023-02-28T23:40:41
{ "title": [ -0.010025024, 0.0035896301, -0.04724121, -0.06329346, 0.0059547424, 0.0014066696, -0.04171753, -0.007598877, -0.018371582, 0.028640747, -0.039855957, 0.04425049, -0.0019454956, -0.017608643, -0.0064849854, -0.0064926147, -0.0043029785, ...
http://arxiv.org/abs/2303.01241v1
PANACEA: An Automated Misinformation Detection System on COVID-19
Runcong Zhao, Miguel Arana-Catania, Lixing Zhu, Elena Kochkina, Lin Gui, Arkaitz Zubiaga, Rob Procter, Maria Liakata, Yulan He
cs.CL, cs.LG
In this demo, we introduce a web-based misinformation detection system PANACEA on COVID-19 related claims, which has two modules, fact-checking and rumour detection. Our fact-checking module, which is supported by novel natural language inference methods with a self-attention network, outperforms state-of-the-art appro...
http://arxiv.org/pdf/2303.01241v1
2023-02-28T21:53:48
2023-02-28T21:53:48
{ "title": [ 0.0073432922, 0.008598328, -0.06781006, -0.058776855, 0.024368286, -0.028121948, -0.0007638931, 0.0010566711, -0.021560669, 0.013008118, -0.020858765, -0.0015258789, 0.03427124, -0.036254883, 0.0044517517, -0.016098022, -0.0014562607, ...
http://arxiv.org/abs/2303.00077v1
Beyond the limitations of any imaginable mechanism: large language models and psycholinguistics
Conor Houghton, Nina Kazanina, Priyanka Sukumaran
cs.CL, cs.AI
Large language models are not detailed models of human linguistic processing. They are, however, extremely successful at their primary task: providing a model for language. For this reason and because there are no animal models for language, large language models are important in psycholinguistics: they are useful as a...
http://arxiv.org/pdf/2303.00077v1
2023-02-28T20:49:38
2023-02-28T20:49:38
{ "title": [ -0.019439697, -0.0103302, -0.029418945, -0.048461914, -0.0023593903, -0.0063552856, -0.028366089, -0.004676819, -0.010597229, 0.023788452, -0.00856781, -0.0049552917, 0.001953125, -0.03036499, -0.03253174, 0.00894928, -0.006427765, 0...
http://arxiv.org/abs/2303.00069v1
ClArTTS: An Open-Source Classical Arabic Text-to-Speech Corpus
Ajinkya Kulkarni, Atharva Kulkarni, Sara Abedalmonem Mohammad Shatnawi, Hanan Aldarmaki
cs.CL, cs.SD, eess.AS, none
At present, Text-to-speech (TTS) systems that are trained with high-quality transcribed speech data using end-to-end neural models can generate speech that is intelligible, natural, and closely resembles human speech. These models are trained with relatively large single-speaker professionally recorded audio, typically...
http://arxiv.org/pdf/2303.00069v1
2023-02-28T20:18:59
2023-02-28T20:18:59
{ "title": [ -0.020690918, 0.002910614, -0.061676025, -0.06976318, 0.002462387, -0.005794525, -0.04135132, 0.015930176, 0.0019798279, 0.035247803, 0.0008945465, -0.0033817291, 0.0013008118, -0.04321289, -0.04232788, -0.0069236755, 0.015731812, 0....
http://arxiv.org/abs/2302.14838v2
EvoPrompting: Language Models for Code-Level Neural Architecture Search
Angelica Chen, David M. Dohan, David R. So
cs.NE, cs.AI, cs.CL, cs.LG
Given the recent impressive accomplishments of language models (LMs) for code generation, we explore the use of LMs as adaptive mutation and crossover operators for an evolutionary neural architecture search (NAS) algorithm. While NAS still proves too difficult a task for LMs to succeed at solely through prompting, we ...
http://arxiv.org/pdf/2302.14838v2
2023-10-02T14:26:42
2023-02-28T18:37:25
{ "title": [ -0.022567749, -0.019714355, -0.009925842, -0.03149414, -0.007423401, 0.0015287399, -0.022247314, -0.014289856, 0.0019950867, 0.031082153, -0.010772705, -0.003074646, -0.0053710938, -0.021759033, -0.052246094, 0.0068588257, -0.012214661, ...
http://arxiv.org/abs/2302.14828v1
Automatic Scoring of Dream Reports' Emotional Content with Large Language Models
Lorenzo Bertolini, Valentina Elce, Adriana Michalak, Giulio Bernardi, Julie Weeds
cs.CL
In the field of dream research, the study of dream content typically relies on the analysis of verbal reports provided by dreamers upon awakening from their sleep. This task is classically performed through manual scoring provided by trained annotators, at a great time expense. While a consistent body of work suggests ...
http://arxiv.org/pdf/2302.14828v1
2023-02-28T18:23:17
2023-02-28T18:23:17
{ "title": [ -0.0368042, -0.015808105, -0.02911377, -0.01828003, -0.02609253, -0.00040245056, -0.031280518, -0.015342712, 0.01373291, 0.030715942, 0.006500244, -0.015899658, 0.021026611, -0.03878784, -0.040802002, 0.019195557, 0.0037879944, 0.012...
http://arxiv.org/abs/2302.14785v1
Joint Representations of Text and Knowledge Graphs for Retrieval and Evaluation
Teven Le Scao, Claire Gardent
cs.CL
A key feature of neural models is that they can produce semantic vector representations of objects (texts, images, speech, etc.) ensuring that similar objects are close to each other in the vector space. While much work has focused on learning representations for other modalities, there are no aligned cross-modal repre...
http://arxiv.org/pdf/2302.14785v1
2023-02-28T17:39:43
2023-02-28T17:39:43
{ "title": [ -0.024291992, -0.0021419525, -0.04812622, -0.062408447, 0.013046265, 0.0017929077, -0.03314209, -0.0038394928, 0.016143799, 0.034362793, -0.025894165, 0.010383606, -0.015350342, -0.0011310577, -0.06640625, -0.0060539246, -0.018249512, ...
http://arxiv.org/abs/2302.14727v1
Automatically Classifying Emotions based on Text: A Comparative Exploration of Different Datasets
Anna Koufakou, Jairo Garciga, Adam Paul, Joseph Morelli, Christopher Frank
cs.CL
Emotion Classification based on text is a task with many applications which has received growing interest in recent years. This paper presents a preliminary study with the goal to help researchers and practitioners gain insight into relatively new datasets as well as emotion classification in general. We focus on three...
http://arxiv.org/pdf/2302.14727v1
2023-02-28T16:34:55
2023-02-28T16:34:55
{ "title": [ -0.01914978, 0.0050468445, -0.03503418, -0.027938843, 0.0138549805, -0.009681702, -0.01007843, -0.022399902, 0.028747559, 0.023910522, -0.027633667, 0.0046653748, 0.00044965744, -0.014816284, -0.06585693, -0.0043182373, -0.012458801, ...
http://arxiv.org/abs/2302.14719v1
Self-training through Classifier Disagreement for Cross-Domain Opinion Target Extraction
Kai Sun, Richong Zhang, Samuel Mensah, Nikolaos Aletras, Yongyi Mao, Xudong Liu
cs.CL, cs.LG
Opinion target extraction (OTE) or aspect extraction (AE) is a fundamental task in opinion mining that aims to extract the targets (or aspects) on which opinions have been expressed. Recent work focus on cross-domain OTE, which is typically encountered in real-world scenarios, where the testing and training distributio...
http://arxiv.org/pdf/2302.14719v1
2023-02-28T16:31:17
2023-02-28T16:31:17
{ "title": [ 0.009757996, 0.016494751, -0.02571106, -0.025497437, 0.015365601, -0.032287598, 0.01486969, -0.024612427, -0.00422287, 0.013633728, -0.017028809, -0.00044345856, -0.040863037, -0.027053833, -0.029937744, -0.020767212, -0.020584106, 0...
http://arxiv.org/abs/2302.14708v1
Is Japanese CCGBank empirically correct? A case study of passive and causative constructions
Daisuke Bekki, Hitomi Yanaka
cs.CL
The Japanese CCGBank serves as training and evaluation data for developing Japanese CCG parsers. However, since it is automatically generated from the Kyoto Corpus, a dependency treebank, its linguistic validity still needs to be sufficiently verified. In this paper, we focus on the analysis of passive/causative constr...
http://arxiv.org/pdf/2302.14708v1
2023-02-28T16:19:24
2023-02-28T16:19:24
{ "title": [ -0.028823853, -0.0035076141, -0.013511658, -0.053344727, -0.008331299, 0.012252808, -0.066223145, -0.005619049, -0.022903442, 0.03225708, -0.014190674, -0.01486969, 0.010940552, -0.020462036, -0.043518066, -0.021591187, -0.034606934, ...
http://arxiv.org/abs/2302.14691v1
In-Context Instruction Learning
Seonghyeon Ye, Hyeonbin Hwang, Sohee Yang, Hyeongu Yun, Yireun Kim, Minjoon Seo
cs.CL, cs.AI
Instruction learning of Large Language Models (LLMs) has enabled zero-shot task generalization. However, instruction learning has been predominantly approached as a fine-tuning problem, including instruction tuning and reinforcement learning from human feedback, where LLMs are multi-task fine-tuned on various tasks wit...
http://arxiv.org/pdf/2302.14691v1
2023-02-28T16:06:35
2023-02-28T16:06:35
{ "title": [ 0.026733398, 0.0020599365, -0.023849487, -0.0073661804, -0.024276733, -0.052886963, 0.00868988, -0.017166138, -0.00945282, 0.04650879, -0.054626465, 0.010398865, -0.016159058, -0.011108398, 0.011146545, -0.014701843, 0.026367188, 0.0...
http://arxiv.org/abs/2302.14680v2
Which One Are You Referring To? Multimodal Object Identification in Situated Dialogue
Holy Lovenia, Samuel Cahyawijaya, Pascale Fung
cs.CL, cs.AI, cs.CV
The demand for multimodal dialogue systems has been rising in various domains, emphasizing the importance of interpreting multimodal inputs from conversational and situational contexts. We explore three methods to tackle this problem and evaluate them on the largest situated dialogue dataset, SIMMC 2.1. Our best method...
http://arxiv.org/pdf/2302.14680v2
2023-03-15T14:38:21
2023-02-28T15:45:20
{ "title": [ 0.00070238113, 0.0035705566, -0.032836914, -0.03881836, 0.0011835098, -0.024642944, -0.018753052, -0.00674057, -0.034576416, 0.031097412, -0.0026359558, 0.007896423, -0.011291504, -0.03262329, -0.0026340485, -0.027923584, -0.01461792, ...
http://arxiv.org/abs/2302.14679v2
Synthesizing Mixed-type Electronic Health Records using Diffusion Models
Taha Ceritli, Ghadeer O. Ghosheh, Vinod Kumar Chauhan, Tingting Zhu, Andrew P. Creagh, David A. Clifton
cs.LG, cs.CL
Electronic Health Records (EHRs) contain sensitive patient information, which presents privacy concerns when sharing such data. Synthetic data generation is a promising solution to mitigate these risks, often relying on deep generative models such as Generative Adversarial Networks (GANs). However, recent studies have ...
http://arxiv.org/pdf/2302.14679v2
2023-08-10T16:46:35
2023-02-28T15:42:30
{ "title": [ -0.021743774, 0.013198853, -0.064453125, -0.076293945, 0.0317688, 0.0019445419, -0.044311523, 0.009666443, -0.051727295, 0.03173828, -0.0368042, 0.028045654, -0.03543091, -0.010971069, -0.015014648, -0.025009155, 0.0043296814, 0.0107...
http://arxiv.org/abs/2302.14635v1
H-AES: Towards Automated Essay Scoring for Hindi
Shubhankar Singh, Anirudh Pupneja, Shivaansh Mital, Cheril Shah, Manish Bawkar, Lakshman Prasad Gupta, Ajit Kumar, Yaman Kumar, Rushali Gupta, Rajiv Ratn Shah
cs.CL
The use of Natural Language Processing (NLP) for Automated Essay Scoring (AES) has been well explored in the English language, with benchmark models exhibiting performance comparable to human scorers. However, AES in Hindi and other low-resource languages remains unexplored. In this study, we reproduce and compare stat...
http://arxiv.org/pdf/2302.14635v1
2023-02-28T15:14:15
2023-02-28T15:14:15
{ "title": [ -0.029724121, 0.0011615753, -0.014183044, -0.016204834, -0.009170532, 0.016143799, -0.010276794, -0.020935059, 0.009391785, 0.015304565, -0.025680542, -0.0017766953, -0.016311646, -0.018798828, -0.051513672, -0.020050049, -0.026565552, ...
http://arxiv.org/abs/2302.14624v1
The 2022 NIST Language Recognition Evaluation
Yooyoung Lee, Craig Greenberg, Eliot Godard, Asad A. Butt, Elliot Singer, Trang Nguyen, Lisa Mason, Douglas Reynolds
cs.CL, cs.LG, cs.SD, eess.AS
In 2022, the U.S. National Institute of Standards and Technology (NIST) conducted the latest Language Recognition Evaluation (LRE) in an ongoing series administered by NIST since 1996 to foster research in language recognition and to measure state-of-the-art technology. Similar to previous LREs, LRE22 focused on conver...
http://arxiv.org/pdf/2302.14624v1
2023-02-28T15:05:33
2023-02-28T15:05:33
{ "title": [ -0.026397705, 0.021560669, -0.035888672, -0.03778076, -0.015312195, 0.0004515648, -0.07165527, -0.0149002075, 0.008003235, 0.025772095, -0.0023536682, -0.026153564, 0.01386261, -0.062805176, -0.008369446, 0.023071289, -0.0021514893, ...
http://arxiv.org/abs/2303.01249v1
Language-Universal Adapter Learning with Knowledge Distillation for End-to-End Multilingual Speech Recognition
Zhijie Shen, Wu Guo, Bin Gu
cs.CL, cs.SD, eess.AS
In this paper, we propose a language-universal adapter learning framework based on a pre-trained model for end-to-end multilingual automatic speech recognition (ASR). For acoustic modeling, the wav2vec 2.0 pre-trained model is fine-tuned by inserting language-specific and language-universal adapters. An online knowledg...
http://arxiv.org/pdf/2303.01249v1
2023-02-28T14:43:49
2023-02-28T14:43:49
{ "title": [ -0.009483337, 0.002779007, -0.057678223, -0.09527588, 0.010513306, 0.0025959015, -0.07788086, 0.033203125, 0.011520386, 0.031921387, -0.022720337, 0.0003080368, -0.0037651062, -0.019348145, -0.029220581, -0.0012931824, -0.0013980865, ...
http://arxiv.org/abs/2302.14534v1
Spacerini: Plug-and-play Search Engines with Pyserini and Hugging Face
Christopher Akiki, Odunayo Ogundepo, Aleksandra Piktus, Xinyu Zhang, Akintunde Oladipo, Jimmy Lin, Martin Potthast
cs.IR, cs.CL
We present Spacerini, a modular framework for seamless building and deployment of interactive search applications, designed to facilitate the qualitative analysis of large scale research datasets. Spacerini integrates features from both the Pyserini toolkit and the Hugging Face ecosystem to ease the indexing text colle...
http://arxiv.org/pdf/2302.14534v1
2023-02-28T12:44:10
2023-02-28T12:44:10
{ "title": [ 0.0018825531, -0.026657104, -0.034973145, 0.016052246, -0.027801514, -0.036712646, -0.00090026855, -0.018707275, -0.027023315, -0.002023697, -0.04473877, -0.020935059, -0.006046295, -0.047851562, 0.0118637085, -0.052337646, 0.026687622, ...
http://arxiv.org/abs/2302.14523v1
Automatic Heteronym Resolution Pipeline Using RAD-TTS Aligners
Jocelyn Huang, Evelina Bakhturina, Oktai Tatanov
cs.CL
Grapheme-to-phoneme (G2P) transduction is part of the standard text-to-speech (TTS) pipeline. However, G2P conversion is difficult for languages that contain heteronyms -- words that have one spelling but can be pronounced in multiple ways. G2P datasets with annotated heteronyms are limited in size and expensive to cre...
http://arxiv.org/pdf/2302.14523v1
2023-02-28T12:33:12
2023-02-28T12:33:12
{ "title": [ -0.010139465, 0.008117676, -0.03363037, -0.07745361, 0.0012044907, 0.013572693, -0.059661865, -0.017913818, 0.010757446, 0.05154419, -0.0011510849, -0.029754639, 0.060546875, -0.057495117, -0.022125244, 0.012031555, 0.0026950836, 0.0...
http://arxiv.org/abs/2302.14520v2
Large Language Models Are State-of-the-Art Evaluators of Translation Quality
Tom Kocmi, Christian Federmann
cs.CL
We describe GEMBA, a GPT-based metric for assessment of translation quality, which works both with a reference translation and without. In our evaluation, we focus on zero-shot prompting, comparing four prompt variants in two modes, based on the availability of the reference. We investigate nine versions of GPT models,...
http://arxiv.org/pdf/2302.14520v2
2023-05-31T19:51:51
2023-02-28T12:23:48
{ "title": [ 0.0014295578, -0.014198303, 0.0059547424, -0.0602417, -0.06286621, 0.002904892, -0.05984497, -0.00034332275, 0.011062622, 0.031555176, -0.00907135, -0.0099105835, 0.012550354, -0.045715332, -0.008331299, 0.0022563934, 0.0016908646, -...
http://arxiv.org/abs/2302.14502v1
A Survey on Long Text Modeling with Transformers
Zican Dong, Tianyi Tang, Lunyi Li, Wayne Xin Zhao
cs.CL
Modeling long texts has been an essential technique in the field of natural language processing (NLP). With the ever-growing number of long documents, it is important to develop effective modeling methods that can process and analyze such texts. However, long texts pose important research challenges for existing text m...
http://arxiv.org/pdf/2302.14502v1
2023-02-28T11:34:30
2023-02-28T11:34:30
{ "title": [ -0.018981934, 0.013092041, -0.004875183, -0.07281494, -0.0032100677, 0.026977539, -0.07513428, -0.013442993, -0.00819397, 0.039978027, -0.031433105, 0.0044670105, 0.012336731, -0.008552551, -0.04321289, -0.004714966, -0.02470398, 0.0...
http://arxiv.org/abs/2302.14494v1
Text classification dataset and analysis for Uzbek language
Elmurod Kuriyozov, Ulugbek Salaev, Sanatbek Matlatipov, Gayrat Matlatipov
cs.CL
Text classification is an important task in Natural Language Processing (NLP), where the goal is to categorize text data into predefined classes. In this study, we analyse the dataset creation steps and evaluation techniques of multi-label news categorisation task as part of text classification. We first present a newl...
http://arxiv.org/pdf/2302.14494v1
2023-02-28T11:21:24
2023-02-28T11:21:24
{ "title": [ 0.002111435, 0.025268555, -0.035949707, -0.028305054, 0.013427734, -0.036254883, -0.018096924, -0.023269653, -0.024810791, 0.066101074, -0.029541016, -0.0070343018, 0.058532715, -0.021408081, -0.021484375, -0.013893127, 0.003932953, ...
http://arxiv.org/abs/2302.14413v1
SMoA: Sparse Mixture of Adapters to Mitigate Multiple Dataset Biases
Yanchen Liu, Jing Yan, Yan Chen, Jing Liu, Hua Wu
cs.CL
Recent studies reveal that various biases exist in different NLP tasks, and over-reliance on biases results in models' poor generalization ability and low adversarial robustness. To mitigate datasets biases, previous works propose lots of debiasing techniques to tackle specific biases, which perform well on respective ...
http://arxiv.org/pdf/2302.14413v1
2023-02-28T08:47:20
2023-02-28T08:47:20
{ "title": [ -0.025314331, -0.0032310486, -0.07446289, -0.059051514, 0.020401001, -0.016708374, -0.02923584, -0.005771637, 0.019500732, 0.032592773, -0.00051784515, 0.00907135, 0.0015478134, -0.026382446, -0.040740967, 0.02168274, 0.008529663, 0....
http://arxiv.org/abs/2302.14406v1
Instruction Clarification Requests in Multimodal Collaborative Dialogue Games: Tasks, and an Analysis of the CoDraw Dataset
Brielen Madureira, David Schlangen
cs.CL
In visual instruction-following dialogue games, players can engage in repair mechanisms in face of an ambiguous or underspecified instruction that cannot be fully mapped to actions in the world. In this work, we annotate Instruction Clarification Requests (iCRs) in CoDraw, an existing dataset of interactions in a multi...
http://arxiv.org/pdf/2302.14406v1
2023-02-28T08:41:53
2023-02-28T08:41:53
{ "title": [ -0.03881836, -0.0043792725, -0.05218506, 0.011062622, -0.01474762, -0.027297974, -0.019683838, -0.008361816, -0.010604858, 0.017166138, 0.0066452026, 0.015029907, -0.042633057, -0.035339355, -0.029937744, -0.016906738, -0.036193848, ...
http://arxiv.org/abs/2302.14401v1
GLM-Dialog: Noise-tolerant Pre-training for Knowledge-grounded Dialogue Generation
Jing Zhang, Xiaokang Zhang, Daniel Zhang-Li, Jifan Yu, Zijun Yao, Zeyao Ma, Yiqi Xu, Haohua Wang, Xiaohan Zhang, Nianyi Lin, Sunrui Lu, Juanzi Li, Jie Tang
cs.CL, cs.AI
We present GLM-Dialog, a large-scale language model (LLM) with 10B parameters capable of knowledge-grounded conversation in Chinese using a search engine to access the Internet knowledge. GLM-Dialog offers a series of applicable techniques for exploiting various external knowledge including both helpful and noisy knowl...
http://arxiv.org/pdf/2302.14401v1
2023-02-28T08:35:28
2023-02-28T08:35:28
{ "title": [ -0.011756897, -0.012313843, -0.0579834, -0.09649658, 0.002714157, -0.012466431, -0.066101074, 0.014595032, 0.0041046143, 0.024963379, 0.0036621094, 0.014053345, -0.006580353, -0.022735596, -0.048309326, 0.0026855469, -0.014175415, 0....
http://arxiv.org/abs/2302.14389v1
Information-Restricted Neural Language Models Reveal Different Brain Regions' Sensitivity to Semantics, Syntax and Context
Alexandre Pasquiou, Yair Lakretz, Bertrand Thirion, Christophe Pallier
cs.CL
A fundamental question in neurolinguistics concerns the brain regions involved in syntactic and semantic processing during speech comprehension, both at the lexical (word processing) and supra-lexical levels (sentence and discourse processing). To what extent are these regions separated or intertwined? To address this ...
http://arxiv.org/pdf/2302.14389v1
2023-02-28T08:16:18
2023-02-28T08:16:18
{ "title": [ -0.023147583, 0.011581421, -0.02798462, 0.0001373291, -0.022415161, -0.029144287, -0.025726318, -0.03137207, -0.014419556, 0.041381836, 0.00016832352, -0.0078125, -0.002275467, -0.029876709, -0.045715332, 0.00793457, 0.0063323975, 0....
http://arxiv.org/abs/2302.14383v2
Linear Spaces of Meanings: Compositional Structures in Vision-Language Models
Matthew Trager, Pramuditha Perera, Luca Zancato, Alessandro Achille, Parminder Bhatia, Stefano Soatto
cs.LG, cs.AI, cs.CL, cs.CV
We investigate compositional structures in data embeddings from pre-trained vision-language models (VLMs). Traditionally, compositionality has been associated with algebraic operations on embeddings of words from a pre-existing vocabulary. In contrast, we seek to approximate representations from an encoder as combinati...
http://arxiv.org/pdf/2302.14383v2
2023-03-27T01:02:17
2023-02-28T08:11:56
{ "title": [ -0.027374268, -0.01625061, -0.019744873, -0.041534424, -0.01966858, -0.0026073456, -0.049224854, -0.01600647, 0.008880615, 0.03564453, -0.015052795, 0.0008740425, 0.00415802, -0.014732361, -0.06121826, -0.0037708282, 0.0035572052, 0....
http://arxiv.org/abs/2302.14337v2
UniFLG: Unified Facial Landmark Generator from Text or Speech
Kentaro Mitsui, Yukiya Hono, Kei Sawada
cs.CV, cs.CL, cs.SD, eess.AS, eess.IV
Talking face generation has been extensively investigated owing to its wide applicability. The two primary frameworks used for talking face generation comprise a text-driven framework, which generates synchronized speech and talking faces from text, and a speech-driven framework, which generates talking faces from spee...
http://arxiv.org/pdf/2302.14337v2
2023-05-19T02:43:32
2023-02-28T06:05:43
{ "title": [ -0.017486572, -0.010879517, -0.05041504, -0.060455322, 0.004573822, -0.012069702, -0.023239136, 0.009407043, -0.00774765, 0.04815674, -0.0065689087, 0.019485474, 0.034179688, -0.030136108, -0.030075073, 0.022247314, 0.014823914, 0.01...
http://arxiv.org/abs/2303.07201v1
An evaluation of Google Translate for Sanskrit to English translation via sentiment and semantic analysis
Akshat Shukla, Chaarvi Bansal, Sushrut Badhe, Mukul Ranjan, Rohitash Chandra
cs.CL, cs.AI
Google Translate has been prominent for language translation; however, limited work has been done in evaluating the quality of translation when compared to human experts. Sanskrit one of the oldest written languages in the world. In 2022, the Sanskrit language was added to the Google Translate engine. Sanskrit is known...
http://arxiv.org/pdf/2303.07201v1
2023-02-28T04:24:55
2023-02-28T04:24:55
{ "title": [ -0.017578125, 0.0058174133, -0.017089844, -0.013252258, -0.0038604736, -0.0010099411, -0.010986328, -0.031097412, -0.011489868, 0.033996582, -0.025039673, 0.015235901, 0.010025024, -0.016708374, -0.043670654, -0.0030841827, 0.006134033, ...
http://arxiv.org/abs/2302.14286v1
HugNLP: A Unified and Comprehensive Library for Natural Language Processing
Jianing Wang, Nuo Chen, Qiushi Sun, Wenkang Huang, Chengyu Wang, Ming Gao
cs.CL
In this paper, we introduce HugNLP, a unified and comprehensive library for natural language processing (NLP) with the prevalent backend of HuggingFace Transformers, which is designed for NLP researchers to easily utilize off-the-shelf algorithms and develop novel methods with user-defined models and tasks in real-worl...
http://arxiv.org/pdf/2302.14286v1
2023-02-28T03:38:26
2023-02-28T03:38:26
{ "title": [ -0.02909851, -0.021743774, -0.038330078, -0.0635376, 0.023757935, 0.011657715, -0.042053223, -0.009155273, 0.013389587, 0.021835327, -0.014953613, 0.0154418945, 0.0035476685, -0.014411926, -0.036468506, 0.018508911, -0.01789856, 0.04...
http://arxiv.org/abs/2302.14261v1
Augmented Transformers with Adaptive n-grams Embedding for Multilingual Scene Text Recognition
Xueming Yan, Zhihang Fang, Yaochu Jin
cs.CL
While vision transformers have been highly successful in improving the performance in image-based tasks, not much work has been reported on applying transformers to multilingual scene text recognition due to the complexities in the visual appearance of multilingual texts. To fill the gap, this paper proposes an augment...
http://arxiv.org/pdf/2302.14261v1
2023-02-28T02:37:30
2023-02-28T02:37:30
{ "title": [ -0.010444641, -0.0048103333, -0.004798889, -0.052642822, -0.020950317, 0.01876831, -0.052642822, -0.008934021, -0.002298355, 0.04095459, -0.05407715, 0.01890564, -0.0060653687, 0.0045776367, -0.03326416, -0.024765015, -0.0033874512, ...
http://arxiv.org/abs/2302.14233v2
Goal Driven Discovery of Distributional Differences via Language Descriptions
Ruiqi Zhong, Peter Zhang, Steve Li, Jinwoo Ahn, Dan Klein, Jacob Steinhardt
cs.CL, cs.AI, cs.LG
Mining large corpora can generate useful discoveries but is time-consuming for humans. We formulate a new task, D5, that automatically discovers differences between two large corpora in a goal-driven way. The task input is a problem comprising a research goal "$\textit{comparing the side effects of drug A and drug B}$"...
http://arxiv.org/pdf/2302.14233v2
2023-10-25T00:43:04
2023-02-28T01:32:32
{ "title": [ -0.03982544, -0.01184082, -0.05819702, -0.038726807, -0.0030059814, -0.01121521, -0.04763794, -0.0007300377, 0.0003669262, 0.017044067, -0.004837036, 0.011909485, -0.012565613, -0.03643799, -0.020492554, 0.006904602, 0.00038409233, 0...
http://arxiv.org/abs/2302.14229v4
Zero-Shot Cross-Lingual Summarization via Large Language Models
Jiaan Wang, Yunlong Liang, Fandong Meng, Beiqi Zou, Zhixu Li, Jianfeng Qu, Jie Zhou
cs.CL, cs.AI
Given a document in a source language, cross-lingual summarization (CLS) aims to generate a summary in a different target language. Recently, the emergence of Large Language Models (LLMs), such as GPT-3.5, ChatGPT and GPT-4, has attracted wide attention from the computational linguistics community. However, it is not y...
http://arxiv.org/pdf/2302.14229v4
2023-10-24T15:34:47
2023-02-28T01:27:37
{ "title": [ -0.0075263977, -0.0082473755, -0.026641846, -0.08337402, -0.005470276, 0.008796692, -0.07244873, 0.0043258667, 0.022003174, 0.04626465, -0.0077705383, 0.010574341, 0.013305664, -0.0017118454, -0.04547119, 0.025314331, -0.0012960434, ...
http://arxiv.org/abs/2302.14225v2
Weighted Sampling for Masked Language Modeling
Linhan Zhang, Qian Chen, Wen Wang, Chong Deng, Xin Cao, Kongzhang Hao, Yuxin Jiang, Wei Wang
cs.CL, cs.AI
Masked Language Modeling (MLM) is widely used to pretrain language models. The standard random masking strategy in MLM causes the pre-trained language models (PLMs) to be biased toward high-frequency tokens. Representation learning of rare tokens is poor and PLMs have limited performance on downstream tasks. To allevia...
http://arxiv.org/pdf/2302.14225v2
2023-05-24T04:31:24
2023-02-28T01:07:39
{ "title": [ -0.054107666, -0.014419556, -0.064819336, -0.05657959, -0.017623901, -0.00047445297, -0.06604004, 0.014854431, 0.0074424744, 0.026763916, -0.018035889, -0.0062713623, 0.019577026, -0.03390503, -0.02432251, 0.01386261, -0.02268982, 0....
http://arxiv.org/abs/2302.14220v2
Are Character-level Translations Worth the Wait? Comparing Character- and Subword-level Models for Machine Translation
Lukas Edman, Gabriele Sarti, Antonio Toral, Gertjan van Noord, Arianna Bisazza
cs.CL
Pretrained character-level language models were recently shown to be competitive with popular subword models across a range of NLP tasks. However, there has been little research on their effectiveness for neural machine translation (NMT). This work performs an extensive comparison across multiple languages and experime...
http://arxiv.org/pdf/2302.14220v2
2023-05-11T14:00:00
2023-02-28T00:50:19
{ "title": [ -0.0513916, -0.018325806, 0.0046958923, -0.025131226, -0.03125, 0.016723633, -0.04208374, -0.0076026917, 0.0048446655, 0.047790527, -0.018844604, 0.009346008, 0.046844482, -0.014205933, -0.042938232, 0.021728516, -0.0028953552, 0.022...
http://arxiv.org/abs/2303.00001v1
Reward Design with Language Models
Minae Kwon, Sang Michael Xie, Kalesha Bullard, Dorsa Sadigh
cs.LG, cs.AI, cs.CL
Reward design in reinforcement learning (RL) is challenging since specifying human notions of desired behavior may be difficult via reward functions or require many expert demonstrations. Can we instead cheaply design rewards using a natural language interface? This paper explores how to simplify reward design by promp...
http://arxiv.org/pdf/2303.00001v1
2023-02-27T22:09:35
2023-02-27T22:09:35
{ "title": [ -0.02848816, -0.012252808, -0.022979736, -0.017196655, -0.018432617, -0.0059432983, -0.034576416, -0.023086548, -0.016830444, 0.025222778, -0.002500534, -0.009613037, -0.024597168, -0.026306152, -0.03616333, -0.0022392273, -0.022659302, ...
http://arxiv.org/abs/2302.14169v1
TabGenie: A Toolkit for Table-to-Text Generation
Zdeněk Kasner, Ekaterina Garanina, Ondřej Plátek, Ondřej Dušek
cs.CL
Heterogenity of data-to-text generation datasets limits the research on data-to-text generation systems. We present TabGenie - a toolkit which enables researchers to explore, preprocess, and analyze a variety of data-to-text generation datasets through the unified framework of table-to-text generation. In TabGenie, all...
http://arxiv.org/pdf/2302.14169v1
2023-02-27T22:05:46
2023-02-27T22:05:46
{ "title": [ -0.033843994, -0.01171875, -0.04437256, -0.054992676, 0.0012979507, 0.011428833, -0.053985596, -0.022079468, 0.006576538, 0.05633545, -0.013465881, 0.02999878, 0.017425537, 0.003868103, -0.05557251, 0.009384155, 0.0012559891, 0.02008...
http://arxiv.org/abs/2302.14132v1
Structured Pruning of Self-Supervised Pre-trained Models for Speech Recognition and Understanding
Yifan Peng, Kwangyoun Kim, Felix Wu, Prashant Sridhar, Shinji Watanabe
cs.CL, cs.SD, eess.AS
Self-supervised speech representation learning (SSL) has shown to be effective in various downstream tasks, but SSL models are usually large and slow. Model compression techniques such as pruning aim to reduce the model size and computation without degradation in accuracy. Prior studies focus on the pruning of Transfor...
http://arxiv.org/pdf/2302.14132v1
2023-02-27T20:39:54
2023-02-27T20:39:54
{ "title": [ -0.009506226, 0.004802704, -0.025543213, -0.0982666, 0.0009646416, 0.012802124, -0.07611084, 0.021118164, -0.011665344, 0.015312195, -0.024978638, 0.0072746277, -0.012680054, -0.020889282, -0.037902832, -0.016311646, 0.0032043457, 0....
http://arxiv.org/abs/2302.14115v2
Vid2Seq: Large-Scale Pretraining of a Visual Language Model for Dense Video Captioning
Antoine Yang, Arsha Nagrani, Paul Hongsuck Seo, Antoine Miech, Jordi Pont-Tuset, Ivan Laptev, Josef Sivic, Cordelia Schmid
cs.CV, cs.AI, cs.CL, cs.LG
In this work, we introduce Vid2Seq, a multi-modal single-stage dense event captioning model pretrained on narrated videos which are readily-available at scale. The Vid2Seq architecture augments a language model with special time tokens, allowing it to seamlessly predict event boundaries and textual descriptions in the ...
http://arxiv.org/pdf/2302.14115v2
2023-03-21T11:01:09
2023-02-27T19:53:49
{ "title": [ -0.0041770935, -0.0035362244, -0.03112793, -0.04522705, -0.008468628, -0.021942139, -0.028564453, -0.012771606, 0.0043907166, 0.048431396, -0.021072388, 0.015960693, 0.011833191, -0.0072288513, -0.042907715, 0.004714966, -0.0017127991, ...
http://arxiv.org/abs/2302.14045v2
Language Is Not All You Need: Aligning Perception with Language Models
Shaohan Huang, Li Dong, Wenhui Wang, Yaru Hao, Saksham Singhal, Shuming Ma, Tengchao Lv, Lei Cui, Owais Khan Mohammed, Barun Patra, Qiang Liu, Kriti Aggarwal, Zewen Chi, Johan Bjorck, Vishrav Chaudhary, Subhojit Som, Xia Song, Furu Wei
cs.CL, cs.CV
A big convergence of language, multimodal perception, action, and world modeling is a key step toward artificial general intelligence. In this work, we introduce Kosmos-1, a Multimodal Large Language Model (MLLM) that can perceive general modalities, learn in context (i.e., few-shot), and follow instructions (i.e., zer...
http://arxiv.org/pdf/2302.14045v2
2023-03-01T11:04:51
2023-02-27T18:55:27
{ "title": [ -0.024398804, -0.014785767, -0.03363037, -0.03704834, -0.0013885498, -0.009635925, -0.034942627, -0.025344849, 0.00025629997, 0.028366089, -0.019088745, -0.015457153, 0.0006303787, -0.022735596, -0.050231934, 0.002325058, -0.0055503845, ...
http://arxiv.org/abs/2302.14036v2
Text-only domain adaptation for end-to-end ASR using integrated text-to-mel-spectrogram generator
Vladimir Bataev, Roman Korostik, Evgeny Shabalin, Vitaly Lavrukhin, Boris Ginsburg
eess.AS, cs.CL, cs.LG, cs.SD
We propose an end-to-end Automatic Speech Recognition (ASR) system that can be trained on transcribed speech data, text-only data, or a mixture of both. The proposed model uses an integrated auxiliary block for text-based training. This block combines a non-autoregressive multi-speaker text-to-mel-spectrogram generator...
http://arxiv.org/pdf/2302.14036v2
2023-08-16T12:03:47
2023-02-27T18:47:55
{ "title": [ -0.01826477, -0.010116577, -0.05203247, -0.059173584, -0.009788513, -0.0014381409, -0.026947021, 0.023742676, 0.008033752, 0.045043945, 0.01826477, 0.006088257, 0.021362305, -0.036132812, -0.046783447, 0.034210205, -0.009513855, 0.04...
http://arxiv.org/abs/2302.14035v1
The ROOTS Search Tool: Data Transparency for LLMs
Aleksandra Piktus, Christopher Akiki, Paulo Villegas, Hugo Laurençon, Gérard Dupont, Alexandra Sasha Luccioni, Yacine Jernite, Anna Rogers
cs.CL, cs.AI
ROOTS is a 1.6TB multilingual text corpus developed for the training of BLOOM, currently the largest language model explicitly accompanied by commensurate data governance efforts. In continuation of these efforts, we present the ROOTS Search Tool: a search engine over the entire ROOTS corpus offering both fuzzy and exa...
http://arxiv.org/pdf/2302.14035v1
2023-02-27T18:45:18
2023-02-27T18:45:18
{ "title": [ -0.029342651, 0.0038967133, -0.049682617, -0.034454346, -0.02279663, -0.007167816, -0.06329346, -0.018692017, -0.0060768127, 0.016143799, -0.048614502, 0.0048332214, 0.018875122, -0.041900635, -0.0071105957, -0.031555176, -0.0068588257, ...
http://arxiv.org/abs/2302.14030v3
Multimodal Speech Recognition for Language-Guided Embodied Agents
Allen Chang, Xiaoyuan Zhu, Aarav Monga, Seoho Ahn, Tejas Srinivasan, Jesse Thomason
cs.CL, cs.SD, eess.AS
Benchmarks for language-guided embodied agents typically assume text-based instructions, but deployed agents will encounter spoken instructions. While Automatic Speech Recognition (ASR) models can bridge the input gap, erroneous ASR transcripts can hurt the agents' ability to complete tasks. In this work, we propose tr...
http://arxiv.org/pdf/2302.14030v3
2023-10-09T22:19:02
2023-02-27T18:41:48
{ "title": [ -0.03704834, -0.0032691956, -0.035095215, -0.04660034, -0.011329651, 0.0003631115, -0.052642822, 0.0079422, -0.0009202957, 0.033691406, -0.036132812, 0.013465881, 0.013389587, -0.0072517395, -0.036132812, 0.0005264282, -0.009971619, ...
http://arxiv.org/abs/2302.14022v1
Diacritic Recognition Performance in Arabic ASR
Hanan Aldarmaki, Ahmad Ghannam
cs.CL
We present an analysis of diacritic recognition performance in Arabic Automatic Speech Recognition (ASR) systems. As most existing Arabic speech corpora do not contain all diacritical marks, which represent short vowels and other phonetic information in Arabic script, current state-of-the-art ASR models do not produce ...
http://arxiv.org/pdf/2302.14022v1
2023-02-27T18:27:42
2023-02-27T18:27:42
{ "title": [ -0.015602112, -0.0011434555, -0.031021118, -0.04888916, -0.006164551, -0.010231018, -0.011184692, 0.021972656, -0.026138306, 0.043792725, -0.0017385483, -0.022399902, 0.0051879883, -0.042755127, -0.03918457, -0.0032520294, 0.01259613, ...
http://arxiv.org/abs/2302.14021v1
Quantifying Valence and Arousal in Text with Multilingual Pre-trained Transformers
Gonçalo Azevedo Mendes, Bruno Martins
cs.CL, cs.AI, cs.IR
The analysis of emotions expressed in text has numerous applications. In contrast to categorical analysis, focused on classifying emotions according to a pre-defined set of common classes, dimensional approaches can offer a more nuanced way to distinguish between different emotions. Still, dimensional methods have been...
http://arxiv.org/pdf/2302.14021v1
2023-02-27T18:25:19
2023-02-27T18:25:19
{ "title": [ 0.017349243, 0.014152527, 0.0029182434, -0.052734375, 0.0064468384, 0.0047187805, -0.030883789, -0.024642944, 0.019943237, 0.04397583, -0.028793335, 0.008071899, 0.02748108, -0.0026073456, -0.052490234, 0.0115356445, -0.004245758, 0....
http://arxiv.org/abs/2302.14017v1
Full Stack Optimization of Transformer Inference: a Survey
Sehoon Kim, Coleman Hooper, Thanakul Wattanawong, Minwoo Kang, Ruohan Yan, Hasan Genc, Grace Dinh, Qijing Huang, Kurt Keutzer, Michael W. Mahoney, Yakun Sophia Shao, Amir Gholami
cs.CL, cs.LG
Recent advances in state-of-the-art DNN architecture design have been moving toward Transformer models. These models achieve superior accuracy across a wide range of applications. This trend has been consistent over the past several years since Transformer models were originally introduced. However, the amount of compu...
http://arxiv.org/pdf/2302.14017v1
2023-02-27T18:18:13
2023-02-27T18:18:13
{ "title": [ -0.023498535, -0.0015563965, -0.028137207, -0.07110596, 0.001405716, 0.02078247, -0.06317139, 0.001540184, 0.0007596016, 0.023727417, -0.007701874, 0.0007305145, 0.028884888, -0.03491211, 0.004676819, 0.011062622, 0.008880615, 0.0268...
http://arxiv.org/abs/2302.14003v1
Systematic Rectification of Language Models via Dead-end Analysis
Meng Cao, Mehdi Fatemi, Jackie Chi Kit Cheung, Samira Shabanian
cs.CL, cs.AI, cs.LG
With adversarial or otherwise normal prompts, existing large language models (LLM) can be pushed to generate toxic discourses. One way to reduce the risk of LLMs generating undesired discourses is to alter the training of the LLM. This can be very restrictive due to demanding computation requirements. Other methods rel...
http://arxiv.org/pdf/2302.14003v1
2023-02-27T17:47:53
2023-02-27T17:47:53
{ "title": [ -0.026779175, -0.0023956299, -0.010383606, -0.055603027, -0.0059890747, 0.020950317, -0.08935547, -0.028839111, 0.017593384, 0.045654297, 0.004207611, -0.008003235, 0.027832031, -0.017349243, -0.056610107, 0.03503418, -0.018966675, 0...
http://arxiv.org/abs/2302.13971v1
LLaMA: Open and Efficient Foundation Language Models
Hugo Touvron, Thibaut Lavril, Gautier Izacard, Xavier Martinet, Marie-Anne Lachaux, Timothée Lacroix, Baptiste Rozière, Naman Goyal, Eric Hambro, Faisal Azhar, Aurelien Rodriguez, Armand Joulin, Edouard Grave, Guillaume Lample
cs.CL
We introduce LLaMA, a collection of foundation language models ranging from 7B to 65B parameters. We train our models on trillions of tokens, and show that it is possible to train state-of-the-art models using publicly available datasets exclusively, without resorting to proprietary and inaccessible datasets. In partic...
http://arxiv.org/pdf/2302.13971v1
2023-02-27T17:11:15
2023-02-27T17:11:15
{ "title": [ -0.04421997, -0.02218628, -0.045196533, -0.084106445, 0.016052246, 0.025680542, -0.07312012, 0.016677856, -0.0050811768, 0.027542114, 0.0045433044, 0.016204834, 0.0044555664, -0.026031494, -0.031173706, 0.018814087, -0.0236969, 0.037...
http://arxiv.org/abs/2302.13959v2
Make Every Example Count: On the Stability and Utility of Self-Influence for Learning from Noisy NLP Datasets
Irina Bejan, Artem Sokolov, Katja Filippova
cs.CL
Increasingly larger datasets have become a standard ingredient to advancing the state-of-the-art in NLP. However, data quality might have already become the bottleneck to unlock further gains. Given the diversity and the sizes of modern datasets, standard data filtering is not straight-forward to apply, because of the ...
http://arxiv.org/pdf/2302.13959v2
2023-10-17T16:03:05
2023-02-27T17:00:06
{ "title": [ 0.0038871765, -0.0040016174, -0.033721924, -0.053619385, -0.0007929802, -0.023468018, -0.010513306, -0.0033168793, -0.017486572, 0.012184143, -0.046569824, 0.03451538, -0.043151855, -0.00071287155, -0.040008545, -0.02859497, 0.0037460327...