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/2301.00665v1
Targeted Phishing Campaigns using Large Scale Language Models
Rabimba Karanjai
cs.CL, cs.CR, cs.LG
In this research, we aim to explore the potential of natural language models (NLMs) such as GPT-3 and GPT-2 to generate effective phishing emails. Phishing emails are fraudulent messages that aim to trick individuals into revealing sensitive information or taking actions that benefit the attackers. We propose a framewo...
http://arxiv.org/pdf/2301.00665v1
2022-12-30T03:18:05
2022-12-30T03:18:05
{ "title": [ -0.011291504, 0.0062332153, -0.037139893, -0.04019165, 0.014755249, -0.005268097, -0.027908325, -0.005332947, -0.037506104, 0.032562256, 0.01600647, 0.012710571, 0.00573349, -0.02420044, -0.019760132, 0.0018939972, -0.028793335, 0.03...
http://arxiv.org/abs/2212.14518v1
ResGrad: Residual Denoising Diffusion Probabilistic Models for Text to Speech
Zehua Chen, Yihan Wu, Yichong Leng, Jiawei Chen, Haohe Liu, Xu Tan, Yang Cui, Ke Wang, Lei He, Sheng Zhao, Jiang Bian, Danilo Mandic
eess.AS, cs.CL, cs.LG, cs.SD, eess.SP
Denoising Diffusion Probabilistic Models (DDPMs) are emerging in text-to-speech (TTS) synthesis because of their strong capability of generating high-fidelity samples. However, their iterative refinement process in high-dimensional data space results in slow inference speed, which restricts their application in real-ti...
http://arxiv.org/pdf/2212.14518v1
2022-12-30T02:31:35
2022-12-30T02:31:35
{ "title": [ -0.01864624, -0.0011539459, -0.045928955, -0.070617676, 0.015464783, 0.0088272095, -0.03137207, 0.021057129, 0.020126343, 0.03012085, -0.018615723, 0.012687683, 0.031173706, -0.02494812, -0.033294678, 0.020721436, -0.018310547, 0.043...
http://arxiv.org/abs/2212.14490v1
Multi-modal deep learning system for depression and anxiety detection
Brian Diep, Marija Stanojevic, Jekaterina Novikova
cs.SD, cs.CL, cs.MM, eess.AS
Traditional screening practices for anxiety and depression pose an impediment to monitoring and treating these conditions effectively. However, recent advances in NLP and speech modelling allow textual, acoustic, and hand-crafted language-based features to jointly form the basis of future mental health screening and co...
http://arxiv.org/pdf/2212.14490v1
2022-12-30T00:02:58
2022-12-30T00:02:58
{ "title": [ 0.010269165, 0.012481689, -0.027252197, -0.053222656, 0.025680542, -0.013114929, -0.02961731, -0.027877808, 0.010398865, 0.03225708, -0.0116119385, -0.008758545, 0.015174866, -0.037261963, -0.048461914, 0.003616333, -0.0047035217, 0....
http://arxiv.org/abs/2212.14486v2
Examining Political Rhetoric with Epistemic Stance Detection
Ankita Gupta, Su Lin Blodgett, Justin H Gross, Brendan O'Connor
cs.CL
Participants in political discourse employ rhetorical strategies -- such as hedging, attributions, or denials -- to display varying degrees of belief commitments to claims proposed by themselves or others. Traditionally, political scientists have studied these epistemic phenomena through labor-intensive manual content ...
http://arxiv.org/pdf/2212.14486v2
2023-01-06T04:03:03
2022-12-29T23:47:14
{ "title": [ 0.016296387, 0.0002503395, -0.010307312, -0.056396484, 0.0063323975, -0.016067505, -0.019699097, -0.012718201, -0.017837524, 0.015090942, -0.026519775, -0.00349617, -0.026382446, -0.010467529, -0.039031982, -0.03488159, -0.015655518, ...
http://arxiv.org/abs/2212.14454v4
MEAformer: Multi-modal Entity Alignment Transformer for Meta Modality Hybrid
Zhuo Chen, Jiaoyan Chen, Wen Zhang, Lingbing Guo, Yin Fang, Yufeng Huang, Yichi Zhang, Yuxia Geng, Jeff Z. Pan, Wenting Song, Huajun Chen
cs.AI, cs.CL
Multi-modal entity alignment (MMEA) aims to discover identical entities across different knowledge graphs (KGs) whose entities are associated with relevant images. However, current MMEA algorithms rely on KG-level modality fusion strategies for multi-modal entity representation, which ignores the variations of modality...
http://arxiv.org/pdf/2212.14454v4
2023-07-30T14:39:36
2022-12-29T20:49:58
{ "title": [ -0.023361206, 0.0051612854, -0.06262207, -0.06201172, 0.019104004, 0.0026779175, -0.060760498, -0.023010254, 0.017501831, 0.059783936, 0.01676941, 0.0069122314, 0.06707764, -0.036499023, -0.029571533, 0.04257202, -0.01727295, 0.02906...
http://arxiv.org/abs/2212.14453v2
Learning Multimodal Data Augmentation in Feature Space
Zichang Liu, Zhiqiang Tang, Xingjian Shi, Aston Zhang, Mu Li, Anshumali Shrivastava, Andrew Gordon Wilson
cs.LG, cs.CL, cs.CV
The ability to jointly learn from multiple modalities, such as text, audio, and visual data, is a defining feature of intelligent systems. While there have been promising advances in designing neural networks to harness multimodal data, the enormous success of data augmentation currently remains limited to single-modal...
http://arxiv.org/pdf/2212.14453v2
2023-04-24T14:48:00
2022-12-29T20:39:36
{ "title": [ -0.015403748, 0.018478394, -0.06549072, -0.053863525, 0.025482178, -0.020324707, -0.032684326, -0.0052337646, 0.010627747, 0.019088745, -0.033355713, 0.010688782, -0.011528015, -0.013824463, -0.028427124, 0.0013275146, -0.0060424805, ...
http://arxiv.org/abs/2301.00671v1
Political representation bias in DBpedia and Wikidata as a challenge for downstream processing
Ozgur Karadeniz, Bettina Berendt, Sercan Kiyak, Stefan Mertens, Leen d'Haenens
cs.CL, cs.AI, cs.CY, cs.DB
Diversity Searcher is a tool originally developed to help analyse diversity in news media texts. It relies on a form of automated content analysis and thus rests on prior assumptions and depends on certain design choices related to diversity and fairness. One such design choice is the external knowledge source(s) used....
http://arxiv.org/pdf/2301.00671v1
2022-12-29T18:21:09
2022-12-29T18:21:09
{ "title": [ 0.0440979, 0.012313843, -0.03918457, -0.06549072, 0.040740967, -0.04168701, 0.011207581, -0.012367249, -0.012367249, 0.040985107, -0.030303955, -0.0011949539, 0.0071372986, -0.008964539, -0.039733887, -0.010955811, -0.015464783, 0.03...
http://arxiv.org/abs/2212.14402v1
GPT Takes the Bar Exam
Michael Bommarito II, Daniel Martin Katz
cs.CL, cs.AI, cs.LG
Nearly all jurisdictions in the United States require a professional license exam, commonly referred to as "the Bar Exam," as a precondition for law practice. To even sit for the exam, most jurisdictions require that an applicant completes at least seven years of post-secondary education, including three years at an ac...
http://arxiv.org/pdf/2212.14402v1
2022-12-29T18:19:43
2022-12-29T18:19:43
{ "title": [ 0.014038086, 0.015609741, -0.014533997, -0.020980835, -0.029891968, -0.014015198, -0.039215088, -0.017333984, -0.01361084, 0.042053223, -0.024108887, -0.011428833, 0.014442444, -0.013755798, 0.016433716, 0.010238647, -0.00027489662, ...
http://arxiv.org/abs/2212.14293v1
Error syntax aware augmentation of feedback comment generation dataset
Nikolay Babakov, Maria Lysyuk, Alexander Shvets, Lilya Kazakova, Alexander Panchenko
cs.CL
This paper presents a solution to the GenChal 2022 shared task dedicated to feedback comment generation for writing learning. In terms of this task given a text with an error and a span of the error, a system generates an explanatory note that helps the writer (language learner) to improve their writing skills. Our sol...
http://arxiv.org/pdf/2212.14293v1
2022-12-29T12:57:23
2022-12-29T12:57:23
{ "title": [ 0.0046691895, -0.017608643, -0.03717041, -0.093566895, -0.008361816, 0.00730896, -0.07092285, 0.02394104, 0.041778564, 0.030792236, -0.066711426, -0.004585266, 0.0003118515, 0.009841919, -0.069885254, -0.017562866, -0.0033874512, 0.0...
http://arxiv.org/abs/2212.14270v1
Reviewing Labels: Label Graph Network with Top-k Prediction Set for Relation Extraction
Bo Li, Wei Ye, Jinglei Zhang, Shikun Zhang
cs.CL
The typical way for relation extraction is fine-tuning large pre-trained language models on task-specific datasets, then selecting the label with the highest probability of the output distribution as the final prediction. However, the usage of the Top-k prediction set for a given sample is commonly overlooked. In this ...
http://arxiv.org/pdf/2212.14270v1
2022-12-29T11:44:42
2022-12-29T11:44:42
{ "title": [ -0.020187378, -0.013755798, -0.052520752, -0.07318115, -0.00869751, 0.012809753, -0.05026245, 0.012435913, 0.010803223, 0.0006532669, 0.0013284683, 0.0073165894, -0.030212402, -0.02142334, -0.038513184, 0.00831604, -0.007873535, 0.04...
http://arxiv.org/abs/2212.14266v2
Sequence Generation with Label Augmentation for Relation Extraction
Bo Li, Dingyao Yu, Wei Ye, Jinglei Zhang, Shikun Zhang
cs.CL
Sequence generation demonstrates promising performance in recent information extraction efforts, by incorporating large-scale pre-trained Seq2Seq models. This paper investigates the merits of employing sequence generation in relation extraction, finding that with relation names or synonyms as generation targets, their ...
http://arxiv.org/pdf/2212.14266v2
2023-02-10T02:07:08
2022-12-29T11:28:05
{ "title": [ 0.0012588501, -0.0059509277, -0.0579834, -0.087768555, 0.0075302124, 0.0011072159, -0.04699707, 0.020614624, 0.021560669, 0.01121521, -0.02017212, 0.011779785, -0.022964478, -0.0049476624, -0.040496826, -0.00032138824, -0.019119263, ...
http://arxiv.org/abs/2212.14206v1
Maximizing Use-Case Specificity through Precision Model Tuning
Pranjali Awasthi, David Recio-Mitter, Yosuke Kyle Sugi
cs.CL, cs.IR, cs.LG, H.3.3
Language models have become increasingly popular in recent years for tasks like information retrieval. As use-cases become oriented toward specific domains, fine-tuning becomes default for standard performance. To fine-tune these models for specific tasks and datasets, it is necessary to carefully tune the model's hype...
http://arxiv.org/pdf/2212.14206v1
2022-12-29T07:50:14
2022-12-29T07:50:14
{ "title": [ -0.0146102905, -0.012504578, -0.034820557, -0.07281494, -0.01058197, 0.00038433075, -0.07757568, -0.020568848, -0.024673462, 0.043640137, 0.01398468, 0.004878998, -0.00819397, -0.02368164, -0.032714844, 0.017501831, -0.008140564, 0.0...
http://arxiv.org/abs/2301.00676v1
Multimodal Sequential Generative Models for Semi-Supervised Language Instruction Following
Kei Akuzawa, Yusuke Iwasawa, Yutaka Matsuo
cs.LG, cs.AI, cs.CL
Agents that can follow language instructions are expected to be useful in a variety of situations such as navigation. However, training neural network-based agents requires numerous paired trajectories and languages. This paper proposes using multimodal generative models for semi-supervised learning in the instruction ...
http://arxiv.org/pdf/2301.00676v1
2022-12-29T03:23:43
2022-12-29T03:23:43
{ "title": [ -0.02130127, -0.0051956177, -0.033294678, -0.067871094, -0.015129089, -0.0049972534, -0.08026123, 0.003074646, -0.0104522705, 0.012802124, 0.005470276, 0.012390137, -0.015777588, -0.034332275, -0.017913818, -0.0022850037, -0.017166138, ...
http://arxiv.org/abs/2212.14149v1
Macro-block dropout for improved regularization in training end-to-end speech recognition models
Chanwoo Kim, Sathish Indurti, Jinhwan Park, Wonyong Sung
cs.LG, cs.CL, cs.SD, eess.AS
This paper proposes a new regularization algorithm referred to as macro-block dropout. The overfitting issue has been a difficult problem in training large neural network models. The dropout technique has proven to be simple yet very effective for regularization by preventing complex co-adaptations during training. In ...
http://arxiv.org/pdf/2212.14149v1
2022-12-29T02:09:49
2022-12-29T02:09:49
{ "title": [ -0.026916504, 0.012031555, -0.043945312, -0.07519531, 0.013946533, 0.012062073, -0.059143066, 0.024261475, 0.008415222, 0.03100586, -0.010437012, 0.021133423, -0.0005850792, -0.02067566, -0.03353882, 0.023925781, -0.0039024353, 0.038...
http://arxiv.org/abs/2212.14117v1
Improving a sequence-to-sequence nlp model using a reinforcement learning policy algorithm
Jabri Ismail, Aboulbichr Ahmed, El ouaazizi Aziza
cs.CL, cs.AI, cs.LG
Nowadays, the current neural network models of dialogue generation(chatbots) show great promise for generating answers for chatty agents. But they are short-sighted in that they predict utterances one at a time while disregarding their impact on future outcomes. Modelling a dialogue's future direction is critical for g...
http://arxiv.org/pdf/2212.14117v1
2022-12-28T22:46:57
2022-12-28T22:46:57
{ "title": [ 0.0044937134, -0.004047394, -0.014976501, -0.05709839, 0.0039520264, -0.004749298, -0.018447876, -0.000808239, -0.024459839, 0.012214661, -0.027816772, 0.02204895, -0.0155181885, -0.017745972, -0.009918213, -0.010482788, -0.013122559, ...
http://arxiv.org/abs/2212.14104v1
Towards automating Codenames spymasters with deep reinforcement learning
Sherman Siu
cs.CL, cs.AI, cs.LG
Although most reinforcement learning research has centered on competitive games, little work has been done on applying it to co-operative multiplayer games or text-based games. Codenames is a board game that involves both asymmetric co-operation and natural language processing, which makes it an excellent candidate for...
http://arxiv.org/pdf/2212.14104v1
2022-12-28T21:45:59
2022-12-28T21:45:59
{ "title": [ -0.030426025, 0.0057678223, -0.016708374, 0.008033752, -0.022323608, -0.02305603, -0.0023727417, -0.040893555, -0.01737976, 0.021865845, 0.002948761, 0.014373779, -0.010383606, -0.022476196, -0.01600647, 0.00037288666, -0.0099487305, ...
http://arxiv.org/abs/2212.14102v1
Customizing Knowledge Graph Embedding to Improve Clinical Study Recommendation
Xiong Liu, Iya Khalil, Murthy Devarakonda
cs.LG, cs.AI, cs.CL, cs.CY, q-bio.QM
Inferring knowledge from clinical trials using knowledge graph embedding is an emerging area. However, customizing graph embeddings for different use cases remains a significant challenge. We propose custom2vec, an algorithmic framework to customize graph embeddings by incorporating user preferences in training the emb...
http://arxiv.org/pdf/2212.14102v1
2022-12-28T21:41:25
2022-12-28T21:41:25
{ "title": [ -0.025909424, -0.007335663, -0.05444336, -0.07458496, -0.009681702, 0.002325058, -0.06384277, 0.019104004, -0.009529114, 0.021652222, -0.03353882, 0.020126343, -0.01638794, -0.014724731, -0.020614624, -0.012001038, 0.0023498535, 0.05...
http://arxiv.org/abs/2212.14100v1
Leveraging World Knowledge in Implicit Hate Speech Detection
Jessica Lin
cs.CL
While much attention has been paid to identifying explicit hate speech, implicit hateful expressions that are disguised in coded or indirect language are pervasive and remain a major challenge for existing hate speech detection systems. This paper presents the first attempt to apply Entity Linking (EL) techniques to bo...
http://arxiv.org/pdf/2212.14100v1
2022-12-28T21:23:55
2022-12-28T21:23:55
{ "title": [ 0.025390625, 0.0026054382, -0.03439331, -0.048034668, 0.018112183, -0.029800415, 0.0057868958, -0.009567261, 0.0030784607, 0.024597168, -0.047943115, 0.0067749023, -0.000017344952, 0.00027370453, -0.053619385, -0.02444458, -0.0058059692,...
http://arxiv.org/abs/2212.14074v2
Choosing the Number of Topics in LDA Models -- A Monte Carlo Comparison of Selection Criteria
Victor Bystrov, Viktoriia Naboka, Anna Staszewska-Bystrova, Peter Winker
cs.CL, cs.LG, stat.ML
Selecting the number of topics in LDA models is considered to be a difficult task, for which alternative approaches have been proposed. The performance of the recently developed singular Bayesian information criterion (sBIC) is evaluated and compared to the performance of alternative model selection criteria. The sBIC ...
http://arxiv.org/pdf/2212.14074v2
2023-02-15T20:38:34
2022-12-28T19:37:04
{ "title": [ -0.040985107, 0.00932312, -0.06890869, -0.026885986, 0.014083862, -0.01789856, -0.034118652, -0.048583984, -0.035125732, 0.0025520325, -0.011619568, 0.043060303, -0.057739258, -0.03024292, 0.0030899048, -0.0075912476, -0.028793335, 0...
http://arxiv.org/abs/2212.14034v1
Cramming: Training a Language Model on a Single GPU in One Day
Jonas Geiping, Tom Goldstein
cs.CL, cs.LG
Recent trends in language modeling have focused on increasing performance through scaling, and have resulted in an environment where training language models is out of reach for most researchers and practitioners. While most in the community are asking how to push the limits of extreme computation, we ask the opposite ...
http://arxiv.org/pdf/2212.14034v1
2022-12-28T18:59:28
2022-12-28T18:59:28
{ "title": [ -0.0013208389, -0.034698486, -0.043182373, -0.03918457, -0.04260254, -0.04006958, -0.015960693, 0.024505615, -0.02949524, 0.03302002, -0.02961731, 0.028411865, 0.016479492, -0.02633667, 0.010116577, -0.017715454, -0.010696411, 0.0570...
http://arxiv.org/abs/2212.14024v2
Demonstrate-Search-Predict: Composing retrieval and language models for knowledge-intensive NLP
Omar Khattab, Keshav Santhanam, Xiang Lisa Li, David Hall, Percy Liang, Christopher Potts, Matei Zaharia
cs.CL, cs.IR
Retrieval-augmented in-context learning has emerged as a powerful approach for addressing knowledge-intensive tasks using frozen language models (LM) and retrieval models (RM). Existing work has combined these in simple "retrieve-then-read" pipelines in which the RM retrieves passages that are inserted into the LM prom...
http://arxiv.org/pdf/2212.14024v2
2023-01-23T17:00:01
2022-12-28T18:52:44
{ "title": [ -0.013687134, -0.017852783, -0.025787354, -0.0635376, -0.013465881, 0.006965637, -0.049224854, -0.0018196106, 0.018203735, 0.031158447, -0.03616333, 0.013244629, 0.010848999, -0.00061511993, -0.05319214, 0.011619568, -0.018325806, 0....
http://arxiv.org/abs/2212.14052v3
Hungry Hungry Hippos: Towards Language Modeling with State Space Models
Daniel Y. Fu, Tri Dao, Khaled K. Saab, Armin W. Thomas, Atri Rudra, Christopher Ré
cs.LG, cs.CL
State space models (SSMs) have demonstrated state-of-the-art sequence modeling performance in some modalities, but underperform attention in language modeling. Moreover, despite scaling nearly linearly in sequence length instead of quadratically, SSMs are still slower than Transformers due to poor hardware utilization....
http://arxiv.org/pdf/2212.14052v3
2023-04-29T03:18:40
2022-12-28T17:56:03
{ "title": [ -0.041778564, -0.009361267, -0.021820068, -0.042633057, -0.009384155, 0.01612854, -0.05239868, -0.0046081543, 0.008300781, 0.011520386, -0.029724121, 0.019729614, -0.0075950623, -0.0140686035, -0.03845215, 0.00048732758, -0.009376526, ...
http://arxiv.org/abs/2212.13939v3
Data Augmentation using Transformers and Similarity Measures for Improving Arabic Text Classification
Dania Refai, Saleh Abo-Soud, Mohammad Abdel-Rahman
cs.CL, cs.AI, cs.LG
The performance of learning models heavily relies on the availability and adequacy of training data. To address the dataset adequacy issue, researchers have extensively explored data augmentation (DA) as a promising approach. DA generates new data instances through transformations applied to the available data, thereby...
http://arxiv.org/pdf/2212.13939v3
2023-08-20T15:32:29
2022-12-28T16:38:43
{ "title": [ -0.008354187, 0.017028809, -0.038513184, -0.049468994, 0.03265381, -0.0033435822, -0.003168106, -0.0077285767, -0.001001358, 0.020812988, -0.046295166, 0.035003662, -0.010513306, -0.0053138733, -0.030075073, -0.020843506, -0.004547119, ...
http://arxiv.org/abs/2212.13860v1
Automatic Recognition and Classification of Future Work Sentences from Academic Articles in a Specific Domain
Chengzhi Zhang, Yi Xiang, Wenke Hao, Zhicheng Li, Yuchen Qian, Yuzhuo Wang
cs.CL, cs.DL
Future work sentences (FWS) are the particular sentences in academic papers that contain the author's description of their proposed follow-up research direction. This paper presents methods to automatically extract FWS from academic papers and classify them according to the different future directions embodied in the p...
http://arxiv.org/pdf/2212.13860v1
2022-12-28T15:26:04
2022-12-28T15:26:04
{ "title": [ -0.016418457, 0.01927185, -0.0574646, -0.024032593, -0.0025348663, -0.022277832, -0.03274536, -0.019546509, -0.01209259, 0.030441284, -0.0031490326, -0.002412796, 0.012008667, -0.030303955, -0.02406311, -0.013069153, -0.012451172, 0....
http://arxiv.org/abs/2212.13738v2
TempCLR: Temporal Alignment Representation with Contrastive Learning
Yuncong Yang, Jiawei Ma, Shiyuan Huang, Long Chen, Xudong Lin, Guangxing Han, Shih-Fu Chang
cs.CV, cs.CL
Video representation learning has been successful in video-text pre-training for zero-shot transfer, where each sentence is trained to be close to the paired video clips in a common feature space. For long videos, given a paragraph of description where the sentences describe different segments of the video, by matching...
http://arxiv.org/pdf/2212.13738v2
2023-03-30T01:42:53
2022-12-28T08:10:31
{ "title": [ -0.0076446533, -0.0024852753, -0.060058594, -0.066711426, 0.0071754456, -0.016647339, -0.033813477, -0.011016846, 0.03302002, 0.040130615, -0.006340027, -0.007911682, 0.002204895, -0.018463135, -0.053955078, 0.017089844, -0.0023422241, ...
http://arxiv.org/abs/2212.14047v1
Using Large Language Models to Generate Engaging Captions for Data Visualizations
Ashley Liew, Klaus Mueller
cs.CL, cs.AI, cs.HC
Creating compelling captions for data visualizations has been a longstanding challenge. Visualization researchers are typically untrained in journalistic reporting and hence the captions that are placed below data visualizations tend to be not overly engaging and rather just stick to basic observations about the data. ...
http://arxiv.org/pdf/2212.14047v1
2022-12-27T23:56:57
2022-12-27T23:56:57
{ "title": [ 0.004878998, -0.0050621033, -0.026870728, -0.051452637, -0.01222229, -0.01524353, -0.035614014, -0.023086548, 0.031463623, 0.0541687, -0.027328491, -0.0054969788, -0.0039634705, 0.001493454, -0.07324219, 0.0017089844, -0.0014982224, ...
http://arxiv.org/abs/2212.14727v1
Countering Malicious Content Moderation Evasion in Online Social Networks: Simulation and Detection of Word Camouflage
Álvaro Huertas-García, Alejandro Martín, Javier Huertas Tato, David Camacho
cs.CL, cs.AI, cs.SI
Content moderation is the process of screening and monitoring user-generated content online. It plays a crucial role in stopping content resulting from unacceptable behaviors such as hate speech, harassment, violence against specific groups, terrorism, racism, xenophobia, homophobia, or misogyny, to mention some few, i...
http://arxiv.org/pdf/2212.14727v1
2022-12-27T16:08:49
2022-12-27T16:08:49
{ "title": [ 0.017974854, 0.02798462, 0.000094115734, -0.009521484, 0.026382446, -0.011016846, 0.020309448, -0.047851562, -0.057647705, 0.053375244, 0.025985718, 0.022567749, 0.028045654, -0.028961182, -0.021194458, -0.008392334, -0.037261963, 0....
http://arxiv.org/abs/2212.13492v1
MultiSpider: Towards Benchmarking Multilingual Text-to-SQL Semantic Parsing
Longxu Dou, Yan Gao, Mingyang Pan, Dingzirui Wang, Wanxiang Che, Dechen Zhan, Jian-Guang Lou
cs.CL
Text-to-SQL semantic parsing is an important NLP task, which greatly facilitates the interaction between users and the database and becomes the key component in many human-computer interaction systems. Much recent progress in text-to-SQL has been driven by large-scale datasets, but most of them are centered on English....
http://arxiv.org/pdf/2212.13492v1
2022-12-27T13:58:30
2022-12-27T13:58:30
{ "title": [ -0.033203125, -0.013160706, -0.043823242, -0.08001709, 0.014785767, 0.020019531, -0.074523926, 0.015777588, 0.005340576, 0.054229736, -0.020904541, 0.019638062, 0.0010213852, 0.00030851364, -0.061279297, 0.00033259392, -0.006450653, ...
http://arxiv.org/abs/2212.13465v2
A Survey on Table-and-Text HybridQA: Concepts, Methods, Challenges and Future Directions
Dingzirui Wang, Longxu Dou, Wanxiang Che
cs.CL, cs.AI
Table-and-text hybrid question answering (HybridQA) is a widely used and challenging NLP task commonly applied in the financial and scientific domain. The early research focuses on migrating other QA task methods to HybridQA, while with further research, more and more HybridQA-specific methods have been present. With t...
http://arxiv.org/pdf/2212.13465v2
2023-02-02T02:44:58
2022-12-27T12:34:57
{ "title": [ -0.018615723, -0.0016784668, -0.043395996, -0.033050537, 0.014274597, -0.010910034, -0.00957489, -0.017578125, -0.0146865845, 0.04296875, -0.03375244, 0.01864624, 0.016571045, -0.01737976, -0.033050537, -0.023605347, -0.009216309, 0....
http://arxiv.org/abs/2212.14674v1
A Comprehensive Gold Standard and Benchmark for Comics Text Detection and Recognition
Gürkan Soykan, Deniz Yuret, Tevfik Metin Sezgin
cs.CL, cs.AI
This study focuses on improving the optical character recognition (OCR) data for panels in the COMICS dataset, the largest dataset containing text and images from comic books. To do this, we developed a pipeline for OCR processing and labeling of comic books and created the first text detection and recognition datasets...
http://arxiv.org/pdf/2212.14674v1
2022-12-27T12:05:23
2022-12-27T12:05:23
{ "title": [ -0.007484436, -0.008071899, -0.03237915, -0.03366089, 0.0104522705, -0.020080566, 0.008430481, -0.0062713623, 0.030090332, 0.036956787, -0.025100708, 0.0020599365, 0.017974854, -0.012252808, -0.057006836, 0.008918762, 0.005519867, 0....
http://arxiv.org/abs/2212.13456v1
TegFormer: Topic-to-Essay Generation with Good Topic Coverage and High Text Coherence
Wang Qi, Rui Liu, Yuan Zuo, Yong Chen, Dell Zhang
cs.CL
Creating an essay based on a few given topics is a challenging NLP task. Although several effective methods for this problem, topic-to-essay generation, have appeared recently, there is still much room for improvement, especially in terms of the coverage of the given topics and the coherence of the generated text. In t...
http://arxiv.org/pdf/2212.13456v1
2022-12-27T11:50:14
2022-12-27T11:50:14
{ "title": [ 0.0008893013, 0.0074539185, -0.031555176, -0.035491943, -0.0025901794, -0.007484436, -0.031982422, -0.04144287, -0.004989624, 0.036376953, 0.0019950867, 0.05899048, 0.0011787415, -0.010025024, -0.006084442, 0.018188477, -0.026138306, ...
http://arxiv.org/abs/2212.13428v1
A Survey on Knowledge-Enhanced Pre-trained Language Models
Chaoqi Zhen, Yanlei Shang, Xiangyu Liu, Yifei Li, Yong Chen, Dell Zhang
cs.CL
Natural Language Processing (NLP) has been revolutionized by the use of Pre-trained Language Models (PLMs) such as BERT. Despite setting new records in nearly every NLP task, PLMs still face a number of challenges including poor interpretability, weak reasoning capability, and the need for a lot of expensive annotated ...
http://arxiv.org/pdf/2212.13428v1
2022-12-27T09:54:14
2022-12-27T09:54:14
{ "title": [ -0.01335907, -0.013626099, -0.031951904, -0.0914917, 0.00027012825, 0.005874634, -0.078063965, 0.016662598, 0.0034694672, 0.035614014, -0.006385803, -0.0026683807, 0.020706177, -0.020187378, -0.036956787, 0.020187378, -0.003025055, 0...
http://arxiv.org/abs/2212.13408v3
NEEDED: Introducing Hierarchical Transformer to Eye Diseases Diagnosis
Xu Ye, Meng Xiao, Zhiyuan Ning, Weiwei Dai, Wenjuan Cui, Yi Du, Yuanchun Zhou
cs.CL, cs.IR
With the development of natural language processing techniques(NLP), automatic diagnosis of eye diseases using ophthalmology electronic medical records (OEMR) has become possible. It aims to evaluate the condition of both eyes of a patient respectively, and we formulate it as a particular multi-label classification tas...
http://arxiv.org/pdf/2212.13408v3
2023-01-11T11:10:59
2022-12-27T08:37:57
{ "title": [ 0.030883789, -0.009117126, -0.05606079, -0.072631836, -0.0016126633, -0.018753052, -0.04248047, 0.008453369, -0.021224976, 0.03793335, -0.022354126, -0.00022602081, 0.031463623, -0.036895752, -0.01322937, -0.02468872, 0.03778076, 0.0...
http://arxiv.org/abs/2212.13392v1
DeepCuts: Single-Shot Interpretability based Pruning for BERT
Jasdeep Singh Grover, Bhavesh Gawri, Ruskin Raj Manku
cs.CL, cs.AI, cs.CV, cs.LG
As language models have grown in parameters and layers, it has become much harder to train and infer with them on single GPUs. This is severely restricting the availability of large language models such as GPT-3, BERT-Large, and many others. A common technique to solve this problem is pruning the network architecture b...
http://arxiv.org/pdf/2212.13392v1
2022-12-27T07:21:41
2022-12-27T07:21:41
{ "title": [ -0.0067634583, -0.031677246, 0.012527466, -0.08874512, -0.04824829, 0.047607422, -0.088256836, 0.010734558, 0.035888672, -0.0016431808, -0.03024292, -0.010345459, 0.017074585, -0.037475586, -0.017532349, -0.000623703, -0.009254456, 0...
http://arxiv.org/abs/2212.13378v1
Don't Be So Sure! Boosting ASR Decoding via Confidence Relaxation
Tomer Wullach, Shlomo E. Chazan
cs.CL, cs.SD, eess.AS
Automatic Speech Recognition (ASR) systems frequently use a search-based decoding strategy aiming to find the best attainable transcript by considering multiple candidates. One prominent speech recognition decoding heuristic is beam search, which seeks the transcript with the greatest likelihood computed using the pred...
http://arxiv.org/pdf/2212.13378v1
2022-12-27T06:42:26
2022-12-27T06:42:26
{ "title": [ -0.044555664, -0.0019493103, -0.05859375, -0.05883789, 0.0019874573, 0.010772705, -0.051574707, 0.0009083748, -0.022033691, 0.05166626, -0.007156372, 0.0044288635, 0.010498047, -0.014335632, -0.050872803, 0.0036849976, 0.0019664764, ...
http://arxiv.org/abs/2212.13317v1
Automatic Text Simplification of News Articles in the Context of Public Broadcasting
Diego Maupomé, Fanny Rancourt, Thomas Soulas, Alexandre Lachance, Marie-Jean Meurs, Desislava Aleksandrova, Olivier Brochu Dufour, Igor Pontes, Rémi Cardon, Michel Simard, Sowmya Vajjala
cs.CL, cs.AI, cs.LG
This report summarizes the work carried out by the authors during the Twelfth Montreal Industrial Problem Solving Workshop, held at Universit\'e de Montr\'eal in August 2022. The team tackled a problem submitted by CBC/Radio-Canada on the theme of Automatic Text Simplification (ATS).
http://arxiv.org/pdf/2212.13317v1
2022-12-26T22:59:57
2022-12-26T22:59:57
{ "title": [ -0.028305054, -0.011039734, -0.047088623, -0.044158936, -0.022598267, -0.0115356445, -0.03274536, 0.0038280487, -0.0021362305, 0.08282471, -0.018234253, 0.029373169, 0.024642944, -0.002521515, -0.057006836, 0.0044288635, 0.008529663, ...
http://arxiv.org/abs/2212.13205v1
Personalized Prediction of Offensive News Comments by Considering the Characteristics of Commenters
Teruki Nakahara, Taketoshi Ushiama
cs.CL
When reading news articles on social networking services and news sites, readers can view comments marked by other people on these articles. By reading these comments, a reader can understand the public opinion about the news, and it is often helpful to grasp the overall picture of the news. However, these comments oft...
http://arxiv.org/pdf/2212.13205v1
2022-12-26T16:19:03
2022-12-26T16:19:03
{ "title": [ 0.008201599, 0.0065994263, -0.04663086, -0.062805176, 0.0206604, -0.01159668, -0.00053596497, 0.0046577454, -0.017532349, 0.033050537, -0.0077171326, 0.009124756, -0.0126953125, -0.008125305, -0.05899048, -0.006515503, -0.0057029724, ...
http://arxiv.org/abs/2212.13201v2
Highlighting Named Entities in Input for Auto-Formulation of Optimization Problems
Neeraj Gangwar, Nickvash Kani
cs.CL
Operations research deals with modeling and solving real-world problems as mathematical optimization problems. While solving mathematical systems is accomplished by analytical software, formulating a problem as a set of mathematical operations has been typically done manually by domain experts. Recent machine learning ...
http://arxiv.org/pdf/2212.13201v2
2023-03-31T22:13:18
2022-12-26T16:13:57
{ "title": [ -0.009521484, 0.002620697, -0.014251709, -0.027450562, -0.013755798, -0.005130768, -0.03930664, -0.031555176, -0.015701294, 0.019515991, 0.009635925, 0.016021729, -0.007133484, -0.033691406, -0.006225586, 0.0069618225, -0.0039482117, ...
http://arxiv.org/abs/2212.13196v1
Biologically Inspired Design Concept Generation Using Generative Pre-Trained Transformers
Qihao Zhu, Xinyu Zhang, Jianxi Luo
cs.CL
Biological systems in nature have evolved for millions of years to adapt and survive the environment. Many features they developed can be inspirational and beneficial for solving technical problems in modern industries. This leads to a specific form of design-by-analogy called bio-inspired design (BID). Although BID as...
http://arxiv.org/pdf/2212.13196v1
2022-12-26T16:06:04
2022-12-26T16:06:04
{ "title": [ -0.03729248, -0.016571045, -0.024261475, -0.017807007, -0.021224976, -0.010368347, -0.040222168, -0.021514893, 0.0040283203, 0.026947021, -0.030303955, 0.008804321, -0.017822266, -0.012626648, -0.03967285, -0.013450623, -0.0017261505, ...
http://arxiv.org/abs/2212.13138v1
Large Language Models Encode Clinical Knowledge
Karan Singhal, Shekoofeh Azizi, Tao Tu, S. Sara Mahdavi, Jason Wei, Hyung Won Chung, Nathan Scales, Ajay Tanwani, Heather Cole-Lewis, Stephen Pfohl, Perry Payne, Martin Seneviratne, Paul Gamble, Chris Kelly, Nathaneal Scharli, Aakanksha Chowdhery, Philip Mansfield, Blaise Aguera y Arcas, Dale Webster, Greg S. Corrado, ...
cs.CL
Large language models (LLMs) have demonstrated impressive capabilities in natural language understanding and generation, but the quality bar for medical and clinical applications is high. Today, attempts to assess models' clinical knowledge typically rely on automated evaluations on limited benchmarks. There is no stan...
http://arxiv.org/pdf/2212.13138v1
2022-12-26T14:28:24
2022-12-26T14:28:24
{ "title": [ -0.019638062, -0.017242432, -0.04663086, -0.09741211, -0.039611816, 0.0015764236, -0.09088135, 0.019104004, -0.004638672, 0.04156494, -0.010566711, -0.018737793, 0.007396698, -0.035186768, -0.011764526, 0.014877319, 0.026123047, 0.01...
http://arxiv.org/abs/2212.13036v1
Improving Complex Knowledge Base Question Answering via Question-to-Action and Question-to-Question Alignment
Yechun Tang, Xiaoxia Cheng, Weiming Lu
cs.CL, cs.AI
Complex knowledge base question answering can be achieved by converting questions into sequences of predefined actions. However, there is a significant semantic and structural gap between natural language and action sequences, which makes this conversion difficult. In this paper, we introduce an alignment-enhanced comp...
http://arxiv.org/pdf/2212.13036v1
2022-12-26T08:12:41
2022-12-26T08:12:41
{ "title": [ -0.005104065, 0.006088257, -0.045288086, -0.055908203, 0.0039749146, -0.00005799532, -0.02053833, -0.00044703484, 0.021438599, 0.03451538, -0.026748657, 0.021087646, 0.0021038055, -0.0066337585, -0.0446167, 0.01108551, 0.00032615662, ...
http://arxiv.org/abs/2212.13015v1
Skit-S2I: An Indian Accented Speech to Intent dataset
Shangeth Rajaa, Swaraj Dalmia, Kumarmanas Nethil
cs.CL, cs.LG, cs.SD, eess.AS
Conventional conversation assistants extract text transcripts from the speech signal using automatic speech recognition (ASR) and then predict intent from the transcriptions. Using end-to-end spoken language understanding (SLU), the intents of the speaker are predicted directly from the speech signal without requiring ...
http://arxiv.org/pdf/2212.13015v1
2022-12-26T05:10:43
2022-12-26T05:10:43
{ "title": [ -0.004547119, 0.010429382, -0.046203613, -0.051513672, 0.017105103, -0.015274048, -0.031311035, 0.007949829, -0.015289307, 0.031921387, -0.00009346008, -0.008728027, 0.027389526, -0.042999268, -0.021316528, 0.00983429, -0.0071487427, ...
http://arxiv.org/abs/2212.13005v1
TextBox 2.0: A Text Generation Library with Pre-trained Language Models
Tianyi Tang, Junyi Li, Zhipeng Chen, Yiwen Hu, Zhuohao Yu, Wenxun Dai, Zican Dong, Xiaoxue Cheng, Yuhao Wang, Wayne Xin Zhao, Jian-Yun Nie, Ji-Rong Wen
cs.CL
To facilitate research on text generation, this paper presents a comprehensive and unified library, TextBox 2.0, focusing on the use of pre-trained language models (PLMs). To be comprehensive, our library covers $13$ common text generation tasks and their corresponding $83$ datasets and further incorporates $45$ PLMs c...
http://arxiv.org/pdf/2212.13005v1
2022-12-26T03:50:36
2022-12-26T03:50:36
{ "title": [ -0.03869629, -0.021224976, -0.04269409, -0.06097412, -0.0022258759, 0.014556885, -0.06304932, -0.002035141, 0.005302429, 0.045166016, -0.0057296753, -0.0006952286, 0.020248413, -0.013343811, -0.051483154, 0.016983032, -0.0154953, 0.0...
http://arxiv.org/abs/2212.12955v1
Search, Structure, and Sentiment: A Comparative Analysis of Network Opinion in Different Query Types on Twitter
Joshua Midha
cs.SI, cs.CL, cs.CY
Understanding the relationship between structure and sentiment is essential in highlighting future operations with online social networks. More specifically, within popular conversation on Twitter. This paper provides a development on the relationship between the two variables: structure, defined as the composition of ...
http://arxiv.org/pdf/2212.12955v1
2022-12-25T20:48:40
2022-12-25T20:48:40
{ "title": [ 0.010650635, -0.00907135, -0.01638794, -0.047058105, 0.007850647, -0.0062217712, -0.007003784, -0.0032215118, -0.0012464523, 0.04147339, -0.056396484, 0.025222778, -0.016143799, 0.018630981, -0.06878662, -0.022201538, -0.006969452, -...
http://arxiv.org/abs/2212.12937v1
GAE-ISumm: Unsupervised Graph-Based Summarization of Indian Languages
Lakshmi Sireesha Vakada, Anudeep Ch, Mounika Marreddy, Subba Reddy Oota, Radhika Mamidi
cs.CL, cs.LG
Document summarization aims to create a precise and coherent summary of a text document. Many deep learning summarization models are developed mainly for English, often requiring a large training corpus and efficient pre-trained language models and tools. However, English summarization models for low-resource Indian la...
http://arxiv.org/pdf/2212.12937v1
2022-12-25T17:20:03
2022-12-25T17:20:03
{ "title": [ -0.028564453, -0.004650116, -0.031066895, -0.07800293, -0.013366699, 0.02583313, -0.09613037, 0.008903503, -0.01763916, 0.040527344, 0.00083494186, 0.02470398, 0.010627747, -0.011650085, -0.038909912, 0.016433716, -0.006629944, 0.043...
http://arxiv.org/abs/2212.12801v1
Linguistic Elements of Engaging Customer Service Discourse on Social Media
Sonam Singh, Anthony Rios
cs.CL
Customers are rapidly turning to social media for customer support. While brand agents on these platforms are motivated and well-intentioned to help and engage with customers, their efforts are often ignored if their initial response to the customer does not match a specific tone, style, or topic the customer is aiming...
http://arxiv.org/pdf/2212.12801v1
2022-12-24T18:49:03
2022-12-24T18:49:03
{ "title": [ 0.013122559, -0.0005955696, -0.0262146, -0.045654297, 0.0023384094, -0.0211792, -0.006980896, 0.0006504059, -0.025634766, 0.047088623, -0.017105103, 0.01574707, 0.0022468567, -0.0079956055, -0.032287598, -0.0040245056, 0.0018033981, ...
http://arxiv.org/abs/2212.12800v1
A Marker-based Neural Network System for Extracting Social Determinants of Health
Xingmeng Zhao, Anthony Rios
cs.CL
Objective. The impact of social determinants of health (SDoH) on patients' healthcare quality and the disparity is well-known. Many SDoH items are not coded in structured forms in electronic health records. These items are often captured in free-text clinical notes, but there are limited methods for automatically extra...
http://arxiv.org/pdf/2212.12800v1
2022-12-24T18:40:23
2022-12-24T18:40:23
{ "title": [ 0.020111084, 0.02078247, -0.019927979, -0.070251465, 0.0317688, -0.017868042, -0.016845703, -0.026184082, -0.030685425, 0.0137786865, 0.015838623, 0.011886597, -0.059570312, -0.027191162, -0.02796936, -0.015396118, 0.0034999847, -0.0...
http://arxiv.org/abs/2212.12799v1
A Comprehensive Study of Gender Bias in Chemical Named Entity Recognition Models
Xingmeng Zhao, Ali Niazi, Anthony Rios
cs.CL
Objective. Chemical named entity recognition (NER) models have the potential to impact a wide range of downstream tasks, from identifying adverse drug reactions to general pharmacoepidemiology. However, it is unknown whether these models work the same for everyone. Performance disparities can potentially cause harm rat...
http://arxiv.org/pdf/2212.12799v1
2022-12-24T18:35:35
2022-12-24T18:35:35
{ "title": [ -0.05429077, -0.010070801, -0.07745361, -0.053649902, -0.003566742, -0.0102005005, -0.051361084, 0.03186035, -0.011695862, 0.028533936, -0.03540039, 0.010658264, 0.00070142746, 0.0024986267, -0.02758789, -0.013214111, -0.03363037, 0....
http://arxiv.org/abs/2212.12795v2
Development of a Thermodynamics of Human Cognition and Human Culture
Diederik Aerts, Jonito Aerts Arguëlles, Lester Beltran, Sandro Sozzo
q-bio.NC, cs.CL, quant-ph
Inspired by foundational studies in classical and quantum physics, and by information retrieval studies in quantum information theory, we prove that the notions of 'energy' and 'entropy' can be consistently introduced in human language and, more generally, in human culture. More explicitly, if energy is attributed to w...
http://arxiv.org/pdf/2212.12795v2
2023-02-24T09:04:08
2022-12-24T18:19:05
{ "title": [ 0.016098022, -0.0032844543, -0.020477295, -0.02230835, -0.022583008, -0.037597656, -0.0075149536, -0.0010509491, -0.030136108, 0.043640137, -0.013206482, 0.0032176971, 0.020996094, -0.03161621, -0.0070152283, -0.00006914139, 0.032684326,...
http://arxiv.org/abs/2212.12672v1
Real or Fake Text?: Investigating Human Ability to Detect Boundaries Between Human-Written and Machine-Generated Text
Liam Dugan, Daphne Ippolito, Arun Kirubarajan, Sherry Shi, Chris Callison-Burch
cs.CL, cs.AI, cs.HC, I.2.7
As text generated by large language models proliferates, it becomes vital to understand how humans engage with such text, and whether or not they are able to detect when the text they are reading did not originate with a human writer. Prior work on human detection of generated text focuses on the case where an entire p...
http://arxiv.org/pdf/2212.12672v1
2022-12-24T06:40:25
2022-12-24T06:40:25
{ "title": [ -0.0026550293, 0.0065727234, -0.009422302, -0.016021729, 0.010253906, -0.005592346, -0.006351471, -0.035339355, -0.046539307, 0.02017212, -0.026367188, 0.019088745, -0.0059776306, -0.008399963, -0.027023315, -0.027816772, -0.02784729, ...
http://arxiv.org/abs/2212.12662v1
Optimizing Deep Transformers for Chinese-Thai Low-Resource Translation
Wenjie Hao, Hongfei Xu, Lingling Mu, Hongying Zan
cs.CL
In this paper, we study the use of deep Transformer translation model for the CCMT 2022 Chinese-Thai low-resource machine translation task. We first explore the experiment settings (including the number of BPE merge operations, dropout probability, embedding size, etc.) for the low-resource scenario with the 6-layer Tr...
http://arxiv.org/pdf/2212.12662v1
2022-12-24T05:35:04
2022-12-24T05:35:04
{ "title": [ -0.022445679, 0.00095272064, -0.00060367584, -0.053833008, -0.025115967, 0.014656067, -0.058776855, -0.014884949, -0.01084137, 0.045288086, -0.03503418, 0.0067596436, 0.025772095, -0.021881104, -0.043121338, -0.0007748604, 0.0078048706, ...
http://arxiv.org/abs/2212.12652v1
STRUDEL: Structured Dialogue Summarization for Dialogue Comprehension
Borui Wang, Chengcheng Feng, Arjun Nair, Madelyn Mao, Jai Desai, Asli Celikyilmaz, Haoran Li, Yashar Mehdad, Dragomir Radev
cs.CL
Abstractive dialogue summarization has long been viewed as an important standalone task in natural language processing, but no previous work has explored the possibility of whether abstractive dialogue summarization can also be used as a means to boost an NLP system's performance on other important dialogue comprehensi...
http://arxiv.org/pdf/2212.12652v1
2022-12-24T04:39:54
2022-12-24T04:39:54
{ "title": [ -0.026397705, -0.019927979, -0.036499023, -0.09893799, -0.015792847, 0.02204895, -0.08807373, 0.010368347, 0.019348145, 0.019622803, 0.0011997223, 0.005432129, -0.008354187, -0.020736694, -0.04626465, 0.015281677, -0.018005371, 0.027...
http://arxiv.org/abs/2212.12643v1
Utilizing Priming to Identify Optimal Class Ordering to Alleviate Catastrophic Forgetting
Gabriel Mantione-Holmes, Justin Leo, Jugal Kalita
cs.LG, cs.AI, cs.CL
In order for artificial neural networks to begin accurately mimicking biological ones, they must be able to adapt to new exigencies without forgetting what they have learned from previous training. Lifelong learning approaches to artificial neural networks attempt to strive towards this goal, yet have not progressed fa...
http://arxiv.org/pdf/2212.12643v1
2022-12-24T03:22:26
2022-12-24T03:22:26
{ "title": [ 0.039733887, 0.008705139, -0.025009155, -0.078308105, 0.014122009, -0.028060913, -0.024459839, -0.02746582, -0.015296936, 0.015472412, -0.011932373, 0.021347046, -0.0054397583, -0.020370483, -0.020965576, -0.010879517, 0.009422302, 0...
http://arxiv.org/abs/2301.00836v1
Kannudi -- A Reference Editor for Kannada
Vishweshwar V. Dixit
cs.HC, cs.CL
Kannudi is a reference editor for Kannada based on OPOK! and OHOK! principles, and domain knowledge. It introduces a method of input for Kannada, called OHOK!, that is, Ottu Haku Ottu Kodu! (apply pressure and give ottu). This is especially suited for pressure sensitive input devices, though the current online implemen...
http://arxiv.org/pdf/2301.00836v1
2022-12-24T01:40:56
2022-12-24T01:40:56
{ "title": [ -0.09564209, -0.0335083, -0.067993164, 0.0038070679, -0.05307007, 0.0019102097, -0.05657959, 0.019332886, -0.008895874, 0.058563232, 0.00089502335, -0.013496399, 0.09667969, -0.042907715, -0.012504578, 0.0017089844, 0.0059814453, 0.0...
http://arxiv.org/abs/2212.12584v1
Neural Transition-based Parsing of Library Deprecations
Petr Babkin, Nacho Navarro, Salwa Alamir, Sameena Shah
cs.CL
This paper tackles the challenging problem of automating code updates to fix deprecated API usages of open source libraries by analyzing their release notes. Our system employs a three-tier architecture: first, a web crawler service retrieves deprecation documentation from the web; then a specially built parser process...
http://arxiv.org/pdf/2212.12584v1
2022-12-23T20:48:33
2022-12-23T20:48:33
{ "title": [ -0.02935791, 0.020462036, -0.033203125, -0.06286621, 0.0014944077, 0.0052375793, -0.07281494, -0.00617218, 0.025238037, 0.022232056, -0.018127441, -0.037719727, -0.04272461, -0.01625061, -0.06567383, -0.04232788, -0.0077400208, 0.015...
http://arxiv.org/abs/2212.12510v2
MicroBERT: Effective Training of Low-resource Monolingual BERTs through Parameter Reduction and Multitask Learning
Luke Gessler, Amir Zeldes
cs.CL
Transformer language models (TLMs) are critical for most NLP tasks, but they are difficult to create for low-resource languages because of how much pretraining data they require. In this work, we investigate two techniques for training monolingual TLMs in a low-resource setting: greatly reducing TLM size, and complemen...
http://arxiv.org/pdf/2212.12510v2
2023-01-05T03:14:58
2022-12-23T18:18:20
{ "title": [ -0.00064611435, -0.021408081, -0.02670288, -0.05996704, -0.018936157, 0.00037407875, -0.05984497, 0.007183075, -0.03466797, 0.049926758, 0.027801514, 0.0022392273, 0.033172607, -0.050476074, -0.011711121, -0.0019187927, -0.017105103, ...
http://arxiv.org/abs/2212.12496v1
Content Rating Classification for Fan Fiction
Yu Qiao, James Pope
cs.CL
Content ratings can enable audiences to determine the suitability of various media products. With the recent advent of fan fiction, the critical issue of fan fiction content ratings has emerged. Whether fan fiction content ratings are done voluntarily or required by regulation, there is the need to automate the content...
http://arxiv.org/pdf/2212.12496v1
2022-12-23T17:40:03
2022-12-23T17:40:03
{ "title": [ 0.012496948, 0.027008057, -0.03945923, 0.015029907, -0.006866455, -0.04083252, 0.034729004, -0.009971619, -0.02658081, 0.05645752, -0.020217896, 0.008338928, 0.041809082, -0.04434204, 0.017822266, 0.0050811768, 0.0048713684, 0.034606...
http://arxiv.org/abs/2212.12454v1
Generalizable Natural Language Processing Framework for Migraine Reporting from Social Media
Yuting Guo, Swati Rajwal, Sahithi Lakamana, Chia-Chun Chiang, Paul C. Menell, Adnan H. Shahid, Yi-Chieh Chen, Nikita Chhabra, Wan-Ju Chao, Chieh-Ju Chao, Todd J. Schwedt, Imon Banerjee, Abeed Sarker
cs.CL
Migraine is a high-prevalence and disabling neurological disorder. However, information migraine management in real-world settings could be limited to traditional health information sources. In this paper, we (i) verify that there is substantial migraine-related chatter available on social media (Twitter and Reddit), s...
http://arxiv.org/pdf/2212.12454v1
2022-12-23T16:29:34
2022-12-23T16:29:34
{ "title": [ 0.013160706, 0.004737854, -0.032226562, -0.07183838, 0.013381958, 0.011062622, -0.03640747, -0.010467529, -0.017211914, 0.02909851, -0.030426025, 0.017990112, 0.00047898293, -0.012771606, -0.020706177, -0.00491333, 0.007259369, 0.019...
http://arxiv.org/abs/2212.12363v1
Discovering Customer-Service Dialog System with Semi-Supervised Learning and Coarse-to-Fine Intent Detection
Zhitong Yang, Xing Ma, Anqi Liu, Zheyu Zhang
cs.CL
Task-oriented dialog(TOD) aims to assist users in achieving specific goals through multi-turn conversation. Recently, good results have been obtained based on large pre-trained models. However, the labeled-data scarcity hinders the efficient development of TOD systems at scale. In this work, we constructed a weakly sup...
http://arxiv.org/pdf/2212.12363v1
2022-12-23T14:36:43
2022-12-23T14:36:43
{ "title": [ 0.009269714, 0.022445679, -0.05328369, -0.0362854, 0.03363037, -0.03604126, -0.011222839, -0.008369446, 0.002565384, 0.030578613, -0.022750854, 0.009742737, -0.009384155, -0.015701294, -0.041412354, -0.016784668, -0.015350342, 0.0215...
http://arxiv.org/abs/2212.12238v1
From Judgement's Premises Towards Key Points
Oren Sultan, Rayen Dhahri, Yauheni Mardan, Tobias Eder, Georg Groh
cs.CL, cs.AI
Key Point Analysis(KPA) is a relatively new task in NLP that combines summarization and classification by extracting argumentative key points (KPs) for a topic from a collection of texts and categorizing their closeness to the different arguments. In our work, we focus on the legal domain and develop methods that ident...
http://arxiv.org/pdf/2212.12238v1
2022-12-23T10:20:58
2022-12-23T10:20:58
{ "title": [ 0.054138184, -0.025726318, 0.02583313, -0.0362854, -0.02229309, -0.014122009, 0.027160645, 0.0055236816, 0.007904053, 0.04763794, -0.014282227, 0.0012788773, -0.00071048737, -0.030319214, -0.015419006, -0.0017547607, 0.0014505386, 0....
http://arxiv.org/abs/2212.12213v1
Finetuning for Sarcasm Detection with a Pruned Dataset
Ishita Goyal, Priyank Bhandia, Sanjana Dulam
cs.CL, cs.AI
Sarcasm is a form of irony that involves saying or writing something that is opposite or opposite to what one really means, often in a humorous or mocking way. It is often used to mock or mock someone or something, or to be humorous or amusing. Sarcasm is usually conveyed through tone of voice, facial expressions, or o...
http://arxiv.org/pdf/2212.12213v1
2022-12-23T08:59:30
2022-12-23T08:59:30
{ "title": [ 0.0034809113, 0.0023097992, -0.039245605, -0.050720215, 0.010719299, -0.014350891, -0.018722534, -0.012466431, 0.027648926, 0.014251709, -0.048431396, -0.042541504, 0.022735596, -0.016326904, -0.062347412, -0.0184021, -0.021987915, 0...
http://arxiv.org/abs/2212.12192v1
Learning to Generate Questions by Enhancing Text Generation with Sentence Selection
Do Hoang Thai Duong, Nguyen Hong Son, Hung Le, Minh-Tien Nguyen
cs.CL
We introduce an approach for the answer-aware question generation problem. Instead of only relying on the capability of strong pre-trained language models, we observe that the information of answers and questions can be found in some relevant sentences in the context. Based on that, we design a model which includes two...
http://arxiv.org/pdf/2212.12192v1
2022-12-23T08:23:32
2022-12-23T08:23:32
{ "title": [ -0.0043411255, -0.002254486, -0.012802124, -0.028121948, -0.011795044, 0.0057868958, -0.017196655, -0.014945984, 0.0121154785, 0.009239197, -0.02017212, 0.023483276, -0.03503418, -0.008659363, -0.04498291, -0.0020618439, -0.01586914, ...
http://arxiv.org/abs/2212.12137v1
Dubbing in Practice: A Large Scale Study of Human Localization With Insights for Automatic Dubbing
William Brannon, Yogesh Virkar, Brian Thompson
cs.CL
We investigate how humans perform the task of dubbing video content from one language into another, leveraging a novel corpus of 319.57 hours of video from 54 professionally produced titles. This is the first such large-scale study we are aware of. The results challenge a number of assumptions commonly made in both qua...
http://arxiv.org/pdf/2212.12137v1
2022-12-23T04:12:52
2022-12-23T04:12:52
{ "title": [ -0.010528564, 0.016143799, -0.018676758, -0.008262634, 0.0066986084, -0.01525116, 0.0047683716, -0.052246094, -0.011856079, 0.0637207, -0.018508911, 0.0019235611, 0.038391113, -0.018737793, -0.049743652, 0.005718231, 0.008804321, 0.0...
http://arxiv.org/abs/2212.12131v1
Why Does Surprisal From Larger Transformer-Based Language Models Provide a Poorer Fit to Human Reading Times?
Byung-Doh Oh, William Schuler
cs.CL
This work presents a detailed linguistic analysis into why larger Transformer-based pre-trained language models with more parameters and lower perplexity nonetheless yield surprisal estimates that are less predictive of human reading times. First, regression analyses show a strictly monotonic, positive log-linear relat...
http://arxiv.org/pdf/2212.12131v1
2022-12-23T03:57:54
2022-12-23T03:57:54
{ "title": [ -0.021194458, -0.015350342, 0.0018072128, 0.0463562, -0.056671143, -0.018463135, -0.0078048706, -0.038360596, -0.023483276, 0.02633667, -0.040283203, 0.00013303757, -0.00806427, -0.018539429, 0.011184692, -0.038085938, 0.0058784485, ...
http://arxiv.org/abs/2212.12061v3
MN-DS: A Multilabeled News Dataset for News Articles Hierarchical Classification
Alina Petukhova, Nuno Fachada
cs.CL, cs.AI, cs.LG
This article presents a dataset of 10,917 news articles with hierarchical news categories collected between 1 January 2019 and 31 December 2019. We manually labeled the articles based on a hierarchical taxonomy with 17 first-level and 109 second-level categories. This dataset can be used to train machine learning model...
http://arxiv.org/pdf/2212.12061v3
2023-04-23T14:49:44
2022-12-22T22:27:26
{ "title": [ -0.014976501, 0.008995056, -0.058563232, -0.06137085, 0.008354187, -0.011657715, -0.021377563, -0.015457153, -0.009490967, 0.048309326, -0.021392822, 0.04360962, -0.0007619858, -0.0018644333, -0.046142578, 0.01121521, -0.02279663, 0....
http://arxiv.org/abs/2212.12048v1
Pushing the performances of ASR models on English and Spanish accents
Pooja Chitkara, Morgane Riviere, Jade Copet, Frank Zhang, Yatharth Saraf
cs.CL, cs.SD, eess.AS
Speech to text models tend to be trained and evaluated against a single target accent. This is especially true for English for which native speakers from the United States became the main benchmark. In this work, we are going to show how two simple methods: pre-trained embeddings and auxiliary classification losses can...
http://arxiv.org/pdf/2212.12048v1
2022-12-22T21:48:29
2022-12-22T21:48:29
{ "title": [ 0.02671814, -0.006385803, -0.008743286, -0.038848877, -0.03488159, -0.009857178, -0.01737976, 0.015525818, -0.046936035, 0.043273926, 0.008255005, -0.020889282, 0.046691895, -0.07678223, 0.029403687, 0.015472412, 0.0087509155, 0.0255...
http://arxiv.org/abs/2212.12043v1
When are Lemons Purple? The Concept Association Bias of CLIP
Yutaro Yamada, Yingtian Tang, Ilker Yildirim
cs.CV, cs.CL, cs.LG
Large-scale vision-language models such as CLIP have shown impressive performance on zero-shot image classification and image-to-text retrieval. However, such zero-shot performance of CLIP-based models does not realize in tasks that require a finer-grained correspondence between vision and language, such as Visual Ques...
http://arxiv.org/pdf/2212.12043v1
2022-12-22T21:27:12
2022-12-22T21:27:12
{ "title": [ 0.03930664, 0.008720398, -0.054840088, -0.023620605, -0.017227173, -0.049560547, -0.010749817, -0.020584106, -0.019638062, 0.024902344, -0.0033607483, -0.010322571, 0.0008225441, -0.041381836, -0.0015878677, -0.03765869, 0.003955841, ...
http://arxiv.org/abs/2212.12017v3
OPT-IML: Scaling Language Model Instruction Meta Learning through the Lens of Generalization
Srinivasan Iyer, Xi Victoria Lin, Ramakanth Pasunuru, Todor Mihaylov, Daniel Simig, Ping Yu, Kurt Shuster, Tianlu Wang, Qing Liu, Punit Singh Koura, Xian Li, Brian O'Horo, Gabriel Pereyra, Jeff Wang, Christopher Dewan, Asli Celikyilmaz, Luke Zettlemoyer, Ves Stoyanov
cs.CL
Recent work has shown that fine-tuning large pre-trained language models on a collection of tasks described via instructions, a.k.a. instruction-tuning, improves their zero and few-shot generalization to unseen tasks. However, there is a limited understanding of the performance trade-offs of different decisions made du...
http://arxiv.org/pdf/2212.12017v3
2023-01-30T05:25:59
2022-12-22T19:56:09
{ "title": [ -0.045074463, -0.0206604, -0.013702393, -0.037597656, -0.040985107, 0.018371582, -0.06774902, -0.022415161, -0.022537231, 0.03274536, -0.0069351196, 0.0109939575, -0.0082092285, -0.025970459, -0.025863647, -0.013198853, -0.01878357, ...
http://arxiv.org/abs/2212.12442v1
Alignment Entropy Regularization
Ehsan Variani, Ke Wu, David Rybach, Cyril Allauzen, Michael Riley
cs.CL, cs.LG
Existing training criteria in automatic speech recognition(ASR) permit the model to freely explore more than one time alignments between the feature and label sequences. In this paper, we use entropy to measure a model's uncertainty, i.e. how it chooses to distribute the probability mass over the set of allowed alignme...
http://arxiv.org/pdf/2212.12442v1
2022-12-22T18:51:02
2022-12-22T18:51:02
{ "title": [ -0.035614014, 0.032836914, -0.08459473, -0.025558472, 0.011726379, -0.0039520264, -0.046173096, -0.015586853, 0.037872314, 0.06378174, -0.008094788, 0.012481689, 0.03656006, -0.026931763, -0.03164673, 0.04949951, 0.034179688, 0.00895...
http://arxiv.org/abs/2212.11937v1
Efficient Induction of Language Models Via Probabilistic Concept Formation
Christopher J. MacLellan, Peter Matsakis, Pat Langley
cs.CL, cs.AI, cs.LG
This paper presents a novel approach to the acquisition of language models from corpora. The framework builds on Cobweb, an early system for constructing taxonomic hierarchies of probabilistic concepts that used a tabular, attribute-value encoding of training cases and concepts, making it unsuitable for sequential inpu...
http://arxiv.org/pdf/2212.11937v1
2022-12-22T18:16:58
2022-12-22T18:16:58
{ "title": [ -0.020431519, -0.011245728, -0.03390503, -0.07550049, 0.0042877197, 0.010688782, -0.06555176, 0.012077332, 0.0107040405, -0.0013237, -0.02368164, 0.016464233, -0.024093628, -0.018157959, -0.026901245, -0.0033836365, -0.026519775, 0.0...
http://arxiv.org/abs/2212.11856v1
Multilingual News Location Detection using an Entity-Based Siamese Network with Semi-Supervised Contrastive Learning and Knowledge Base
Víctor Suárez-Paniagua, Steven Derby, Tri Kurniawan Wijaya
cs.CL, cs.AI
Early detection of relevant locations in a piece of news is especially important in extreme events such as environmental disasters, war conflicts, disease outbreaks, or political turmoils. Additionally, this detection also helps recommender systems to promote relevant news based on user locations. Note that, when the r...
http://arxiv.org/pdf/2212.11856v1
2022-12-22T16:42:21
2022-12-22T16:42:21
{ "title": [ 0.009010315, 0.012321472, -0.020721436, -0.0088272095, -0.00036859512, -0.014717102, 0.006149292, -0.038879395, 0.016159058, 0.046722412, -0.0032901764, 0.004814148, 0.006286621, -0.030593872, -0.026535034, 0.004322052, -0.008094788, ...
http://arxiv.org/abs/2212.11791v1
Training Integer-Only Deep Recurrent Neural Networks
Vahid Partovi Nia, Eyyüb Sari, Vanessa Courville, Masoud Asgharian
cs.LG, cs.CL, I.2.6
Recurrent neural networks (RNN) are the backbone of many text and speech applications. These architectures are typically made up of several computationally complex components such as; non-linear activation functions, normalization, bi-directional dependence and attention. In order to maintain good accuracy, these compo...
http://arxiv.org/pdf/2212.11791v1
2022-12-22T15:22:36
2022-12-22T15:22:36
{ "title": [ -0.020629883, -0.0031604767, -0.054473877, -0.07458496, 0.018234253, -0.009353638, -0.034423828, 0.018951416, 0.013710022, 0.04248047, -0.0048103333, 0.010551453, -0.011924744, -0.008377075, -0.059020996, 0.012336731, -0.0107040405, ...
http://arxiv.org/abs/2301.06859v1
Methodological reflections for AI alignment research using human feedback
Thilo Hagendorff, Sarah Fabi
cs.HC, cs.AI, cs.CL, cs.LG
The field of artificial intelligence (AI) alignment aims to investigate whether AI technologies align with human interests and values and function in a safe and ethical manner. AI alignment is particularly relevant for large language models (LLMs), which have the potential to exhibit unintended behavior due to their ab...
http://arxiv.org/pdf/2301.06859v1
2022-12-22T14:27:33
2022-12-22T14:27:33
{ "title": [ -0.05517578, 0.008346558, -0.046417236, 0.018737793, -0.012451172, -0.025497437, 0.0035686493, -0.04144287, 0.0030517578, 0.036132812, -0.035247803, -0.025024414, 0.008163452, -0.014625549, -0.055480957, -0.015312195, -0.0032424927, ...
http://arxiv.org/abs/2212.11685v2
Text Generation with Diffusion Language Models: A Pre-training Approach with Continuous Paragraph Denoise
Zhenghao Lin, Yeyun Gong, Yelong Shen, Tong Wu, Zhihao Fan, Chen Lin, Nan Duan, Weizhu Chen
cs.CL, cs.LG
In this paper, we introduce a novel dIffusion language modEl pre-training framework for text generation, which we call GENIE. GENIE is a large-scale pretrained diffusion language model that consists of an encoder and a diffusion-based decoder, which can generate text by gradually transforming a random noise sequence in...
http://arxiv.org/pdf/2212.11685v2
2023-02-17T17:14:52
2022-12-22T13:17:11
{ "title": [ -0.03967285, -0.0003297329, -0.04776001, -0.025909424, -0.0006842613, -0.003921509, -0.02671814, -0.0030021667, 0.0178833, 0.033599854, -0.015289307, 0.0076065063, 0.010986328, -0.013717651, -0.054779053, 0.0036792755, -0.023834229, ...
http://arxiv.org/abs/2212.11985v1
Multi-Lingual DALL-E Storytime
Noga Mudrik, Adam S. Charles
cs.CL
While recent advancements in artificial intelligence (AI) language models demonstrate cutting-edge performance when working with English texts, equivalent models do not exist in other languages or do not reach the same performance level. This undesired effect of AI advancements increases the gap between access to new t...
http://arxiv.org/pdf/2212.11985v1
2022-12-22T07:06:35
2022-12-22T07:06:35
{ "title": [ 0.005924225, -0.0050201416, -0.029251099, -0.028060913, -0.04724121, -0.012557983, -0.038970947, 0.0044288635, -0.01965332, 0.043151855, -0.011817932, -0.010604858, 0.021026611, -0.023422241, 0.045135498, -0.008293152, -0.010650635, ...
http://arxiv.org/abs/2212.11456v1
CAMeMBERT: Cascading Assistant-Mediated Multilingual BERT
Dan DeGenaro, Jugal Kalita
cs.CL
Large language models having hundreds of millions, and even billions, of parameters have performed extremely well on a variety of natural language processing (NLP) tasks. Their widespread use and adoption, however, is hindered by the lack of availability and portability of sufficiently large computational resources. Th...
http://arxiv.org/pdf/2212.11456v1
2022-12-22T02:19:25
2022-12-22T02:19:25
{ "title": [ -0.021713257, -0.038085938, -0.011833191, -0.07122803, -0.033081055, 0.020980835, -0.07800293, 0.0072021484, 0.0116119385, 0.07873535, 0.019515991, -0.0018720627, 0.08728027, -0.029449463, -0.022155762, 0.047210693, -0.019821167, 0.0...
http://arxiv.org/abs/2212.11455v1
Understanding Postpartum Parents' Experiences via Two Digital Platforms
Xuewen Yao, Miriam Mikhelson, Megan Micheletti, Eunsol Choi, S Craig Watkins, Edison Thomaz, Kaya De Barbaro
cs.CL
Digital platforms, including online forums and helplines, have emerged as avenues of support for caregivers suffering from postpartum mental health distress. Understanding support seekers' experiences as shared on these platforms could provide crucial insight into caregivers' needs during this vulnerable time. In the c...
http://arxiv.org/pdf/2212.11455v1
2022-12-22T02:13:30
2022-12-22T02:13:30
{ "title": [ 0.0059890747, 0.021408081, -0.03274536, -0.011146545, 0.0013065338, -0.033050537, -0.005657196, -0.008621216, -0.041992188, 0.03314209, -0.014328003, 0.010597229, 0.0021572113, -0.044921875, 0.005405426, -0.030654907, 0.029174805, -0...
http://arxiv.org/abs/2212.11382v1
Automatic Emotion Modelling in Written Stories
Lukas Christ, Shahin Amiriparian, Manuel Milling, Ilhan Aslan, Björn W. Schuller
cs.CL
Telling stories is an integral part of human communication which can evoke emotions and influence the affective states of the audience. Automatically modelling emotional trajectories in stories has thus attracted considerable scholarly interest. However, as most existing works have been limited to unsupervised dictiona...
http://arxiv.org/pdf/2212.11382v1
2022-12-21T21:46:01
2022-12-21T21:46:01
{ "title": [ -0.0357666, -0.0056762695, -0.023330688, -0.02835083, -0.017440796, 0.0024738312, -0.01763916, -0.010971069, 0.038269043, 0.032958984, -0.021026611, -0.020355225, 0.015029907, -0.029159546, -0.059509277, 0.013435364, -0.015052795, 0....
http://arxiv.org/abs/2212.11353v1
Contrastive Distillation Is a Sample-Efficient Self-Supervised Loss Policy for Transfer Learning
Chris Lengerich, Gabriel Synnaeve, Amy Zhang, Hugh Leather, Kurt Shuster, François Charton, Charysse Redwood
cs.CL, cs.LG
Traditional approaches to RL have focused on learning decision policies directly from episodic decisions, while slowly and implicitly learning the semantics of compositional representations needed for generalization. While some approaches have been adopted to refine representations via auxiliary self-supervised losses ...
http://arxiv.org/pdf/2212.11353v1
2022-12-21T20:43:46
2022-12-21T20:43:46
{ "title": [ 0.058166504, 0.022079468, -0.008255005, -0.03463745, -0.0018453598, -0.029266357, -0.022628784, -0.055419922, 0.032684326, -0.0062065125, -0.050872803, -0.004043579, -0.003522873, -0.027496338, 0.0028095245, -0.027816772, -0.0008506775, ...
http://arxiv.org/abs/2212.11311v1
What do LLMs Know about Financial Markets? A Case Study on Reddit Market Sentiment Analysis
Xiang Deng, Vasilisa Bashlovkina, Feng Han, Simon Baumgartner, Michael Bendersky
cs.CL, cs.AI, cs.LG, cs.SI
Market sentiment analysis on social media content requires knowledge of both financial markets and social media jargon, which makes it a challenging task for human raters. The resulting lack of high-quality labeled data stands in the way of conventional supervised learning methods. Instead, we approach this problem usi...
http://arxiv.org/pdf/2212.11311v1
2022-12-21T19:11:19
2022-12-21T19:11:19
{ "title": [ -0.01007843, 0.00907135, -0.046905518, 0.012931824, -0.0064849854, -0.031829834, -0.0016489029, -0.0309906, -0.02041626, 0.028030396, -0.03427124, 0.007972717, -0.0048103333, -0.029037476, -0.029815674, -0.03186035, 0.007259369, 0.02...
http://arxiv.org/abs/2212.11270v1
Generalized Decoding for Pixel, Image, and Language
Xueyan Zou, Zi-Yi Dou, Jianwei Yang, Zhe Gan, Linjie Li, Chunyuan Li, Xiyang Dai, Harkirat Behl, Jianfeng Wang, Lu Yuan, Nanyun Peng, Lijuan Wang, Yong Jae Lee, Jianfeng Gao
cs.CV, cs.CL
We present X-Decoder, a generalized decoding model that can predict pixel-level segmentation and language tokens seamlessly. X-Decodert takes as input two types of queries: (i) generic non-semantic queries and (ii) semantic queries induced from text inputs, to decode different pixel-level and token-level outputs in the...
http://arxiv.org/pdf/2212.11270v1
2022-12-21T18:58:41
2022-12-21T18:58:41
{ "title": [ -0.03665161, -0.010116577, -0.027786255, -0.0519104, 0.000043094158, 0.0021247864, -0.031555176, 0.0069122314, 0.018096924, 0.04953003, -0.026290894, -0.007282257, 0.028137207, -0.01727295, -0.04034424, 0.036895752, 0.023162842, 0.01...
http://arxiv.org/abs/2212.11261v2
Contrastive Language-Vision AI Models Pretrained on Web-Scraped Multimodal Data Exhibit Sexual Objectification Bias
Robert Wolfe, Yiwei Yang, Bill Howe, Aylin Caliskan
cs.CY, cs.AI, cs.CL, cs.CV, cs.LG
Nine language-vision AI models trained on web scrapes with the Contrastive Language-Image Pretraining (CLIP) objective are evaluated for evidence of a bias studied by psychologists: the sexual objectification of girls and women, which occurs when a person's human characteristics, such as emotions, are disregarded and t...
http://arxiv.org/pdf/2212.11261v2
2023-05-15T23:49:27
2022-12-21T18:54:19
{ "title": [ 0.028045654, 0.012199402, -0.01436615, -0.003921509, -0.014465332, -0.042388916, 0.018630981, -0.019165039, 0.033813477, 0.033294678, -0.038848877, -0.018035889, 0.024017334, -0.02116394, -0.039001465, -0.023208618, -0.004032135, 0.0...
http://arxiv.org/abs/2212.11234v1
Improving Narrative Relationship Embeddings by Training with Additional Inverse-Relationship Constraints
Mikolaj Figurski
cs.CL, cs.LG, I.2.7
We consider the problem of embedding character-entity relationships from the reduced semantic space of narratives, proposing and evaluating the assumption that these relationships hold under a reflection operation. We analyze this assumption and compare the approach to a baseline state-of-the-art model with a unique ev...
http://arxiv.org/pdf/2212.11234v1
2022-12-21T17:59:11
2022-12-21T17:59:11
{ "title": [ -0.0061836243, -0.003616333, -0.02432251, -0.046844482, -0.008880615, 0.0014371872, -0.025863647, -0.011764526, 0.0045700073, 0.0035495758, -0.023345947, 0.019561768, -0.00818634, -0.013366699, -0.0014381409, -0.0049324036, -0.033935547,...
http://arxiv.org/abs/2212.11281v1
Language models are better than humans at next-token prediction
Buck Shlegeris, Fabien Roger, Lawrence Chan, Euan McLean
cs.CL, cs.AI, cs.LG
Current language models are considered to have sub-human capabilities at natural language tasks like question-answering or writing code. However, language models are not trained to perform well at these tasks, they are trained to accurately predict the next token given previous tokes in tokenized text. It is not clear ...
http://arxiv.org/pdf/2212.11281v1
2022-12-21T17:58:01
2022-12-21T17:58:01
{ "title": [ 0.02279663, 0.002544403, -0.00932312, -0.05532837, -0.06317139, -0.014228821, -0.06890869, -0.017440796, -0.023483276, 0.048431396, -0.046569824, -0.007701874, 0.025421143, -0.012199402, -0.020690918, -0.014785767, -0.005218506, 0.02...
http://arxiv.org/abs/2212.11185v1
Entropy- and Distance-Based Predictors From GPT-2 Attention Patterns Predict Reading Times Over and Above GPT-2 Surprisal
Byung-Doh Oh, William Schuler
cs.CL
Transformer-based large language models are trained to make predictions about the next word by aggregating representations of previous tokens through their self-attention mechanism. In the field of cognitive modeling, such attention patterns have recently been interpreted as embodying the process of cue-based retrieval...
http://arxiv.org/pdf/2212.11185v1
2022-12-21T16:56:07
2022-12-21T16:56:07
{ "title": [ 0.012107849, 0.040100098, -0.04043579, 0.020065308, -0.0024204254, -0.051971436, -0.007888794, -0.06524658, -0.0004992485, 0.03491211, -0.0065612793, -0.011642456, 0.017028809, -0.035095215, -0.0039787292, 0.0017004013, -0.0019550323, ...
http://arxiv.org/abs/2212.11182v1
Universal versus system-specific features of punctuation usage patterns in~major Western~languages
Tomasz Stanisz, Stanislaw Drozdz, Jaroslaw Kwapien
cs.CL, physics.soc-ph
The celebrated proverb that "speech is silver, silence is golden" has a long multinational history and multiple specific meanings. In written texts punctuation can in fact be considered one of its manifestations. Indeed, the virtue of effectively speaking and writing involves - often decisively - the capacity to apply ...
http://arxiv.org/pdf/2212.11182v1
2022-12-21T16:52:10
2022-12-21T16:52:10
{ "title": [ 0.00057792664, 0.0020256042, -0.041015625, -0.042663574, -0.015930176, -0.019836426, -0.031021118, 0.0015115738, -0.041900635, 0.052978516, -0.011138916, 0.017074585, 0.02104187, -0.026885986, -0.0046195984, 0.00472641, -0.017211914, ...
http://arxiv.org/abs/2212.13917v1
Multimodal Emotion Recognition among Couples from Lab Settings to Daily Life using Smartwatches
George Boateng
cs.HC, cs.CL, cs.LG, cs.SD, eess.SP
Couples generally manage chronic diseases together and the management takes an emotional toll on both patients and their romantic partners. Consequently, recognizing the emotions of each partner in daily life could provide an insight into their emotional well-being in chronic disease management. The emotions of partner...
http://arxiv.org/pdf/2212.13917v1
2022-12-21T16:41:11
2022-12-21T16:41:11
{ "title": [ -0.049835205, 0.016464233, -0.056884766, 0.016799927, 0.015380859, -0.022109985, -0.0022945404, -0.013809204, 0.011352539, 0.030166626, -0.009117126, -0.013076782, 0.0015897751, -0.030548096, -0.053833008, -0.0048332214, 0.0067977905, ...
http://arxiv.org/abs/2212.12407v1
Text classification in shipping industry using unsupervised models and Transformer based supervised models
Ying Xie, Dongping Song
cs.CL, cs.LG, ACM-class: J.m
Obtaining labelled data in a particular context could be expensive and time consuming. Although different algorithms, including unsupervised learning, semi-supervised learning, self-learning have been adopted, the performance of text classification varies with context. Given the lack of labelled dataset, we proposed a ...
http://arxiv.org/pdf/2212.12407v1
2022-12-21T16:00:44
2022-12-21T16:00:44
{ "title": [ -0.025360107, 0.021591187, -0.050323486, -0.0054473877, 0.005531311, -0.025878906, -0.029876709, -0.04534912, 0.0016355515, 0.020629883, -0.06750488, 0.021087646, -0.0074691772, -0.0036716461, -0.02760315, -0.023010254, -0.016845703, ...
http://arxiv.org/abs/2212.10983v2
Computer says "No": The Case Against Empathetic Conversational AI
Alba Curry, Amanda Cercas Curry
cs.CL, cs.HC
Emotions are an integral part of human cognition and they guide not only our understanding of the world but also our actions within it. As such, whether we soothe or flame an emotion is not inconsequential. Recent work in conversational AI has focused on responding empathetically to users, validating and soothing their...
http://arxiv.org/pdf/2212.10983v2
2023-07-06T10:52:49
2022-12-21T12:36:17
{ "title": [ -0.0121154785, -0.011123657, -0.018844604, -0.017547607, -0.010475159, -0.023742676, 0.019683838, -0.017150879, -0.01474762, 0.018371582, -0.023666382, 0.0041160583, 0.005504608, -0.028076172, -0.029373169, -0.018676758, -0.0070648193, ...
http://arxiv.org/abs/2212.10947v3
Parallel Context Windows for Large Language Models
Nir Ratner, Yoav Levine, Yonatan Belinkov, Ori Ram, Inbal Magar, Omri Abend, Ehud Karpas, Amnon Shashua, Kevin Leyton-Brown, Yoav Shoham
cs.CL
When applied to processing long text, Large Language Models (LLMs) are limited by their context window. Existing efforts to address this limitation involve training specialized architectures, and cannot be easily applied to off-the-shelf LLMs. We present Parallel Context Windows (PCW), a method that alleviates the cont...
http://arxiv.org/pdf/2212.10947v3
2023-08-01T16:48:47
2022-12-21T11:38:51
{ "title": [ -0.0418396, -0.019302368, -0.041137695, -0.08959961, -0.004787445, 0.022094727, -0.06713867, 0.03527832, 0.013580322, 0.06665039, -0.019851685, -0.001367569, 0.033050537, -0.008110046, -0.05340576, 0.0259552, 0.0023880005, 0.01940918...
http://arxiv.org/abs/2212.10938v1
Critic-Guided Decoding for Controlled Text Generation
Minbeom Kim, Hwanhee Lee, Kang Min Yoo, Joonsuk Park, Hwaran Lee, Kyomin Jung
cs.CL
Steering language generation towards objectives or away from undesired content has been a long-standing goal in utilizing language models (LM). Recent work has demonstrated reinforcement learning and weighted decoding as effective approaches to achieve a higher level of language control and quality with pros and cons. ...
http://arxiv.org/pdf/2212.10938v1
2022-12-21T11:25:41
2022-12-21T11:25:41
{ "title": [ -0.07086182, -0.027954102, -0.029266357, -0.02230835, -0.045013428, 0.024795532, -0.06500244, -0.0027484894, -0.00075387955, 0.045562744, -0.009735107, 0.015213013, 0.011070251, -0.01184845, -0.05178833, 0.013366699, -0.010353088, 0....
http://arxiv.org/abs/2212.10935v1
Esports Data-to-commentary Generation on Large-scale Data-to-text Dataset
Zihan Wang, Naoki Yoshinaga
cs.CL
Esports, a sports competition using video games, has become one of the most important sporting events in recent years. Although the amount of esports data is increasing than ever, only a small fraction of those data accompanies text commentaries for the audience to retrieve and understand the plays. Therefore, in this ...
http://arxiv.org/pdf/2212.10935v1
2022-12-21T11:23:31
2022-12-21T11:23:31
{ "title": [ 0.007167816, -0.021652222, -0.017227173, -0.049743652, -0.0126953125, -0.01259613, -0.021514893, 0.0025901794, 0.024490356, 0.043762207, -0.05596924, 0.005771637, 0.0008382797, 0.0064735413, -0.06085205, -0.014587402, -0.001405716, 0...
http://arxiv.org/abs/2212.10933v2
Resolving Indirect Referring Expressions for Entity Selection
Mohammad Javad Hosseini, Filip Radlinski, Silvia Pareti, Annie Louis
cs.CL
Recent advances in language modeling have enabled new conversational systems. In particular, it is often desirable for people to make choices among specified options when using such systems. We address this problem of reference resolution, when people use natural expressions to choose between the entities. For example,...
http://arxiv.org/pdf/2212.10933v2
2023-05-26T20:17:15
2022-12-21T11:22:38
{ "title": [ -0.0019760132, 0.011039734, -0.032318115, -0.0524292, 0.027557373, -0.0057792664, -0.016708374, -0.02947998, 0.010734558, 0.033447266, -0.017211914, 0.011428833, 0.039031982, -0.025268555, -0.012046814, 0.014678955, -0.021255493, 0.0...
http://arxiv.org/abs/2212.10929v1
SPT: Semi-Parametric Prompt Tuning for Multitask Prompted Learning
M Saiful Bari, Aston Zhang, Shuai Zheng, Xingjian Shi, Yi Zhu, Shafiq Joty, Mu Li
cs.CL, cs.AI, cs.LG
Pre-trained large language models can efficiently interpolate human-written prompts in a natural way. Multitask prompted learning can help generalization through a diverse set of tasks at once, thus enhancing the potential for more effective downstream fine-tuning. To perform efficient multitask-inference in the same b...
http://arxiv.org/pdf/2212.10929v1
2022-12-21T11:18:09
2022-12-21T11:18:09
{ "title": [ -0.017532349, -0.0132751465, -0.03527832, -0.052947998, -0.01335907, -0.015098572, -0.04272461, 0.0009198189, -0.034240723, -0.00028896332, -0.018463135, 0.016052246, -0.029663086, -0.038726807, 0.0043563843, -0.015777588, 0.0031585693, ...
http://arxiv.org/abs/2212.10923v1
Language Models as Inductive Reasoners
Zonglin Yang, Li Dong, Xinya Du, Hao Cheng, Erik Cambria, Xiaodong Liu, Jianfeng Gao, Furu Wei
cs.CL, cs.AI
Inductive reasoning is a core component of human intelligence. In the past research of inductive reasoning within computer science, logic language is used as representations of knowledge (facts and rules, more specifically). However, logic language can cause systematic problems for inductive reasoning such as disabilit...
http://arxiv.org/pdf/2212.10923v1
2022-12-21T11:12:14
2022-12-21T11:12:14
{ "title": [ -0.021759033, -0.018829346, -0.019134521, -0.06137085, -0.0052948, 0.008178711, -0.047698975, 0.001168251, -0.00198555, 0.012664795, -0.019805908, -0.0062332153, -0.018203735, -0.024429321, -0.032287598, 0.0013523102, -0.017074585, 0...
http://arxiv.org/abs/2212.10901v3
ALCAP: Alignment-Augmented Music Captioner
Zihao He, Weituo Hao, Wei-Tsung Lu, Changyou Chen, Kristina Lerman, Xuchen Song
cs.SD, cs.CL, cs.IR, cs.MM, eess.AS
Music captioning has gained significant attention in the wake of the rising prominence of streaming media platforms. Traditional approaches often prioritize either the audio or lyrics aspect of the music, inadvertently ignoring the intricate interplay between the two. However, a comprehensive understanding of music nec...
http://arxiv.org/pdf/2212.10901v3
2023-10-22T02:12:16
2022-12-21T10:20:54
{ "title": [ -0.03250122, -0.008575439, -0.06829834, -0.04840088, -0.013832092, -0.018051147, -0.039123535, 0.017150879, -0.0018491745, 0.06561279, -0.022521973, 0.010650635, 0.04559326, -0.013290405, -0.045898438, 0.003250122, -0.013267517, 0.03...
http://arxiv.org/abs/2212.10898v2
Training language models to summarize narratives improves brain alignment
Khai Loong Aw, Mariya Toneva
cs.CL, q-bio.NC
Building systems that achieve a deeper understanding of language is one of the central goals of natural language processing (NLP). Towards this goal, recent works have begun to train language models on narrative datasets which require extracting the most critical information by integrating across long contexts. However...
http://arxiv.org/pdf/2212.10898v2
2023-03-01T03:20:08
2022-12-21T10:15:19
{ "title": [ 0.0003736019, -0.003627777, -0.024261475, -0.018951416, -0.05886841, -0.015357971, -0.062683105, -0.027770996, -0.036621094, 0.062805176, 0.010307312, -0.0020885468, 0.021118164, -0.031066895, -0.03173828, 0.007549286, 0.0028133392, ...
http://arxiv.org/abs/2212.10888v1
A Survey of Mix-based Data Augmentation: Taxonomy, Methods, Applications, and Explainability
Chengtai Cao, Fan Zhou, Yurou Dai, Jianping Wang
cs.LG, cs.CL, cs.CV
Data augmentation (DA) is indispensable in modern machine learning and deep neural networks. The basic idea of DA is to construct new training data to improve the model's generalization by adding slightly disturbed versions of existing data or synthesizing new data. In this work, we review a small but essential subset ...
http://arxiv.org/pdf/2212.10888v1
2022-12-21T09:58:14
2022-12-21T09:58:14
{ "title": [ -0.000026464462, 0.0038871765, -0.082092285, -0.07757568, 0.03781128, -0.014823914, -0.03112793, 0.006427765, -0.020202637, 0.025543213, -0.022460938, 0.03668213, -0.020355225, -0.030227661, -0.011802673, 0.0033512115, -0.018615723, ...
http://arxiv.org/abs/2212.10879v1
Cross-Linguistic Syntactic Difference in Multilingual BERT: How Good is It and How Does It Affect Transfer?
Ningyu Xu, Tao Gui, Ruotian Ma, Qi Zhang, Jingting Ye, Menghan Zhang, Xuanjing Huang
cs.CL
Multilingual BERT (mBERT) has demonstrated considerable cross-lingual syntactic ability, whereby it enables effective zero-shot cross-lingual transfer of syntactic knowledge. The transfer is more successful between some languages, but it is not well understood what leads to this variation and whether it fairly reflects...
http://arxiv.org/pdf/2212.10879v1
2022-12-21T09:44:08
2022-12-21T09:44:08
{ "title": [ -0.028900146, -0.0074882507, -0.020492554, -0.046691895, -0.015701294, 0.004650116, -0.060424805, -0.0016498566, -0.028717041, 0.043762207, 0.00072574615, 0.038360596, 0.020553589, -0.022659302, -0.011932373, 0.021652222, -0.02897644, ...
http://arxiv.org/abs/2212.10873v3
Prompt-Augmented Linear Probing: Scaling beyond the Limit of Few-shot In-Context Learners
Hyunsoo Cho, Hyuhng Joon Kim, Junyeob Kim, Sang-Woo Lee, Sang-goo Lee, Kang Min Yoo, Taeuk Kim
cs.CL, cs.LG
Through in-context learning (ICL), large-scale language models are effective few-shot learners without additional model fine-tuning. However, the ICL performance does not scale well with the number of available training samples as it is limited by the inherent input length constraint of the underlying language model. M...
http://arxiv.org/pdf/2212.10873v3
2023-06-14T03:23:44
2022-12-21T09:37:05
{ "title": [ 0.024215698, 0.0030288696, -0.0012607574, -0.025238037, 0.0011663437, -0.022460938, 0.0070381165, -0.031433105, -0.015304565, 0.00043916702, -0.045928955, 0.01084137, -0.04550171, -0.032196045, -0.010559082, -0.038085938, 0.0031795502, ...