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/2212.07530v3
Causes and Cures for Interference in Multilingual Translation
Uri Shaham, Maha Elbayad, Vedanuj Goswami, Omer Levy, Shruti Bhosale
cs.CL, cs.AI, cs.LG
Multilingual machine translation models can benefit from synergy between different language pairs, but also suffer from interference. While there is a growing number of sophisticated methods that aim to eliminate interference, our understanding of interference as a phenomenon is still limited. This work identifies the ...
http://arxiv.org/pdf/2212.07530v3
2023-05-19T12:26:50
2022-12-14T22:30:55
{ "title": [ -0.02003479, -0.0010757446, -0.02166748, -0.046020508, -0.03488159, 0.0038528442, -0.047546387, 0.003665924, -0.045776367, 0.069885254, 0.004108429, 0.003211975, 0.04953003, -0.04071045, -0.006576538, 0.024765015, 0.0029468536, 0.035...
http://arxiv.org/abs/2212.07526v1
Relationship Between Online Harmful Behaviors and Social Network Message Writing Style
Talia Sanchez Viera, Richard Khoury
cs.CL, cs.SI
In this paper, we explore the relationship between an individual's writing style and the risk that they will engage in online harmful behaviors (such as cyberbullying). In particular, we consider whether measurable differences in writing style relate to different personality types, as modeled by the Big-Five personalit...
http://arxiv.org/pdf/2212.07526v1
2022-12-14T22:13:55
2022-12-14T22:13:55
{ "title": [ 0.0022659302, 0.01335907, -0.042388916, -0.034606934, 0.0060653687, -0.026550293, -0.0025234222, 0.0039138794, -0.054504395, 0.03741455, 0.002216339, 0.012611389, 0.031555176, -0.030838013, 0.013053894, -0.011962891, 0.0028476715, 0....
http://arxiv.org/abs/2212.07525v2
Efficient Self-supervised Learning with Contextualized Target Representations for Vision, Speech and Language
Alexei Baevski, Arun Babu, Wei-Ning Hsu, Michael Auli
cs.LG, cs.CL, cs.SD, eess.AS
Current self-supervised learning algorithms are often modality-specific and require large amounts of computational resources. To address these issues, we increase the training efficiency of data2vec, a learning objective that generalizes across several modalities. We do not encode masked tokens, use a fast convolutiona...
http://arxiv.org/pdf/2212.07525v2
2023-06-15T15:19:22
2022-12-14T22:13:11
{ "title": [ 0.00025105476, 0.01007843, -0.040130615, -0.05947876, 0.01600647, -0.014480591, -0.023590088, -0.004016876, -0.015403748, 0.030715942, -0.012031555, 0.013793945, -0.019302368, -0.017105103, -0.028259277, -0.0021820068, 0.008598328, 0...
http://arxiv.org/abs/2212.07507v1
Artificial Intelligence for Health Message Generation: Theory, Method, and an Empirical Study Using Prompt Engineering
Sue Lim, Ralf Schmälzle
cs.CL
This study introduces and examines the potential of an AI system to generate health awareness messages. The topic of folic acid, a vitamin that is critical during pregnancy, served as a test case. Using prompt engineering, we generated messages that could be used to raise awareness and compared them to retweeted human-...
http://arxiv.org/pdf/2212.07507v1
2022-12-14T21:13:08
2022-12-14T21:13:08
{ "title": [ -0.02027893, -0.006160736, -0.024017334, -0.03677368, 0.011047363, -0.0036373138, -0.021606445, -0.01625061, -0.027908325, 0.026184082, -0.00699234, 0.024597168, -0.021011353, -0.03338623, -0.024154663, -0.006038666, 0.003452301, 0.0...
http://arxiv.org/abs/2212.07476v2
The Infinite Index: Information Retrieval on Generative Text-To-Image Models
Niklas Deckers, Maik Fröbe, Johannes Kiesel, Gianluca Pandolfo, Christopher Schröder, Benno Stein, Martin Potthast
cs.IR, cs.CL, cs.CV
Conditional generative models such as DALL-E and Stable Diffusion generate images based on a user-defined text, the prompt. Finding and refining prompts that produce a desired image has become the art of prompt engineering. Generative models do not provide a built-in retrieval model for a user's information need expres...
http://arxiv.org/pdf/2212.07476v2
2023-01-21T18:16:14
2022-12-14T19:50:35
{ "title": [ -0.00078868866, 0.01927185, -0.01727295, -0.03643799, 0.008712769, 0.0026607513, -0.015419006, -0.034362793, 0.045532227, 0.050445557, -0.041656494, 0.017700195, -0.003019333, 0.004058838, -0.07556152, 0.007820129, -0.0026493073, 0.0...
http://arxiv.org/abs/2212.07284v1
MANTa: Efficient Gradient-Based Tokenization for Robust End-to-End Language Modeling
Nathan Godey, Roman Castagné, Éric de la Clergerie, Benoît Sagot
cs.CL
Static subword tokenization algorithms have been an essential component of recent works on language modeling. However, their static nature results in important flaws that degrade the models' downstream performance and robustness. In this work, we propose MANTa, a Module for Adaptive Neural TokenizAtion. MANTa is a diff...
http://arxiv.org/pdf/2212.07284v1
2022-12-14T15:33:44
2022-12-14T15:33:44
{ "title": [ -0.027938843, 0.0011873245, -0.048919678, -0.08691406, 0.021972656, 0.013542175, -0.071777344, 0.003522873, -0.0024814606, 0.028335571, -0.0075569153, 0.018066406, 0.021102905, -0.018997192, -0.032226562, 0.019607544, -0.016235352, 0...
http://arxiv.org/abs/2212.07249v2
APOLLO: An Optimized Training Approach for Long-form Numerical Reasoning
Jiashuo Sun, Hang Zhang, Chen Lin, Yeyun Gong, Jian Guo, Nan Duan
cs.CL, cs.LG
Long-form numerical reasoning in financial analysis aims to generate a reasoning program to calculate the correct answer for a given question. Previous work followed a retriever-generator framework, where the retriever selects key facts from a long-form document, and the generator generates a reasoning program based on...
http://arxiv.org/pdf/2212.07249v2
2023-02-13T16:06:43
2022-12-14T14:34:15
{ "title": [ -0.015045166, -0.00068330765, -0.040039062, -0.04714966, 0.009628296, -0.015235901, -0.010795593, -0.0062828064, -0.010368347, 0.006652832, -0.0014362335, -0.004085541, -0.015014648, -0.041503906, -0.012565613, -0.02178955, -0.004234314,...
http://arxiv.org/abs/2212.07223v1
Evaluating Byte and Wordpiece Level Models for Massively Multilingual Semantic Parsing
Massimo Nicosia, Francesco Piccinno
cs.CL
Token free approaches have been successfully applied to a series of word and span level tasks. In this work, we compare a byte-level (ByT5) and a wordpiece based (mT5) sequence to sequence model on the 51 languages of the MASSIVE multilingual semantic parsing dataset. We examine multiple experimental settings: (i) zero...
http://arxiv.org/pdf/2212.07223v1
2022-12-14T13:48:32
2022-12-14T13:48:32
{ "title": [ -0.02784729, -0.007335663, -0.03475952, -0.08276367, -0.0054473877, 0.026046753, -0.0848999, 0.0154800415, -0.0016374588, 0.040985107, -0.023010254, 0.0026111603, 0.0027179718, 0.001115799, -0.0541687, 0.0032253265, -0.01777649, 0.03...
http://arxiv.org/abs/2212.07220v2
Understanding Translationese in Cross-Lingual Summarization
Jiaan Wang, Fandong Meng, Yunlong Liang, Tingyi Zhang, Jiarong Xu, Zhixu Li, Jie Zhou
cs.CL, cs.AI
Given a document in a source language, cross-lingual summarization (CLS) aims at generating a concise summary in a different target language. Unlike monolingual summarization (MS), naturally occurring source-language documents paired with target-language summaries are rare. To collect large-scale CLS data, existing dat...
http://arxiv.org/pdf/2212.07220v2
2023-10-10T03:00:35
2022-12-14T13:41:49
{ "title": [ -0.02973938, -0.011749268, -0.007019043, -0.0635376, -0.02015686, 0.01096344, -0.05770874, -0.006931305, -0.009216309, 0.063964844, -0.011131287, 0.024353027, 0.023742676, 0.0036678314, -0.06121826, 0.017349243, 0.004245758, 0.035247...
http://arxiv.org/abs/2212.07219v1
VTCC-NLP at NL4Opt competition subtask 1: An Ensemble Pre-trained language models for Named Entity Recognition
Xuan-Dung Doan
cs.CL
We propose a combined three pre-trained language models (XLM-R, BART, and DeBERTa-V3) as an empower of contextualized embedding for named entity recognition. Our model achieves a 92.9% F1 score on the test set and ranks 5th on the leaderboard at NL4Opt competition subtask 1.
http://arxiv.org/pdf/2212.07219v1
2022-12-14T13:41:36
2022-12-14T13:41:36
{ "title": [ -0.013679504, -0.004142761, -0.04925537, -0.05303955, 0.0031776428, -0.014289856, -0.058380127, 0.0025482178, 0.009613037, 0.030899048, -0.043762207, 0.004131317, 0.011489868, -0.009605408, -0.03253174, 0.0011072159, -0.016403198, 0....
http://arxiv.org/abs/2212.07183v1
Mitigating Negative Style Transfer in Hybrid Dialogue System
Shimin Li, Qinyuan Cheng, Linyang Li, Xipeng Qiu
cs.CL
As the functionality of dialogue systems evolves, hybrid dialogue systems that accomplish user-specific goals and participate in open-topic chitchat with users are attracting growing attention. Existing research learns both tasks concurrently utilizing a multi-task fusion technique but ignores the negative transfer phe...
http://arxiv.org/pdf/2212.07183v1
2022-12-14T12:13:34
2022-12-14T12:13:34
{ "title": [ 0.008956909, -0.0008573532, -0.044799805, -0.030654907, 0.016052246, -0.029220581, 0.001253128, 0.0031471252, -0.015113831, 0.04196167, 0.005847931, 0.010055542, 0.019622803, -0.028640747, -0.024353027, -0.0041618347, -0.008178711, 0...
http://arxiv.org/abs/2212.07172v1
Quotations, Coreference Resolution, and Sentiment Annotations in Croatian News Articles: An Exploratory Study
Jelena Sarajlić, Gaurish Thakkar, Diego Alves, Nives Mikelic Preradović
cs.CL
This paper presents a corpus annotated for the task of direct-speech extraction in Croatian. The paper focuses on the annotation of the quotation, co-reference resolution, and sentiment annotation in SETimes news corpus in Croatian and on the analysis of its language-specific differences compared to English. From this,...
http://arxiv.org/pdf/2212.07172v1
2022-12-14T11:54:12
2022-12-14T11:54:12
{ "title": [ 0.012481689, -0.011741638, -0.010719299, -0.03353882, -0.0046691895, -0.002336502, -0.004917145, -0.017242432, -0.018310547, 0.06970215, -0.027420044, 0.014915466, 0.020599365, 0.017196655, -0.07879639, -0.004501343, -0.0236969, 0.02...
http://arxiv.org/abs/2212.07429v1
Building Multilingual Corpora for a Complex Named Entity Recognition and Classification Hierarchy using Wikipedia and DBpedia
Diego Alves, Gaurish Thakkar, Gabriel Amaral, Tin Kuculo, Marko Tadić
cs.CL
With the ever-growing popularity of the field of NLP, the demand for datasets in low resourced-languages follows suit. Following a previously established framework, in this paper, we present the UNER dataset, a multilingual and hierarchical parallel corpus annotated for named-entities. We describe in detail the develop...
http://arxiv.org/pdf/2212.07429v1
2022-12-14T11:38:48
2022-12-14T11:38:48
{ "title": [ -0.00819397, 0.0033740997, -0.05429077, -0.076660156, 0.024536133, 0.0041160583, -0.042785645, -0.0016298294, -0.023284912, 0.028274536, -0.024261475, 0.022140503, -0.0005235672, -0.0033473969, -0.01852417, -0.010108948, -0.009025574, ...
http://arxiv.org/abs/2212.07164v1
Speech and Natural Language Processing Technologies for Pseudo-Pilot Simulator
Amrutha Prasad, Juan Zuluaga-Gomez, Petr Motlicek, Saeed Sarfjoo, Iuliia Nigmatulina, Karel Vesely
cs.CL, cs.AI, cs.LG, eess.AS
This paper describes a simple yet efficient repetition-based modular system for speeding up air-traffic controllers (ATCos) training. E.g., a human pilot is still required in EUROCONTROL's ESCAPE lite simulator (see https://www.eurocontrol.int/simulator/escape) during ATCo training. However, this need can be substitute...
http://arxiv.org/pdf/2212.07164v1
2022-12-14T11:34:59
2022-12-14T11:34:59
{ "title": [ 0.0021438599, -0.010528564, -0.031463623, -0.05899048, -0.0025119781, -0.015853882, -0.02734375, -0.013084412, -0.038848877, 0.048187256, -0.011360168, -0.014854431, 0.0044441223, -0.023590088, -0.017333984, -0.0071868896, -0.014511108, ...
http://arxiv.org/abs/2212.07162v1
Building and Evaluating Universal Named-Entity Recognition English corpus
Diego Alves, Gaurish Thakkar, Marko Tadić
cs.CL
This article presents the application of the Universal Named Entity framework to generate automatically annotated corpora. By using a workflow that extracts Wikipedia data and meta-data and DBpedia information, we generated an English dataset which is described and evaluated. Furthermore, we conducted a set of experime...
http://arxiv.org/pdf/2212.07162v1
2022-12-14T11:32:24
2022-12-14T11:32:24
{ "title": [ -0.0127334595, 0.00434494, -0.040618896, -0.066833496, 0.026168823, 0.001824379, -0.028381348, 0.002149582, 0.0021800995, 0.026809692, -0.03491211, 0.023147583, 0.0018501282, 0.0006303787, -0.03643799, 0.010314941, -0.020690918, 0.05...
http://arxiv.org/abs/2212.07160v1
Multi-task Learning for Cross-Lingual Sentiment Analysis
Gaurish Thakkar, Nives Mikelic Preradovic, Marko Tadic
cs.CL
This paper presents a cross-lingual sentiment analysis of news articles using zero-shot and few-shot learning. The study aims to classify the Croatian news articles with positive, negative, and neutral sentiments using the Slovene dataset. The system is based on a trilingual BERT-based model trained in three languages:...
http://arxiv.org/pdf/2212.07160v1
2022-12-14T11:29:03
2022-12-14T11:29:03
{ "title": [ -0.009887695, -0.0016508102, -0.036376953, -0.05343628, 0.0069770813, -0.0084991455, -0.046875, -0.0113220215, 0.020736694, 0.053497314, -0.01953125, 0.006881714, 0.0042533875, 0.0069122314, -0.068115234, 0.009803772, -0.009643555, 0...
http://arxiv.org/abs/2212.07156v1
MIST: a Large-Scale Annotated Resource and Neural Models for Functions of Modal Verbs in English Scientific Text
Sophie Henning, Nicole Macher, Stefan Grünewald, Annemarie Friedrich
cs.CL, cs.AI
Modal verbs (e.g., "can", "should", or "must") occur highly frequently in scientific articles. Decoding their function is not straightforward: they are often used for hedging, but they may also denote abilities and restrictions. Understanding their meaning is important for various NLP tasks such as writing assistance o...
http://arxiv.org/pdf/2212.07156v1
2022-12-14T11:10:03
2022-12-14T11:10:03
{ "title": [ -0.0317688, -0.0029850006, -0.046325684, -0.028076172, -0.008979797, -0.007255554, -0.040985107, -0.008476257, 0.0025577545, 0.043273926, 0.018356323, 0.0066986084, 0.03463745, -0.03378296, -0.033843994, 0.040740967, -0.0084991455, 0...
http://arxiv.org/abs/2212.07428v5
Towards Linguistically Informed Multi-Objective Pre-Training for Natural Language Inference
Maren Pielka, Svetlana Schmidt, Lisa Pucknat, Rafet Sifa
cs.CL, cs.LG
We introduce a linguistically enhanced combination of pre-training methods for transformers. The pre-training objectives include POS-tagging, synset prediction based on semantic knowledge graphs, and parent prediction based on dependency parse trees. Our approach achieves competitive results on the Natural Language Inf...
http://arxiv.org/pdf/2212.07428v5
2022-12-30T09:15:11
2022-12-14T10:50:13
{ "title": [ -0.01626587, -0.023971558, -0.012290955, -0.06896973, -0.023391724, 0.00019443035, -0.056854248, 0.001709938, -0.0042037964, 0.024719238, -0.008682251, 0.007873535, 0.009216309, -0.01776123, -0.024459839, 0.00472641, -0.01486969, 0.0...
http://arxiv.org/abs/2212.07127v4
Towards mapping the contemporary art world with ArtLM: an art-specific NLP model
Qinkai Chen, Mohamed El-Mennaoui, Antoine Fosset, Amine Rebei, Haoyang Cao, Philine Bouscasse, Christy Eóin O'Beirne, Sasha Shevchenko, Mathieu Rosenbaum
cs.CL, cs.LG
With an increasing amount of data in the art world, discovering artists and artworks suitable to collectors' tastes becomes a challenge. It is no longer enough to use visual information, as contextual information about the artist has become just as important in contemporary art. In this work, we present a generic Natur...
http://arxiv.org/pdf/2212.07127v4
2022-12-22T17:10:46
2022-12-14T09:26:07
{ "title": [ 0.00095033646, -0.0141067505, -0.015975952, -0.02848816, -0.006954193, -0.0010576248, -0.011230469, -0.013504028, 0.013671875, 0.011024475, -0.011665344, -0.0054779053, -0.0074501038, -0.030578613, -0.029296875, -0.0048561096, -0.0095520...
http://arxiv.org/abs/2212.07126v1
Explainability of Text Processing and Retrieval Methods: A Critical Survey
Sourav Saha, Debapriyo Majumdar, Mandar Mitra
cs.IR, cs.AI, cs.CL
Deep Learning and Machine Learning based models have become extremely popular in text processing and information retrieval. However, the non-linear structures present inside the networks make these models largely inscrutable. A significant body of research has focused on increasing the transparency of these models. Thi...
http://arxiv.org/pdf/2212.07126v1
2022-12-14T09:25:49
2022-12-14T09:25:49
{ "title": [ -0.018844604, 0.0037231445, 0.0070724487, -0.04940796, 0.008216858, 0.029693604, -0.013572693, -0.02079773, 0.003545761, 0.009353638, -0.017990112, -0.0021781921, 0.0026245117, -0.014251709, -0.047180176, -0.010375977, -0.03262329, 0...
http://arxiv.org/abs/2212.07112v1
DialogQAE: N-to-N Question Answer Pair Extraction from Customer Service Chatlog
Xin Zheng, Tianyu Liu, Haoran Meng, Xu Wang, Yufan Jiang, Mengliang Rao, Binghuai Lin, Zhifang Sui, Yunbo Cao
cs.CL, cs.AI, cs.IR
Harvesting question-answer (QA) pairs from customer service chatlog in the wild is an efficient way to enrich the knowledge base for customer service chatbots in the cold start or continuous integration scenarios. Prior work attempts to obtain 1-to-1 QA pairs from growing customer service chatlog, which fails to integr...
http://arxiv.org/pdf/2212.07112v1
2022-12-14T09:05:14
2022-12-14T09:05:14
{ "title": [ -0.012184143, -0.02319336, -0.032806396, -0.0574646, 0.0042533875, -0.01146698, -0.0124053955, 0.016174316, 0.01939392, 0.031982422, -0.018341064, 0.012580872, 0.009269714, -0.009567261, -0.056396484, 0.0016727448, -0.009536743, 0.01...
http://arxiv.org/abs/2212.07075v1
Cross-Modal Similarity-Based Curriculum Learning for Image Captioning
Hongkuan Zhang, Saku Sugawara, Akiko Aizawa, Lei Zhou, Ryohei Sasano, Koichi Takeda
cs.CV, cs.CL
Image captioning models require the high-level generalization ability to describe the contents of various images in words. Most existing approaches treat the image-caption pairs equally in their training without considering the differences in their learning difficulties. Several image captioning approaches introduce cu...
http://arxiv.org/pdf/2212.07075v1
2022-12-14T07:52:36
2022-12-14T07:52:36
{ "title": [ -0.013259888, 0.0088272095, -0.024810791, -0.020324707, 0.012023926, -0.02784729, 0.0018529892, -0.028915405, 0.01398468, 0.03564453, -0.030181885, 0.04434204, 0.0021820068, -0.002861023, -0.0116119385, -0.00028657913, -0.016479492, ...
http://arxiv.org/abs/2212.07072v2
SMSMix: Sense-Maintained Sentence Mixup for Word Sense Disambiguation
Hee Suk Yoon, Eunseop Yoon, John Harvill, Sunjae Yoon, Mark Hasegawa-Johnson, Chang D. Yoo
cs.CL, cs.LG
Word Sense Disambiguation (WSD) is an NLP task aimed at determining the correct sense of a word in a sentence from discrete sense choices. Although current systems have attained unprecedented performances for such tasks, the nonuniform distribution of word senses during training generally results in systems performing ...
http://arxiv.org/pdf/2212.07072v2
2022-12-21T07:36:49
2022-12-14T07:48:42
{ "title": [ -0.01889038, -0.013046265, -0.021636963, -0.034240723, 0.0017232895, 0.001663208, -0.029052734, -0.012634277, 0.012451172, 0.04547119, -0.014694214, 0.018356323, 0.025512695, -0.0031929016, -0.061065674, 0.01386261, -0.02381897, 0.02...
http://arxiv.org/abs/2301.00652v1
Efficient Speech Representation Learning with Low-Bit Quantization
Ching-Feng Yeh, Wei-Ning Hsu, Paden Tomasello, Abdelrahman Mohamed
eess.AS, cs.CL
With the development of hardware for machine learning, newer models often come at the cost of both increased sizes and computational complexity. In effort to improve the efficiency for these models, we apply and investigate recent quantization techniques on speech representation learning models. The quantization techni...
http://arxiv.org/pdf/2301.00652v1
2022-12-14T06:09:08
2022-12-14T06:09:08
{ "title": [ -0.006462097, 0.006126404, -0.059448242, -0.06286621, 0.01361084, -0.013542175, -0.03137207, 0.007987976, -0.0022144318, 0.06652832, -0.00504303, -0.0030670166, 0.0096588135, -0.01322937, -0.061309814, 0.010307312, 0.0055732727, 0.04...
http://arxiv.org/abs/2212.07043v1
AsPOS: Assamese Part of Speech Tagger using Deep Learning Approach
Dhrubajyoti Pathak, Sukumar Nandi, Priyankoo Sarmah
cs.CL, cs.AI, cs.LG, I.2.7
Part of Speech (POS) tagging is crucial to Natural Language Processing (NLP). It is a well-studied topic in several resource-rich languages. However, the development of computational linguistic resources is still in its infancy despite the existence of numerous languages that are historically and literary rich. Assames...
http://arxiv.org/pdf/2212.07043v1
2022-12-14T05:36:18
2022-12-14T05:36:18
{ "title": [ -0.024139404, 0.00022816658, -0.03515625, -0.03152466, -0.012374878, -0.010391235, -0.011985779, 0.015357971, -0.0103302, 0.02468872, -0.045654297, 0.0026626587, 0.025939941, -0.026016235, -0.012275696, -0.020980835, -0.018005371, 0....
http://arxiv.org/abs/2212.06972v2
Disentangling Prosody Representations with Unsupervised Speech Reconstruction
Leyuan Qu, Taihao Li, Cornelius Weber, Theresa Pekarek-Rosin, Fuji Ren, Stefan Wermter
cs.SD, cs.CL, eess.AS
Human speech can be characterized by different components, including semantic content, speaker identity and prosodic information. Significant progress has been made in disentangling representations for semantic content and speaker identity in Automatic Speech Recognition (ASR) and speaker verification tasks respectivel...
http://arxiv.org/pdf/2212.06972v2
2023-09-26T02:42:11
2022-12-14T01:37:35
{ "title": [ -0.007419586, 0.004714966, -0.024108887, -0.081604004, -0.009460449, 0.0051727295, -0.052459717, 0.015792847, 0.0041160583, 0.042175293, 0.008323669, -0.013206482, 0.02520752, -0.032989502, -0.041656494, 0.012176514, -0.0041885376, 0...
http://arxiv.org/abs/2212.06971v1
Find Someone Who: Visual Commonsense Understanding in Human-Centric Grounding
Haoxuan You, Rui Sun, Zhecan Wang, Kai-Wei Chang, Shih-Fu Chang
cs.CV, cs.CL
From a visual scene containing multiple people, human is able to distinguish each individual given the context descriptions about what happened before, their mental/physical states or intentions, etc. Above ability heavily relies on human-centric commonsense knowledge and reasoning. For example, if asked to identify th...
http://arxiv.org/pdf/2212.06971v1
2022-12-14T01:37:16
2022-12-14T01:37:16
{ "title": [ 0.049835205, -0.0022010803, -0.025543213, -0.030044556, -0.02885437, -0.030563354, -0.007003784, -0.029205322, 0.0096588135, 0.033325195, -0.04248047, 0.01828003, 0.028701782, -0.017562866, -0.026184082, -0.019638062, -0.0009675026, ...
http://arxiv.org/abs/2212.06950v2
Pre-trained Language Models Can be Fully Zero-Shot Learners
Xuandong Zhao, Siqi Ouyang, Zhiguo Yu, Ming Wu, Lei Li
cs.CL
How can we extend a pre-trained model to many language understanding tasks, without labeled or additional unlabeled data? Pre-trained language models (PLMs) have been effective for a wide range of NLP tasks. However, existing approaches either require fine-tuning on downstream labeled datasets or manually constructing ...
http://arxiv.org/pdf/2212.06950v2
2023-05-26T06:05:50
2022-12-14T00:03:52
{ "title": [ 0.050750732, -0.014884949, -0.006248474, -0.06817627, -0.05532837, -0.016830444, -0.0524292, 0.016281128, 0.008575439, 0.031555176, -0.018600464, -0.025848389, 0.012786865, -0.048339844, 0.010055542, -0.0017538071, 0.019943237, 0.044...
http://arxiv.org/abs/2212.06933v1
Paraphrase Identification with Deep Learning: A Review of Datasets and Methods
Chao Zhou, Cheng Qiu, Daniel E. Acuna
cs.CL, cs.AI, cs.IR
The rapid advancement of AI technology has made text generation tools like GPT-3 and ChatGPT increasingly accessible, scalable, and effective. This can pose serious threat to the credibility of various forms of media if these technologies are used for plagiarism, including scientific literature and news sources. Despit...
http://arxiv.org/pdf/2212.06933v1
2022-12-13T23:06:20
2022-12-13T23:06:20
{ "title": [ 0.0013093948, 0.008834839, -0.022964478, -0.050994873, 0.001364708, -0.007095337, -0.034851074, -0.012107849, -0.011398315, 0.048553467, -0.0073051453, -0.0025482178, 0.030273438, -0.02293396, -0.03567505, 0.0050315857, -0.03062439, ...
http://arxiv.org/abs/2212.06868v1
Deep Image Style Transfer from Freeform Text
Tejas Santanam, Mengyang Liu, Jiangyue Yu, Zhaodong Yang
cs.CV, cs.CL, cs.LG
This paper creates a novel method of deep neural style transfer by generating style images from freeform user text input. The language model and style transfer model form a seamless pipeline that can create output images with similar losses and improved quality when compared to baseline style transfer methods. The lang...
http://arxiv.org/pdf/2212.06868v1
2022-12-13T19:24:08
2022-12-13T19:24:08
{ "title": [ -0.0284729, 0.005756378, -0.042755127, -0.04937744, 0.0037994385, -0.0067443848, -0.017486572, -0.0010509491, 0.005203247, 0.050445557, -0.0061912537, 0.007785797, 0.02848816, -0.017074585, -0.022003174, 0.014778137, 0.017227173, 0.0...
http://arxiv.org/abs/2212.07796v3
CREPE: Can Vision-Language Foundation Models Reason Compositionally?
Zixian Ma, Jerry Hong, Mustafa Omer Gul, Mona Gandhi, Irena Gao, Ranjay Krishna
cs.CL, cs.CV
A fundamental characteristic common to both human vision and natural language is their compositional nature. Yet, despite the performance gains contributed by large vision and language pretraining, we find that: across 7 architectures trained with 4 algorithms on massive datasets, they struggle at compositionality. To ...
http://arxiv.org/pdf/2212.07796v3
2023-05-16T16:27:08
2022-12-13T19:17:36
{ "title": [ -0.036376953, -0.037261963, -0.06994629, -0.08685303, -0.025817871, -0.009346008, -0.064697266, 0.043304443, 0.015350342, 0.0057678223, -0.030899048, 0.0031166077, -0.000182271, -0.022155762, -0.008125305, -0.017562866, -0.020690918, ...
http://arxiv.org/abs/2212.08072v2
Foresight -- Generative Pretrained Transformer (GPT) for Modelling of Patient Timelines using EHRs
Zeljko Kraljevic, Dan Bean, Anthony Shek, Rebecca Bendayan, Harry Hemingway, Joshua Au Yeung, Alexander Deng, Alfie Baston, Jack Ross, Esther Idowu, James T Teo, Richard J Dobson
cs.CL, cs.AI, cs.LG
Background: Electronic Health Records hold detailed longitudinal information about each patient's health status and general clinical history, a large portion of which is stored within the unstructured text. Existing approaches focus mostly on structured data and a subset of single-domain outcomes. We explore how tempor...
http://arxiv.org/pdf/2212.08072v2
2023-01-24T09:24:23
2022-12-13T19:06:00
{ "title": [ 0.043701172, 0.003376007, -0.041625977, -0.08227539, -0.002204895, -0.01486969, -0.066589355, -0.01953125, -0.0013942719, 0.044891357, -0.002603531, 0.020126343, 0.025939941, -0.04586792, -0.0040893555, 0.010955811, 0.020431519, 0.01...
http://arxiv.org/abs/2212.06817v2
RT-1: Robotics Transformer for Real-World Control at Scale
Anthony Brohan, Noah Brown, Justice Carbajal, Yevgen Chebotar, Joseph Dabis, Chelsea Finn, Keerthana Gopalakrishnan, Karol Hausman, Alex Herzog, Jasmine Hsu, Julian Ibarz, Brian Ichter, Alex Irpan, Tomas Jackson, Sally Jesmonth, Nikhil J Joshi, Ryan Julian, Dmitry Kalashnikov, Yuheng Kuang, Isabel Leal, Kuang-Huei Lee,...
cs.RO, cs.AI, cs.CL, cs.CV, cs.LG
By transferring knowledge from large, diverse, task-agnostic datasets, modern machine learning models can solve specific downstream tasks either zero-shot or with small task-specific datasets to a high level of performance. While this capability has been demonstrated in other fields such as computer vision, natural lan...
http://arxiv.org/pdf/2212.06817v2
2023-08-11T17:45:27
2022-12-13T18:55:15
{ "title": [ -0.009269714, 0.0070228577, -0.050109863, -0.023132324, -0.003452301, -0.03439331, -0.03237915, -0.037109375, -0.04473877, 0.026992798, -0.030044556, 0.010253906, 0.020812988, -0.02168274, 0.008682251, -0.021713257, 0.018844604, 0.03...
http://arxiv.org/abs/2212.06801v2
A fine-grained comparison of pragmatic language understanding in humans and language models
Jennifer Hu, Sammy Floyd, Olessia Jouravlev, Evelina Fedorenko, Edward Gibson
cs.CL, cs.AI
Pragmatics and non-literal language understanding are essential to human communication, and present a long-standing challenge for artificial language models. We perform a fine-grained comparison of language models and humans on seven pragmatic phenomena, using zero-shot prompting on an expert-curated set of English mat...
http://arxiv.org/pdf/2212.06801v2
2023-05-23T18:35:34
2022-12-13T18:34:59
{ "title": [ -0.015617371, 0.0021419525, -0.024017334, -0.0519104, -0.0045051575, -0.01184082, -0.05593872, -0.0053977966, 0.011169434, 0.015449524, -0.025634766, -0.018676758, 0.009376526, -0.020629883, -0.04623413, 0.00242424, -0.011917114, 0.0...
http://arxiv.org/abs/2212.06800v3
Diverse Demonstrations Improve In-context Compositional Generalization
Itay Levy, Ben Bogin, Jonathan Berant
cs.CL
In-context learning has shown great success in i.i.d semantic parsing splits, where the training and test sets are drawn from the same distribution. In this setup, models are typically prompted with demonstrations that are similar to the input utterance. However, in the setup of compositional generalization, where mode...
http://arxiv.org/pdf/2212.06800v3
2023-06-24T15:49:31
2022-12-13T18:34:15
{ "title": [ 0.048980713, 0.014556885, -0.0362854, -0.11053467, 0.01134491, -0.021560669, -0.06573486, 0.007095337, -0.003616333, 0.015899658, -0.028198242, 0.03265381, -0.042175293, -0.019958496, -0.013893127, -0.016189575, -0.014709473, 0.06842...
http://arxiv.org/abs/2212.06742v2
ERNIE-Code: Beyond English-Centric Cross-lingual Pretraining for Programming Languages
Yekun Chai, Shuohuan Wang, Chao Pang, Yu Sun, Hao Tian, Hua Wu
cs.CL, cs.LG, cs.PL, cs.SE
Software engineers working with the same programming language (PL) may speak different natural languages (NLs) and vice versa, erecting huge barriers to communication and working efficiency. Recent studies have demonstrated the effectiveness of generative pre-training in computer programs, yet they are always English-c...
http://arxiv.org/pdf/2212.06742v2
2023-05-19T14:14:30
2022-12-13T17:21:44
{ "title": [ -0.020996094, -0.030075073, -0.030380249, -0.046539307, -0.0056152344, -0.0061454773, -0.029129028, 0.0073928833, -0.018127441, 0.03527832, -0.006164551, 0.015380859, 0.0043640137, -0.0309906, -0.013130188, 0.0031223297, -0.00699234, ...
http://arxiv.org/abs/2212.06713v1
Structured Prompting: Scaling In-Context Learning to 1,000 Examples
Yaru Hao, Yutao Sun, Li Dong, Zhixiong Han, Yuxian Gu, Furu Wei
cs.CL
Large language models have exhibited intriguing in-context learning capability, achieving promising zero- and few-shot performance without updating the parameters. However, conventional in-context learning is usually restricted by length constraints, rendering it ineffective to absorb supervision from a large number of...
http://arxiv.org/pdf/2212.06713v1
2022-12-13T16:31:21
2022-12-13T16:31:21
{ "title": [ 0.027694702, -0.007972717, -0.035064697, -0.07312012, 0.011756897, -0.017196655, -0.020126343, -0.004749298, 0.0033817291, -0.0029659271, -0.05404663, 0.01133728, -0.03793335, -0.021881104, -0.016204834, -0.02746582, 0.020217896, -0....
http://arxiv.org/abs/2212.06711v4
On Text-based Personality Computing: Challenges and Future Directions
Qixiang Fang, Anastasia Giachanou, Ayoub Bagheri, Laura Boeschoten, Erik-Jan van Kesteren, Mahdi Shafiee Kamalabad, Daniel L Oberski
cs.CL, cs.CY
Text-based personality computing (TPC) has gained many research interests in NLP. In this paper, we describe 15 challenges that we consider deserving the attention of the research community. These challenges are organized by the following topics: personality taxonomies, measurement quality, datasets, performance evalua...
http://arxiv.org/pdf/2212.06711v4
2023-05-22T17:40:14
2022-12-13T16:29:51
{ "title": [ -0.031799316, -0.0052490234, -0.017547607, -0.030090332, 0.007648468, 0.013832092, -0.0049552917, -0.005420685, -0.002922058, 0.05206299, -0.009384155, 0.015419006, 0.027267456, -0.015670776, -0.04611206, 0.029586792, -0.011489868, 0...
http://arxiv.org/abs/2212.06645v1
Do Text-to-Text Multi-Task Learners Suffer from Task Conflict?
David Mueller, Nicholas Andrews, Mark Dredze
cs.CL
Traditional multi-task learning architectures train a single model across multiple tasks through a shared encoder followed by task-specific decoders. Learning these models often requires specialized training algorithms that address task-conflict in the shared parameter updates, which otherwise can lead to negative tran...
http://arxiv.org/pdf/2212.06645v1
2022-12-13T15:28:57
2022-12-13T15:28:57
{ "title": [ -0.0033512115, 0.0041122437, -0.03161621, -0.007171631, -0.008636475, -0.028701782, -0.0010089874, -0.019607544, -0.01940918, 0.037078857, -0.019485474, 0.005809784, -0.006298065, -0.010276794, -0.042816162, -0.038391113, -0.019683838, ...
http://arxiv.org/abs/2212.06636v1
Categorical Tools for Natural Language Processing
Giovanni de Felice
cs.CL, math.CT
This thesis develops the translation between category theory and computational linguistics as a foundation for natural language processing. The three chapters deal with syntax, semantics and pragmatics. First, string diagrams provide a unified model of syntactic structures in formal grammars. Second, functors compute s...
http://arxiv.org/pdf/2212.06636v1
2022-12-13T15:12:37
2022-12-13T15:12:37
{ "title": [ -0.025939941, -0.01701355, -0.03164673, -0.07849121, -0.000119149685, 0.02017212, -0.05355835, 0.0031967163, -0.0033874512, 0.03463745, -0.017944336, 0.0005712509, -0.0011196136, -0.012863159, -0.04849243, 0.007472992, -0.021118164, ...
http://arxiv.org/abs/2212.06615v1
Category Theory for Quantum Natural Language Processing
Alexis Toumi
math.CT, cs.CL, quant-ph
This thesis introduces quantum natural language processing (QNLP) models based on a simple yet powerful analogy between computational linguistics and quantum mechanics: grammar as entanglement. The grammatical structure of text and sentences connects the meaning of words in the same way that entanglement structure conn...
http://arxiv.org/pdf/2212.06615v1
2022-12-13T14:38:57
2022-12-13T14:38:57
{ "title": [ -0.020111084, -0.00699234, -0.03564453, -0.05935669, 0.020462036, -0.00409317, -0.010879517, 0.014572144, -0.031051636, 0.0206604, 0.02911377, -0.0047721863, -0.0038337708, -0.03237915, -0.029052734, 0.015525818, -0.037261963, 0.0583...
http://arxiv.org/abs/2212.06560v1
Exploring Fake News Detection with Heterogeneous Social Media Context Graphs
Gregor Donabauer, Udo Kruschwitz
cs.CL, cs.IR
Fake news detection has become a research area that goes way beyond a purely academic interest as it has direct implications on our society as a whole. Recent advances have primarily focused on textbased approaches. However, it has become clear that to be effective one needs to incorporate additional, contextual inform...
http://arxiv.org/pdf/2212.06560v1
2022-12-13T13:29:47
2022-12-13T13:29:47
{ "title": [ 0.0069885254, 0.030761719, -0.059173584, -0.04324341, 0.045837402, -0.01626587, 0.0057258606, -0.021408081, -0.03842163, 0.03829956, -0.043060303, 0.044891357, -0.010414124, 0.001613617, -0.009300232, -0.020217896, -0.039367676, 0.07...
http://arxiv.org/abs/2212.06556v1
Localized Latent Updates for Fine-Tuning Vision-Language Models
Moritz Ibing, Isaak Lim, Leif Kobbelt
cs.CV, cs.CL, cs.LG
Although massive pre-trained vision-language models like CLIP show impressive generalization capabilities for many tasks, still it often remains necessary to fine-tune them for improved performance on specific datasets. When doing so, it is desirable that updating the model is fast and that the model does not lose its ...
http://arxiv.org/pdf/2212.06556v1
2022-12-13T13:15:20
2022-12-13T13:15:20
{ "title": [ -0.017288208, -0.0065460205, -0.02809143, -0.060028076, 0.0012969971, -0.007461548, -0.03994751, 0.00085401535, 0.019134521, 0.04421997, -0.017074585, -0.00054597855, 0.008148193, -0.0074310303, -0.04925537, 0.011421204, 0.030792236, ...
http://arxiv.org/abs/2212.06543v1
Modelling Stance Detection as Textual Entailment Recognition and Leveraging Measurement Knowledge from Social Sciences
Qixiang Fang, Anastasia Giachanou, Ayoub Bagheri
cs.CL, cs.CY, cs.IR
Stance detection (SD) can be considered a special case of textual entailment recognition (TER), a generic natural language task. Modelling SD as TER may offer benefits like more training data and a more general learning scheme. In this paper, we present an initial empirical analysis of this approach. We apply it to a d...
http://arxiv.org/pdf/2212.06543v1
2022-12-13T12:56:55
2022-12-13T12:56:55
{ "title": [ -0.020828247, -0.0019683838, -0.032470703, -0.055999756, -0.013771057, -0.0006799698, -0.047943115, -0.015548706, 0.025878906, 0.023025513, -0.043792725, -0.0005130768, -0.003540039, 0.0006227493, -0.064453125, -0.009384155, -0.008773804...
http://arxiv.org/abs/2212.06522v1
Distantly-Supervised Named Entity Recognition with Adaptive Teacher Learning and Fine-grained Student Ensemble
Xiaoye Qu, Jun Zeng, Daizong Liu, Zhefeng Wang, Baoxing Huai, Pan Zhou
cs.CL
Distantly-Supervised Named Entity Recognition (DS-NER) effectively alleviates the data scarcity problem in NER by automatically generating training samples. Unfortunately, the distant supervision may induce noisy labels, thus undermining the robustness of the learned models and restricting the practical application. To...
http://arxiv.org/pdf/2212.06522v1
2022-12-13T12:14:09
2022-12-13T12:14:09
{ "title": [ -0.0038661957, 0.0074501038, -0.050201416, -0.06109619, 0.0289917, -0.01763916, -0.020431519, 0.013648987, -0.0045433044, 0.005958557, -0.068115234, 0.020523071, -0.0034980774, 0.010879517, -0.012512207, -0.032562256, -0.007835388, 0...
http://arxiv.org/abs/2212.06468v2
Lisan: Yemeni, Iraqi, Libyan, and Sudanese Arabic Dialect Copora with Morphological Annotations
Mustafa Jarrar, Fadi A Zaraket, Tymaa Hammouda, Daanish Masood Alavi, Martin Waahlisch
cs.CL, cs.DL
This article presents morphologically-annotated Yemeni, Sudanese, Iraqi, and Libyan Arabic dialects Lisan corpora. Lisan features around 1.2 million tokens. We collected the content of the corpora from several social media platforms. The Yemeni corpus (~ 1.05M tokens) was collected automatically from Twitter. The corpo...
http://arxiv.org/pdf/2212.06468v2
2022-12-17T12:37:29
2022-12-13T10:37:10
{ "title": [ -0.01878357, -0.0009407997, -0.044281006, -0.031173706, -0.03967285, 0.005569458, -0.07507324, 0.01727295, -0.061462402, 0.04598999, 0.02935791, 0.0098724365, 0.03050232, -0.05532837, 0.0335083, 0.0025787354, 0.01449585, 0.0037555695...
http://arxiv.org/abs/2212.06397v1
Style-Label-Free: Cross-Speaker Style Transfer by Quantized VAE and Speaker-wise Normalization in Speech Synthesis
Chunyu Qiang, Peng Yang, Hao Che, Xiaorui Wang, Zhongyuan Wang
cs.SD, cs.AI, cs.CL, eess.AS
Cross-speaker style transfer in speech synthesis aims at transferring a style from source speaker to synthesised speech of a target speaker's timbre. Most previous approaches rely on data with style labels, but manually-annotated labels are expensive and not always reliable. In response to this problem, we propose Styl...
http://arxiv.org/pdf/2212.06397v1
2022-12-13T06:26:25
2022-12-13T06:26:25
{ "title": [ -0.024765015, 0.00040125847, -0.0703125, -0.051239014, 0.0052223206, -0.024902344, -0.026657104, 0.03173828, -0.037506104, 0.07409668, 0.01927185, 0.032806396, 0.047698975, -0.035186768, -0.004447937, 0.04043579, -0.00242424, 0.03741...
http://arxiv.org/abs/2212.06385v2
TencentPretrain: A Scalable and Flexible Toolkit for Pre-training Models of Different Modalities
Zhe Zhao, Yudong Li, Cheng Hou, Jing Zhao, Rong Tian, Weijie Liu, Yiren Chen, Ningyuan Sun, Haoyan Liu, Weiquan Mao, Han Guo, Weigang Guo, Taiqiang Wu, Tao Zhu, Wenhang Shi, Chen Chen, Shan Huang, Sihong Chen, Liqun Liu, Feifei Li, Xiaoshuai Chen, Xingwu Sun, Zhanhui Kang, Xiaoyong Du, Linlin Shen, Kimmo Yan
cs.CL
Recently, the success of pre-training in text domain has been fully extended to vision, audio, and cross-modal scenarios. The proposed pre-training models of different modalities are showing a rising trend of homogeneity in their model structures, which brings the opportunity to implement different pre-training models ...
http://arxiv.org/pdf/2212.06385v2
2023-07-11T09:49:50
2022-12-13T05:46:40
{ "title": [ 0.0022010803, 0.0014448166, -0.044403076, -0.06732178, 0.0017223358, -0.014602661, -0.07159424, -0.043884277, -0.024795532, 0.018173218, 0.013137817, -0.01499176, -0.007820129, -0.062469482, -0.012397766, -0.0115356445, 0.01109314, 0...
http://arxiv.org/abs/2212.06383v1
Towards a general purpose machine translation system for Sranantongo
Just Zwennicker, David Stap
cs.CL
Machine translation for Sranantongo (Sranan, srn), a low-resource Creole language spoken predominantly in Surinam, is virgin territory. In this study we create a general purpose machine translation system for srn. In order to facilitate this research, we introduce the SRNcorpus, a collection of parallel Dutch (nl) to s...
http://arxiv.org/pdf/2212.06383v1
2022-12-13T05:36:18
2022-12-13T05:36:18
{ "title": [ -0.015464783, -0.017242432, -0.008590698, -0.05230713, -0.014297485, 0.010871887, -0.05307007, -0.0049438477, -0.0031490326, 0.055603027, -0.0006489754, -0.011543274, 0.03643799, -0.040283203, -0.059570312, 0.020324707, -0.0035381317, ...
http://arxiv.org/abs/2212.06373v6
InferEM: Inferring the Speaker's Intention for Empathetic Dialogue Generation
Guoqing Lv, Jiang Li, Xiaoping Wang, Zhigang Zeng
cs.CL, cs.HC
Current approaches to empathetic response generation typically encode the entire dialogue history directly and put the output into a decoder to generate friendly feedback. These methods focus on modelling contextual information but neglect capturing the direct intention of the speaker. We argue that the last utterance ...
http://arxiv.org/pdf/2212.06373v6
2023-04-25T08:58:41
2022-12-13T05:12:40
{ "title": [ -0.019332886, -0.0099105835, -0.05722046, -0.044403076, -0.0101623535, -0.021484375, -0.039794922, 0.0033073425, 0.006252289, 0.00995636, -0.00062799454, -0.0020561218, 0.010536194, -0.04095459, -0.02935791, -0.0053482056, -0.02519226, ...
http://arxiv.org/abs/2212.06369v3
Technical Report -- Competition Solution for Prompt Tuning using Pretrained Language Model
Jiang-Long Song, Wu-He Zou, Feng Li, Xiao-Lei Qin, Wei-Dong Zhang
cs.CL
Prompt tuning recently becomes a hot-spot in the applications of large pretrained language models on specific downstream tasks. Regarding the Language Model as a Service (LMaaS), black-box tuning using derivative-free optimization (DFO) provides a novel approach to expand the practical scenarios of pretrained models an...
http://arxiv.org/pdf/2212.06369v3
2022-12-20T07:40:37
2022-12-13T04:57:04
{ "title": [ 0.00047683716, -0.021514893, -0.01651001, -0.051513672, -0.0234375, -0.025787354, -0.055023193, 0.0031166077, -0.057800293, 0.034820557, 0.003993988, -0.011627197, 0.010574341, -0.04827881, -0.007671356, -0.02281189, 0.0026626587, 0....
http://arxiv.org/abs/2212.06346v1
The Massively Multilingual Natural Language Understanding 2022 (MMNLU-22) Workshop and Competition
Christopher Hench, Charith Peris, Jack FitzGerald, Kay Rottmann
cs.CL
Despite recent progress in Natural Language Understanding (NLU), the creation of multilingual NLU systems remains a challenge. It is common to have NLU systems limited to a subset of languages due to lack of available data. They also often vary widely in performance. We launch a three-phase approach to address the limi...
http://arxiv.org/pdf/2212.06346v1
2022-12-13T03:00:36
2022-12-13T03:00:36
{ "title": [ -0.01524353, -0.0357666, -0.04156494, -0.059906006, -0.02998352, -0.0027389526, -0.05493164, 0.03555298, -0.03250122, 0.01626587, 0.032928467, -0.023483276, 0.026123047, -0.06188965, 0.040283203, 0.0005993843, -0.0085372925, 0.003721...
http://arxiv.org/abs/2212.13899v1
Attentive Deep Neural Networks for Legal Document Retrieval
Ha-Thanh Nguyen, Manh-Kien Phi, Xuan-Bach Ngo, Vu Tran, Le-Minh Nguyen, Minh-Phuong Tu
cs.IR, cs.CL
Legal text retrieval serves as a key component in a wide range of legal text processing tasks such as legal question answering, legal case entailment, and statute law retrieval. The performance of legal text retrieval depends, to a large extent, on the representation of text, both query and legal documents. Based on go...
http://arxiv.org/pdf/2212.13899v1
2022-12-13T01:37:27
2022-12-13T01:37:27
{ "title": [ -0.015525818, -0.005836487, -0.032592773, -0.03515625, 0.0047187805, -0.0036582947, -0.02281189, -0.006126404, 0.038635254, 0.042633057, -0.035858154, -0.009643555, 0.037353516, 0.0061912537, -0.058532715, 0.01108551, -0.011634827, 0...
http://arxiv.org/abs/2212.06295v1
Despite "super-human" performance, current LLMs are unsuited for decisions about ethics and safety
Joshua Albrecht, Ellie Kitanidis, Abraham J. Fetterman
cs.CL, cs.AI
Large language models (LLMs) have exploded in popularity in the past few years and have achieved undeniably impressive results on benchmarks as varied as question answering and text summarization. We provide a simple new prompting strategy that leads to yet another supposedly "super-human" result, this time outperformi...
http://arxiv.org/pdf/2212.06295v1
2022-12-13T00:29:45
2022-12-13T00:29:45
{ "title": [ 0.00043606758, -0.012428284, 0.0015878677, 0.018707275, -0.039031982, -0.016098022, 0.045410156, -0.018753052, -0.020309448, 0.027252197, -0.039794922, 0.017730713, -0.022949219, -0.020553589, -0.009010315, -0.03503418, -0.018585205, ...
http://arxiv.org/abs/2212.06267v1
Mortality Prediction Models with Clinical Notes Using Sparse Attention at the Word and Sentence Levels
Miguel Rios, Ameen Abu-Hanna
cs.CL, cs.LG
Intensive Care in-hospital mortality prediction has various clinical applications. Neural prediction models, especially when capitalising on clinical notes, have been put forward as improvement on currently existing models. However, to be acceptable these models should be performant and transparent. This work studies d...
http://arxiv.org/pdf/2212.06267v1
2022-12-12T22:08:45
2022-12-12T22:08:45
{ "title": [ -0.006511688, -0.015350342, -0.010383606, -0.057556152, -0.03164673, 0.012168884, -0.054779053, -0.007675171, 0.02079773, 0.039520264, -0.035003662, -0.0024433136, 0.015670776, -0.0104904175, -0.04421997, 0.00064468384, 0.011116028, ...
http://arxiv.org/abs/2212.07425v3
Robust and Explainable Identification of Logical Fallacies in Natural Language Arguments
Zhivar Sourati, Vishnu Priya Prasanna Venkatesh, Darshan Deshpande, Himanshu Rawlani, Filip Ilievski, Hông-Ân Sandlin, Alain Mermoud
cs.CL, cs.AI
The spread of misinformation, propaganda, and flawed argumentation has been amplified in the Internet era. Given the volume of data and the subtlety of identifying violations of argumentation norms, supporting information analytics tasks, like content moderation, with trustworthy methods that can identify logical falla...
http://arxiv.org/pdf/2212.07425v3
2023-09-25T04:47:16
2022-12-12T20:27:17
{ "title": [ 0.038024902, 0.0068511963, -0.022750854, -0.06793213, 0.023284912, -0.022109985, -0.017059326, -0.022399902, -0.019302368, 0.014915466, -0.012969971, 0.018585205, -0.031173706, -0.015960693, -0.037322998, -0.021087646, -0.03616333, 0...
http://arxiv.org/abs/2212.08167v1
Evaluation of Synthetic Datasets for Conversational Recommender Systems
Harsh Lara, Manoj Tiwari
cs.CL, cs.AI, cs.IR, cs.LG
For researchers leveraging Large-Language Models (LLMs) in the generation of training datasets, especially for conversational recommender systems - the absence of robust evaluation frameworks has been a long-standing problem. The efficiency brought about by LLMs in the data generation phase is impeded during the proces...
http://arxiv.org/pdf/2212.08167v1
2022-12-12T18:53:10
2022-12-12T18:53:10
{ "title": [ 0.012039185, 0.008773804, -0.04083252, -0.050048828, 0.016967773, -0.013290405, -0.02078247, -0.0035209656, -0.0068511963, 0.029281616, -0.03729248, 0.015914917, -0.038513184, -0.0033740997, -0.053466797, -0.022338867, -0.029052734, ...
http://arxiv.org/abs/2212.06121v1
In Defense of Cross-Encoders for Zero-Shot Retrieval
Guilherme Rosa, Luiz Bonifacio, Vitor Jeronymo, Hugo Abonizio, Marzieh Fadaee, Roberto Lotufo, Rodrigo Nogueira
cs.IR, cs.CL
Bi-encoders and cross-encoders are widely used in many state-of-the-art retrieval pipelines. In this work we study the generalization ability of these two types of architectures on a wide range of parameter count on both in-domain and out-of-domain scenarios. We find that the number of parameters and early query-docume...
http://arxiv.org/pdf/2212.06121v1
2022-12-12T18:50:03
2022-12-12T18:50:03
{ "title": [ 0.030792236, -0.0026073456, -0.009895325, -0.040985107, -0.012626648, -0.021224976, 0.019165039, -0.018478394, 0.028030396, 0.05206299, -0.021728516, -0.0014896393, -0.012672424, -0.020309448, -0.05532837, 0.024673462, 0.007255554, 0...
http://arxiv.org/abs/2212.06094v3
Prompting Is Programming: A Query Language for Large Language Models
Luca Beurer-Kellner, Marc Fischer, Martin Vechev
cs.CL, cs.AI
Large language models have demonstrated outstanding performance on a wide range of tasks such as question answering and code generation. On a high level, given an input, a language model can be used to automatically complete the sequence in a statistically-likely way. Based on this, users prompt these models with langu...
http://arxiv.org/pdf/2212.06094v3
2023-05-30T12:56:41
2022-12-12T18:09:09
{ "title": [ -0.02659607, -0.031585693, -0.027297974, -0.08642578, -0.0052719116, 0.020263672, -0.07244873, 0.012145996, 0.0052223206, 0.022628784, -0.028823853, -0.0034770966, 0.0033473969, -0.020324707, -0.04547119, -0.0016307831, 0.0019817352, ...
http://arxiv.org/abs/2212.06002v2
Effective Seed-Guided Topic Discovery by Integrating Multiple Types of Contexts
Yu Zhang, Yunyi Zhang, Martin Michalski, Yucheng Jiang, Yu Meng, Jiawei Han
cs.CL, cs.IR
Instead of mining coherent topics from a given text corpus in a completely unsupervised manner, seed-guided topic discovery methods leverage user-provided seed words to extract distinctive and coherent topics so that the mined topics can better cater to the user's interest. To model the semantic correlation between wor...
http://arxiv.org/pdf/2212.06002v2
2023-01-11T04:22:38
2022-12-12T16:03:38
{ "title": [ -0.002445221, 0.0044021606, -0.042266846, -0.040985107, 0.021728516, -0.007896423, 0.0018634796, -0.021514893, 0.009689331, 0.026275635, -0.0385437, 0.041992188, -0.0033988953, -0.0030593872, -0.007980347, 0.008300781, -0.0017356873, ...
http://arxiv.org/abs/2212.05998v1
Continuation KD: Improved Knowledge Distillation through the Lens of Continuation Optimization
Aref Jafari, Ivan Kobyzev, Mehdi Rezagholizadeh, Pascal Poupart, Ali Ghodsi
cs.LG, cs.CL
Knowledge Distillation (KD) has been extensively used for natural language understanding (NLU) tasks to improve a small model's (a student) generalization by transferring the knowledge from a larger model (a teacher). Although KD methods achieve state-of-the-art performance in numerous settings, they suffer from severa...
http://arxiv.org/pdf/2212.05998v1
2022-12-12T16:00:20
2022-12-12T16:00:20
{ "title": [ -0.026550293, -0.014694214, -0.025299072, -0.07458496, -0.0031013489, 0.009185791, -0.05871582, -0.020568848, 0.01675415, 0.043823242, -0.008346558, 0.011940002, 0.00039625168, -0.010772705, -0.052581787, 0.018203735, 0.010414124, 0....
http://arxiv.org/abs/2212.05982v1
Real-World Compositional Generalization with Disentangled Sequence-to-Sequence Learning
Hao Zheng, Mirella Lapata
cs.CL, cs.AI
Compositional generalization is a basic mechanism in human language learning, which current neural networks struggle with. A recently proposed Disentangled sequence-to-sequence model (Dangle) shows promising generalization capability by learning specialized encodings for each decoding step. We introduce two key modific...
http://arxiv.org/pdf/2212.05982v1
2022-12-12T15:40:30
2022-12-12T15:40:30
{ "title": [ 0.0022830963, -0.0051116943, -0.052001953, -0.11053467, 0.010528564, -0.0059280396, -0.057891846, 0.03994751, 0.017593384, 0.031082153, -0.026672363, 0.025375366, -0.010551453, -0.008110046, -0.029647827, 0.004562378, -0.010009766, 0...
http://arxiv.org/abs/2212.05974v2
Federated Few-Shot Learning for Mobile NLP
Dongqi Cai, Shangguang Wang, Yaozong Wu, Felix Xiaozhu Lin, Mengwei Xu
cs.LG, cs.CL
Natural language processing (NLP) sees rich mobile applications. To support various language understanding tasks, a foundation NLP model is often fine-tuned in a federated, privacy-preserving setting (FL). This process currently relies on at least hundreds of thousands of labeled training samples from mobile clients; y...
http://arxiv.org/pdf/2212.05974v2
2023-08-19T07:31:01
2022-12-12T15:29:48
{ "title": [ -0.008735657, -0.023071289, -0.022537231, -0.04937744, -0.012512207, 0.020523071, -0.023788452, 0.020355225, -0.015464783, 0.021774292, -0.04208374, 0.02079773, -0.024932861, -0.00024867058, -0.029647827, -0.02180481, -0.007881165, -...
http://arxiv.org/abs/2212.05961v3
RPN: A Word Vector Level Data Augmentation Algorithm in Deep Learning for Language Understanding
Zhengqing Yuan, Xiaolong Zhang, Yue Wang, Xuecong Hou, Huiwen Xue, Zhuanzhe Zhao, Yongming Liu
cs.CL, cs.AI
Data augmentation is a widely used technique in machine learning to improve model performance. However, existing data augmentation techniques in natural language understanding (NLU) may not fully capture the complexity of natural language variations, and they can be challenging to apply to large datasets. This paper pr...
http://arxiv.org/pdf/2212.05961v3
2023-07-05T01:18:04
2022-12-12T15:16:12
{ "title": [ -0.014328003, -0.014404297, -0.04071045, -0.062438965, -0.003396988, -0.018508911, -0.03869629, 0.005180359, 0.008964539, 0.002292633, -0.030441284, 0.0059509277, 0.0005288124, -0.02319336, -0.016571045, -0.0002065897, -0.009033203, ...
http://arxiv.org/abs/2212.05956v2
Improving Generalization of Pre-trained Language Models via Stochastic Weight Averaging
Peng Lu, Ivan Kobyzev, Mehdi Rezagholizadeh, Ahmad Rashid, Ali Ghodsi, Philippe Langlais
cs.CL, cs.LG
Knowledge Distillation (KD) is a commonly used technique for improving the generalization of compact Pre-trained Language Models (PLMs) on downstream tasks. However, such methods impose the additional burden of training a separate teacher model for every new dataset. Alternatively, one may directly work on the improvem...
http://arxiv.org/pdf/2212.05956v2
2022-12-16T22:25:28
2022-12-12T15:09:56
{ "title": [ -0.02204895, 0.00062179565, -0.034729004, -0.07598877, 0.0022277832, 0.006965637, -0.072387695, 0.007801056, 0.0050697327, 0.03164673, -0.008796692, 0.015342712, -0.00856781, -0.014221191, -0.044006348, 0.008399963, -0.0008149147, 0....
http://arxiv.org/abs/2301.05395v1
MaNLP@SMM4H22: BERT for Classification of Twitter Posts
Keshav Kapur, Rajitha Harikrishnan
cs.CL
The reported work is our straightforward approach for the shared task Classification of tweets self-reporting age organized by the Social Media Mining for Health Applications (SMM4H) workshop. This literature describes the approach that was used to build a binary classification system, that classifies the tweets relate...
http://arxiv.org/pdf/2301.05395v1
2022-12-12T14:43:46
2022-12-12T14:43:46
{ "title": [ 0.01158905, -0.017166138, 0.006149292, -0.02557373, -0.024734497, 0.012901306, -0.036132812, 0.0012702942, -0.058532715, 0.045288086, 0.015327454, 0.025482178, 0.02204895, -0.05831909, 0.022888184, -0.020309448, -0.03137207, 0.040466...
http://arxiv.org/abs/2212.05901v1
Parameter-Efficient Finetuning of Transformers for Source Code
Shamil Ayupov, Nadezhda Chirkova
cs.CL, cs.LG, cs.SE
Pretrained Transformers achieve state-of-the-art performance in various code-processing tasks but may be too large to be deployed. As software development tools often incorporate modules for various purposes which may potentially use a single instance of the pretrained model, it appears relevant to utilize parameter-ef...
http://arxiv.org/pdf/2212.05901v1
2022-12-12T14:00:57
2022-12-12T14:00:57
{ "title": [ 0.017608643, -0.0043525696, -0.004119873, -0.119506836, 0.022964478, 0.0178833, -0.04220581, 0.014343262, -0.010787964, 0.029083252, -0.024520874, 0.019302368, -0.006084442, -0.016281128, -0.018310547, 0.0069122314, 0.0051002502, 0.0...
http://arxiv.org/abs/2212.06765v1
Earthquake Impact Analysis Based on Text Mining and Social Media Analytics
Zhe Zheng, Hong-Zheng Shi, Yu-Cheng Zhou, Xin-Zheng Lu, Jia-Rui Lin
cs.CL, cs.LG, cs.SI
Earthquakes have a deep impact on wide areas, and emergency rescue operations may benefit from social media information about the scope and extent of the disaster. Therefore, this work presents a text miningbased approach to collect and analyze social media data for early earthquake impact analysis. First, disasterrela...
http://arxiv.org/pdf/2212.06765v1
2022-12-12T13:51:07
2022-12-12T13:51:07
{ "title": [ -0.012840271, 0.016204834, -0.044677734, -0.043518066, 0.020568848, -0.0051193237, -0.0055122375, -0.0052604675, -0.02810669, 0.042144775, -0.03540039, 0.027679443, -0.02255249, -0.0067710876, -0.051330566, -0.020584106, -0.0029258728, ...
http://arxiv.org/abs/2212.05891v1
Text Mining-Based Patent Analysis for Automated Rule Checking in AEC
Zhe Zheng, Bo-Rui Kang, Qi-Tian Yuan, Yu-Cheng Zhou, Xin-Zheng Lu, Jia-Rui Lin
cs.IR, cs.CL, cs.LG
Automated rule checking (ARC), which is expected to promote the efficiency of the compliance checking process in the architecture, engineering, and construction (AEC) industry, is gaining increasing attention. Throwing light on the ARC application hotspots and forecasting its trends are useful to the related research a...
http://arxiv.org/pdf/2212.05891v1
2022-12-12T13:48:38
2022-12-12T13:48:38
{ "title": [ -0.03274536, 0.007007599, -0.038757324, -0.0362854, -0.005908966, 0.004585266, -0.045074463, -0.0158844, -0.036132812, 0.04574585, -0.056030273, 0.037506104, 0.008842468, -0.0064582825, -0.031341553, -0.021591187, -0.021316528, 0.037...
http://arxiv.org/abs/2212.05857v2
Automated ICD Coding using Extreme Multi-label Long Text Transformer-based Models
Leibo Liu, Oscar Perez-Concha, Anthony Nguyen, Vicki Bennett, Louisa Jorm
cs.CL
Background: Encouraged by the success of pretrained Transformer models in many natural language processing tasks, their use for International Classification of Diseases (ICD) coding tasks is now actively being explored. In this study, we investigate three types of Transformer-based models, aiming to address the extreme...
http://arxiv.org/pdf/2212.05857v2
2022-12-13T04:42:44
2022-12-12T12:48:33
{ "title": [ -0.006099701, 0.0059547424, -0.015258789, -0.04586792, -0.0051078796, 0.0034103394, -0.03945923, -0.022979736, -0.01977539, 0.05630493, 0.0021533966, 0.027374268, 0.0597229, -0.036895752, 0.0061302185, 0.03466797, 0.005176544, 0.0241...
http://arxiv.org/abs/2212.05856v1
"I think this is the most disruptive technology": Exploring Sentiments of ChatGPT Early Adopters using Twitter Data
Mubin Ul Haque, Isuru Dharmadasa, Zarrin Tasnim Sworna, Roshan Namal Rajapakse, Hussain Ahmad
cs.CL
Large language models have recently attracted significant attention due to their impressive performance on a variety of tasks. ChatGPT developed by OpenAI is one such implementation of a large, pre-trained language model that has gained immense popularity among early adopters, where certain users go to the extent of ch...
http://arxiv.org/pdf/2212.05856v1
2022-12-12T12:41:24
2022-12-12T12:41:24
{ "title": [ -0.016403198, -0.003545761, -0.009094238, 0.021774292, -0.019958496, -0.0069885254, 0.025558472, -0.021057129, -0.007534027, 0.02116394, -0.0435791, 0.035186768, -0.027557373, 0.006729126, -0.041015625, -0.020050049, -0.020553589, 0....
http://arxiv.org/abs/2212.05830v1
P-Transformer: Towards Better Document-to-Document Neural Machine Translation
Yachao Li, Junhui Li, Jing Jiang, Shimin Tao, Hao Yang, Min Zhang
cs.CL
Directly training a document-to-document (Doc2Doc) neural machine translation (NMT) via Transformer from scratch, especially on small datasets usually fails to converge. Our dedicated probing tasks show that 1) both the absolute position and relative position information gets gradually weakened or even vanished once it...
http://arxiv.org/pdf/2212.05830v1
2022-12-12T11:19:05
2022-12-12T11:19:05
{ "title": [ -0.02243042, 0.0007405281, -0.008895874, -0.05218506, 0.002576828, 0.017990112, -0.050048828, -0.015487671, -0.003376007, 0.049224854, -0.009613037, 0.005645752, 0.036315918, -0.0065193176, -0.034057617, 0.02330017, -0.0075569153, 0....
http://arxiv.org/abs/2212.13924v1
Page Layout Analysis of Text-heavy Historical Documents: a Comparison of Textual and Visual Approaches
Najem-Meyer Sven, Romanello Matteo
cs.IR, cs.AI, cs.CL, cs.CV
Page layout analysis is a fundamental step in document processing which enables to segment a page into regions of interest. With highly complex layouts and mixed scripts, scholarly commentaries are text-heavy documents which remain challenging for state-of-the-art models. Their layout considerably varies across edition...
http://arxiv.org/pdf/2212.13924v1
2022-12-12T10:10:29
2022-12-12T10:10:29
{ "title": [ -0.029693604, 0.008865356, -0.0413208, -0.032409668, 0.0052261353, -0.014526367, -0.03475952, -0.019042969, 0.028182983, 0.030975342, -0.0362854, 0.020019531, 0.017486572, -0.004142761, -0.03945923, 0.008903503, 0.010353088, 0.003330...
http://arxiv.org/abs/2212.05805v1
Direct Speech-to-speech Translation without Textual Annotation using Bottleneck Features
Junhui Zhang, Junjie Pan, Xiang Yin, Zejun Ma
cs.CL, cs.SD, eess.AS
Speech-to-speech translation directly translates a speech utterance to another between different languages, and has great potential in tasks such as simultaneous interpretation. State-of-art models usually contains an auxiliary module for phoneme sequences prediction, and this requires textual annotation of the trainin...
http://arxiv.org/pdf/2212.05805v1
2022-12-12T10:03:10
2022-12-12T10:03:10
{ "title": [ -0.023620605, 0.009887695, -0.035827637, -0.0637207, 0.007446289, -0.000869751, -0.039642334, 0.0016736984, -0.004501343, 0.04864502, -0.006137848, 0.004798889, 0.008377075, -0.022384644, -0.0592041, 0.008583069, 0.004322052, 0.02612...
http://arxiv.org/abs/2212.05798v2
BigText-QA: Question Answering over a Large-Scale Hybrid Knowledge Graph
Jingjing Xu, Maria Biryukov, Martin Theobald, Vinu Ellampallil Venugopal
cs.CL, cs.AI
Answering complex questions over textual resources remains a challenge, particularly when dealing with nuanced relationships between multiple entities expressed within natural-language sentences. To this end, curated knowledge bases (KBs) like YAGO, DBpedia, Freebase, and Wikidata have been widely used and gained great...
http://arxiv.org/pdf/2212.05798v2
2023-09-07T12:22:43
2022-12-12T09:49:02
{ "title": [ -0.039978027, -0.015335083, -0.043273926, -0.059631348, -0.010559082, 0.008201599, -0.042785645, 0.011924744, 0.019348145, 0.03353882, -0.032714844, -0.0069732666, -0.0011959076, -0.002128601, -0.059051514, -0.017623901, 0.0016717911, ...
http://arxiv.org/abs/2212.05789v1
Collaborating Heterogeneous Natural Language Processing Tasks via Federated Learning
Chenhe Dong, Yuexiang Xie, Bolin Ding, Ying Shen, Yaliang Li
cs.CL
The increasing privacy concerns on personal private text data promote the development of federated learning (FL) in recent years. However, the existing studies on applying FL in NLP are not suitable to coordinate participants with heterogeneous or private learning objectives. In this study, we further broaden the appli...
http://arxiv.org/pdf/2212.05789v1
2022-12-12T09:27:50
2022-12-12T09:27:50
{ "title": [ -0.012840271, 0.01574707, -0.07745361, -0.083496094, 0.03704834, -0.0032291412, -0.036468506, 0.015045166, -0.009902954, 0.03778076, -0.043762207, 0.021194458, -0.021957397, 0.011764526, -0.035491943, -0.028259277, -0.023269653, 0.02...
http://arxiv.org/abs/2212.05773v2
A Survey on Natural Language Processing for Programming
Qingfu Zhu, Xianzhen Luo, Fang Liu, Cuiyun Gao, Wanxiang Che
cs.CL
Natural language processing for programming aims to use NLP techniques to assist programming. It is increasingly prevalent for its effectiveness in improving productivity. Distinct from natural language, a programming language is highly structured and functional. Constructing a structure-based representation and a func...
http://arxiv.org/pdf/2212.05773v2
2023-08-06T02:10:07
2022-12-12T08:51:30
{ "title": [ -0.03363037, -0.030426025, -0.023223877, -0.055877686, 0.0055618286, 0.0115356445, -0.03845215, 0.009849548, 0.006839752, 0.014411926, -0.034454346, 0.00063323975, -0.01675415, -0.011940002, -0.04827881, -0.0053596497, -0.023208618, ...
http://arxiv.org/abs/2212.05767v7
A Survey of Knowledge Graph Reasoning on Graph Types: Static, Dynamic, and Multimodal
Ke Liang, Lingyuan Meng, Meng Liu, Yue Liu, Wenxuan Tu, Siwei Wang, Sihang Zhou, Xinwang Liu, Fuchun Sun
cs.AI, cs.CL, cs.IR
Knowledge graph reasoning (KGR), aiming to deduce new facts from existing facts based on mined logic rules underlying knowledge graphs (KGs), has become a fast-growing research direction. It has been proven to significantly benefit the usage of KGs in many AI applications, such as question answering, recommendation sys...
http://arxiv.org/pdf/2212.05767v7
2023-07-22T13:05:21
2022-12-12T08:40:04
{ "title": [ -0.043640137, -0.019302368, -0.071777344, -0.08288574, -0.0020713806, 0.0056419373, -0.06628418, 0.03817749, -0.0014648438, -0.0014591217, -0.0037841797, 0.0017948151, -0.01802063, -0.03274536, 0.003829956, 0.0035076141, 0.0066375732, ...
http://arxiv.org/abs/2212.05765v1
Information-Theoretic Text Hallucination Reduction for Video-grounded Dialogue
Sunjae Yoon, Eunseop Yoon, Hee Suk Yoon, Junyeong Kim, Chang D. Yoo
cs.CL, cs.CV
Video-grounded Dialogue (VGD) aims to decode an answer sentence to a question regarding a given video and dialogue context. Despite the recent success of multi-modal reasoning to generate answer sentences, existing dialogue systems still suffer from a text hallucination problem, which denotes indiscriminate text-copyin...
http://arxiv.org/pdf/2212.05765v1
2022-12-12T08:38:28
2022-12-12T08:38:28
{ "title": [ -0.0021572113, 0.02015686, -0.038330078, -0.053833008, 0.00919342, -0.00084877014, -0.023162842, -0.0009007454, 0.0001629591, 0.046875, -0.009017944, 0.012268066, 0.011962891, -0.012962341, -0.055633545, -0.003921509, -0.012336731, 0...
http://arxiv.org/abs/2212.05764v2
Domain Adaptation of Transformer-Based Models using Unlabeled Data for Relevance and Polarity Classification of German Customer Feedback
Ahmad Idrissi-Yaghir, Henning Schäfer, Nadja Bauer, Christoph M. Friedrich
cs.CL, cs.LG
Understanding customer feedback is becoming a necessity for companies to identify problems and improve their products and services. Text classification and sentiment analysis can play a major role in analyzing this data by using a variety of machine and deep learning approaches. In this work, different transformer-base...
http://arxiv.org/pdf/2212.05764v2
2023-03-08T12:29:29
2022-12-12T08:32:28
{ "title": [ -0.017532349, 0.019424438, -0.035247803, 0.0036621094, -0.014709473, -0.011711121, -0.04119873, -0.05718994, 0.00028324127, 0.044281006, 0.0047836304, -0.014839172, 0.042053223, -0.029815674, -0.038024902, 0.005367279, -0.0059051514, ...
http://arxiv.org/abs/2212.05762v3
Momentum Contrastive Pre-training for Question Answering
Minda Hu, Muzhi Li, Yasheng Wang, Irwin King
cs.CL, cs.AI
Existing pre-training methods for extractive Question Answering (QA) generate cloze-like queries different from natural questions in syntax structure, which could overfit pre-trained models to simple keyword matching. In order to address this problem, we propose a novel Momentum Contrastive pRe-training fOr queStion an...
http://arxiv.org/pdf/2212.05762v3
2023-10-14T08:44:35
2022-12-12T08:28:22
{ "title": [ 0.014793396, 0.010032654, -0.014862061, -0.054748535, 0.0071525574, -0.018081665, -0.007358551, -0.0129776, 0.011169434, 0.028945923, -0.026870728, 0.014572144, -0.0051727295, -0.005126953, -0.041137695, -0.0041122437, 0.002023697, 0...
http://arxiv.org/abs/2212.05740v1
Searching for Effective Multilingual Fine-Tuning Methods: A Case Study in Summarization
Yiwei Qin, Graham Neubig, Pengfei Liu
cs.CL
Recently, a large number of tuning strategies have been proposed to adapt pre-trained language models to downstream tasks. In this paper, we perform an extensive empirical evaluation of various tuning strategies for multilingual learning, particularly in the context of text summarization. Specifically, we explore the r...
http://arxiv.org/pdf/2212.05740v1
2022-12-12T07:37:45
2022-12-12T07:37:45
{ "title": [ -0.016448975, -0.012557983, -0.010482788, -0.068847656, -0.009384155, 0.019012451, -0.06567383, -0.008575439, -0.015129089, 0.06542969, 0.0063056946, 0.026977539, 0.001042366, 0.006374359, -0.05947876, 0.012390137, -0.010360718, 0.03...
http://arxiv.org/abs/2212.05726v1
T5Score: Discriminative Fine-tuning of Generative Evaluation Metrics
Yiwei Qin, Weizhe Yuan, Graham Neubig, Pengfei Liu
cs.CL
Modern embedding-based metrics for evaluation of generated text generally fall into one of two paradigms: discriminative metrics that are trained to directly predict which outputs are of higher quality according to supervised human annotations, and generative metrics that are trained to evaluate text based on the proba...
http://arxiv.org/pdf/2212.05726v1
2022-12-12T06:29:04
2022-12-12T06:29:04
{ "title": [ -0.018951416, -0.013450623, -0.029663086, -0.03543091, -0.0022296906, -0.0069847107, -0.009742737, -0.03479004, 0.002122879, 0.0036182404, -0.01436615, 0.0063438416, -0.028533936, -0.044555664, -0.02909851, -0.009552002, -0.014633179, ...
http://arxiv.org/abs/2301.00656v2
TriNet: stabilizing self-supervised learning from complete or slow collapse on ASR
Lixin Cao, Jun Wang, Ben Yang, Dan Su, Dong Yu
eess.AS, cs.CL, cs.LG
Self-supervised learning (SSL) models confront challenges of abrupt informational collapse or slow dimensional collapse. We propose TriNet, which introduces a novel triple-branch architecture for preventing collapse and stabilizing the pre-training. TriNet learns the SSL latent embedding space and incorporates it to a ...
http://arxiv.org/pdf/2301.00656v2
2023-03-14T12:23:33
2022-12-12T05:55:07
{ "title": [ 0.0062026978, 0.0010251999, -0.03375244, -0.06317139, 0.003768921, -0.007785797, -0.034332275, -0.014160156, -0.024810791, 0.028366089, -0.021240234, 0.011604309, -0.023406982, -0.017868042, -0.029891968, -0.019577026, 0.012420654, 0...
http://arxiv.org/abs/2212.05702v1
Implementing Deep Learning-Based Approaches for Article Summarization in Indian Languages
Rahul Tangsali, Aabha Pingle, Aditya Vyawahare, Isha Joshi, Raviraj Joshi
cs.CL, cs.LG
The research on text summarization for low-resource Indian languages has been limited due to the availability of relevant datasets. This paper presents a summary of various deep-learning approaches used for the ILSUM 2022 Indic language summarization datasets. The ISUM 2022 dataset consists of news articles written in ...
http://arxiv.org/pdf/2212.05702v1
2022-12-12T04:50:43
2022-12-12T04:50:43
{ "title": [ -0.022338867, 0.008049011, -0.025878906, -0.05279541, -0.005897522, 0.006313324, -0.05908203, -0.017852783, -0.010276794, 0.04043579, -0.018844604, 0.025375366, 0.0049209595, 0.006298065, -0.045410156, 0.01134491, -0.0033664703, 0.03...
http://arxiv.org/abs/2212.05696v1
Ensembling Transformers for Cross-domain Automatic Term Extraction
Hanh Thi Hong Tran, Matej Martinc, Andraz Pelicon, Antoine Doucet, Senja Pollak
cs.CL, cs.IR
Automatic term extraction plays an essential role in domain language understanding and several natural language processing downstream tasks. In this paper, we propose a comparative study on the predictive power of Transformers-based pretrained language models toward term extraction in a multi-language cross-domain sett...
http://arxiv.org/pdf/2212.05696v1
2022-12-12T04:20:11
2022-12-12T04:20:11
{ "title": [ -0.005584717, 0.009590149, -0.047729492, -0.070617676, 0.035247803, -0.0075531006, -0.022766113, -0.014968872, 0.0074806213, 0.031707764, -0.021469116, 0.025924683, 0.007282257, -0.0009689331, -0.034454346, 0.019805908, -0.018707275, ...
http://arxiv.org/abs/2212.05613v3
A Study of Slang Representation Methods
Aravinda Kolla, Filip Ilievski, Hông-Ân Sandlin, Alain Mermoud
cs.CL, cs.AI
Considering the large amount of content created online by the minute, slang-aware automatic tools are critically needed to promote social good, and assist policymakers and moderators in restricting the spread of offensive language, abuse, and hate speech. Despite the success of large language models and the spontaneous...
http://arxiv.org/pdf/2212.05613v3
2023-01-31T19:37:52
2022-12-11T21:56:44
{ "title": [ 0.01361084, -0.0076408386, -0.025405884, -0.005596161, -0.026290894, -0.038238525, 0.01499176, 0.006996155, 0.003376007, 0.07910156, -0.04107666, -0.0031700134, 0.06317139, -0.0067863464, -0.026824951, 0.014732361, 0.012016296, 0.037...
http://arxiv.org/abs/2212.05612v3
Multimodal and Explainable Internet Meme Classification
Abhinav Kumar Thakur, Filip Ilievski, Hông-Ân Sandlin, Zhivar Sourati, Luca Luceri, Riccardo Tommasini, Alain Mermoud
cs.AI, cs.CL, cs.LG
In the current context where online platforms have been effectively weaponized in a variety of geo-political events and social issues, Internet memes make fair content moderation at scale even more difficult. Existing work on meme classification and tracking has focused on black-box methods that do not explicitly consi...
http://arxiv.org/pdf/2212.05612v3
2023-04-07T00:57:16
2022-12-11T21:52:21
{ "title": [ 0.021408081, 0.02204895, -0.034057617, -0.033233643, 0.019012451, -0.027557373, 0.0013141632, -0.016937256, 0.020446777, 0.029388428, -0.0011806488, 0.012207031, -0.007003784, -0.019134521, -0.04144287, 0.0076408386, -0.024230957, 0....
http://arxiv.org/abs/2212.11109v1
MAViC: Multimodal Active Learning for Video Captioning
Gyanendra Das, Xavier Thomas, Anant Raj, Vikram Gupta
cs.CV, cs.CL, cs.LG
A large number of annotated video-caption pairs are required for training video captioning models, resulting in high annotation costs. Active learning can be instrumental in reducing these annotation requirements. However, active learning for video captioning is challenging because multiple semantically similar caption...
http://arxiv.org/pdf/2212.11109v1
2022-12-11T18:51:57
2022-12-11T18:51:57
{ "title": [ -0.018112183, 0.0036621094, -0.029281616, -0.038024902, 0.003250122, -0.018707275, -0.0013017654, -0.005630493, 0.005455017, 0.050079346, -0.017288208, 0.023025513, -0.0037212372, -0.007987976, -0.037628174, 0.010131836, -0.016479492, ...
http://arxiv.org/abs/2212.11121v1
Religion and Spirituality on Social Media in the Aftermath of the Global Pandemic
Olanrewaju Tahir Aduragba, Alexandra I. Cristea, Pete Phillips, Jonas Kurlberg, Jialin Yu
cs.CY, cs.AI, cs.CL, cs.SI
During the COVID-19 pandemic, the Church closed its physical doors for the first time in about 800 years, which is, arguably, a cataclysmic event. Other religions have found themselves in a similar situation, and they were practically forced to move online, which is an unprecedented occasion. In this paper, we analyse ...
http://arxiv.org/pdf/2212.11121v1
2022-12-11T18:41:02
2022-12-11T18:41:02
{ "title": [ 0.021026611, 0.006839752, -0.06451416, -0.0368042, 0.028533936, -0.05529785, 0.02204895, 0.015083313, -0.039886475, 0.044891357, -0.0053100586, 0.031829834, 0.030517578, -0.029266357, -0.0018634796, -0.009605408, 0.025161743, 0.01695...
http://arxiv.org/abs/2212.08041v1
Can REF output quality scores be assigned by AI? Experimental evidence
Mike Thelwall, Kayvan Kousha, Mahshid Abdoli, Emma Stuart, Meiko Makita, Paul Wilson, Jonathan Levitt
cs.CY, cs.CL, cs.GL, cs.IR, cs.LG
This document describes strategies for using Artificial Intelligence (AI) to predict some journal article scores in future research assessment exercises. Five strategies have been assessed.
http://arxiv.org/pdf/2212.08041v1
2022-12-11T18:32:00
2022-12-11T18:32:00
{ "title": [ -0.032714844, 0.00806427, -0.033813477, -0.0036067963, -0.03555298, -0.015014648, -0.018615723, -0.041534424, -0.00026631355, 0.029769897, -0.029571533, -0.013206482, 0.0035133362, -0.023971558, -0.044006348, -0.009254456, -0.020111084, ...
http://arxiv.org/abs/2212.05546v3
Associations Between Natural Language Processing (NLP) Enriched Social Determinants of Health and Suicide Death among US Veterans
Avijit Mitra, Richeek Pradhan, Rachel D Melamed, Kun Chen, David C Hoaglin, Katherine L Tucker, Joel I Reisman, Zhichao Yang, Weisong Liu, Jack Tsai, Hong Yu
cs.CL
Importance: Social determinants of health (SDOH) are known to be associated with increased risk of suicidal behaviors, but few studies utilized SDOH from unstructured electronic health record (EHR) notes. Objective: To investigate associations between suicide and recent SDOH, identified using structured and unstructu...
http://arxiv.org/pdf/2212.05546v3
2022-12-28T05:43:52
2022-12-11T17:15:02
{ "title": [ 0.04425049, 0.0048942566, -0.02748108, -0.076538086, 0.02255249, -0.034118652, -0.011413574, -0.0012874603, 0.0064315796, 0.016815186, -0.031433105, 0.030715942, 0.017440796, -0.03378296, -0.001074791, -0.006538391, -0.02607727, 0.03...
http://arxiv.org/abs/2212.05525v3
Extending TrOCR for Text Localization-Free OCR of Full-Page Scanned Receipt Images
Hongkuan Zhang, Edward Whittaker, Ikuo Kitagishi
cs.CL, cs.CV
Digitization of scanned receipts aims to extract text from receipt images and save it into structured documents. This is usually split into two sub-tasks: text localization and optical character recognition (OCR). Most existing OCR models only focus on the cropped text instance images, which require the bounding box in...
http://arxiv.org/pdf/2212.05525v3
2023-10-16T16:11:58
2022-12-11T15:45:26
{ "title": [ -0.017333984, -0.016174316, -0.010673523, -0.013061523, -0.04095459, -0.009025574, -0.023208618, -0.023986816, 0.017120361, 0.04046631, -0.057159424, 0.006969452, 0.01423645, -0.013023376, -0.022705078, -0.004573822, 0.0009083748, 0....
http://arxiv.org/abs/2301.00657v1
MnTTS2: An Open-Source Multi-Speaker Mongolian Text-to-Speech Synthesis Dataset
Kailin Liang, Bin Liu, Yifan Hu, Rui Liu, Feilong Bao, Guanglai Gao
eess.AS, cs.AI, cs.CL
Text-to-Speech (TTS) synthesis for low-resource languages is an attractive research issue in academia and industry nowadays. Mongolian is the official language of the Inner Mongolia Autonomous Region and a representative low-resource language spoken by over 10 million people worldwide. However, there is a relative lack...
http://arxiv.org/pdf/2301.00657v1
2022-12-11T14:55:02
2022-12-11T14:55:02
{ "title": [ -0.005695343, 0.045043945, -0.06524658, -0.07672119, 0.036193848, 0.0018148422, -0.064453125, -0.0018472672, -0.011833191, 0.047454834, -0.023757935, 0.0017290115, 0.060028076, -0.023284912, -0.026855469, -0.0016183853, -0.019989014, ...
http://arxiv.org/abs/2212.05506v2
FastClass: A Time-Efficient Approach to Weakly-Supervised Text Classification
Tingyu Xia, Yue Wang, Yuan Tian, Yi Chang
cs.CL
Weakly-supervised text classification aims to train a classifier using only class descriptions and unlabeled data. Recent research shows that keyword-driven methods can achieve state-of-the-art performance on various tasks. However, these methods not only rely on carefully-crafted class descriptions to obtain class-spe...
http://arxiv.org/pdf/2212.05506v2
2022-12-15T01:07:43
2022-12-11T13:43:22
{ "title": [ 0.002773285, -0.008834839, -0.033172607, -0.0690918, 0.025115967, -0.0064315796, -0.016494751, -0.014717102, 0.012268066, 0.021530151, -0.03515625, 0.027572632, -0.020935059, -0.009277344, -0.044189453, -0.013168335, -0.0058555603, 0...
http://arxiv.org/abs/2212.05479v1
End-to-End Speech Translation of Arabic to English Broadcast News
Fethi Bougares, Salim Jouili
cs.CL, cs.AI, cs.LG
Speech translation (ST) is the task of directly translating acoustic speech signals in a source language into text in a foreign language. ST task has been addressed, for a long time, using a pipeline approach with two modules : first an Automatic Speech Recognition (ASR) in the source language followed by a text-to-tex...
http://arxiv.org/pdf/2212.05479v1
2022-12-11T11:35:46
2022-12-11T11:35:46
{ "title": [ -0.02659607, -0.008682251, -0.03552246, -0.010292053, -0.021072388, -0.02104187, -0.00459671, 0.027053833, -0.031677246, 0.06951904, 0.028564453, 0.0008702278, 0.050323486, -0.057250977, 0.00598526, 0.024368286, 0.023223877, -0.00267...
http://arxiv.org/abs/2212.05429v1
MORTY: Structured Summarization for Targeted Information Extraction from Scholarly Articles
Mohamad Yaser Jaradeh, Markus Stocker, Sören Auer
cs.CL, cs.IR
Information extraction from scholarly articles is a challenging task due to the sizable document length and implicit information hidden in text, figures, and citations. Scholarly information extraction has various applications in exploration, archival, and curation services for digital libraries and knowledge managemen...
http://arxiv.org/pdf/2212.05429v1
2022-12-11T06:49:29
2022-12-11T06:49:29
{ "title": [ -0.01979065, 0.011146545, -0.032104492, -0.07434082, 0.015823364, 0.017333984, -0.03503418, -0.015556335, -0.004055023, 0.036621094, -0.0014190674, 0.030044556, -0.016220093, -0.0044441223, -0.052856445, 0.015556335, -0.011833191, 0....
http://arxiv.org/abs/2212.05421v3
Feature-Level Debiased Natural Language Understanding
Yougang Lyu, Piji Li, Yechang Yang, Maarten de Rijke, Pengjie Ren, Yukun Zhao, Dawei Yin, Zhaochun Ren
cs.CL
Natural language understanding (NLU) models often rely on dataset biases rather than intended task-relevant features to achieve high performance on specific datasets. As a result, these models perform poorly on datasets outside the training distribution. Some recent studies address this issue by reducing the weights of...
http://arxiv.org/pdf/2212.05421v3
2023-01-04T06:56:15
2022-12-11T06:16:14
{ "title": [ -0.018096924, -0.00089883804, -0.037231445, -0.09210205, 0.017730713, 0.00869751, -0.07299805, -0.004962921, 0.0146102905, 0.0072135925, -0.025283813, -0.02508545, -0.0056762695, -0.013504028, -0.043762207, -0.012390137, -0.007080078, ...
http://arxiv.org/abs/2212.05409v3
Towards Leaving No Indic Language Behind: Building Monolingual Corpora, Benchmark and Models for Indic Languages
Sumanth Doddapaneni, Rahul Aralikatte, Gowtham Ramesh, Shreya Goyal, Mitesh M. Khapra, Anoop Kunchukuttan, Pratyush Kumar
cs.CL
Building Natural Language Understanding (NLU) capabilities for Indic languages, which have a collective speaker base of more than one billion speakers is absolutely crucial. In this work, we aim to improve the NLU capabilities of Indic languages by making contributions along 3 important axes (i) monolingual corpora (ii...
http://arxiv.org/pdf/2212.05409v3
2023-05-24T17:05:16
2022-12-11T04:45:50
{ "title": [ -0.018951416, -0.00057935715, -0.053955078, -0.060150146, -0.007095337, -0.005088806, -0.06915283, 0.013404846, -0.027816772, 0.01940918, -0.023803711, 0.022781372, -0.012649536, -0.01626587, -0.022445679, -0.016586304, -0.008613586, ...
http://arxiv.org/abs/2301.04120v1
BASPRO: a balanced script producer for speech corpus collection based on the genetic algorithm
Yu-Wen Chen, Hsin-Min Wang, Yu Tsao
cs.NE, cs.AI, cs.CL, cs.LG, eess.AS
The performance of speech-processing models is heavily influenced by the speech corpus that is used for training and evaluation. In this study, we propose BAlanced Script PROducer (BASPRO) system, which can automatically construct a phonetically balanced and rich set of Chinese sentences for collecting Mandarin Chinese...
http://arxiv.org/pdf/2301.04120v1
2022-12-11T02:05:30
2022-12-11T02:05:30
{ "title": [ -0.024520874, -0.008369446, -0.057373047, -0.069885254, -0.0010032654, -0.0044441223, -0.05102539, 0.030517578, 0.011764526, 0.028884888, -0.029388428, 0.024749756, -0.02230835, -0.020690918, -0.056274414, -0.0110321045, -0.017044067, ...