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/2302.01626v1
Modeling Sequential Sentence Relation to Improve Cross-lingual Dense Retrieval
Shunyu Zhang, Yaobo Liang, Ming Gong, Daxin Jiang, Nan Duan
cs.CL, cs.IR
Recently multi-lingual pre-trained language models (PLM) such as mBERT and XLM-R have achieved impressive strides in cross-lingual dense retrieval. Despite its successes, they are general-purpose PLM while the multilingual PLM tailored for cross-lingual retrieval is still unexplored. Motivated by an observation that th...
http://arxiv.org/pdf/2302.01626v1
2023-02-03T09:54:27
2023-02-03T09:54:27
{ "title": [ 0.0154953, 0.0014324188, -0.010139465, -0.07543945, 0.0020427704, 0.010215759, -0.044952393, -0.016189575, 0.0362854, 0.048675537, -0.01083374, 0.013763428, 0.033966064, -0.003540039, -0.051940918, 0.033447266, -0.006690979, 0.027694...
http://arxiv.org/abs/2302.01614v1
Around the world in 60 words: A generative vocabulary test for online research
Pol van Rijn, Yue Sun, Harin Lee, Raja Marjieh, Ilia Sucholutsky, Francesca Lanzarini, Elisabeth André, Nori Jacoby
cs.CL
Conducting experiments with diverse participants in their native languages can uncover insights into culture, cognition, and language that may not be revealed otherwise. However, conducting these experiments online makes it difficult to validate self-reported language proficiency. Furthermore, existing proficiency test...
http://arxiv.org/pdf/2302.01614v1
2023-02-03T09:27:12
2023-02-03T09:27:12
{ "title": [ -0.025497437, -0.0013599396, -0.06762695, -0.025299072, 0.012641907, -0.032836914, -0.017425537, -0.009422302, -0.0041236877, 0.0044174194, -0.045410156, -0.004699707, 0.00047659874, -0.027557373, -0.008026123, -0.035217285, 0.0060539246...
http://arxiv.org/abs/2302.07138v1
The Heritage Digital Twin: a bicycle made for two. The integration of digital methodologies into cultural heritage research
Franco Niccolucci, Béatrice Markhoff, Maria Theodoridou, Achille Felicetti, Sorin Hermon
cs.CY, cs.CL, cs.IR, E.1; H.2.3; H.3.1; J.5
The paper concerns the definition of a novel ontology for cultural heritage based on the concept of digital twin. The ontology, called Heritage Digital Twin ontology, is a compatible extension of the well-known CIDOC CRM ISO standard for cultural heritage documentation and incorporates all the different documentation s...
http://arxiv.org/pdf/2302.07138v1
2023-02-03T08:36:37
2023-02-03T08:36:37
{ "title": [ -0.0045700073, 0.0033111572, -0.04159546, 0.019088745, 0.0014505386, -0.04537964, 0.041656494, -0.009422302, -0.0069618225, 0.027175903, -0.01965332, 0.02128601, 0.01939392, -0.0395813, -0.004016876, -0.00806427, 0.013046265, 0.01828...
http://arxiv.org/abs/2302.01588v1
Bioformer: an efficient transformer language model for biomedical text mining
Li Fang, Qingyu Chen, Chih-Hsuan Wei, Zhiyong Lu, Kai Wang
cs.CL
Pretrained language models such as Bidirectional Encoder Representations from Transformers (BERT) have achieved state-of-the-art performance in natural language processing (NLP) tasks. Recently, BERT has been adapted to the biomedical domain. Despite the effectiveness, these models have hundreds of millions of paramete...
http://arxiv.org/pdf/2302.01588v1
2023-02-03T08:04:59
2023-02-03T08:04:59
{ "title": [ -0.025772095, -0.013145447, -0.046020508, -0.055358887, 0.0068626404, 0.018569946, -0.06427002, -0.0030517578, 0.020217896, 0.031433105, -0.022415161, 0.0019702911, 0.018234253, -0.0076446533, -0.053375244, 0.0063171387, -0.017074585, ...
http://arxiv.org/abs/2302.01582v1
Controlling for Stereotypes in Multimodal Language Model Evaluation
Manuj Malik, Richard Johansson
cs.CL
We propose a methodology and design two benchmark sets for measuring to what extent language-and-vision language models use the visual signal in the presence or absence of stereotypes. The first benchmark is designed to test for stereotypical colors of common objects, while the second benchmark considers gender stereot...
http://arxiv.org/pdf/2302.01582v1
2023-02-03T07:27:50
2023-02-03T07:27:50
{ "title": [ -0.022872925, 0.012748718, -0.041503906, -0.027999878, -0.0058174133, -0.01763916, -0.039367676, -0.0211792, -0.018493652, 0.043670654, -0.0079422, -0.0058937073, 0.008865356, -0.021942139, -0.049835205, -0.0058898926, -0.04156494, 0...
http://arxiv.org/abs/2302.01570v1
Witgenstein's influence on artificial intelligence
Piero Molino, Jacopo Tagliabue
cs.AI, cs.CL
We examine how much of the contemporary progress in artificial intelligence (and, specifically, in natural language processing), can be, more or less directly, traced back to the seminal work and ideas of the Austrian-British philosopher Ludwig Wittgenstein, with particular focus on his late views. Discussing Wittgenst...
http://arxiv.org/pdf/2302.01570v1
2023-02-03T06:47:20
2023-02-03T06:47:20
{ "title": [ 0.005138397, 0.01159668, -0.024337769, 0.005001068, -0.017700195, -0.018554688, 0.04107666, -0.021102905, 0.0112838745, 0.041229248, -0.037231445, -0.02229309, 0.015617371, -0.036712646, -0.019760132, -0.0014858246, 0.0038337708, 0.0...
http://arxiv.org/abs/2302.02759v1
Detecting Reddit Users with Depression Using a Hybrid Neural Network
Ziyi Chen, Ren Yang, Sunyang Fu, Nansu Zong, Hongfang Liu, Ming Huang
cs.CL, cs.AI, cs.LG
Depression is a widespread mental health issue, affecting an estimated 3.8% of the global population. It is also one of the main contributors to disability worldwide. Recently it is becoming popular for individuals to use social media platforms (e.g., Reddit) to express their difficulties and health issues (e.g., depre...
http://arxiv.org/pdf/2302.02759v1
2023-02-03T06:22:18
2023-02-03T06:22:18
{ "title": [ 0.035614014, 0.018875122, -0.004852295, -0.021728516, 0.02532959, -0.022506714, 0.010925293, -0.043914795, -0.008148193, 0.046813965, -0.007827759, -0.00059986115, 0.019439697, -0.021438599, -0.052886963, 0.0064964294, -0.008522034, ...
http://arxiv.org/abs/2302.01536v1
Using natural language processing and structured medical data to phenotype patients hospitalized due to COVID-19
Feier Chang, Jay Krishnan, Jillian H Hurst, Michael E Yarrington, Deverick J Anderson, Emily C O'Brien, Benjamin A Goldstein
cs.CL, cs.LG, stat.ML
To identify patients who are hospitalized because of COVID-19 as opposed to those who were admitted for other indications, we compared the performance of different computable phenotype definitions for COVID-19 hospitalizations that use different types of data from the electronic health records (EHR), including structur...
http://arxiv.org/pdf/2302.01536v1
2023-02-03T04:22:29
2023-02-03T04:22:29
{ "title": [ 0.009277344, -0.0065193176, -0.05630493, -0.07421875, -0.0052337646, -0.011367798, -0.020584106, 0.042877197, -0.014579773, 0.009788513, -0.017410278, -0.0025806427, 0.018692017, -0.030441284, 0.033233643, -0.02861023, 0.0025997162, ...
http://arxiv.org/abs/2302.01530v1
Revisiting Intermediate Layer Distillation for Compressing Language Models: An Overfitting Perspective
Jongwoo Ko, Seungjoon Park, Minchan Jeong, Sukjin Hong, Euijai Ahn, Du-Seong Chang, Se-Young Yun
cs.CL, cs.AI, cs.LG
Knowledge distillation (KD) is a highly promising method for mitigating the computational problems of pre-trained language models (PLMs). Among various KD approaches, Intermediate Layer Distillation (ILD) has been a de facto standard KD method with its performance efficacy in the NLP field. In this paper, we find that ...
http://arxiv.org/pdf/2302.01530v1
2023-02-03T04:09:22
2023-02-03T04:09:22
{ "title": [ -0.017654419, -0.016830444, -0.035736084, -0.06347656, -0.008354187, 0.008125305, -0.059173584, 0.021408081, 0.014984131, 0.028839111, 0.0010213852, -0.035736084, 0.0016880035, -0.035308838, -0.057373047, -0.0028076172, -0.00843811, ...
http://arxiv.org/abs/2302.01496v1
Efficient Domain Adaptation for Speech Foundation Models
Bo Li, Dongseong Hwang, Zhouyuan Huo, Junwen Bai, Guru Prakash, Tara N. Sainath, Khe Chai Sim, Yu Zhang, Wei Han, Trevor Strohman, Francoise Beaufays
cs.CL, cs.LG, cs.SD, eess.AS
Foundation models (FMs), that are trained on broad data at scale and are adaptable to a wide range of downstream tasks, have brought large interest in the research community. Benefiting from the diverse data sources such as different modalities, languages and application domains, foundation models have demonstrated str...
http://arxiv.org/pdf/2302.01496v1
2023-02-03T02:10:35
2023-02-03T02:10:35
{ "title": [ -0.024108887, -0.00793457, -0.05508423, -0.08911133, 0.008934021, -0.0074920654, -0.06774902, 0.02760315, 0.0025138855, 0.04095459, 0.012817383, 0.0010108948, 0.019577026, -0.028793335, -0.029769897, 0.020355225, -0.020324707, 0.0427...
http://arxiv.org/abs/2302.01451v2
CTE: A Dataset for Contextualized Table Extraction
Andrea Gemelli, Emanuele Vivoli, Simone Marinai
cs.CL, cs.CV
Relevant information in documents is often summarized in tables, helping the reader to identify useful facts. Most benchmark datasets support either document layout analysis or table understanding, but lack in providing data to apply both tasks in a unified way. We define the task of Contextualized Table Extraction (CT...
http://arxiv.org/pdf/2302.01451v2
2023-02-13T18:22:57
2023-02-02T22:38:23
{ "title": [ -0.00932312, 0.0131073, -0.07092285, -0.07104492, 0.03427124, -0.012191772, -0.044708252, -0.0124053955, 0.02015686, 0.05581665, -0.031021118, 0.027511597, 0.00028681755, 0.0126953125, -0.06335449, 0.011398315, 0.0006451607, 0.027023...
http://arxiv.org/abs/2302.01441v1
Commonsense-Aware Prompting for Controllable Empathetic Dialogue Generation
Yiren Liu, Halil Kilicoglu
cs.CL, cs.LG
Improving the emotional awareness of pre-trained language models is an emerging important problem for dialogue generation tasks. Although prior studies have introduced methods to improve empathetic dialogue generation, few have discussed how to incorporate commonsense knowledge into pre-trained language models for cont...
http://arxiv.org/pdf/2302.01441v1
2023-02-02T22:04:07
2023-02-02T22:04:07
{ "title": [ 0.010261536, -0.028884888, -0.020309448, -0.06048584, -0.015930176, -0.007160187, -0.0284729, 0.011711121, 0.014228821, 0.005508423, -0.0017719269, -0.0044822693, -0.0027236938, -0.0413208, -0.025512695, -0.0038585663, -0.019851685, ...
http://arxiv.org/abs/2302.01398v1
The unreasonable effectiveness of few-shot learning for machine translation
Xavier Garcia, Yamini Bansal, Colin Cherry, George Foster, Maxim Krikun, Fangxiaoyu Feng, Melvin Johnson, Orhan Firat
cs.CL
We demonstrate the potential of few-shot translation systems, trained with unpaired language data, for both high and low-resource language pairs. We show that with only 5 examples of high-quality translation data shown at inference, a transformer decoder-only model trained solely with self-supervised learning, is able ...
http://arxiv.org/pdf/2302.01398v1
2023-02-02T20:19:46
2023-02-02T20:19:46
{ "title": [ -0.027816772, -0.027770996, 0.0006046295, -0.014976501, -0.033111572, 0.002878189, -0.017166138, -0.00038456917, -0.004589081, 0.037597656, -0.0154418945, 0.011169434, 0.008514404, -0.021957397, -0.054016113, 0.0030174255, 0.009399414, ...
http://arxiv.org/abs/2302.01328v3
IC3: Image Captioning by Committee Consensus
David M. Chan, Austin Myers, Sudheendra Vijayanarasimhan, David A. Ross, John Canny
cs.CV, cs.AI, cs.CL, cs.LG
If you ask a human to describe an image, they might do so in a thousand different ways. Traditionally, image captioning models are trained to generate a single "best" (most like a reference) image caption. Unfortunately, doing so encourages captions that are "informationally impoverished," and focus on only a subset of...
http://arxiv.org/pdf/2302.01328v3
2023-10-19T17:58:05
2023-02-02T18:58:05
{ "title": [ -0.021453857, 0.0012569427, -0.039093018, -0.013595581, -0.022506714, -0.02746582, -0.026824951, -0.02999878, 0.011260986, 0.049987793, -0.0385437, 0.024810791, 0.019805908, -0.012214661, -0.033325195, -0.0014867783, -0.0022697449, 0...
http://arxiv.org/abs/2302.01318v1
Accelerating Large Language Model Decoding with Speculative Sampling
Charlie Chen, Sebastian Borgeaud, Geoffrey Irving, Jean-Baptiste Lespiau, Laurent Sifre, John Jumper
cs.CL
We present speculative sampling, an algorithm for accelerating transformer decoding by enabling the generation of multiple tokens from each transformer call. Our algorithm relies on the observation that the latency of parallel scoring of short continuations, generated by a faster but less powerful draft model, is compa...
http://arxiv.org/pdf/2302.01318v1
2023-02-02T18:44:11
2023-02-02T18:44:11
{ "title": [ -0.057495117, -0.024642944, -0.031982422, -0.072753906, -0.027877808, 0.020217896, -0.07513428, 0.026138306, 0.008644104, 0.052215576, -0.019973755, -0.008811951, 0.015975952, -0.0045280457, -0.061157227, 0.021774292, -0.002746582, 0...
http://arxiv.org/abs/2302.01313v6
Double Equivariance for Inductive Link Prediction for Both New Nodes and New Relation Types
Jianfei Gao, Yangze Zhou, Jincheng Zhou, Bruno Ribeiro
cs.LG, cs.AI, cs.CL
The task of inductive link prediction in discrete attributed multigraphs (e.g., knowledge graphs, multilayer networks, heterogeneous networks, etc.) generally focuses on test predictions with solely new nodes but not both new nodes and new relation types. In this work, we formally define the task of predicting (complet...
http://arxiv.org/pdf/2302.01313v6
2023-10-03T19:31:48
2023-02-02T18:39:30
{ "title": [ -0.003692627, -0.021560669, -0.059448242, -0.050628662, -0.013084412, -0.030136108, -0.04232788, 0.011199951, -0.012229919, 0.0020008087, -0.026489258, 0.023284912, 0.020019531, -0.03717041, 0.037017822, 0.006965637, -0.0059661865, 0...
http://arxiv.org/abs/2302.01308v2
Large language models predict human sensory judgments across six modalities
Raja Marjieh, Ilia Sucholutsky, Pol van Rijn, Nori Jacoby, Thomas L. Griffiths
cs.CL, cs.LG, stat.ML
Determining the extent to which the perceptual world can be recovered from language is a longstanding problem in philosophy and cognitive science. We show that state-of-the-art large language models can unlock new insights into this problem by providing a lower bound on the amount of perceptual information that can be ...
http://arxiv.org/pdf/2302.01308v2
2023-06-15T17:18:04
2023-02-02T18:32:46
{ "title": [ -0.015007019, 0.0071983337, -0.04019165, -0.06506348, -0.006046295, -0.015472412, -0.053222656, 0.0011444092, 0.01360321, 0.04168701, 0.005268097, -0.024230957, 0.033081055, -0.033721924, -0.039001465, 0.014137268, 0.007511139, 0.035...
http://arxiv.org/abs/2302.01148v1
Combining Deep Neural Reranking and Unsupervised Extraction for Multi-Query Focused Summarization
Philipp Seeberger, Korbinian Riedhammer
cs.CL
The CrisisFACTS Track aims to tackle challenges such as multi-stream fact-finding in the domain of event tracking; participants' systems extract important facts from several disaster-related events while incorporating the temporal order. We propose a combination of retrieval, reranking, and the well-known Integer Linea...
http://arxiv.org/pdf/2302.01148v1
2023-02-02T15:08:25
2023-02-02T15:08:25
{ "title": [ 0.00995636, 0.015449524, -0.036224365, -0.09106445, 0.014060974, 0.0034160614, -0.05114746, -0.013366699, 0.010414124, 0.045684814, 0.0039863586, 0.028060913, -0.0061073303, -0.003967285, -0.03945923, 0.01991272, -0.0037174225, -0.00...
http://arxiv.org/abs/2302.01097v1
New Linear-time Algorithm for SubTree Kernel Computation based on Root-Weighted Tree Automata
Ludovic Mignot, Faissal Ouardi, Djelloul Ziadi
cs.CL
Tree kernels have been proposed to be used in many areas as the automatic learning of natural language applications. In this paper, we propose a new linear time algorithm based on the concept of weighted tree automata for SubTree kernel computation. First, we introduce a new class of weighted tree automata, called Root...
http://arxiv.org/pdf/2302.01097v1
2023-02-02T13:37:48
2023-02-02T13:37:48
{ "title": [ -0.053863525, -0.014144897, -0.029388428, -0.06866455, -0.022338867, 0.013153076, -0.06604004, 0.012306213, 0.018585205, 0.04373169, -0.015304565, -0.018249512, -0.017990112, -0.0178833, -0.07873535, -0.018859863, 0.0010404587, 0.004...
http://arxiv.org/abs/2302.01025v1
Semantic Coherence Markers for the Early Diagnosis of the Alzheimer Disease
Davide Colla, Matteo Delsanto, Marco Agosto, Benedetto Vitiello, Daniele Paolo Radicioni
cs.CL, cs.AI
In this work we explore how language models can be employed to analyze language and discriminate between mentally impaired and healthy subjects through the perplexity metric. Perplexity was originally conceived as an information-theoretic measure to assess how much a given language model is suited to predict a text seq...
http://arxiv.org/pdf/2302.01025v1
2023-02-02T11:40:16
2023-02-02T11:40:16
{ "title": [ -0.011650085, 0.0017595291, -0.05142212, -0.04916382, -0.0044937134, -0.008087158, -0.037506104, -0.015167236, -0.024932861, 0.042053223, 0.0016222, -0.0033931732, 0.02758789, -0.041259766, -0.0115356445, 0.017700195, 0.0067749023, 0...
http://arxiv.org/abs/2302.01342v2
Curriculum-Guided Abstractive Summarization
Sajad Sotudeh, Hanieh Deilamsalehy, Franck Dernoncourt, Nazli Goharian
cs.CL
Recent Transformer-based summarization models have provided a promising approach to abstractive summarization. They go beyond sentence selection and extractive strategies to deal with more complicated tasks such as novel word generation and sentence paraphrasing. Nonetheless, these models have two shortcomings: (1) the...
http://arxiv.org/pdf/2302.01342v2
2023-02-08T05:58:11
2023-02-02T11:09:37
{ "title": [ -0.003730774, -0.007850647, -0.030685425, -0.07067871, -0.040008545, -0.009048462, -0.0513916, -0.0065460205, -0.00541687, 0.038360596, -0.027938843, 0.027404785, -0.025177002, -0.021759033, 0.00076055527, 0.0014047623, 0.0046043396, ...
http://arxiv.org/abs/2302.01005v1
Predefined domain specific embeddings of food concepts and recipes: A case study on heterogeneous recipe datasets
Gordana Ispirova, Tome Eftimov, Barbara Koroušić Seljak
cs.CL, cs.AI
Although recipe data are very easy to come by nowadays, it is really hard to find a complete recipe dataset - with a list of ingredients, nutrient values per ingredient, and per recipe, allergens, etc. Recipe datasets are usually collected from social media websites where users post and publish recipes. Usually written...
http://arxiv.org/pdf/2302.01005v1
2023-02-02T10:49:06
2023-02-02T10:49:06
{ "title": [ 0.030532837, 0.005645752, -0.024917603, -0.054718018, 0.014785767, -0.022628784, -0.009529114, -0.00091362, 0.023727417, 0.0287323, -0.026275635, 0.0158844, 0.017837524, -0.0099487305, -0.012039185, 0.0064849854, -0.008255005, 0.0477...
http://arxiv.org/abs/2302.00954v1
Curriculum-guided Abstractive Summarization for Mental Health Online Posts
Sajad Sotudeh, Nazli Goharian, Hanieh Deilamsalehy, Franck Dernoncourt
cs.CL, cs.AI
Automatically generating short summaries from users' online mental health posts could save counselors' reading time and reduce their fatigue so that they can provide timely responses to those seeking help for improving their mental state. Recent Transformers-based summarization models have presented a promising approac...
http://arxiv.org/pdf/2302.00954v1
2023-02-02T08:48:26
2023-02-02T08:48:26
{ "title": [ 0.048553467, -0.009666443, -0.005443573, -0.06933594, -0.013122559, -0.018310547, -0.029281616, -0.010040283, -0.016036987, 0.038391113, -0.01776123, 0.033111572, 0.027252197, -0.027435303, 0.004825592, 0.005493164, 0.019546509, 0.01...
http://arxiv.org/abs/2302.00944v2
TransFool: An Adversarial Attack against Neural Machine Translation Models
Sahar Sadrizadeh, Ljiljana Dolamic, Pascal Frossard
cs.CL
Deep neural networks have been shown to be vulnerable to small perturbations of their inputs, known as adversarial attacks. In this paper, we investigate the vulnerability of Neural Machine Translation (NMT) models to adversarial attacks and propose a new attack algorithm called TransFool. To fool NMT models, TransFool...
http://arxiv.org/pdf/2302.00944v2
2023-06-16T13:24:15
2023-02-02T08:35:34
{ "title": [ -0.026153564, -0.016662598, -0.018844604, -0.031921387, 0.005382538, 0.0046539307, -0.006526947, 0.003320694, -0.041748047, 0.043548584, 0.010475159, 0.01436615, 0.004005432, -0.019638062, -0.036743164, -0.0016946793, -0.017410278, 0...
http://arxiv.org/abs/2302.00937v1
The Fewer Splits are Better: Deconstructing Readability in Sentence Splitting
Tadashi Nomoto
cs.CL, cs.AI
In this work, we focus on sentence splitting, a subfield of text simplification, motivated largely by an unproven idea that if you divide a sentence in pieces, it should become easier to understand. Our primary goal in this paper is to find out whether this is true. In particular, we ask, does it matter whether we brea...
http://arxiv.org/pdf/2302.00937v1
2023-02-02T08:25:48
2023-02-02T08:25:48
{ "title": [ -0.028152466, -0.0022220612, -0.01914978, -0.046173096, -0.008934021, 0.011398315, -0.049224854, -0.025909424, -0.012504578, 0.02217102, -0.020477295, 0.010215759, 0.0038642883, -0.007297516, -0.042633057, -0.0060424805, -0.01033783, ...
http://arxiv.org/abs/2302.00923v4
Multimodal Chain-of-Thought Reasoning in Language Models
Zhuosheng Zhang, Aston Zhang, Mu Li, Hai Zhao, George Karypis, Alex Smola
cs.CL, cs.AI, cs.CV
Large language models (LLMs) have shown impressive performance on complex reasoning by leveraging chain-of-thought (CoT) prompting to generate intermediate reasoning chains as the rationale to infer the answer. However, existing CoT studies have focused on the language modality. We propose Multimodal-CoT that incorpora...
http://arxiv.org/pdf/2302.00923v4
2023-02-17T04:35:55
2023-02-02T07:51:19
{ "title": [ -0.026275635, -0.010055542, -0.027130127, -0.05618286, -0.01928711, 0.0028305054, -0.062561035, 0.0052604675, 0.017211914, 0.017730713, -0.015380859, -0.017318726, 0.016479492, -0.027008057, -0.0362854, 0.002155304, -0.021224976, 0.0...
http://arxiv.org/abs/2302.00907v1
History-Aware Hierarchical Transformer for Multi-session Open-domain Dialogue System
Tong Zhang, Yong Liu, Boyang Li, Zhiwei Zeng, Pengwei Wang, Yuan You, Chunyan Miao, Lizhen Cui
cs.CL
With the evolution of pre-trained language models, current open-domain dialogue systems have achieved great progress in conducting one-session conversations. In contrast, Multi-Session Conversation (MSC), which consists of multiple sessions over a long term with the same user, is under-investigated. In this paper, we p...
http://arxiv.org/pdf/2302.00907v1
2023-02-02T06:54:33
2023-02-02T06:54:33
{ "title": [ 0.036834717, 0.027175903, -0.042297363, -0.09967041, 0.033233643, -0.0042648315, -0.041381836, 0.0034236908, 0.0012779236, 0.036834717, 0.005393982, 0.0126953125, 0.020233154, -0.030532837, -0.030303955, -0.0043144226, -0.018295288, ...
http://arxiv.org/abs/2302.00902v2
Language Quantized AutoEncoders: Towards Unsupervised Text-Image Alignment
Hao Liu, Wilson Yan, Pieter Abbeel
cs.LG, cs.CL, cs.CV
Recent progress in scaling up large language models has shown impressive capabilities in performing few-shot learning across a wide range of text-based tasks. However, a key limitation is that these language models fundamentally lack visual perception - a crucial attribute needed to extend these models to be able to in...
http://arxiv.org/pdf/2302.00902v2
2023-02-03T05:06:46
2023-02-02T06:38:44
{ "title": [ -0.031951904, 0.022659302, -0.044555664, -0.014160156, -0.005306244, -0.017623901, -0.053009033, -0.032714844, 0.013305664, 0.07775879, -0.0064430237, 0.012687683, 0.037597656, 0.0022678375, -0.058380127, 0.029678345, 0.0062179565, 0...
http://arxiv.org/abs/2302.00894v1
How to choose "Good" Samples for Text Data Augmentation
Xiaotian Lin, Nankai Lin, Yingwen Fu, Ziyu Yang, Shengyi Jiang
cs.CL
Deep learning-based text classification models need abundant labeled data to obtain competitive performance. Unfortunately, annotating large-size corpus is time-consuming and laborious. To tackle this, multiple researches try to use data augmentation to expand the corpus size. However, data augmentation may potentially...
http://arxiv.org/pdf/2302.00894v1
2023-02-02T06:01:50
2023-02-02T06:01:50
{ "title": [ 0.028671265, 0.0025959015, -0.015281677, -0.028335571, -0.029434204, -0.018478394, -0.026794434, -0.013900757, -0.025741577, 0.04321289, -0.053497314, 0.0043182373, 0.011024475, -0.0030002594, -0.005882263, -0.02772522, -0.0073127747, ...
http://arxiv.org/abs/2302.00871v3
Using In-Context Learning to Improve Dialogue Safety
Nicholas Meade, Spandana Gella, Devamanyu Hazarika, Prakhar Gupta, Di Jin, Siva Reddy, Yang Liu, Dilek Hakkani-Tür
cs.CL
While large neural-based conversational models have become increasingly proficient dialogue agents, recent work has highlighted safety issues with these systems. For example, these systems can be goaded into generating toxic content, which often perpetuates social biases or stereotypes. We investigate a retrieval-based...
http://arxiv.org/pdf/2302.00871v3
2023-10-22T19:28:24
2023-02-02T04:46:03
{ "title": [ 0.033294678, 0.00009536743, -0.043121338, -0.084350586, 0.027374268, -0.019989014, -0.02217102, 0.008354187, -0.02645874, 0.037322998, -0.017669678, 0.022781372, -0.0062713623, -0.021575928, -0.017913818, -0.013175964, -0.013435364, ...
http://arxiv.org/abs/2302.00856v2
idT5: Indonesian Version of Multilingual T5 Transformer
Mukhlish Fuadi, Adhi Dharma Wibawa, Surya Sumpeno
cs.CL, I.2.7
Indonesian language is spoken by almost 200 million people and is the 10th most spoken language in the world, but it is under-represented in NLP (Natural Language Processing) research. A sparsity of language resources has hampered previous work on Indonesian. The Transformer is a new architecture rapidly becoming domin...
http://arxiv.org/pdf/2302.00856v2
2023-11-09T08:47:19
2023-02-02T03:56:16
{ "title": [ -0.0006313324, -0.005680084, -0.012832642, -0.012992859, -0.056488037, 0.01134491, -0.056915283, -0.017364502, -0.034118652, 0.05029297, 0.006526947, -0.0062179565, 0.048217773, -0.05581665, 0.022491455, -0.013168335, 0.010375977, 0....
http://arxiv.org/abs/2302.00836v1
Improving Rare Words Recognition through Homophone Extension and Unified Writing for Low-resource Cantonese Speech Recognition
HoLam Chung, Junan Li, Pengfei Liu1, Wai-Kim Leung, Xixin Wu, Helen Meng
cs.CL, cs.SD, eess.AS
Homophone characters are common in tonal syllable-based languages, such as Mandarin and Cantonese. The data-intensive end-to-end Automatic Speech Recognition (ASR) systems are more likely to mis-recognize homophone characters and rare words under low-resource settings. For the problem of lowresource Cantonese speech re...
http://arxiv.org/pdf/2302.00836v1
2023-02-02T02:46:32
2023-02-02T02:46:32
{ "title": [ -0.005142212, 0.0072898865, -0.041046143, -0.043884277, -0.011016846, 0.0023841858, -0.049713135, -0.0050315857, -0.03463745, 0.06213379, -0.06707764, 0.029052734, 0.018432617, 0.0043525696, -0.038238525, -0.023345947, -0.0178833, 0....
http://arxiv.org/abs/2302.01339v2
Creating a Large Language Model of a Philosopher
Eric Schwitzgebel, David Schwitzgebel, Anna Strasser
cs.CL, cs.AI
Can large language models be trained to produce philosophical texts that are difficult to distinguish from texts produced by human philosophers? To address this question, we fine-tuned OpenAI's GPT-3 with the works of philosopher Daniel C. Dennett as additional training data. To explore the Dennett model, we asked the ...
http://arxiv.org/pdf/2302.01339v2
2023-05-09T17:06:43
2023-02-02T01:10:26
{ "title": [ -0.018630981, -0.016983032, -0.041015625, -0.032104492, -0.053253174, -0.012077332, -0.052825928, 0.025375366, -0.007972717, 0.044647217, -0.024993896, -0.023025513, 0.05114746, -0.037902832, 0.013938904, -0.00006765127, 0.020523071, ...
http://arxiv.org/abs/2302.00796v2
Unsupervised Entity Alignment for Temporal Knowledge Graphs
Xiaoze Liu, Junyang Wu, Tianyi Li, Lu Chen, Yunjun Gao
cs.IR, cs.AI, cs.CL
Entity alignment (EA) is a fundamental data integration task that identifies equivalent entities between different knowledge graphs (KGs). Temporal Knowledge graphs (TKGs) extend traditional knowledge graphs by introducing timestamps, which have received increasing attention. State-of-the-art time-aware EA studies have...
http://arxiv.org/pdf/2302.00796v2
2023-02-08T08:33:25
2023-02-01T23:03:22
{ "title": [ -0.023376465, 0.0025672913, -0.10357666, -0.10076904, 0.031799316, -0.023254395, -0.05581665, 0.035705566, 0.027923584, 0.044525146, -0.013710022, -0.00030231476, 0.021331787, -0.0055007935, -0.052642822, 0.011894226, 0.002506256, 0....
http://arxiv.org/abs/2302.00778v1
User Study for Improving Tools for Bible Translation
Joel Mathew, Ulf Hermjakob
cs.CL, cs.HC
Technology has increasingly become an integral part of the Bible translation process. Over time, both the translation process and relevant technology have evolved greatly. More recently, the field of Natural Language Processing (NLP) has made great progress in solving some problems previously thought impenetrable. Thro...
http://arxiv.org/pdf/2302.00778v1
2023-02-01T22:22:03
2023-02-01T22:22:03
{ "title": [ -0.010658264, -0.017410278, -0.021759033, 0.015823364, -0.033966064, -0.030960083, 0.017593384, -0.009361267, -0.028808594, 0.09442139, -0.0029773712, 0.008460999, 0.11212158, -0.016296387, -0.011291504, 0.037994385, 0.018341064, -0....
http://arxiv.org/abs/2302.00768v3
Leveraging Task Dependency and Contrastive Learning for Case Outcome Classification on European Court of Human Rights Cases
T. Y. S. S Santosh, Marcel Perez San Blas, Phillip Kemper, Matthias Grabmair
cs.CL
We report on an experiment in case outcome classification on European Court of Human Rights cases where our model first learns to identify the convention articles allegedly violated by the state from case facts descriptions, and subsequently uses that information to classify whether the court finds a violation of those...
http://arxiv.org/pdf/2302.00768v3
2023-02-13T14:19:02
2023-02-01T21:38:47
{ "title": [ 0.027252197, 0.014862061, -0.04650879, -0.056365967, 0.018112183, -0.030380249, -0.009147644, -0.0065231323, 0.016494751, 0.023468018, -0.040863037, 0.015945435, 0.0015954971, -0.016662598, -0.041900635, -0.015716553, -0.01751709, 0....
http://arxiv.org/abs/2302.00765v1
Visually Grounded Keyword Detection and Localisation for Low-Resource Languages
Kayode Kolawole Olaleye
cs.CL, cs.SD, eess.AS
This study investigates the use of Visually Grounded Speech (VGS) models for keyword localisation in speech. The study focusses on two main research questions: (1) Is keyword localisation possible with VGS models and (2) Can keyword localisation be done cross-lingually in a real low-resource setting? Four methods for l...
http://arxiv.org/pdf/2302.00765v1
2023-02-01T21:32:15
2023-02-01T21:32:15
{ "title": [ -0.021072388, 0.0001385212, -0.007221222, -0.020248413, -0.01309967, -0.00050878525, -0.021728516, -0.030456543, 0.029022217, 0.044189453, -0.033599854, 0.0047569275, 0.017944336, -0.004173279, -0.06311035, 0.0018262863, -0.005207062, ...
http://arxiv.org/abs/2302.00763v1
Collaborating with language models for embodied reasoning
Ishita Dasgupta, Christine Kaeser-Chen, Kenneth Marino, Arun Ahuja, Sheila Babayan, Felix Hill, Rob Fergus
cs.LG, cs.AI, cs.CL
Reasoning in a complex and ambiguous environment is a key goal for Reinforcement Learning (RL) agents. While some sophisticated RL agents can successfully solve difficult tasks, they require a large amount of training data and often struggle to generalize to new unseen environments and new tasks. On the other hand, Lar...
http://arxiv.org/pdf/2302.00763v1
2023-02-01T21:26:32
2023-02-01T21:26:32
{ "title": [ -0.043060303, -0.018112183, -0.011886597, -0.03439331, -0.032104492, 0.018234253, -0.05407715, -0.01461792, 0.0013189316, 0.022872925, -0.024810791, -0.009033203, -0.00017821789, -0.022262573, -0.03237915, -0.0031929016, -0.0020542145, ...
http://arxiv.org/abs/2302.00762v2
AmbiCoref: Evaluating Human and Model Sensitivity to Ambiguous Coreference
Yuewei Yuan, Chaitanya Malaviya, Mark Yatskar
cs.CL
Given a sentence "Abby told Brittney that she upset Courtney", one would struggle to understand who "she" refers to, and ask for clarification. However, if the word "upset" were replaced with "hugged", "she" unambiguously refers to Abby. We study if modern coreference resolution models are sensitive to such pronominal ...
http://arxiv.org/pdf/2302.00762v2
2023-02-03T16:07:53
2023-02-01T21:25:34
{ "title": [ -0.01373291, 0.011749268, -0.036010742, -0.029327393, -0.007381439, -0.017715454, -0.024490356, -0.030563354, 0.0053901672, 0.052520752, -0.016723633, 0.012046814, 0.012504578, 0.0017738342, -0.07122803, 0.016967773, -0.019836426, 0....
http://arxiv.org/abs/2302.00744v1
A Formal Algebraic Framework for DSL Composition
Zachary Flores, Angelo Taranto, Eric Bond
math.CT, cs.CL, cs.LO, cs.PL
We discuss a formal framework for using algebraic structures to model a meta-language that can write, compose, and provide interoperability between abstractions of DSLs. The purpose of this formal framework is to provide a verification of compositional properties of the meta-language. Throughout our paper we discuss th...
http://arxiv.org/pdf/2302.00744v1
2023-02-01T20:39:37
2023-02-01T20:39:37
{ "title": [ -0.029022217, 0.002796173, -0.05731201, -0.13232422, 0.023239136, 0.037017822, -0.107543945, 0.022827148, -0.0036239624, 0.035705566, 0.033203125, -0.02128601, -0.013412476, -0.039093018, -0.05026245, 0.010116577, -0.022659302, 0.012...
http://arxiv.org/abs/2302.00739v1
Inference of Partial Colexifications from Multilingual Wordlists
Johann-Mattis List
cs.CL
The past years have seen a drastic rise in studies devoted to the investigation of colexification patterns in individual languages families in particular and the languages of the world in specific. Specifically computational studies have profited from the fact that colexification as a scientific construct is easy to op...
http://arxiv.org/pdf/2302.00739v1
2023-02-01T20:22:20
2023-02-01T20:22:20
{ "title": [ -0.031799316, -0.021148682, -0.035949707, -0.060302734, -0.0076065063, 0.012863159, -0.061676025, 0.009246826, -0.02003479, 0.031921387, -0.022155762, 0.016326904, -0.009109497, -0.016113281, -0.027954102, -0.01739502, -0.008575439, ...
http://arxiv.org/abs/2302.01756v1
DANES: Deep Neural Network Ensemble Architecture for Social and Textual Context-aware Fake News Detection
Ciprian-Octavian Truică, Elena-Simona Apostol, Panagiotis Karras
cs.AI, cs.CL, cs.NE, cs.SI, 68T50
The growing popularity of social media platforms has simplified the creation and distribution of news articles but also creates a conduit for spreading fake news. In consequence, the need arises for effective context-aware fake news detection mechanisms, where the contextual information can be built either from the tex...
http://arxiv.org/pdf/2302.01756v1
2023-02-01T20:05:53
2023-02-01T20:05:53
{ "title": [ 0.0037574768, 0.013885498, -0.03387451, -0.032684326, 0.028244019, -0.0124435425, 0.008651733, -0.024276733, -0.020446777, 0.040527344, -0.031677246, 0.022064209, 0.023361206, -0.005596161, -0.02557373, -0.00074338913, -0.040039062, ...
http://arxiv.org/abs/2302.00674v4
Improving Few-Shot Generalization by Exploring and Exploiting Auxiliary Data
Alon Albalak, Colin Raffel, William Yang Wang
cs.LG, cs.CL
Few-shot learning is valuable in many real-world applications, but learning a generalizable model without overfitting to the few labeled datapoints is challenging. In this work, we focus on Few-shot Learning with Auxiliary Data (FLAD), a training paradigm that assumes access to auxiliary data during few-shot learning i...
http://arxiv.org/pdf/2302.00674v4
2023-10-03T15:50:30
2023-02-01T18:59:36
{ "title": [ 0.0020046234, -0.021102905, -0.025924683, -0.038360596, -0.0110321045, 0.0024795532, -0.011444092, -0.007194519, 0.011352539, 0.02017212, -0.03189087, 0.024398804, -0.04660034, -0.009735107, -0.036743164, -0.015205383, 0.008544922, 0...
http://arxiv.org/abs/2302.00672v1
'Generative CI' through Collective Response Systems
Aviv Ovadya
cs.HC, cs.CL, cs.IR
How can many people (who may disagree) come together to answer a question or make a decision? "Collective response systems" are a type of generative collective intelligence (CI) facilitation process meant to address this challenge. They enable a form of "generative voting", where both the votes, and the choices of what...
http://arxiv.org/pdf/2302.00672v1
2023-02-01T18:59:02
2023-02-01T18:59:02
{ "title": [ 0.014152527, -0.013923645, -0.0158844, -0.026397705, -0.045410156, -0.018829346, -0.024032593, -0.009742737, -0.0019187927, 0.035949707, -0.03552246, 0.009979248, -0.016525269, -0.017410278, -0.03604126, -0.023086548, 0.027435303, 0....
http://arxiv.org/abs/2302.00667v1
Does Vision Accelerate Hierarchical Generalization of Neural Language Learners?
Tatsuki Kuribayashi
cs.CL
Neural language models (LMs) are arguably less data-efficient than humans -- why does this gap occur? In this study, we hypothesize that this gap stems from the learners' accessibility to modalities other than text, specifically, vision. We conducted two complementary experiments (using noisy, realistic data and a simp...
http://arxiv.org/pdf/2302.00667v1
2023-02-01T18:53:42
2023-02-01T18:53:42
{ "title": [ -0.020401001, 0.012138367, -0.041107178, -0.038116455, -0.006336212, -0.01965332, 0.00059461594, -0.0071640015, -0.010528564, 0.035614014, -0.029296875, 0.015731812, -0.0052948, -0.009529114, -0.0440979, -0.011230469, -0.008666992, 0...
http://arxiv.org/abs/2302.00636v1
Are UD Treebanks Getting More Consistent? A Report Card for English UD
Amir Zeldes, Nathan Schneider
cs.CL
Recent efforts to consolidate guidelines and treebanks in the Universal Dependencies project raise the expectation that joint training and dataset comparison is increasingly possible for high-resource languages such as English, which have multiple corpora. Focusing on the two largest UD English treebanks, we examine pr...
http://arxiv.org/pdf/2302.00636v1
2023-02-01T17:58:28
2023-02-01T17:58:28
{ "title": [ 0.017303467, 0.012519836, -0.030792236, -0.06524658, -0.0093688965, -0.011512756, -0.04525757, -0.007835388, -0.01689148, 0.050109863, -0.034362793, -0.00724411, 0.014137268, -0.030578613, -0.039093018, -0.010795593, -0.026809692, 0....
http://arxiv.org/abs/2302.00618v1
Synthetic Prompting: Generating Chain-of-Thought Demonstrations for Large Language Models
Zhihong Shao, Yeyun Gong, Yelong Shen, Minlie Huang, Nan Duan, Weizhu Chen
cs.CL
Large language models can perform various reasoning tasks by using chain-of-thought prompting, which guides them to find answers through step-by-step demonstrations. However, the quality of the prompts depends on the demonstrations given to the models, and creating many of them by hand is costly. We introduce Synthetic...
http://arxiv.org/pdf/2302.00618v1
2023-02-01T17:33:12
2023-02-01T17:33:12
{ "title": [ -0.0026512146, -0.018981934, 0.0032100677, -0.06506348, -0.02508545, 0.015930176, -0.0579834, -0.010803223, 0.018035889, -0.002254486, -0.039611816, 0.003250122, -0.0137786865, -0.014427185, -0.03189087, -0.013221741, -0.0076446533, ...
http://arxiv.org/abs/2302.00609v3
Zero-shot Transfer of Article-aware Legal Outcome Classification for European Court of Human Rights Cases
T. Y. S. S Santosh, Oana Ichim, Matthias Grabmair
cs.CL
In this paper, we cast Legal Judgment Prediction on European Court of Human Rights cases into an article-aware classification task, where the case outcome is classified from a combined input of case facts and convention articles. This configuration facilitates the model learning some legal reasoning ability in mapping ...
http://arxiv.org/pdf/2302.00609v3
2023-02-13T14:23:05
2023-02-01T17:20:52
{ "title": [ 0.030410767, 0.0021247864, -0.04650879, -0.025436401, 0.005996704, -0.0362854, 0.00970459, -0.004558563, 0.008392334, 0.021484375, -0.024475098, 0.037322998, 0.02281189, -0.02508545, 0.0030403137, 0.01222229, 0.0042304993, 0.02719116...
http://arxiv.org/abs/2302.00606v1
The RW3D: A multi-modal panel dataset to understand the psychological impact of the pandemic
Isabelle van der Vegt, Bennett Kleinberg
cs.CL
Besides far-reaching public health consequences, the COVID-19 pandemic had a significant psychological impact on people around the world. To gain further insight into this matter, we introduce the Real World Worry Waves Dataset (RW3D). The dataset combines rich open-ended free-text responses with survey data on emotion...
http://arxiv.org/pdf/2302.00606v1
2023-02-01T17:13:06
2023-02-01T17:13:06
{ "title": [ -0.0018177032, -0.005340576, -0.055725098, -0.06274414, 0.017364502, -0.029785156, -0.03555298, 0.009307861, -0.016983032, 0.02267456, -0.004585266, 0.016281128, -0.0054092407, -0.03262329, -0.0158844, -0.016708374, 0.0181427, -0.002...
http://arxiv.org/abs/2302.00560v1
Co-Writing with Opinionated Language Models Affects Users' Views
Maurice Jakesch, Advait Bhat, Daniel Buschek, Lior Zalmanson, Mor Naaman
cs.HC, cs.AI, cs.CL
If large language models like GPT-3 preferably produce a particular point of view, they may influence people's opinions on an unknown scale. This study investigates whether a language-model-powered writing assistant that generates some opinions more often than others impacts what users write - and what they think. In a...
http://arxiv.org/pdf/2302.00560v1
2023-02-01T16:26:32
2023-02-01T16:26:32
{ "title": [ 0.009880066, 0.0104599, -0.03479004, -0.004638672, -0.051727295, -0.034423828, -0.03062439, -0.018966675, -0.037231445, 0.0435791, -0.046051025, -0.0008468628, 0.0021476746, -0.021636963, -0.018005371, -0.028762817, -0.0079574585, 0....
http://arxiv.org/abs/2302.00509v1
Exploring Semantic Perturbations on Grover
Pranav Kulkarni, Ziqing Ji, Yan Xu, Marko Neskovic, Kevin Nolan
cs.LG, cs.CL, cs.CV
With news and information being as easy to access as they currently are, it is more important than ever to ensure that people are not mislead by what they read. Recently, the rise of neural fake news (AI-generated fake news) and its demonstrated effectiveness at fooling humans has prompted the development of models to ...
http://arxiv.org/pdf/2302.00509v1
2023-02-01T15:28:55
2023-02-01T15:28:55
{ "title": [ -0.012809753, 0.0033187866, -0.043060303, -0.070373535, 0.0057258606, -0.0023994446, -0.030090332, -0.0027198792, -0.00042796135, 0.04272461, -0.017349243, 0.019180298, -0.0018882751, -0.01272583, -0.019256592, 0.020935059, 0.015052795, ...
http://arxiv.org/abs/2302.00493v1
You Are What You Talk About: Inducing Evaluative Topics for Personality Analysis
Josip Jukić, Iva Vukojević, Jan Šnajder
cs.CL
Expressing attitude or stance toward entities and concepts is an integral part of human behavior and personality. Recently, evaluative language data has become more accessible with social media's rapid growth, enabling large-scale opinion analysis. However, surprisingly little research examines the relationship between...
http://arxiv.org/pdf/2302.00493v1
2023-02-01T15:04:04
2023-02-01T15:04:04
{ "title": [ -0.004135132, 0.0053253174, -0.034088135, -0.03237915, 0.016647339, -0.001786232, 0.0017127991, -0.023101807, -0.0049095154, 0.022277832, 0.007637024, 0.03390503, -0.011558533, -0.023406982, -0.019088745, 0.020858765, -0.019989014, 0...
http://arxiv.org/abs/2302.00456v2
Analyzing Feed-Forward Blocks in Transformers through the Lens of Attention Map
Goro Kobayashi, Tatsuki Kuribayashi, Sho Yokoi, Kentaro Inui
cs.CL
Given that Transformers are ubiquitous in wide tasks, interpreting their internals is a pivotal issue. Still, their particular components, feed-forward (FF) blocks, have typically been less analyzed despite their substantial parameter amounts. We analyze the input contextualization effects of FF blocks by rendering the...
http://arxiv.org/pdf/2302.00456v2
2023-10-01T10:27:24
2023-02-01T13:59:47
{ "title": [ 0.0178833, 0.015312195, -0.015716553, -0.041107178, 0.01096344, -0.017028809, -0.010276794, -0.029159546, -0.015655518, 0.04257202, -0.02519226, 0.023803711, -0.002040863, -0.006401062, -0.03050232, 0.0044059753, -0.0065612793, 0.033...
http://arxiv.org/abs/2302.00455v1
HunSum-1: an Abstractive Summarization Dataset for Hungarian
Botond Barta, Dorina Lakatos, Attila Nagy, Milán Konor Nyist, Judit Ács
cs.CL
We introduce HunSum-1: a dataset for Hungarian abstractive summarization, consisting of 1.14M news articles. The dataset is built by collecting, cleaning and deduplicating data from 9 major Hungarian news sites through CommonCrawl. Using this dataset, we build abstractive summarizer models based on huBERT and mT5. We d...
http://arxiv.org/pdf/2302.00455v1
2023-02-01T13:59:45
2023-02-01T13:59:45
{ "title": [ -0.006626129, 0.015991211, -0.041412354, -0.09057617, 0.009635925, 0.013572693, -0.08013916, -0.013145447, 0.0014705658, 0.07434082, -0.022659302, 0.014755249, 0.054718018, 0.0038528442, -0.05126953, 0.034088135, -0.01084137, 0.03347...
http://arxiv.org/abs/2302.00444v1
Improved Knowledge Distillation for Pre-trained Language Models via Knowledge Selection
Chenglong Wang, Yi Lu, Yongyu Mu, Yimin Hu, Tong Xiao, Jingbo Zhu
cs.CL
Knowledge distillation addresses the problem of transferring knowledge from a teacher model to a student model. In this process, we typically have multiple types of knowledge extracted from the teacher model. The problem is to make full use of them to train the student model. Our preliminary study shows that: (1) not a...
http://arxiv.org/pdf/2302.00444v1
2023-02-01T13:40:19
2023-02-01T13:40:19
{ "title": [ -0.022476196, -0.00617218, -0.06060791, -0.08984375, -0.000043570995, -0.0062294006, -0.09466553, 0.015396118, 0.020263672, 0.016342163, -0.02897644, -0.008262634, 0.005832672, -0.022140503, -0.033050537, -0.0039367676, 0.0032615662, ...
http://arxiv.org/abs/2302.00412v2
KNNs of Semantic Encodings for Rating Prediction
Léo Laugier, Raghuram Vadapalli, Thomas Bonald, Lucas Dixon
cs.CL
This paper explores a novel application of textual semantic similarity to user-preference representation for rating prediction. The approach represents a user's preferences as a graph of textual snippets from review text, where the edges are defined by semantic similarity. This textual, memory-based approach to rating ...
http://arxiv.org/pdf/2302.00412v2
2023-03-28T12:01:21
2023-02-01T12:53:31
{ "title": [ -0.025985718, -0.014770508, -0.011543274, -0.009681702, -0.024536133, 0.0014686584, -0.00907135, -0.020690918, 0.021636963, 0.04336548, -0.0019102097, -0.0012254715, 0.02192688, -0.015129089, -0.04486084, 0.02947998, -0.0005354881, 0...
http://arxiv.org/abs/2302.00407v3
On the Role of Morphological Information for Contextual Lemmatization
Olia Toporkov, Rodrigo Agerri
cs.CL, cs.AI
Lemmatization is a natural language processing (NLP) task which consists of producing, from a given inflected word, its canonical form or lemma. Lemmatization is one of the basic tasks that facilitate downstream NLP applications, and is of particular importance for high-inflected languages. Given that the process to ob...
http://arxiv.org/pdf/2302.00407v3
2023-10-20T15:31:57
2023-02-01T12:47:09
{ "title": [ 0.009925842, 0.015960693, -0.015838623, -0.08483887, 0.011451721, 0.011207581, -0.04949951, -0.016937256, -0.019470215, 0.040802002, -0.012641907, -0.0069999695, 0.018295288, -0.027114868, -0.026870728, 0.012069702, -0.012794495, 0.0...
http://arxiv.org/abs/2302.00402v1
mPLUG-2: A Modularized Multi-modal Foundation Model Across Text, Image and Video
Haiyang Xu, Qinghao Ye, Ming Yan, Yaya Shi, Jiabo Ye, Yuanhong Xu, Chenliang Li, Bin Bi, Qi Qian, Wei Wang, Guohai Xu, Ji Zhang, Songfang Huang, Fei Huang, Jingren Zhou
cs.CV, cs.CL, cs.MM
Recent years have witnessed a big convergence of language, vision, and multi-modal pretraining. In this work, we present mPLUG-2, a new unified paradigm with modularized design for multi-modal pretraining, which can benefit from modality collaboration while addressing the problem of modality entanglement. In contrast t...
http://arxiv.org/pdf/2302.00402v1
2023-02-01T12:40:03
2023-02-01T12:40:03
{ "title": [ -0.012031555, 0.0036792755, -0.059753418, -0.04989624, -0.009643555, -0.014099121, -0.06274414, -0.0141067505, -0.015182495, 0.044921875, 0.0021209717, 0.036590576, 0.003479004, -0.031341553, -0.010726929, 0.003004074, -0.016235352, ...
http://arxiv.org/abs/2302.00378v2
An Empirical Study on the Transferability of Transformer Modules in Parameter-Efficient Fine-Tuning
Mohammad Akbar-Tajari, Sara Rajaee, Mohammad Taher Pilehvar
cs.CL
Parameter-efficient fine-tuning approaches have recently garnered a lot of attention. Having considerably lower number of trainable weights, these methods can bring about scalability and computational effectiveness. In this paper, we look for optimal sub-networks and investigate the capability of different transformer ...
http://arxiv.org/pdf/2302.00378v2
2023-02-22T16:56:58
2023-02-01T11:20:18
{ "title": [ 0.013504028, 0.0056648254, -0.0035915375, -0.06323242, 0.011688232, -0.00045204163, -0.019256592, -0.01574707, -0.03778076, 0.03326416, -0.00044751167, 0.024963379, 0.010498047, -0.023269653, -0.001247406, 0.007472992, -0.000364542, ...
http://arxiv.org/abs/2302.00340v1
Attention Link: An Efficient Attention-Based Low Resource Machine Translation Architecture
Zeping Min
cs.CL, cs.AI
Transformers have achieved great success in machine translation, but transformer-based NMT models often require millions of bilingual parallel corpus for training. In this paper, we propose a novel architecture named as attention link (AL) to help improve transformer models' performance, especially in low training reso...
http://arxiv.org/pdf/2302.00340v1
2023-02-01T09:52:36
2023-02-01T09:52:36
{ "title": [ -0.029785156, -0.010551453, -0.013465881, -0.06365967, -0.0052108765, 0.021133423, -0.03717041, 0.014434814, 0.020477295, 0.065979004, -0.016693115, -0.0049552917, 0.051513672, -0.001958847, -0.061553955, 0.02420044, -0.0037651062, 0...
http://arxiv.org/abs/2302.00326v1
Evaluating TCFD Reporting: A New Application of Zero-Shot Analysis to Climate-Related Financial Disclosures
Alix Auzepy, Elena Tönjes, David Lenz, Christoph Funk
cs.CY, cs.CL
We examine climate-related disclosures in a large sample of reports published by banks that officially endorsed the recommendations of the Task Force for Climate-related Financial Disclosures (TCFD). In doing so, we introduce a new application of the zero-shot text classification. By developing a set of fine-grained TC...
http://arxiv.org/pdf/2302.00326v1
2023-02-01T09:11:05
2023-02-01T09:11:05
{ "title": [ -0.015823364, -0.023635864, -0.065979004, -0.05227661, -0.023254395, -0.016998291, -0.015602112, 0.030212402, -0.022262573, 0.034240723, -0.027145386, 0.017440796, -0.043914795, -0.019195557, -0.053833008, -0.030395508, -0.01689148, ...
http://arxiv.org/abs/2302.00321v1
An Evaluation of Persian-English Machine Translation Datasets with Transformers
Amir Sartipi, Meghdad Dehghan, Afsaneh Fatemi
cs.CL, cs.AI
Nowadays, many researchers are focusing their attention on the subject of machine translation (MT). However, Persian machine translation has remained unexplored despite a vast amount of research being conducted in languages with high resources, such as English. Moreover, while a substantial amount of research has been ...
http://arxiv.org/pdf/2302.00321v1
2023-02-01T08:55:08
2023-02-01T08:55:08
{ "title": [ -0.008110046, 0.007007599, -0.004005432, -0.04949951, 0.002960205, 0.0027389526, -0.036376953, -0.014030457, -0.0055389404, 0.036102295, -0.045806885, 0.008728027, 0.013694763, -0.01838684, -0.045532227, 0.00017929077, -0.013153076, ...
http://arxiv.org/abs/2302.00239v1
Filtering Context Mitigates Scarcity and Selection Bias in Political Ideology Prediction
Chen Chen, Dylan Walker, Venkatesh Saligrama
cs.LG, cs.CL, stat.ML
We propose a novel supervised learning approach for political ideology prediction (PIP) that is capable of predicting out-of-distribution inputs. This problem is motivated by the fact that manual data-labeling is expensive, while self-reported labels are often scarce and exhibit significant selection bias. We propose a...
http://arxiv.org/pdf/2302.00239v1
2023-02-01T04:34:48
2023-02-01T04:34:48
{ "title": [ 0.04525757, 0.008644104, -0.025634766, -0.072265625, 0.04031372, -0.021896362, 0.0056533813, -0.029632568, -0.009147644, 0.032043457, -0.0209198, 0.008270264, -0.0017576218, -0.017242432, -0.027709961, -0.0005812645, -0.0015125275, 0...
http://arxiv.org/abs/2302.00200v1
A Transaction Represented with Weighted Finite-State Transducers
J. Nathaniel Holmes, Homayoon Beigi
cs.FL, cs.AI, cs.CL, cs.LG, math.RA
Not all contracts are good, but all good contracts can be expressed as a finite-state transition system ("State-Transition Contracts"). Contracts that can be represented as State-Transition Contracts discretize fat-tailed risk to foreseeable, managed risk, define the boundary of relevant events governed by the relation...
http://arxiv.org/pdf/2302.00200v1
2023-02-01T03:05:14
2023-02-01T03:05:14
{ "title": [ 0.054260254, -0.00046730042, -0.033843994, -0.11065674, -0.003921509, -0.020324707, -0.044128418, 0.03274536, 0.010650635, 0.02192688, -0.07244873, -0.00356102, -0.002035141, -0.015586853, 0.0099487305, -0.033325195, 0.0060691833, 0....
http://arxiv.org/abs/2303.12038v1
Grading Conversational Responses Of Chatbots
Grant Rosario, David Noever
cs.CL, cs.LG
Chatbots have long been capable of answering basic questions and even responding to obscure prompts, but recently their improvements have been far more significant. Modern chatbots like Open AIs ChatGPT3 not only have the ability to answer basic questions but can write code and movie scripts and imitate well-known peop...
http://arxiv.org/pdf/2303.12038v1
2023-02-01T02:54:43
2023-02-01T02:54:43
{ "title": [ -0.008644104, -0.011665344, -0.01576233, -0.03314209, -0.01763916, -0.0067863464, -0.017074585, -0.0016050339, 0.010528564, 0.04019165, -0.03616333, -0.0056114197, -0.019714355, -0.00198555, -0.072753906, -0.016098022, -0.023483276, ...
http://arxiv.org/abs/2302.00189v2
Detecting Lexical Borrowings from Dominant Languages in Multilingual Wordlists
John E. Miller, Johann-Mattis List
cs.CL
Language contact is a pervasive phenomenon reflected in the borrowing of words from donor to recipient languages. Most computational approaches to borrowing detection treat all languages under study as equally important, even though dominant languages have a stronger impact on heritage languages than vice versa. We tes...
http://arxiv.org/pdf/2302.00189v2
2023-02-21T08:13:54
2023-02-01T02:44:28
{ "title": [ -0.0016822815, 0.0008196831, -0.017944336, -0.030639648, 0.009170532, -0.010879517, -0.016967773, -0.019699097, -0.028244019, 0.044769287, -0.027786255, 0.034576416, -0.014984131, -0.0037899017, -0.037902832, -0.030227661, -0.023910522, ...
http://arxiv.org/abs/2302.00178v1
Program Generation from Diverse Video Demonstrations
Anthony Manchin, Jamie Sherrah, Qi Wu, Anton van den Hengel
cs.CV, cs.CL, cs.LG
The ability to use inductive reasoning to extract general rules from multiple observations is a vital indicator of intelligence. As humans, we use this ability to not only interpret the world around us, but also to predict the outcomes of the various interactions we experience. Generalising over multiple observations i...
http://arxiv.org/pdf/2302.00178v1
2023-02-01T01:51:45
2023-02-01T01:51:45
{ "title": [ -0.02406311, 0.012893677, -0.04244995, -0.06414795, 0.01826477, 0.0012197495, -0.055023193, -0.007358551, -0.013298035, 0.008354187, -0.030639648, 0.0085372925, -0.057128906, -0.016708374, -0.035064697, -0.027404785, -0.004966736, 0....
http://arxiv.org/abs/2302.00129v1
Universal Topological Regularities of Syntactic Structures: Decoupling Efficiency from Optimization
Fermín Moscoso del Prado Martín
cs.CL, q-bio.NC
Human syntactic structures are usually represented as graphs. Much research has focused on the mapping between such graphs and linguistic sequences, but less attention has been paid to the shapes of the graphs themselves: their topologies. This study investigates how the topologies of syntactic graphs reveal traces of ...
http://arxiv.org/pdf/2302.00129v1
2023-01-31T22:35:11
2023-01-31T22:35:11
{ "title": [ -0.036315918, -0.00058460236, -0.014801025, -0.08679199, 0.0019006729, 0.018371582, -0.06335449, 0.007297516, 0.007019043, 0.023361206, 0.0073280334, 0.007030487, -0.017196655, -0.024490356, -0.03253174, 0.011177063, -0.009384155, 0....
http://arxiv.org/abs/2302.00119v1
Machine Translation Impact in E-commerce Multilingual Search
Bryan Zhang, Amita Misra
cs.CL, cs.IR
Previous work suggests that performance of cross-lingual information retrieval correlates highly with the quality of Machine Translation. However, there may be a threshold beyond which improving query translation quality yields little or no benefit to further improve the retrieval performance. This threshold may depend...
http://arxiv.org/pdf/2302.00119v1
2023-01-31T21:59:35
2023-01-31T21:59:35
{ "title": [ -0.019088745, -0.011711121, -0.031829834, -0.0284729, -0.021377563, -0.020874023, -0.026367188, -0.007949829, -0.018051147, 0.07952881, 0.0070610046, -0.020126343, 0.061157227, -0.028396606, -0.030929565, 0.012680054, 0.0010166168, 0...
http://arxiv.org/abs/2302.00102v3
Towards Detecting Harmful Agendas in News Articles
Melanie Subbiah, Amrita Bhattacharjee, Yilun Hua, Tharindu Kumarage, Huan Liu, Kathleen McKeown
cs.CL, cs.LG
Manipulated news online is a growing problem which necessitates the use of automated systems to curtail its spread. We argue that while misinformation and disinformation detection have been studied, there has been a lack of investment in the important open challenge of detecting harmful agendas in news articles; identi...
http://arxiv.org/pdf/2302.00102v3
2023-08-02T17:16:48
2023-01-31T21:08:58
{ "title": [ 0.012428284, 0.014167786, -0.025787354, -0.038146973, 0.026672363, -0.016723633, 0.0071372986, -0.038635254, -0.028869629, 0.021240234, -0.019332886, 0.017318726, -0.0134887695, -0.005252838, -0.029083252, -0.012268066, -0.024459839, ...
http://arxiv.org/abs/2302.00093v3
Large Language Models Can Be Easily Distracted by Irrelevant Context
Freda Shi, Xinyun Chen, Kanishka Misra, Nathan Scales, David Dohan, Ed Chi, Nathanael Schärli, Denny Zhou
cs.CL, cs.AI
Large language models have achieved impressive performance on various natural language processing tasks. However, so far they have been evaluated primarily on benchmarks where all information in the input context is relevant for solving the task. In this work, we investigate the distractibility of large language models...
http://arxiv.org/pdf/2302.00093v3
2023-06-06T08:36:20
2023-01-31T20:48:57
{ "title": [ 0.03414917, -0.005302429, 0.01184845, -0.049468994, -0.020095825, -0.0035266876, 0.003730774, 0.0063171387, 0.027297974, 0.05609131, -0.042938232, -0.0546875, 0.02696228, -0.020385742, -0.045532227, -0.01197052, -0.0059013367, 0.0534...
http://arxiv.org/abs/2302.00083v3
In-Context Retrieval-Augmented Language Models
Ori Ram, Yoav Levine, Itay Dalmedigos, Dor Muhlgay, Amnon Shashua, Kevin Leyton-Brown, Yoav Shoham
cs.CL, cs.IR
Retrieval-Augmented Language Modeling (RALM) methods, which condition a language model (LM) on relevant documents from a grounding corpus during generation, were shown to significantly improve language modeling performance. In addition, they can mitigate the problem of factually inaccurate text generation and provide n...
http://arxiv.org/pdf/2302.00083v3
2023-08-01T12:10:15
2023-01-31T20:26:16
{ "title": [ -0.004348755, 0.0025730133, -0.030639648, -0.05355835, 0.0053901672, -0.0014877319, -0.034301758, -0.005760193, 0.020370483, 0.062042236, -0.019302368, -0.0005130768, 0.031707764, -0.006046295, -0.058532715, 0.025360107, -0.0074501038, ...
http://arxiv.org/abs/2301.13868v1
PADL: Language-Directed Physics-Based Character Control
Jordan Juravsky, Yunrong Guo, Sanja Fidler, Xue Bin Peng
cs.LG, cs.AI, cs.CL, cs.GR
Developing systems that can synthesize natural and life-like motions for simulated characters has long been a focus for computer animation. But in order for these systems to be useful for downstream applications, they need not only produce high-quality motions, but must also provide an accessible and versatile interfac...
http://arxiv.org/pdf/2301.13868v1
2023-01-31T18:59:22
2023-01-31T18:59:22
{ "title": [ -0.037078857, -0.026748657, -0.018051147, -0.04660034, -0.023147583, 0.011306763, -0.043029785, -0.0067977905, -0.0131073, 0.04486084, 0.038360596, 0.012535095, 0.025650024, -0.03692627, -0.023391724, 0.029220581, -0.007270813, 0.009...
http://arxiv.org/abs/2301.13867v2
Mathematical Capabilities of ChatGPT
Simon Frieder, Luca Pinchetti, Alexis Chevalier, Ryan-Rhys Griffiths, Tommaso Salvatori, Thomas Lukasiewicz, Philipp Christian Petersen, Julius Berner
cs.LG, cs.AI, cs.CL
We investigate the mathematical capabilities of two iterations of ChatGPT (released 9-January-2023 and 30-January-2023) and of GPT-4 by testing them on publicly available datasets, as well as hand-crafted ones, using a novel methodology. In contrast to formal mathematics, where large databases of formal proofs are avai...
http://arxiv.org/pdf/2301.13867v2
2023-07-20T17:59:14
2023-01-31T18:59:03
{ "title": [ 0.022247314, -0.002866745, 0.0009636879, -0.006313324, -0.034179688, -0.017608643, 0.0027923584, -0.018447876, 0.004673004, 0.040771484, -0.0033607483, 0.012466431, 0.0058135986, -0.044677734, 0.012641907, 0.032409668, 0.014129639, 0...
http://arxiv.org/abs/2301.13848v1
Benchmarking Large Language Models for News Summarization
Tianyi Zhang, Faisal Ladhak, Esin Durmus, Percy Liang, Kathleen McKeown, Tatsunori B. Hashimoto
cs.CL, cs.AI, cs.LG
Large language models (LLMs) have shown promise for automatic summarization but the reasons behind their successes are poorly understood. By conducting a human evaluation on ten LLMs across different pretraining methods, prompts, and model scales, we make two important observations. First, we find instruction tuning, a...
http://arxiv.org/pdf/2301.13848v1
2023-01-31T18:46:19
2023-01-31T18:46:19
{ "title": [ -0.037719727, -0.008842468, -0.023590088, -0.06726074, -0.015579224, 0.019119263, -0.07086182, 0.000115692616, 0.0032577515, 0.057495117, -0.008460999, 0.016586304, -0.0010519028, 0.0031795502, -0.068115234, 0.016342163, -0.0013275146, ...
http://arxiv.org/abs/2301.13844v1
Do Multi-Document Summarization Models Synthesize?
Jay DeYoung, Stephanie C. Martinez, Iain J. Marshall, Byron C. Wallace
cs.CL
Multi-document summarization entails producing concise synopses of collections of inputs. For some applications, the synopsis should accurately \emph{synthesize} inputs with respect to a key property or aspect. For example, a synopsis of film reviews all written about a particular movie should reflect the average criti...
http://arxiv.org/pdf/2301.13844v1
2023-01-31T18:40:46
2023-01-31T18:40:46
{ "title": [ -0.003452301, 0.0024490356, -0.010284424, -0.0657959, -0.046905518, 0.013938904, -0.09710693, -0.0103302, -0.023422241, 0.060638428, -0.030075073, 0.013313293, -0.006313324, 0.0039634705, -0.034332275, -0.016418457, -0.0015468597, 0....
http://arxiv.org/abs/2301.13823v4
Grounding Language Models to Images for Multimodal Inputs and Outputs
Jing Yu Koh, Ruslan Salakhutdinov, Daniel Fried
cs.CL, cs.AI, cs.CV, cs.LG
We propose an efficient method to ground pretrained text-only language models to the visual domain, enabling them to process arbitrarily interleaved image-and-text data, and generate text interleaved with retrieved images. Our method leverages the abilities of language models learnt from large scale text-only pretraini...
http://arxiv.org/pdf/2301.13823v4
2023-06-13T21:54:58
2023-01-31T18:33:44
{ "title": [ -0.003025055, 0.005050659, -0.024795532, -0.04611206, -0.0014066696, -0.010513306, -0.04147339, -0.027267456, 0.021820068, 0.032043457, -0.013153076, 0.0048828125, 0.0008702278, -0.022399902, -0.040618896, 0.010375977, 0.0049324036, ...
http://arxiv.org/abs/2301.13826v2
Attend-and-Excite: Attention-Based Semantic Guidance for Text-to-Image Diffusion Models
Hila Chefer, Yuval Alaluf, Yael Vinker, Lior Wolf, Daniel Cohen-Or
cs.CV, cs.CL, cs.GR, cs.LG
Recent text-to-image generative models have demonstrated an unparalleled ability to generate diverse and creative imagery guided by a target text prompt. While revolutionary, current state-of-the-art diffusion models may still fail in generating images that fully convey the semantics in the given text prompt. We analyz...
http://arxiv.org/pdf/2301.13826v2
2023-05-31T15:42:00
2023-01-31T18:10:38
{ "title": [ 0.00982666, 0.0066871643, -0.013168335, -0.030548096, -0.003255844, -0.0032176971, -0.010345459, -0.017471313, 0.03555298, 0.042633057, -0.028030396, 0.025756836, 0.027786255, -0.009796143, -0.032684326, 0.016784668, -0.0063667297, 0...
http://arxiv.org/abs/2301.13816v4
Execution-based Code Generation using Deep Reinforcement Learning
Parshin Shojaee, Aneesh Jain, Sindhu Tipirneni, Chandan K. Reddy
cs.LG, cs.AI, cs.CL, cs.PL
The utilization of programming language (PL) models, pre-trained on large-scale code corpora, as a means of automating software engineering processes has demonstrated considerable potential in streamlining various code generation tasks such as code completion, code translation, and program synthesis. However, current a...
http://arxiv.org/pdf/2301.13816v4
2023-07-19T19:55:31
2023-01-31T18:02:26
{ "title": [ -0.033996582, -0.01739502, -0.039031982, -0.044555664, -0.003768921, -0.011703491, -0.03857422, 0.011924744, -0.027404785, 0.012199402, -0.01625061, 0.02053833, -0.04119873, -0.020431519, -0.012802124, -0.020736694, -0.013687134, 0.0...
http://arxiv.org/abs/2301.13808v3
Large Language Models are Versatile Decomposers: Decompose Evidence and Questions for Table-based Reasoning
Yunhu Ye, Binyuan Hui, Min Yang, Binhua Li, Fei Huang, Yongbin Li
cs.CL
Table-based reasoning has shown remarkable progress in combining deep models with discrete reasoning, which requires reasoning over both free-form natural language (NL) questions and structured tabular data. However, previous table-based reasoning solutions usually suffer from significant performance degradation on hug...
http://arxiv.org/pdf/2301.13808v3
2023-04-27T11:24:10
2023-01-31T17:51:45
{ "title": [ -0.02520752, -0.02027893, -0.024902344, -0.048980713, -0.039398193, 0.013053894, -0.06402588, 0.008636475, -0.0021686554, 0.028152466, -0.028244019, -0.009010315, 0.016143799, -0.02519226, -0.023376465, -0.011878967, 0.014381409, 0.0...
http://arxiv.org/abs/2301.13771v1
The Touché23-ValueEval Dataset for Identifying Human Values behind Arguments
Nailia Mirzakhmedova, Johannes Kiesel, Milad Alshomary, Maximilian Heinrich, Nicolas Handke, Xiaoni Cai, Barriere Valentin, Doratossadat Dastgheib, Omid Ghahroodi, Mohammad Ali Sadraei, Ehsaneddin Asgari, Lea Kawaletz, Henning Wachsmuth, Benno Stein
cs.CL
We present the Touch\'e23-ValueEval Dataset for Identifying Human Values behind Arguments. To investigate approaches for the automated detection of human values behind arguments, we collected 9324 arguments from 6 diverse sources, covering religious texts, political discussions, free-text arguments, newspaper editorial...
http://arxiv.org/pdf/2301.13771v1
2023-01-31T17:15:33
2023-01-31T17:15:33
{ "title": [ 0.039215088, -0.0026626587, -0.028823853, -0.05407715, 0.017654419, -0.049041748, 0.008384705, -0.011550903, -0.02355957, 0.017303467, 0.004798889, -0.016815186, -0.012382507, -0.04876709, -0.00030469894, -0.0075683594, -0.008995056, ...
http://arxiv.org/abs/2301.13753v1
Dynamic Scheduled Sampling with Imitation Loss for Neural Text Generation
Xiang Lin, Prathyusha Jwalapuram, Shafiq Joty
cs.CL
State-of-the-art neural text generation models are typically trained to maximize the likelihood of each token in the ground-truth sequence conditioned on the previous target tokens. However, during inference, the model needs to make a prediction conditioned on the tokens generated by itself. This train-test discrepancy...
http://arxiv.org/pdf/2301.13753v1
2023-01-31T16:41:06
2023-01-31T16:41:06
{ "title": [ -0.03665161, -0.018920898, -0.02142334, -0.031463623, -0.031951904, 0.017471313, -0.025726318, -0.0031356812, -0.012069702, 0.048553467, -0.016143799, 0.022445679, -0.006095886, 0.0004131794, -0.042022705, 0.011741638, -0.0026226044, ...
http://arxiv.org/abs/2301.13741v3
UPop: Unified and Progressive Pruning for Compressing Vision-Language Transformers
Dachuan Shi, Chaofan Tao, Ying Jin, Zhendong Yang, Chun Yuan, Jiaqi Wang
cs.CV, cs.CL, cs.LG
Real-world data contains a vast amount of multimodal information, among which vision and language are the two most representative modalities. Moreover, increasingly heavier models, \textit{e}.\textit{g}., Transformers, have attracted the attention of researchers to model compression. However, how to compress multimodal...
http://arxiv.org/pdf/2301.13741v3
2023-06-30T03:25:27
2023-01-31T16:18:52
{ "title": [ -0.026138306, -0.017303467, -0.03265381, -0.08630371, 0.0073242188, 0.017944336, -0.052368164, 0.022201538, 0.04269409, 0.01878357, -0.014205933, -0.015716553, 0.022079468, -0.029525757, -0.041870117, 0.0075798035, 0.02734375, 0.0309...
http://arxiv.org/abs/2301.13720v1
Zero-shot cross-lingual transfer language selection using linguistic similarity
Juuso Eronen, Michal Ptaszynski, Fumito Masui
cs.CL
We study the selection of transfer languages for different Natural Language Processing tasks, specifically sentiment analysis, named entity recognition and dependency parsing. In order to select an optimal transfer language, we propose to utilize different linguistic similarity metrics to measure the distance between l...
http://arxiv.org/pdf/2301.13720v1
2023-01-31T15:56:40
2023-01-31T15:56:40
{ "title": [ -0.01914978, -0.004432678, -0.041381836, -0.029785156, 0.009689331, -0.015625, -0.024047852, -0.009841919, 0.005016327, 0.036224365, -0.0028438568, 0.036590576, 0.015960693, -0.02293396, -0.014450073, 0.014808655, -0.014137268, 0.040...
http://arxiv.org/abs/2301.13714v1
Recursive Neural Networks with Bottlenecks Diagnose (Non-)Compositionality
Verna Dankers, Ivan Titov
cs.CL
A recent line of work in NLP focuses on the (dis)ability of models to generalise compositionally for artificial languages. However, when considering natural language tasks, the data involved is not strictly, or locally, compositional. Quantifying the compositionality of data is a challenging task, which has been invest...
http://arxiv.org/pdf/2301.13714v1
2023-01-31T15:46:39
2023-01-31T15:46:39
{ "title": [ 0.0014562607, 0.018127441, -0.033325195, -0.061706543, -0.008529663, -0.020141602, -0.057617188, -0.013595581, 0.0025100708, 0.027542114, -0.030456543, -0.013572693, -0.018753052, -0.023956299, -0.033233643, -0.015686035, -0.013366699, ...
http://arxiv.org/abs/2301.13688v2
The Flan Collection: Designing Data and Methods for Effective Instruction Tuning
Shayne Longpre, Le Hou, Tu Vu, Albert Webson, Hyung Won Chung, Yi Tay, Denny Zhou, Quoc V. Le, Barret Zoph, Jason Wei, Adam Roberts
cs.AI, cs.CL, cs.LG
We study the design decisions of publicly available instruction tuning methods, and break down the development of Flan 2022 (Chung et al., 2022). Through careful ablation studies on the Flan Collection of tasks and methods, we tease apart the effect of design decisions which enable Flan-T5 to outperform prior work by 3...
http://arxiv.org/pdf/2301.13688v2
2023-02-14T16:33:33
2023-01-31T15:03:44
{ "title": [ -0.05645752, 0.0012168884, -0.05770874, -0.015853882, 0.0036907196, -0.025939941, -0.008872986, 0.00015580654, -0.0059509277, 0.004798889, -0.04006958, 0.012619019, -0.021255493, -0.02519226, -0.013114929, -0.008804321, 0.006690979, ...
http://arxiv.org/abs/2301.13683v1
Friend-training: Learning from Models of Different but Related Tasks
Mian Zhang, Lifeng Jin, Linfeng Song, Haitao Mi, Xiabing Zhou, Dong Yu
cs.CL
Current self-training methods such as standard self-training, co-training, tri-training, and others often focus on improving model performance on a single task, utilizing differences in input features, model architectures, and training processes. However, many tasks in natural language processing are about different bu...
http://arxiv.org/pdf/2301.13683v1
2023-01-31T15:00:56
2023-01-31T15:00:56
{ "title": [ -0.0075302124, -0.012062073, -0.03277588, -0.034179688, -0.0035095215, -0.019119263, -0.035186768, -0.032562256, -0.024520874, 0.017242432, -0.014640808, 0.02571106, -0.04244995, -0.019866943, -0.024276733, -0.016983032, -0.008346558, ...
http://arxiv.org/abs/2301.13668v1
Automated Sentiment and Hate Speech Analysis of Facebook Data by Employing Multilingual Transformer Models
Ritumbra Manuvie, Saikat Chatterjee
cs.CL, cs.LG
In recent years, there has been a heightened consensus within academia and in the public discourse that Social Media Platforms (SMPs), amplify the spread of hateful and negative sentiment content. Researchers have identified how hateful content, political propaganda, and targeted messaging contributed to real-world har...
http://arxiv.org/pdf/2301.13668v1
2023-01-31T14:37:04
2023-01-31T14:37:04
{ "title": [ 0.008605957, 0.016921997, -0.038391113, -0.04849243, 0.013870239, -0.01600647, -0.024505615, -0.017913818, -0.012382507, 0.048065186, -0.02848816, 0.017990112, 0.020706177, -0.0054626465, -0.03933716, -0.0015497208, -0.005050659, 0.0...
http://arxiv.org/abs/2301.13631v2
TopoBERT: Plug and Play Toponym Recognition Module Harnessing Fine-tuned BERT
Bing Zhou, Lei Zou, Yingjie Hu, Yi Qiang, Daniel Goldberg
cs.CL, cs.AI
Extracting precise geographical information from textual contents is crucial in a plethora of applications. For example, during hazardous events, a robust and unbiased toponym extraction framework can provide an avenue to tie the location concerned to the topic discussed by news media posts and pinpoint humanitarian he...
http://arxiv.org/pdf/2301.13631v2
2023-02-03T17:45:11
2023-01-31T13:44:34
{ "title": [ 0.02748108, -0.026885986, 0.0046195984, -0.04638672, -0.033233643, -0.009902954, -0.017837524, -0.021652222, -0.0119018555, 0.029464722, -0.017440796, 0.018249512, 0.015182495, -0.029830933, -0.0052337646, 0.002878189, -0.014038086, ...
http://arxiv.org/abs/2301.13479v1
Archive TimeLine Summarization (ATLS): Conceptual Framework for Timeline Generation over Historical Document Collections
Nicolas Gutehrlé, Antoine Doucet, Adam Jatowt
cs.CL
Archive collections are nowadays mostly available through search engines interfaces, which allow a user to retrieve documents by issuing queries. The study of these collections may be, however, impaired by some aspects of search engines, such as the overwhelming number of documents returned or the lack of contextual kn...
http://arxiv.org/pdf/2301.13479v1
2023-01-31T08:58:47
2023-01-31T08:58:47
{ "title": [ 0.0020179749, -0.014129639, -0.033325195, -0.057281494, -0.011627197, -0.002336502, -0.044677734, 0.00039625168, 0.007850647, 0.03829956, 0.017547607, 0.011550903, 0.032165527, -0.03543091, -0.010391235, 0.01940918, 0.010192871, 0.01...
http://arxiv.org/abs/2301.13455v1
ZhichunRoad at Amazon KDD Cup 2022: MultiTask Pre-Training for E-Commerce Product Search
Xuange Cui, Wei Xiong, Songlin Wang
cs.CL
In this paper, we propose a robust multilingual model to improve the quality of search results. Our model not only leverage the processed class-balanced dataset, but also benefit from multitask pre-training that leads to more general representations. In pre-training stage, we adopt mlm task, classification task and con...
http://arxiv.org/pdf/2301.13455v1
2023-01-31T07:31:34
2023-01-31T07:31:34
{ "title": [ -0.018676758, -0.018234253, -0.012275696, -0.0034713745, -0.022262573, -0.0029888153, -0.016067505, 0.003932953, -0.023147583, 0.051483154, -0.019729614, 0.015350342, 0.03857422, -0.019592285, 0.014793396, -0.022216797, 0.03668213, -...
http://arxiv.org/abs/2301.13382v1
Numeracy from Literacy: Data Science as an Emergent Skill from Large Language Models
David Noever, Forrest McKee
cs.CL
Large language models (LLM) such as OpenAI's ChatGPT and GPT-3 offer unique testbeds for exploring the translation challenges of turning literacy into numeracy. Previous publicly-available transformer models from eighteen months prior and 1000 times smaller failed to provide basic arithmetic. The statistical analysis o...
http://arxiv.org/pdf/2301.13382v1
2023-01-31T03:14:57
2023-01-31T03:14:57
{ "title": [ -0.022827148, -0.014030457, -0.045013428, -0.034301758, -0.016159058, -0.024597168, -0.013031006, 0.0013828278, 0.0107803345, 0.008743286, -0.024246216, -0.023864746, -0.008262634, -0.03540039, -0.028839111, -0.019210815, 0.00044322014, ...
http://arxiv.org/abs/2301.13379v3
Faithful Chain-of-Thought Reasoning
Qing Lyu, Shreya Havaldar, Adam Stein, Li Zhang, Delip Rao, Eric Wong, Marianna Apidianaki, Chris Callison-Burch
cs.CL
While Chain-of-Thought (CoT) prompting boosts Language Models' (LM) performance on a gamut of complex reasoning tasks, the generated reasoning chain does not necessarily reflect how the model arrives at the answer (aka. faithfulness). We propose Faithful CoT, a reasoning framework involving two stages: Translation (Nat...
http://arxiv.org/pdf/2301.13379v3
2023-09-20T22:19:30
2023-01-31T03:04:26
{ "title": [ 0.04348755, -0.0015048981, -0.03579712, -0.07208252, -0.01675415, -0.022903442, -0.027755737, -0.0036678314, -0.011878967, 0.035705566, -0.0017442703, -0.016479492, 0.023483276, -0.062347412, 0.009597778, 0.019577026, 0.016906738, 0....
http://arxiv.org/abs/2301.13372v1
Improving Open-Domain Dialogue Evaluation with a Causal Inference Model
Cat P. Le, Luke Dai, Michael Johnston, Yang Liu, Marilyn Walker, Reza Ghanadan
cs.CL, cs.AI
Effective evaluation methods remain a significant challenge for research on open-domain conversational dialogue systems. Explicit satisfaction ratings can be elicited from users, but users often do not provide ratings when asked, and those they give can be highly subjective. Post-hoc ratings by experts are an alternati...
http://arxiv.org/pdf/2301.13372v1
2023-01-31T02:31:42
2023-01-31T02:31:42
{ "title": [ 0.01121521, 0.006427765, -0.03491211, -0.08343506, 0.010292053, 0.0011396408, -0.049926758, -0.0003528595, -0.028533936, -0.0021915436, 0.013908386, 0.006877899, -0.031585693, -0.03930664, -0.016586304, -0.021743774, -0.04296875, 0.0...
http://arxiv.org/abs/2302.10307v1
ViewCo: Discovering Text-Supervised Segmentation Masks via Multi-View Semantic Consistency
Pengzhen Ren, Changlin Li, Hang Xu, Yi Zhu, Guangrun Wang, Jianzhuang Liu, Xiaojun Chang, Xiaodan Liang
cs.CV, cs.AI, cs.CL
Recently, great success has been made in learning visual representations from text supervision, facilitating the emergence of text-supervised semantic segmentation. However, existing works focus on pixel grouping and cross-modal semantic alignment, while ignoring the correspondence among multiple augmented views of the...
http://arxiv.org/pdf/2302.10307v1
2023-01-31T01:57:52
2023-01-31T01:57:52
{ "title": [ -0.047912598, 0.0029640198, -0.06726074, -0.03616333, -0.0064353943, -0.004798889, -0.044311523, 0.013656616, -0.006137848, 0.023117065, -0.049224854, 0.024673462, -0.0077171326, -0.004562378, -0.026229858, -0.010826111, -0.008285522, ...
http://arxiv.org/abs/2301.13352v1
Sentence Identification with BOS and EOS Label Combinations
Takuma Udagawa, Hiroshi Kanayama, Issei Yoshida
cs.CL, cs.AI
The sentence is a fundamental unit in many NLP applications. Sentence segmentation is widely used as the first preprocessing task, where an input text is split into consecutive sentences considering the end of the sentence (EOS) as their boundaries. This task formulation relies on a strong assumption that the input tex...
http://arxiv.org/pdf/2301.13352v1
2023-01-31T01:03:07
2023-01-31T01:03:07
{ "title": [ 0.007980347, 0.03656006, -0.039001465, -0.038635254, -0.0007390976, -0.02519226, -0.04196167, -0.026412964, -0.016952515, 0.023406982, -0.032440186, 0.0023326874, -0.0010156631, -0.03717041, -0.023483276, -0.025314331, -0.026351929, ...
http://arxiv.org/abs/2301.13345v1
Differentiable Entailment for Parameter Efficient Few Shot Learning
Ethan Kim, Jerry Yang
cs.CL
Few-shot learning allows pre-trained language models to adapt to downstream tasks while using a limited number of training examples. However, practical applications are limited when all model parameters must be optimized. In this work we apply a new technique for parameter efficient few shot learning while adopting a s...
http://arxiv.org/pdf/2301.13345v1
2023-01-31T00:31:11
2023-01-31T00:31:11
{ "title": [ 0.0061149597, -0.020339966, -0.014503479, -0.070007324, -0.0017461777, 0.001742363, 0.003709793, 0.027313232, -0.004261017, -0.02015686, -0.0357666, 0.030258179, -0.03744507, -0.019592285, 0.024154663, -0.014381409, 0.007686615, 0.02...
http://arxiv.org/abs/2302.00003v1
The Power of External Memory in Increasing Predictive Model Capacity
Cenk Baykal, Dylan J Cutler, Nishanth Dikkala, Nikhil Ghosh, Rina Panigrahy, Xin Wang
cs.LG, cs.CL
One way of introducing sparsity into deep networks is by attaching an external table of parameters that is sparsely looked up at different layers of the network. By storing the bulk of the parameters in the external table, one can increase the capacity of the model without necessarily increasing the inference time. Two...
http://arxiv.org/pdf/2302.00003v1
2023-01-31T00:29:39
2023-01-31T00:29:39
{ "title": [ 0.05404663, -0.021224976, 0.00049066544, -0.05697632, -0.035705566, -0.026031494, -0.031677246, -0.021850586, -0.006351471, 0.02748108, -0.03967285, 0.0037727356, -0.015129089, -0.03274536, -0.0062294006, -0.0076293945, 0.037506104, ...
http://arxiv.org/abs/2301.13318v1
Proxy-based Zero-Shot Entity Linking by Effective Candidate Retrieval
Maciej Wiatrak, Eirini Arvaniti, Angus Brayne, Jonas Vetterle, Aaron Sim
cs.LG, cs.CL, cs.IR
A recent advancement in the domain of biomedical Entity Linking is the development of powerful two-stage algorithms, an initial candidate retrieval stage that generates a shortlist of entities for each mention, followed by a candidate ranking stage. However, the effectiveness of both stages are inextricably dependent o...
http://arxiv.org/pdf/2301.13318v1
2023-01-30T22:43:21
2023-01-30T22:43:21
{ "title": [ 0.036712646, -0.002544403, -0.020950317, -0.082458496, 0.02267456, -0.014221191, -0.0006046295, -0.0053749084, 0.008964539, 0.047668457, 0.018814087, -0.00024485588, 0.037200928, -0.029083252, -0.032806396, 0.026672363, -0.0055732727, ...
http://arxiv.org/abs/2301.13310v2
Alternating Updates for Efficient Transformers
Cenk Baykal, Dylan Cutler, Nishanth Dikkala, Nikhil Ghosh, Rina Panigrahy, Xin Wang
cs.LG, cs.CL
It has been well established that increasing scale in deep transformer networks leads to improved quality and performance. However, this increase in scale often comes with prohibitive increases in compute cost and inference latency. We introduce Alternating Updates (AltUp), a simple-to-implement method to increase a mo...
http://arxiv.org/pdf/2301.13310v2
2023-10-03T21:40:41
2023-01-30T22:06:05
{ "title": [ 0.016357422, 0.019256592, -0.0178833, -0.09246826, 0.026748657, 0.015037537, -0.020706177, 0.0035476685, -0.01763916, 0.07086182, -0.021911621, 0.004714966, 0.02671814, -0.014602661, -0.025466919, 0.009757996, 0.045776367, 0.00564193...