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/2302.09664v3
Semantic Uncertainty: Linguistic Invariances for Uncertainty Estimation in Natural Language Generation
Lorenz Kuhn, Yarin Gal, Sebastian Farquhar
cs.CL, cs.AI, cs.LG
We introduce a method to measure uncertainty in large language models. For tasks like question answering, it is essential to know when we can trust the natural language outputs of foundation models. We show that measuring uncertainty in natural language is challenging because of "semantic equivalence" -- different sent...
http://arxiv.org/pdf/2302.09664v3
2023-04-15T12:55:45
2023-02-19T20:10:07
{ "title": [ -0.06262207, -0.017868042, -0.06488037, -0.04446411, -0.012359619, -0.0022220612, -0.06707764, -0.0037231445, 0.017425537, 0.019821167, -0.015594482, 0.0124435425, 0.0033550262, -0.027999878, -0.039886475, 0.0118637085, -0.012413025, ...
http://arxiv.org/abs/2302.09650v1
Scaling Laws for Multilingual Neural Machine Translation
Patrick Fernandes, Behrooz Ghorbani, Xavier Garcia, Markus Freitag, Orhan Firat
cs.CL, cs.LG
In this work, we provide a large-scale empirical study of the scaling properties of multilingual neural machine translation models. We examine how increases in the model size affect the model performance and investigate the role of the training mixture composition on the scaling behavior. We find that changing the weig...
http://arxiv.org/pdf/2302.09650v1
2023-02-19T18:43:24
2023-02-19T18:43:24
{ "title": [ -0.030288696, -0.0025253296, -0.013298035, -0.046569824, -0.010665894, 0.0155181885, -0.049591064, 0.0054512024, -0.021209717, 0.05911255, -0.0037631989, -0.0011329651, 0.023483276, -0.011299133, -0.04135132, 0.014282227, -0.009178162, ...
http://arxiv.org/abs/2302.10199v1
Evaluating the Effectiveness of Pre-trained Language Models in Predicting the Helpfulness of Online Product Reviews
Ali Boluki, Javad Pourmostafa Roshan Sharami, Dimitar Shterionov
cs.CL, cs.LG
Businesses and customers can gain valuable information from product reviews. The sheer number of reviews often necessitates ranking them based on their potential helpfulness. However, only a few reviews ever receive any helpfulness votes on online marketplaces. Sorting all reviews based on the few existing votes can ca...
http://arxiv.org/pdf/2302.10199v1
2023-02-19T18:22:59
2023-02-19T18:22:59
{ "title": [ -0.018310547, -0.017532349, -0.019226074, -0.04333496, -0.019515991, -0.0073547363, -0.03768921, -0.008781433, -0.017654419, 0.04147339, -0.020965576, 0.012672424, 0.0149383545, -0.0000359416, -0.04937744, -0.0069274902, 0.0030326843, ...
http://arxiv.org/abs/2302.09632v1
HomoDistil: Homotopic Task-Agnostic Distillation of Pre-trained Transformers
Chen Liang, Haoming Jiang, Zheng Li, Xianfeng Tang, Bin Yin, Tuo Zhao
cs.CL, cs.LG
Knowledge distillation has been shown to be a powerful model compression approach to facilitate the deployment of pre-trained language models in practice. This paper focuses on task-agnostic distillation. It produces a compact pre-trained model that can be easily fine-tuned on various tasks with small computational cos...
http://arxiv.org/pdf/2302.09632v1
2023-02-19T17:37:24
2023-02-19T17:37:24
{ "title": [ 0.019104004, 0.02168274, -0.033081055, -0.09442139, 0.014923096, 0.00843811, -0.06555176, -0.026016235, -0.0016555786, 0.039123535, -0.02645874, -0.009819031, 0.031311035, -0.022613525, -0.023162842, 0.0030994415, 0.026382446, 0.0256...
http://arxiv.org/abs/2302.09618v1
Multilingual Content Moderation: A Case Study on Reddit
Meng Ye, Karan Sikka, Katherine Atwell, Sabit Hassan, Ajay Divakaran, Malihe Alikhani
cs.CL
Content moderation is the process of flagging content based on pre-defined platform rules. There has been a growing need for AI moderators to safeguard users as well as protect the mental health of human moderators from traumatic content. While prior works have focused on identifying hateful/offensive language, they ar...
http://arxiv.org/pdf/2302.09618v1
2023-02-19T16:36:33
2023-02-19T16:36:33
{ "title": [ 0.0135650635, 0.004219055, -0.027023315, -0.02772522, 0.021835327, -0.018585205, -0.0011129379, -0.020141602, -0.032958984, 0.05847168, 0.012428284, 0.012161255, 0.02973938, -0.024673462, -0.02557373, 0.021728516, -0.014099121, 0.051...
http://arxiv.org/abs/2302.09611v1
Exploring the Potential of Machine Translation for Generating Named Entity Datasets: A Case Study between Persian and English
Amir Sartipi, Afsaneh Fatemi
cs.CL, cs.AI
This study focuses on the generation of Persian named entity datasets through the application of machine translation on English datasets. The generated datasets were evaluated by experimenting with one monolingual and one multilingual transformer model. Notably, the CoNLL 2003 dataset has achieved the highest F1 score ...
http://arxiv.org/pdf/2302.09611v1
2023-02-19T16:12:21
2023-02-19T16:12:21
{ "title": [ -0.0037441254, -0.01574707, -0.023803711, -0.053619385, 0.0090408325, -0.00349617, -0.030151367, -0.0017108917, -0.0048561096, 0.026931763, -0.01701355, 0.007911682, 0.023910522, -0.021621704, -0.024551392, 0.010887146, -0.026367188, ...
http://arxiv.org/abs/2302.09582v4
Language-Specific Representation of Emotion-Concept Knowledge Causally Supports Emotion Inference
Ming Li, Yusheng Su, Hsiu-Yuan Huang, Jiali Cheng, Xin Hu, Xinmiao Zhang, Huadong Wang, Yujia Qin, Xiaozhi Wang, Zhiyuan Liu, Dan Zhang
cs.AI, cs.CL
Understanding how language supports emotion inference remains a topic of debate in emotion science. The present study investigated whether language-derived emotion-concept knowledge would causally support emotion inference by manipulating the language-specific knowledge representations in large language models. Using t...
http://arxiv.org/pdf/2302.09582v4
2023-08-21T09:44:19
2023-02-19T14:21:33
{ "title": [ 0.010429382, -0.015472412, -0.02394104, -0.07946777, -0.03062439, -0.017807007, -0.0524292, 0.011474609, 0.028656006, 0.011932373, -0.027328491, -0.01247406, 0.049743652, -0.055358887, -0.0057525635, 0.018096924, 0.025283813, 0.01252...
http://arxiv.org/abs/2302.10198v2
Can ChatGPT Understand Too? A Comparative Study on ChatGPT and Fine-tuned BERT
Qihuang Zhong, Liang Ding, Juhua Liu, Bo Du, Dacheng Tao
cs.CL
Recently, ChatGPT has attracted great attention, as it can generate fluent and high-quality responses to human inquiries. Several prior studies have shown that ChatGPT attains remarkable generation ability compared with existing models. However, the quantitative analysis of ChatGPT's understanding ability has been give...
http://arxiv.org/pdf/2302.10198v2
2023-03-02T14:33:12
2023-02-19T12:29:33
{ "title": [ -0.0044136047, -0.032562256, 0.003730774, -0.054473877, -0.013000488, 0.013320923, -0.026229858, 0.0064048767, -0.010475159, 0.0135269165, -0.025131226, 0.0033073425, 0.0019779205, -0.02053833, -0.025131226, 0.0058174133, -0.023788452, ...
http://arxiv.org/abs/2302.09540v1
Upvotes? Downvotes? No Votes? Understanding the relationship between reaction mechanisms and political discourse on Reddit
Orestis Papakyriakopoulos, Severin Engelmann, Amy Winecoff
cs.CY, cs.CL, cs.SI
A significant share of political discourse occurs online on social media platforms. Policymakers and researchers try to understand the role of social media design in shaping the quality of political discourse around the globe. In the past decades, scholarship on political discourse theory has produced distinct characte...
http://arxiv.org/pdf/2302.09540v1
2023-02-19T11:12:45
2023-02-19T11:12:45
{ "title": [ 0.022369385, 0.0036411285, -0.020736694, -0.03302002, 0.0178833, -0.043304443, 0.00038218498, -0.021240234, -0.02796936, 0.026184082, 0.008918762, 0.024505615, -0.02355957, -0.024215698, -0.0368042, -0.003145218, 0.011688232, 0.00880...
http://arxiv.org/abs/2302.09527v2
SanskritShala: A Neural Sanskrit NLP Toolkit with Web-Based Interface for Pedagogical and Annotation Purposes
Jivnesh Sandhan, Anshul Agarwal, Laxmidhar Behera, Tushar Sandhan, Pawan Goyal
cs.CL
We present a neural Sanskrit Natural Language Processing (NLP) toolkit named SanskritShala (a school of Sanskrit) to facilitate computational linguistic analyses for several tasks such as word segmentation, morphological tagging, dependency parsing, and compound type identification. Our systems currently report state-o...
http://arxiv.org/pdf/2302.09527v2
2023-05-29T07:36:21
2023-02-19T09:58:55
{ "title": [ -0.01777649, 0.0014133453, -0.030014038, -0.02760315, -0.0067863464, -0.0020885468, -0.024093628, -0.023284912, 0.0033035278, 0.010528564, -0.0131073, -0.0011730194, -0.004673004, -0.025314331, -0.025146484, -0.0037117004, 0.0014028549, ...
http://arxiv.org/abs/2302.09509v1
Text Classification in the Wild: a Large-scale Long-tailed Name Normalization Dataset
Jiexing Qi, Shuhao Li, Zhixin Guo, Yusheng Huang, Chenghu Zhou, Weinan Zhang, Xinbing Wang, Zhouhan Lin
cs.CL, cs.AI
Real-world data usually exhibits a long-tailed distribution,with a few frequent labels and a lot of few-shot labels. The study of institution name normalization is a perfect application case showing this phenomenon. There are many institutions worldwide with enormous variations of their names in the publicly available ...
http://arxiv.org/pdf/2302.09509v1
2023-02-19T08:44:21
2023-02-19T08:44:21
{ "title": [ -0.026626587, 0.0138549805, -0.054138184, -0.04812622, 0.026473999, -0.005493164, -0.030303955, -0.011520386, -0.0062713623, 0.017959595, -0.022521973, 0.02960205, -0.02229309, 0.0035762787, -0.037109375, -0.00762558, -0.015151978, 0...
http://arxiv.org/abs/2302.09473v2
Video-Text Retrieval by Supervised Sparse Multi-Grained Learning
Yimu Wang, Peng Shi
cs.CV, cs.CL, cs.IR, cs.LG, cs.MM
While recent progress in video-text retrieval has been advanced by the exploration of better representation learning, in this paper, we present a novel multi-grained sparse learning framework, S3MA, to learn an aligned sparse space shared between the video and the text for video-text retrieval. The shared sparse space ...
http://arxiv.org/pdf/2302.09473v2
2023-10-17T22:01:00
2023-02-19T04:03:22
{ "title": [ -0.011276245, 0.017471313, -0.049865723, -0.03137207, 0.021270752, -0.011985779, -0.0045661926, -0.009757996, 0.024734497, 0.043182373, -0.05581665, -0.0067214966, -0.0011930466, 0.004421234, -0.07458496, -0.021469116, 0.0019264221, ...
http://arxiv.org/abs/2303.12489v1
Few-shot Multimodal Multitask Multilingual Learning
Aman Chadha, Vinija Jain
cs.LG, cs.AI, cs.CL, cs.CV, cs.MM
While few-shot learning as a transfer learning paradigm has gained significant traction for scenarios with limited data, it has primarily been explored in the context of building unimodal and unilingual models. Furthermore, a significant part of the existing literature in the domain of few-shot multitask learning perfo...
http://arxiv.org/pdf/2303.12489v1
2023-02-19T03:48:46
2023-02-19T03:48:46
{ "title": [ -0.012428284, -0.0090789795, -0.05380249, -0.01222229, -0.01701355, -0.025985718, -0.031707764, 0.006580353, -0.011764526, 0.033477783, 0.0069885254, 0.0060691833, -0.002298355, -0.043945312, 0.014556885, -0.0066375732, 0.0035076141, ...
http://arxiv.org/abs/2302.09458v1
Learning Language Representations with Logical Inductive Bias
Jianshu Chen
cs.CL, cs.LG, cs.LO
Transformer architectures have achieved great success in solving natural language tasks, which learn strong language representations from large-scale unlabeled texts. In this paper, we seek to go further beyond and explore a new logical inductive bias for better language representation learning. Logic reasoning is know...
http://arxiv.org/pdf/2302.09458v1
2023-02-19T02:21:32
2023-02-19T02:21:32
{ "title": [ -0.021026611, -0.007850647, -0.03970337, -0.04623413, 0.003063202, -0.013336182, -0.038513184, -0.0023326874, -0.0067634583, 0.017105103, -0.043273926, 0.015472412, -0.011993408, -0.011734009, -0.021957397, -0.009941101, -0.01713562, ...
http://arxiv.org/abs/2304.12411v1
ChatGPT (Feb 13 Version) is a Chinese Room
Maurice HT Ling
cs.CL
ChatGPT has gained both positive and negative publicity after reports suggesting that it is able to pass various professional and licensing examinations. This suggests that ChatGPT may pass Turing Test in the near future. However, a computer program that passing Turing Test can either mean that it is a Chinese Room or ...
http://arxiv.org/pdf/2304.12411v1
2023-02-19T01:52:06
2023-02-19T01:52:06
{ "title": [ -0.006324768, -0.0026817322, -0.027496338, 0.024932861, -0.03466797, -0.022277832, -0.008071899, -0.026290894, -0.003917694, 0.011878967, 0.011627197, 0.025863647, -0.0289917, -0.04724121, 0.01878357, 0.012321472, 0.005756378, -0.016...
http://arxiv.org/abs/2302.10195v1
Uncertainty-Aware Reward-based Deep Reinforcement Learning for Intent Analysis of Social Media Information
Zhen Guo, Qi Zhang, Xinwei An, Qisheng Zhang, Audun Jøsang, Lance M. Kaplan, Feng Chen, Dong H. Jeong, Jin-Hee Cho
cs.CL, cs.AI, cs.LG
Due to various and serious adverse impacts of spreading fake news, it is often known that only people with malicious intent would propagate fake news. However, it is not necessarily true based on social science studies. Distinguishing the types of fake news spreaders based on their intent is critical because it will ef...
http://arxiv.org/pdf/2302.10195v1
2023-02-19T00:54:33
2023-02-19T00:54:33
{ "title": [ 0.020553589, 0.035064697, -0.044281006, -0.030715942, 0.03665161, -0.039031982, 0.008628845, -0.028808594, -0.022018433, 0.007789612, -0.028884888, 0.035064697, -0.025054932, -0.022277832, -0.003376007, -0.013076782, -0.008293152, 0....
http://arxiv.org/abs/2302.09432v2
BBT-Fin: Comprehensive Construction of Chinese Financial Domain Pre-trained Language Model, Corpus and Benchmark
Dakuan Lu, Hengkui Wu, Jiaqing Liang, Yipei Xu, Qianyu He, Yipeng Geng, Mengkun Han, Yingsi Xin, Yanghua Xiao
cs.CL
To advance Chinese financial natural language processing (NLP), we introduce BBT-FinT5, a new Chinese financial pre-training language model based on the T5 model. To support this effort, we have built BBT-FinCorpus, a large-scale financial corpus with approximately 300GB of raw text from four different sources. In gene...
http://arxiv.org/pdf/2302.09432v2
2023-02-26T10:50:09
2023-02-18T22:20:37
{ "title": [ -0.015579224, -0.0014019012, -0.04309082, -0.056396484, 0.00004017353, -0.012336731, -0.043914795, -0.0027656555, -0.0006170273, 0.03427124, 0.0037384033, -0.016586304, 0.025375366, -0.033843994, -0.04638672, 0.0062294006, -0.032409668, ...
http://arxiv.org/abs/2302.09424v1
Zero and Few-Shot Localization of Task-Oriented Dialogue Agents with a Distilled Representation
Mehrad Moradshahi, Sina J. Semnani, Monica S. Lam
cs.CL
Task-oriented Dialogue (ToD) agents are mostly limited to a few widely-spoken languages, mainly due to the high cost of acquiring training data for each language. Existing low-cost approaches that rely on cross-lingual embeddings or naive machine translation sacrifice a lot of accuracy for data efficiency, and largely ...
http://arxiv.org/pdf/2302.09424v1
2023-02-18T21:30:36
2023-02-18T21:30:36
{ "title": [ 0.004623413, -0.0034713745, -0.0071487427, -0.015686035, -0.011131287, -0.0034179688, -0.008872986, -0.019805908, 0.038116455, 0.039398193, -0.0231781, -0.023803711, 0.019134521, -0.020584106, -0.05895996, -0.005935669, -0.013122559, ...
http://arxiv.org/abs/2302.09419v3
A Comprehensive Survey on Pretrained Foundation Models: A History from BERT to ChatGPT
Ce Zhou, Qian Li, Chen Li, Jun Yu, Yixin Liu, Guangjing Wang, Kai Zhang, Cheng Ji, Qiben Yan, Lifang He, Hao Peng, Jianxin Li, Jia Wu, Ziwei Liu, Pengtao Xie, Caiming Xiong, Jian Pei, Philip S. Yu, Lichao Sun
cs.AI, cs.CL, cs.LG
Pretrained Foundation Models (PFMs) are regarded as the foundation for various downstream tasks with different data modalities. A PFM (e.g., BERT, ChatGPT, and GPT-4) is trained on large-scale data which provides a reasonable parameter initialization for a wide range of downstream applications. BERT learns bidirectiona...
http://arxiv.org/pdf/2302.09419v3
2023-05-01T07:48:05
2023-02-18T20:51:09
{ "title": [ -0.0088272095, -0.020065308, -0.010215759, -0.03378296, -0.028518677, 0.0053520203, -0.03375244, -0.010169983, 0.0020866394, 0.006324768, 0.001871109, 0.02067566, -0.0071640015, -0.038024902, 0.0026779175, 0.018356323, -0.016082764, ...
http://arxiv.org/abs/2302.09418v1
M-SENSE: Modeling Narrative Structure in Short Personal Narratives Using Protagonist's Mental Representations
Prashanth Vijayaraghavan, Deb Roy
cs.CL, cs.AI
Narrative is a ubiquitous component of human communication. Understanding its structure plays a critical role in a wide variety of applications, ranging from simple comparative analyses to enhanced narrative retrieval, comprehension, or reasoning capabilities. Prior research in narratology has highlighted the importanc...
http://arxiv.org/pdf/2302.09418v1
2023-02-18T20:48:02
2023-02-18T20:48:02
{ "title": [ -0.031311035, -0.020889282, -0.008560181, -0.01876831, -0.040740967, 0.0057525635, -0.031173706, -0.018508911, -0.02722168, 0.052215576, 0.021911621, 0.0058555603, 0.011329651, -0.036956787, -0.030471802, 0.0045166016, -0.031951904, ...
http://arxiv.org/abs/2302.09381v1
Speaker and Language Change Detection using Wav2vec2 and Whisper
Tijn Berns, Nik Vaessen, David A. van Leeuwen
eess.AS, cs.CL, cs.LG, cs.SD
We investigate recent transformer networks pre-trained for automatic speech recognition for their ability to detect speaker and language changes in speech. We do this by simply adding speaker (change) or language targets to the labels. For Wav2vec2 pre-trained networks, we also investigate if the representation for the...
http://arxiv.org/pdf/2302.09381v1
2023-02-18T16:45:30
2023-02-18T16:45:30
{ "title": [ 0.0028705597, 0.0076293945, -0.053588867, -0.0413208, 0.0010662079, -0.023513794, -0.034820557, -0.014404297, -0.018615723, 0.041046143, -0.002374649, -0.010536194, 0.029769897, -0.037719727, -0.020980835, -0.00019216537, 0.012413025, ...
http://arxiv.org/abs/2302.09368v1
Natural Language-conditioned Reinforcement Learning with Inside-out Task Language Development and Translation
Jing-Cheng Pang, Xin-Yu Yang, Si-Hang Yang, Yang Yu
cs.CL, cs.LG
Natural Language-conditioned reinforcement learning (RL) enables the agents to follow human instructions. Previous approaches generally implemented language-conditioned RL by providing human instructions in natural language (NL) and training a following policy. In this outside-in approach, the policy needs to comprehen...
http://arxiv.org/pdf/2302.09368v1
2023-02-18T15:49:09
2023-02-18T15:49:09
{ "title": [ -0.040222168, -0.010643005, -0.04159546, 0.00013673306, -0.013420105, -0.023895264, -0.0036697388, 0.021209717, -0.04812622, 0.024902344, -0.023803711, 0.018997192, -0.013908386, -0.03060913, 0.018600464, -0.020980835, 0.0018291473, ...
http://arxiv.org/abs/2302.09350v1
BERT is not The Count: Learning to Match Mathematical Statements with Proofs
Weixian Waylon Li, Yftah Ziser, Maximin Coavoux, Shay B. Cohen
cs.CL
We introduce a task consisting in matching a proof to a given mathematical statement. The task fits well within current research on Mathematical Information Retrieval and, more generally, mathematical article analysis (Mathematical Sciences, 2014). We present a dataset for the task (the MATcH dataset) consisting of ove...
http://arxiv.org/pdf/2302.09350v1
2023-02-18T14:48:20
2023-02-18T14:48:20
{ "title": [ -0.009307861, -0.005744934, -0.030532837, -0.039764404, 0.0052452087, -0.020446777, -0.001036644, -0.008140564, -0.015380859, -0.0002682209, -0.024902344, 0.00793457, -0.013771057, -0.020141602, -0.01550293, -0.020126343, -0.008010864, ...
http://arxiv.org/abs/2302.09345v1
Improving the Out-Of-Distribution Generalization Capability of Language Models: Counterfactually-Augmented Data is not Enough
Caoyun Fan, Wenqing Chen, Jidong Tian, Yitian Li, Hao He, Yaohui Jin
cs.CL
Counterfactually-Augmented Data (CAD) has the potential to improve language models' Out-Of-Distribution (OOD) generalization capability, as CAD induces language models to exploit causal features and exclude spurious correlations. However, the empirical results of OOD generalization on CAD are not as efficient as expect...
http://arxiv.org/pdf/2302.09345v1
2023-02-18T14:39:03
2023-02-18T14:39:03
{ "title": [ -0.006587982, -0.0013780594, -0.025497437, -0.04940796, -0.006374359, 0.00008755922, -0.047302246, -0.017242432, 0.019439697, 0.039764404, -0.03100586, 0.009407043, 0.0061950684, -0.013587952, -0.06414795, 0.008369446, -0.024230957, ...
http://arxiv.org/abs/2302.09331v1
Front-End Adapter: Adapting Front-End Input of Speech based Self-Supervised Learning for Speech Recognition
Xie Chen, Ziyang Ma, Changli Tang, Yujin Wang, Zhisheng Zheng
eess.AS, cs.CL, cs.SD
Recent years have witnessed a boom in self-supervised learning (SSL) in various areas including speech processing. Speech based SSL models present promising performance in a range of speech related tasks. However, the training of SSL models is computationally expensive and a common practice is to fine-tune a released S...
http://arxiv.org/pdf/2302.09331v1
2023-02-18T13:46:12
2023-02-18T13:46:12
{ "title": [ -0.008613586, 0.0024032593, -0.07299805, -0.06500244, 0.0055122375, -0.029388428, -0.032409668, 0.023666382, -0.041412354, 0.031082153, -0.018554688, -0.0038757324, -0.031311035, -0.021255493, -0.04006958, -0.03945923, 0.0063095093, ...
http://arxiv.org/abs/2302.09327v1
Transformadores: Fundamentos teoricos y Aplicaciones
Jordi de la Torre
cs.CL, cs.AI, 68T01, I.2
Transformers are a neural network architecture originally designed for natural language processing that it is now a mainstream tool for solving a wide variety of problems, including natural language processing, sound, image, reinforcement learning, and other problems with heterogeneous input data. Its distinctive featu...
http://arxiv.org/pdf/2302.09327v1
2023-02-18T13:30:32
2023-02-18T13:30:32
{ "title": [ 0.0063705444, 0.013114929, -0.022262573, -0.03201294, 0.00092697144, -0.00014984608, -0.041046143, -0.036590576, -0.02218628, 0.04006958, 0.009109497, -0.02293396, 0.018585205, -0.07104492, -0.021331787, -0.0070762634, 0.014404297, 0...
http://arxiv.org/abs/2302.09324v1
Optimising Human-Machine Collaboration for Efficient High-Precision Information Extraction from Text Documents
Bradley Butcher, Miri Zilka, Darren Cook, Jiri Hron, Adrian Weller
cs.CL, cs.HC, cs.IR, cs.LG
While humans can extract information from unstructured text with high precision and recall, this is often too time-consuming to be practical. Automated approaches, on the other hand, produce nearly-immediate results, but may not be reliable enough for high-stakes applications where precision is essential. In this work,...
http://arxiv.org/pdf/2302.09324v1
2023-02-18T13:07:22
2023-02-18T13:07:22
{ "title": [ -0.040283203, 0.0030937195, -0.045776367, -0.04336548, -0.0076408386, -0.016357422, -0.034362793, -0.0119018555, 0.012374878, 0.039489746, -0.029144287, -0.0017671585, 0.004131317, 0.004917145, -0.06414795, -0.007850647, -0.013694763, ...
http://arxiv.org/abs/2302.09268v1
Bag of Tricks for Effective Language Model Pretraining and Downstream Adaptation: A Case Study on GLUE
Qihuang Zhong, Liang Ding, Keqin Peng, Juhua Liu, Bo Du, Li Shen, Yibing Zhan, Dacheng Tao
cs.CL
This technical report briefly describes our JDExplore d-team's submission Vega v1 on the General Language Understanding Evaluation (GLUE) leaderboard, where GLUE is a collection of nine natural language understanding tasks, including question answering, linguistic acceptability, sentiment analysis, text similarity, par...
http://arxiv.org/pdf/2302.09268v1
2023-02-18T09:26:35
2023-02-18T09:26:35
{ "title": [ -0.023086548, -0.013160706, -0.04135132, -0.045898438, -0.02609253, -0.013061523, -0.05239868, 0.0074882507, -0.03842163, 0.03024292, 0.0209198, 0.00283432, 0.014533997, -0.04800415, -0.0031795502, 0.0023593903, -0.001789093, 0.04638...
http://arxiv.org/abs/2302.09243v1
A Federated Approach for Hate Speech Detection
Jay Gala, Deep Gandhi, Jash Mehta, Zeerak Talat
cs.LG, cs.AI, cs.CL
Hate speech detection has been the subject of high research attention, due to the scale of content created on social media. In spite of the attention and the sensitive nature of the task, privacy preservation in hate speech detection has remained under-studied. The majority of research has focused on centralised machin...
http://arxiv.org/pdf/2302.09243v1
2023-02-18T06:08:04
2023-02-18T06:08:04
{ "title": [ 0.020187378, 0.020126343, -0.046325684, -0.06732178, 0.04385376, -0.021377563, -0.009719849, -0.0047302246, -0.0038337708, 0.046051025, -0.026229858, 0.00819397, 0.009933472, 0.001364708, -0.058288574, -0.0004580021, -0.01751709, 0.0...
http://arxiv.org/abs/2302.09236v1
Scalable Prompt Generation for Semi-supervised Learning with Language Models
Yuhang Zhou, Suraj Maharjan, Beiye Liu
cs.CL, cs.AI
Prompt-based learning methods in semi-supervised learning (SSL) settings have been shown to be effective on multiple natural language understanding (NLU) datasets and tasks in the literature. However, manually designing multiple prompts and verbalizers requires domain knowledge and human effort, making it difficult and...
http://arxiv.org/pdf/2302.09236v1
2023-02-18T05:06:28
2023-02-18T05:06:28
{ "title": [ -0.022888184, -0.009094238, -0.046051025, -0.056488037, -0.012184143, -0.002298355, -0.072631836, 0.009933472, -0.0042037964, 0.00945282, -0.039855957, 0.005279541, -0.0113220215, -0.012481689, -0.03289795, -0.019500732, -0.0040016174, ...
http://arxiv.org/abs/2302.09230v1
VLN-Trans: Translator for the Vision and Language Navigation Agent
Yue Zhang, Parisa Kordjamshidi
cs.CL
Language understanding is essential for the navigation agent to follow instructions. We observe two kinds of issues in the instructions that can make the navigation task challenging: 1. The mentioned landmarks are not recognizable by the navigation agent due to the different vision abilities of the instructor and the m...
http://arxiv.org/pdf/2302.09230v1
2023-02-18T04:19:51
2023-02-18T04:19:51
{ "title": [ -0.030838013, -0.026504517, -0.024459839, -0.009002686, -0.031280518, -0.014434814, -0.006275177, -0.019470215, -0.015457153, 0.033416748, -0.0064735413, -0.03427124, 0.053009033, -0.048950195, -0.028884888, -0.0119018555, 0.0063705444, ...
http://arxiv.org/abs/2302.09210v1
How Good Are GPT Models at Machine Translation? A Comprehensive Evaluation
Amr Hendy, Mohamed Abdelrehim, Amr Sharaf, Vikas Raunak, Mohamed Gabr, Hitokazu Matsushita, Young Jin Kim, Mohamed Afify, Hany Hassan Awadalla
cs.CL
Generative Pre-trained Transformer (GPT) models have shown remarkable capabilities for natural language generation, but their performance for machine translation has not been thoroughly investigated. In this paper, we present a comprehensive evaluation of GPT models for machine translation, covering various aspects suc...
http://arxiv.org/pdf/2302.09210v1
2023-02-18T02:11:36
2023-02-18T02:11:36
{ "title": [ -0.03515625, -0.0061569214, -0.0010251999, -0.038208008, -0.01763916, 0.018432617, -0.054870605, -0.022415161, -0.0008029938, 0.03237915, -0.013755798, 0.009544373, 0.014854431, -0.015167236, -0.05053711, 0.020065308, -0.017105103, 0...
http://arxiv.org/abs/2302.09207v2
RETVec: Resilient and Efficient Text Vectorizer
Elie Bursztein, Marina Zhang, Owen Vallis, Xinyu Jia, Alexey Kurakin
cs.CL, cs.AI
This paper describes RETVec, an efficient, resilient, and multilingual text vectorizer designed for neural-based text processing. RETVec combines a novel character encoding with an optional small embedding model to embed words into a 256-dimensional vector space. The RETVec embedding model is pre-trained using pair-wis...
http://arxiv.org/pdf/2302.09207v2
2023-10-06T21:36:32
2023-02-18T02:06:52
{ "title": [ -0.01953125, -0.018951416, -0.019607544, -0.059631348, 0.0025177002, 0.0093688965, -0.047027588, -0.029342651, 0.008796692, 0.05630493, -0.0016822815, 0.009208679, 0.025024414, -0.008964539, -0.05908203, 0.023117065, 0.001663208, 0.0...
http://arxiv.org/abs/2302.09189v1
Extraction of Constituent Factors of Digestion Efficiency in Information Transfer by Media Composed of Texts and Images
Koike Hiroaki, Teruaki Hayashi
cs.CL
The development and spread of information and communication technologies have increased and diversified information. However, the increase in the volume and the selection of information does not necessarily promote understanding. In addition, conventional evaluations of information transfer have focused only on the arr...
http://arxiv.org/pdf/2302.09189v1
2023-02-17T23:45:02
2023-02-17T23:45:02
{ "title": [ -0.0063667297, 0.0063056946, -0.042510986, 0.022720337, -0.02041626, -0.042022705, 0.024841309, -0.043945312, 0.027832031, 0.021148682, -0.016113281, 0.0069732666, 0.03768921, -0.055419922, -0.0048942566, 0.005531311, -0.008010864, 0...
http://arxiv.org/abs/2302.09185v1
Bounding the Capabilities of Large Language Models in Open Text Generation with Prompt Constraints
Albert Lu, Hongxin Zhang, Yanzhe Zhang, Xuezhi Wang, Diyi Yang
cs.CL, cs.AI, cs.LG
The limits of open-ended generative models are unclear, yet increasingly important. What causes them to succeed and what causes them to fail? In this paper, we take a prompt-centric approach to analyzing and bounding the abilities of open-ended generative models. We present a generic methodology of analysis with two ch...
http://arxiv.org/pdf/2302.09185v1
2023-02-17T23:30:28
2023-02-17T23:30:28
{ "title": [ -0.041992188, -0.006843567, -0.041137695, -0.05279541, -0.012557983, 0.022476196, -0.08337402, -0.0051116943, -0.003156662, 0.030303955, -0.023361206, 0.0068244934, 0.011749268, -0.009414673, -0.031341553, 0.0026760101, -0.005809784, ...
http://arxiv.org/abs/2302.09173v2
Unsupervised Task Graph Generation from Instructional Video Transcripts
Lajanugen Logeswaran, Sungryull Sohn, Yunseok Jang, Moontae Lee, Honglak Lee
cs.AI, cs.CL, cs.LG
This work explores the problem of generating task graphs of real-world activities. Different from prior formulations, we consider a setting where text transcripts of instructional videos performing a real-world activity (e.g., making coffee) are provided and the goal is to identify the key steps relevant to the task as...
http://arxiv.org/pdf/2302.09173v2
2023-05-02T19:46:14
2023-02-17T22:50:08
{ "title": [ -0.035491943, -0.0072669983, -0.057800293, -0.046813965, -0.015716553, -0.009819031, -0.052764893, 0.005332947, -0.014373779, 0.026855469, -0.011390686, 0.012672424, -0.041259766, -0.0087509155, -0.039276123, -0.03491211, -0.019714355, ...
http://arxiv.org/abs/2302.09170v1
KILM: Knowledge Injection into Encoder-Decoder Language Models
Yan Xu, Mahdi Namazifar, Devamanyu Hazarika, Aishwarya Padmakumar, Yang Liu, Dilek Hakkani-Tür
cs.CL, cs.AI
Large pre-trained language models (PLMs) have been shown to retain implicit knowledge within their parameters. To enhance this implicit knowledge, we propose Knowledge Injection into Language Models (KILM), a novel approach that injects entity-related knowledge into encoder-decoder PLMs, via a generative knowledge infi...
http://arxiv.org/pdf/2302.09170v1
2023-02-17T22:48:07
2023-02-17T22:48:07
{ "title": [ -0.03591919, -0.031188965, -0.04119873, -0.054748535, -0.019134521, -0.00052165985, -0.049743652, 0.0054893494, -0.0036239624, 0.026367188, -0.026000977, 0.007522583, 0.00699234, -0.011810303, -0.04034424, 0.0014772415, -0.009986877, ...
http://arxiv.org/abs/2302.09155v1
Med-EASi: Finely Annotated Dataset and Models for Controllable Simplification of Medical Texts
Chandrayee Basu, Rosni Vasu, Michihiro Yasunaga, Qian Yang
cs.CL
Automatic medical text simplification can assist providers with patient-friendly communication and make medical texts more accessible, thereby improving health literacy. But curating a quality corpus for this task requires the supervision of medical experts. In this work, we present $\textbf{Med-EASi}$ ($\underline{\te...
http://arxiv.org/pdf/2302.09155v1
2023-02-17T21:50:13
2023-02-17T21:50:13
{ "title": [ -0.044891357, -0.010276794, -0.05090332, -0.06677246, -0.0007081032, 0.02041626, -0.058746338, 0.0030956268, 0.004650116, 0.05114746, 0.002494812, 0.014312744, 0.03277588, -0.01953125, -0.044799805, 0.011558533, 0.013702393, 0.012168...
http://arxiv.org/abs/2302.09150v1
Cluster-Guided Label Generation in Extreme Multi-Label Classification
Taehee Jung, Joo-Kyung Kim, Sungjin Lee, Dongyeop Kang
cs.CL, cs.LG
For extreme multi-label classification (XMC), existing classification-based models poorly perform for tail labels and often ignore the semantic relations among labels, like treating "Wikipedia" and "Wiki" as independent and separate labels. In this paper, we cast XMC as a generation task (XLGen), where we benefit from ...
http://arxiv.org/pdf/2302.09150v1
2023-02-17T21:20:36
2023-02-17T21:20:36
{ "title": [ -0.006702423, 0.003326416, -0.030761719, -0.035095215, 0.0102005005, -0.0064048767, -0.002998352, -0.00655365, -0.012908936, 0.019699097, -0.009803772, 0.03540039, -0.019378662, -0.023376465, -0.020889282, 0.013008118, -0.01071167, 0...
http://arxiv.org/abs/2302.09122v1
Conveying the Predicted Future to Users: A Case Study of Story Plot Prediction
Chieh-Yang Huang, Saniya Naphade, Kavya Laalasa Karanam, Ting-Hao 'Kenneth' Huang
cs.CL, cs.HC
Creative writing is hard: Novelists struggle with writer's block daily. While automatic story generation has advanced recently, it is treated as a "toy task" for advancing artificial intelligence rather than helping people. In this paper, we create a system that produces a short description that narrates a predicted pl...
http://arxiv.org/pdf/2302.09122v1
2023-02-17T20:10:55
2023-02-17T20:10:55
{ "title": [ 0.0009422302, -0.009109497, -0.0262146, -0.051818848, -0.00008916855, -0.014831543, -0.0116119385, -0.012527466, -0.0002784729, 0.034057617, 0.0029792786, -0.008850098, 0.009162903, -0.01890564, -0.05947876, 0.0037975311, -0.017578125, ...
http://arxiv.org/abs/2302.09051v4
Complex QA and language models hybrid architectures, Survey
Xavier Daull, Patrice Bellot, Emmanuel Bruno, Vincent Martin, Elisabeth Murisasco
cs.CL, cs.AI, cs.IR, cs.LG
This paper reviews the state-of-the-art of language models architectures and strategies for "complex" question-answering (QA, CQA, CPS) with a focus on hybridization. Large Language Models (LLM) are good at leveraging public data on standard problems but once you want to tackle more specific complex questions or proble...
http://arxiv.org/pdf/2302.09051v4
2023-04-07T16:37:35
2023-02-17T18:31:31
{ "title": [ -0.024932861, -0.0045433044, -0.04260254, -0.04260254, -0.019577026, -0.025634766, -0.03933716, 0.016601562, -0.028945923, 0.06573486, -0.041046143, -0.004825592, 0.04748535, -0.018737793, -0.022079468, -0.010047913, -0.00390625, 0.0...
http://arxiv.org/abs/2302.09027v1
CK-Transformer: Commonsense Knowledge Enhanced Transformers for Referring Expression Comprehension
Zhi Zhang, Helen Yannakoudakis, Xiantong Zhen, Ekaterina Shutova
cs.CV, cs.AI, cs.CL, cs.MM
The task of multimodal referring expression comprehension (REC), aiming at localizing an image region described by a natural language expression, has recently received increasing attention within the research comminity. In this paper, we specifically focus on referring expression comprehension with commonsense knowledg...
http://arxiv.org/pdf/2302.09027v1
2023-02-17T17:49:26
2023-02-17T17:49:26
{ "title": [ -0.0044174194, -0.017181396, -0.0105896, -0.06188965, -0.013435364, 0.018203735, -0.045135498, -0.011146545, 0.029647827, 0.024169922, -0.009788513, -0.007701874, 0.038116455, -0.025238037, -0.03390503, 0.020645142, -0.0014486313, 0....
http://arxiv.org/abs/2302.08997v1
Designing and Evaluating Interfaces that Highlight News Coverage Diversity Using Discord Questions
Philippe Laban, Chien-Sheng Wu, Lidiya Murakhovs'ka, Xiang 'Anthony' Chen, Caiming Xiong
cs.HC, cs.CL
Modern news aggregators do the hard work of organizing a large news stream, creating collections for a given news story with tens of source options. This paper shows that navigating large source collections for a news story can be challenging without further guidance. In this work, we design three interfaces -- the Ann...
http://arxiv.org/pdf/2302.08997v1
2023-02-17T16:59:31
2023-02-17T16:59:31
{ "title": [ 0.005756378, -0.004497528, -0.04776001, -0.022445679, 0.0059165955, -0.03250122, 0.016738892, 0.00982666, 0.019500732, 0.024490356, -0.039245605, -0.022994995, -0.007167816, -0.021743774, -0.06060791, -0.02897644, -0.029647827, 0.021...
http://arxiv.org/abs/2302.08975v1
Towards Fine-Grained Information: Identifying the Type and Location of Translation Errors
Keqin Bao, Yu Wan, Dayiheng Liu, Baosong Yang, Wenqiang Lei, Xiangnan He, Derek F. Wong, Jun Xie
cs.CL
Fine-grained information on translation errors is helpful for the translation evaluation community. Existing approaches can not synchronously consider error position and type, failing to integrate the error information of both. In this paper, we propose Fine-Grained Translation Error Detection (FG-TED) task, aiming at ...
http://arxiv.org/pdf/2302.08975v1
2023-02-17T16:20:33
2023-02-17T16:20:33
{ "title": [ -0.0039901733, -0.005794525, -0.015419006, -0.052886963, -0.0028533936, -0.0020103455, -0.033996582, -0.011917114, 0.001376152, 0.036956787, -0.016494751, 0.004863739, 0.019134521, -0.028778076, -0.047698975, 0.0036792755, -0.0006246567,...
http://arxiv.org/abs/2302.08961v2
Grimm in Wonderland: Prompt Engineering with Midjourney to Illustrate Fairytales
Martin Ruskov
cs.CL, cs.AI, cs.HC, I.2
The quality of text-to-image generation is continuously improving, yet the boundaries of its applicability are still unclear. In particular, refinement of the text input with the objective of achieving better results - commonly called prompt engineering - so far seems to have not been geared towards work with pre-exist...
http://arxiv.org/pdf/2302.08961v2
2023-08-25T14:12:43
2023-02-17T15:49:19
{ "title": [ -0.00712204, -0.011070251, -0.031585693, 0.037261963, -0.032989502, -0.0368042, 0.016815186, -0.0345459, -0.01020813, 0.03010559, -0.0016345978, 0.008422852, 0.05343628, -0.060638428, 0.013084412, -0.0005888939, 0.005115509, 0.016571...
http://arxiv.org/abs/2302.08957v2
Like a Good Nearest Neighbor: Practical Content Moderation with Sentence Transformers
Luke Bates, Iryna Gurevych
cs.CL
Modern text classification systems have impressive capabilities but are infeasible to deploy and use reliably due to their dependence on prompting and billion-parameter language models. SetFit (Tunstall et al., 2022) is a recent, practical approach that fine-tunes a Sentence Transformer under a contrastive learning par...
http://arxiv.org/pdf/2302.08957v2
2023-03-02T14:23:39
2023-02-17T15:43:29
{ "title": [ 0.014976501, 0.00085163116, 0.014656067, -0.052825928, 0.0107421875, 0.011238098, -0.01940918, -0.038726807, -0.013473511, 0.02583313, -0.020584106, 0.016296387, 0.02418518, -0.017745972, -0.020706177, 0.004798889, -0.019363403, 0.05...
http://arxiv.org/abs/2302.08956v5
AfriSenti: A Twitter Sentiment Analysis Benchmark for African Languages
Shamsuddeen Hassan Muhammad, Idris Abdulmumin, Abinew Ali Ayele, Nedjma Ousidhoum, David Ifeoluwa Adelani, Seid Muhie Yimam, Ibrahim Sa'id Ahmad, Meriem Beloucif, Saif M. Mohammad, Sebastian Ruder, Oumaima Hourrane, Pavel Brazdil, Felermino Dário Mário António Ali, Davis David, Salomey Osei, Bello Shehu Bello, Falalu I...
cs.CL
Africa is home to over 2,000 languages from more than six language families and has the highest linguistic diversity among all continents. These include 75 languages with at least one million speakers each. Yet, there is little NLP research conducted on African languages. Crucial to enabling such research is the availa...
http://arxiv.org/pdf/2302.08956v5
2023-11-04T19:48:38
2023-02-17T15:40:12
{ "title": [ -0.016784668, -0.024627686, 0.006034851, -0.04763794, -0.022521973, 0.023391724, -0.054260254, -0.013084412, 0.0132751465, 0.049468994, -0.02078247, 0.02670288, 0.020767212, -0.004005432, -0.061767578, 0.01966858, 0.00016820431, 0.02...
http://arxiv.org/abs/2302.08950v3
Handling the Alignment for Wake Word Detection: A Comparison Between Alignment-Based, Alignment-Free and Hybrid Approaches
Vinicius Ribeiro, Yiteng Huang, Yuan Shangguan, Zhaojun Yang, Li Wan, Ming Sun
cs.CL, cs.SD, eess.AS
Wake word detection exists in most intelligent homes and portable devices. It offers these devices the ability to "wake up" when summoned at a low cost of power and computing. This paper focuses on understanding alignment's role in developing a wake-word system that answers a generic phrase. We discuss three approaches...
http://arxiv.org/pdf/2302.08950v3
2023-06-07T15:04:41
2023-02-17T15:33:47
{ "title": [ -0.039031982, -0.004386902, -0.05529785, 0.0020446777, -0.0028934479, -0.023956299, -0.0066337585, -0.020339966, -0.031188965, 0.06982422, -0.023117065, -0.009460449, 0.020248413, -0.013999939, -0.07128906, -0.01776123, -0.0036563873, ...
http://arxiv.org/abs/2302.08948v1
Entry Separation using a Mixed Visual and Textual Language Model: Application to 19th century French Trade Directories
Bertrand Duménieu, Edwin Carlinet, Nathalie Abadie, Joseph Chazalon
cs.CL, cs.CV
When extracting structured data from repetitively organized documents, such as dictionaries, directories, or even newspapers, a key challenge is to correctly segment what constitutes the basic text regions for the target database. Traditionally, such a problem was tackled as part of the layout analysis and was mostly b...
http://arxiv.org/pdf/2302.08948v1
2023-02-17T15:30:44
2023-02-17T15:30:44
{ "title": [ 0.0002539158, 0.005218506, -0.0262146, -0.016799927, -0.0048294067, -0.012397766, -0.022415161, -0.024215698, 0.031951904, 0.02166748, -0.043029785, 0.0044555664, 0.0071640015, -0.017196655, -0.031051636, -0.019866943, 0.0076293945, ...
http://arxiv.org/abs/2302.08936v1
More Data Types More Problems: A Temporal Analysis of Complexity, Stability, and Sensitivity in Privacy Policies
Juniper Lovato, Philip Mueller, Parisa Suchdev, Peter S. Dodds
cs.CL, cs.SI
Collecting personally identifiable information (PII) on data subjects has become big business. Data brokers and data processors are part of a multi-billion-dollar industry that profits from collecting, buying, and selling consumer data. Yet there is little transparency in the data collection industry which makes it dif...
http://arxiv.org/pdf/2302.08936v1
2023-02-17T15:21:24
2023-02-17T15:21:24
{ "title": [ 0.06677246, 0.0075187683, -0.02760315, -0.07885742, 0.034942627, -0.054382324, 0.016662598, -0.00016570091, -0.08557129, 0.04345703, -0.02519226, 0.03894043, -0.018295288, -0.0129776, 0.0009880066, -0.046447754, -0.0020923615, 0.0305...
http://arxiv.org/abs/2302.08921v1
Deep Implicit Distribution Alignment Networks for Cross-Corpus Speech Emotion Recognition
Yan Zhao, Jincen Wang, Yuan Zong, Wenming Zheng, Hailun Lian, Li Zhao
cs.SD, cs.CL, eess.AS
In this paper, we propose a novel deep transfer learning method called deep implicit distribution alignment networks (DIDAN) to deal with cross-corpus speech emotion recognition (SER) problem, in which the labeled training (source) and unlabeled testing (target) speech signals come from different corpora. Specifically,...
http://arxiv.org/pdf/2302.08921v1
2023-02-17T14:51:37
2023-02-17T14:51:37
{ "title": [ -0.008850098, 0.00062561035, -0.04248047, -0.066589355, 0.0037708282, -0.000074744225, -0.037841797, 0.018554688, 0.04333496, 0.049438477, -0.03302002, -0.02355957, 0.036315918, -0.012420654, -0.0647583, 0.01625061, -0.010116577, 0.0...
http://arxiv.org/abs/2302.08917v1
Massively Multilingual Shallow Fusion with Large Language Models
Ke Hu, Tara N. Sainath, Bo Li, Nan Du, Yanping Huang, Andrew M. Dai, Yu Zhang, Rodrigo Cabrera, Zhifeng Chen, Trevor Strohman
cs.CL, cs.LG
While large language models (LLM) have made impressive progress in natural language processing, it remains unclear how to utilize them in improving automatic speech recognition (ASR). In this work, we propose to train a single multilingual language model (LM) for shallow fusion in multiple languages. We push the limits...
http://arxiv.org/pdf/2302.08917v1
2023-02-17T14:46:38
2023-02-17T14:46:38
{ "title": [ -0.016708374, -0.0010871887, -0.03591919, -0.07531738, 0.018356323, 0.013717651, -0.06829834, 0.0042381287, 0.020080566, 0.03161621, -0.017059326, -0.0026245117, 0.022918701, -0.020401001, -0.016616821, 0.014266968, -0.016937256, 0.0...
http://arxiv.org/abs/2302.08822v1
False perspectives on human language: why statistics needs linguistics
Matteo Greco, Andrea Cometa, Fiorenzo Artoni, Robert Frank, Andrea Moro
cs.CL
A sharp tension exists about the nature of human language between two opposite parties: those who believe that statistical surface distributions, in particular using measures like surprisal, provide a better understanding of language processing, vs. those who believe that discrete hierarchical structures implementing l...
http://arxiv.org/pdf/2302.08822v1
2023-02-17T11:40:32
2023-02-17T11:40:32
{ "title": [ -0.03161621, -0.0151901245, -0.02180481, -0.018295288, -0.017303467, -0.010139465, 0.0024089813, -0.015037537, -0.031677246, 0.021408081, -0.011268616, -0.015533447, -0.018218994, -0.03466797, -0.03439331, -0.013221741, -0.012557983, ...
http://arxiv.org/abs/2302.08817v2
Natural Response Generation for Chinese Reading Comprehension
Nuo Chen, Hongguang Li, Yinan Bao, Baoyuan Wang, Jia Li
cs.CL, cs.AI
Machine reading comprehension (MRC) is an important area of conversation agents and draws a lot of attention. However, there is a notable limitation to current MRC benchmarks: The labeled answers are mostly either spans extracted from the target corpus or the choices of the given candidates, ignoring the natural aspect...
http://arxiv.org/pdf/2302.08817v2
2023-10-09T04:57:56
2023-02-17T11:31:05
{ "title": [ -0.044921875, -0.016952515, -0.003332138, -0.012809753, -0.042816162, 0.0025997162, -0.03289795, -0.00086545944, 0.0025959015, 0.046081543, -0.02532959, -0.0004849434, -0.0032520294, -0.009376526, -0.07080078, -0.0020370483, 0.0003075599...
http://arxiv.org/abs/2302.08777v1
Hate Speech and Offensive Language Detection using an Emotion-aware Shared Encoder
Khouloud Mnassri, Praboda Rajapaksha, Reza Farahbakhsh, Noel Crespi
cs.CL, cs.IT, cs.LG, cs.SI, math.IT
The rise of emergence of social media platforms has fundamentally altered how people communicate, and among the results of these developments is an increase in online use of abusive content. Therefore, automatically detecting this content is essential for banning inappropriate information, and reducing toxicity and vio...
http://arxiv.org/pdf/2302.08777v1
2023-02-17T09:31:06
2023-02-17T09:31:06
{ "title": [ 0.0044670105, 0.016693115, -0.035491943, -0.030532837, 0.0158844, -0.028457642, -0.0013666153, -0.013954163, 0.015960693, 0.040283203, -0.011932373, -0.0021648407, 0.023620605, -0.017364502, -0.045562744, 0.008636475, -0.009117126, 0...
http://arxiv.org/abs/2302.08709v1
Multimodal Propaganda Processing
Vincent Ng, Shengjie Li
cs.CL, cs.CV
Propaganda campaigns have long been used to influence public opinion via disseminating biased and/or misleading information. Despite the increasing prevalence of propaganda content on the Internet, few attempts have been made by AI researchers to analyze such content. We introduce the task of multimodal propaganda proc...
http://arxiv.org/pdf/2302.08709v1
2023-02-17T05:49:55
2023-02-17T05:49:55
{ "title": [ 0.041259766, 0.026550293, -0.011062622, -0.03955078, -0.0021438599, -0.023910522, -0.0042686462, -0.02897644, -0.032440186, 0.038482666, -0.005844116, 0.000248909, 0.0012931824, -0.04763794, 0.009460449, -0.0009775162, -0.016906738, ...
http://arxiv.org/abs/2302.08675v1
DREEAM: Guiding Attention with Evidence for Improving Document-Level Relation Extraction
Youmi Ma, An Wang, Naoaki Okazaki
cs.CL
Document-level relation extraction (DocRE) is the task of identifying all relations between each entity pair in a document. Evidence, defined as sentences containing clues for the relationship between an entity pair, has been shown to help DocRE systems focus on relevant texts, thus improving relation extraction. Howev...
http://arxiv.org/pdf/2302.08675v1
2023-02-17T03:54:31
2023-02-17T03:54:31
{ "title": [ 0.0054473877, -0.0026988983, -0.04812622, -0.08404541, 0.011276245, -0.0087890625, -0.041900635, 0.002714157, 0.024276733, 0.012451172, -0.023864746, 0.0060195923, 0.0061836243, -0.004753113, -0.035827637, 0.0031967163, -0.011421204, ...
http://arxiv.org/abs/2302.08672v1
Multimodal Subtask Graph Generation from Instructional Videos
Yunseok Jang, Sungryull Sohn, Lajanugen Logeswaran, Tiange Luo, Moontae Lee, Honglak Lee
cs.LG, cs.AI, cs.CL, cs.CV
Real-world tasks consist of multiple inter-dependent subtasks (e.g., a dirty pan needs to be washed before it can be used for cooking). In this work, we aim to model the causal dependencies between such subtasks from instructional videos describing the task. This is a challenging problem since complete information abou...
http://arxiv.org/pdf/2302.08672v1
2023-02-17T03:41:38
2023-02-17T03:41:38
{ "title": [ -0.025039673, -0.00050354004, -0.06237793, -0.04550171, -0.01890564, -0.019210815, -0.06738281, -0.0012960434, -0.008163452, 0.015464783, 0.0014858246, 0.017654419, -0.031082153, -0.037109375, -0.0058937073, -0.023666382, -0.0068359375, ...
http://arxiv.org/abs/2302.08659v1
Uncertainty-aware Self-training for Low-resource Neural Sequence Labeling
Jianing Wang, Chengyu Wang, Jun Huang, Ming Gao, Aoying Zhou
cs.CL
Neural sequence labeling (NSL) aims at assigning labels for input language tokens, which covers a broad range of applications, such as named entity recognition (NER) and slot filling, etc. However, the satisfying results achieved by traditional supervised-based approaches heavily depend on the large amounts of human an...
http://arxiv.org/pdf/2302.08659v1
2023-02-17T02:40:04
2023-02-17T02:40:04
{ "title": [ -0.0067253113, 0.018295288, -0.04397583, -0.07312012, 0.0072517395, -0.01348114, -0.043121338, 0.012138367, -0.0096588135, 0.038635254, -0.010429382, 0.0115737915, -0.028427124, -0.025497437, -0.04840088, -0.002910614, 0.013427734, 0...
http://arxiv.org/abs/2302.10915v1
Conformers are All You Need for Visual Speech Recogntion
Oscar Chang, Hank Liao, Dmitriy Serdyuk, Ankit Shah, Olivier Siohan
cs.LG, cs.CL, cs.CV, cs.SD, eess.AS
Visual speech recognition models extract visual features in a hierarchical manner. At the lower level, there is a visual front-end with a limited temporal receptive field that processes the raw pixels depicting the lips or faces. At the higher level, there is an encoder that attends to the embeddings produced by the fr...
http://arxiv.org/pdf/2302.10915v1
2023-02-17T01:31:55
2023-02-17T01:31:55
{ "title": [ 0.009796143, -0.016036987, -0.05319214, -0.087646484, -0.033935547, 0.004470825, -0.078430176, 0.04046631, -0.0025157928, 0.06878662, -0.020889282, -0.018463135, 0.049346924, -0.031585693, -0.025177002, -0.0008921623, 0.0005784035, 0...
http://arxiv.org/abs/2302.08624v6
InstructABSA: Instruction Learning for Aspect Based Sentiment Analysis
Kevin Scaria, Himanshu Gupta, Siddharth Goyal, Saurabh Arjun Sawant, Swaroop Mishra, Chitta Baral
cs.CL, cs.LG
We introduce InstructABSA, an instruction learning paradigm for Aspect-Based Sentiment Analysis (ABSA) subtasks. Our method introduces positive, negative, and neutral examples to each training sample, and instruction tune the model (Tk-Instruct) for ABSA subtasks, yielding significant performance improvements. Experime...
http://arxiv.org/pdf/2302.08624v6
2023-11-13T17:56:19
2023-02-16T23:29:22
{ "title": [ -0.0034255981, -0.026306152, -0.012901306, -0.051086426, -0.0053482056, 0.0027103424, -0.021499634, 0.0036449432, 0.03616333, 0.04611206, -0.02720642, -0.013427734, 0.008041382, -0.0025177002, -0.074279785, 0.0020008087, -0.008132935, ...
http://arxiv.org/abs/2302.10186v7
E2E Spoken Entity Extraction for Virtual Agents
Karan Singla, Yeon-Jun Kim, Srinivas Bangalore
eess.AS, cs.CL, cs.SD
In human-computer conversations, extracting entities such as names, street addresses and email addresses from speech is a challenging task. In this paper, we study the impact of fine-tuning pre-trained speech encoders on extracting spoken entities in human-readable form directly from speech without the need for text tr...
http://arxiv.org/pdf/2302.10186v7
2023-11-09T20:51:27
2023-02-16T21:45:35
{ "title": [ -0.019302368, 0.008926392, -0.068603516, -0.062072754, 0.031585693, -0.018325806, -0.02394104, 0.022628784, 0.004322052, 0.04397583, -0.014923096, -0.01625061, 0.03945923, -0.018920898, -0.049835205, 0.013473511, -0.02142334, 0.03369...
http://arxiv.org/abs/2302.08590v1
What A Situated Language-Using Agent Must be Able to Do: A Top-Down Analysis
David Schlangen
cs.CL
Even in our increasingly text-intensive times, the primary site of language use is situated, co-present interaction. It is primary ontogenetically and phylogenetically, and it is arguably also still primary in negotiating everyday social situations. Situated interaction is also the final frontier of Natural Language Pr...
http://arxiv.org/pdf/2302.08590v1
2023-02-16T21:30:26
2023-02-16T21:30:26
{ "title": [ 0.021759033, -0.020568848, -0.008850098, -0.039642334, -0.029403687, -0.015296936, -0.0044403076, -0.014587402, -0.027282715, 0.01713562, -0.017532349, 0.0026283264, 0.003528595, -0.020507812, -0.007457733, -0.018356323, 0.011756897, ...
http://arxiv.org/abs/2302.08589v1
Syntactic Structure Processing in the Brain while Listening
Subba Reddy Oota, Mounika Marreddy, Manish Gupta, Bapi Raju Surampud
cs.CL, q-bio.NC
Syntactic parsing is the task of assigning a syntactic structure to a sentence. There are two popular syntactic parsing methods: constituency and dependency parsing. Recent works have used syntactic embeddings based on constituency trees, incremental top-down parsing, and other word syntactic features for brain activit...
http://arxiv.org/pdf/2302.08589v1
2023-02-16T21:28:11
2023-02-16T21:28:11
{ "title": [ 0.013885498, -0.0020198822, -0.024475098, -0.053131104, -0.041107178, -0.018035889, -0.046813965, 0.0049972534, -0.027999878, 0.046966553, 0.015090942, -0.020828247, 0.01109314, -0.06036377, -0.012023926, 0.009361267, 0.0028648376, 0...
http://arxiv.org/abs/2302.08583v1
JEIT: Joint End-to-End Model and Internal Language Model Training for Speech Recognition
Zhong Meng, Weiran Wang, Rohit Prabhavalkar, Tara N. Sainath, Tongzhou Chen, Ehsan Variani, Yu Zhang, Bo Li, Andrew Rosenberg, Bhuvana Ramabhadran
eess.AS, cs.AI, cs.CL, cs.LG, cs.SD
We propose JEIT, a joint end-to-end (E2E) model and internal language model (ILM) training method to inject large-scale unpaired text into ILM during E2E training which improves rare-word speech recognition. With JEIT, the E2E model computes an E2E loss on audio-transcript pairs while its ILM estimates a cross-entropy ...
http://arxiv.org/pdf/2302.08583v1
2023-02-16T21:07:38
2023-02-16T21:07:38
{ "title": [ -0.027145386, -0.014785767, -0.045562744, -0.060668945, 0.0015678406, 0.003206253, -0.058563232, 0.01309967, -0.023040771, 0.025344849, -0.018310547, 0.014221191, 0.012123108, -0.0335083, -0.014831543, 0.011726379, -0.014434814, 0.06...
http://arxiv.org/abs/2302.08582v2
Pretraining Language Models with Human Preferences
Tomasz Korbak, Kejian Shi, Angelica Chen, Rasika Bhalerao, Christopher L. Buckley, Jason Phang, Samuel R. Bowman, Ethan Perez
cs.CL, cs.LG
Language models (LMs) are pretrained to imitate internet text, including content that would violate human preferences if generated by an LM: falsehoods, offensive comments, personally identifiable information, low-quality or buggy code, and more. Here, we explore alternative objectives for pretraining LMs in a way that...
http://arxiv.org/pdf/2302.08582v2
2023-06-14T13:27:58
2023-02-16T21:03:33
{ "title": [ -0.009689331, -0.00079488754, -0.022872925, -0.050079346, -0.0031089783, -0.004585266, -0.04321289, 0.00007891655, -0.01184845, 0.024230957, -0.017456055, -0.0024814606, 0.0044059753, -0.018676758, -0.026397705, 0.0015525818, -0.01168823...
http://arxiv.org/abs/2302.08577v2
Keep it Neutral: Using Natural Language Inference to Improve Generation
Michail Mersinias, Kyle Mahowald
cs.CL
We explore incorporating natural language inference (NLI) into the text generative pipeline by using a pre-trained NLI model to assess whether a generated sentence entails, contradicts, or is neutral to the prompt and preceding text. First, we show that the NLI task is predictive of generation errors made by GPT-3. We ...
http://arxiv.org/pdf/2302.08577v2
2023-10-27T23:48:01
2023-02-16T20:46:36
{ "title": [ -0.024093628, -0.03048706, -0.038970947, -0.06677246, -0.008583069, 0.004196167, -0.06365967, -0.007209778, 0.013870239, 0.0056419373, -0.0115737915, -0.0031814575, 0.000562191, -0.033081055, -0.019317627, 0.007331848, -0.015945435, ...
http://arxiv.org/abs/2302.08575v1
Foundation Models for Natural Language Processing -- Pre-trained Language Models Integrating Media
Gerhard Paaß, Sven Giesselbach
cs.CL, cs.CV, cs.LG, cs.MM, 68W20, 68W25, I.2.6; I.2.7; I.2.8; I.2.10; I.4.8; I.4.10; I.5.2; I.5.4; I.7.0; J.1; J.3; K.4.1; K.4.2; K.5.0
This open access book provides a comprehensive overview of the state of the art in research and applications of Foundation Models and is intended for readers familiar with basic Natural Language Processing (NLP) concepts. Over the recent years, a revolutionary new paradigm has been developed for training models for NLP...
http://arxiv.org/pdf/2302.08575v1
2023-02-16T20:42:04
2023-02-16T20:42:04
{ "title": [ -0.011054993, 0.0024604797, -0.046844482, -0.06210327, -0.00034976006, -0.015304565, -0.046661377, -0.0059051514, -0.024002075, 0.031951904, -0.029266357, 0.008033752, 0.013031006, -0.028335571, -0.0069847107, -0.009857178, -0.002998352,...
http://arxiv.org/abs/2302.08531v1
Learning with Rejection for Abstractive Text Summarization
Meng Cao, Yue Dong, Jingyi He, Jackie Chi Kit Cheung
cs.CL
State-of-the-art abstractive summarization systems frequently hallucinate content that is not supported by the source document, mainly due to noise in the training dataset. Existing methods opt to drop the noisy samples or tokens from the training set entirely, reducing the effective training set size and creating an a...
http://arxiv.org/pdf/2302.08531v1
2023-02-16T19:07:08
2023-02-16T19:07:08
{ "title": [ -0.012542725, 0.0034160614, -0.013832092, -0.037628174, -0.0131073, 0.0052757263, -0.028945923, -0.01928711, 0.03289795, 0.029891968, -0.037017822, 0.023025513, -0.018798828, 0.0025539398, -0.064331055, -0.0038719177, -0.015487671, 0...
http://arxiv.org/abs/2302.08500v2
Auditing large language models: a three-layered approach
Jakob Mökander, Jonas Schuett, Hannah Rose Kirk, Luciano Floridi
cs.CL, cs.AI, cs.CY, K.4; K.6
Large language models (LLMs) represent a major advance in artificial intelligence (AI) research. However, the widespread use of LLMs is also coupled with significant ethical and social challenges. Previous research has pointed towards auditing as a promising governance mechanism to help ensure that AI systems are desig...
http://arxiv.org/pdf/2302.08500v2
2023-06-27T07:40:15
2023-02-16T18:55:21
{ "title": [ -0.045837402, -0.02671814, -0.054992676, -0.07745361, -0.006374359, 0.013771057, -0.07244873, 0.020553589, 0.0022239685, 0.031280518, -0.027786255, -0.0042648315, 0.003955841, -0.007080078, -0.048217773, -0.006465912, -0.019424438, 0...
http://arxiv.org/abs/2302.08468v3
LEVER: Learning to Verify Language-to-Code Generation with Execution
Ansong Ni, Srini Iyer, Dragomir Radev, Ves Stoyanov, Wen-tau Yih, Sida I. Wang, Xi Victoria Lin
cs.LG, cs.CL, cs.PL, cs.SE
The advent of large language models trained on code (code LLMs) has led to significant progress in language-to-code generation. State-of-the-art approaches in this area combine LLM decoding with sample pruning and reranking using test cases or heuristics based on the execution results. However, it is challenging to obt...
http://arxiv.org/pdf/2302.08468v3
2023-09-01T17:37:42
2023-02-16T18:23:22
{ "title": [ -0.03717041, -0.02029419, -0.04043579, -0.068115234, -0.0015487671, 0.011192322, -0.050964355, 0.029159546, -0.027526855, 0.012367249, -0.024337769, 0.045532227, -0.022216797, -0.01828003, -0.0005578995, -0.010292053, -0.026870728, 0...
http://arxiv.org/abs/2302.08464v1
Evaluating and Improving the Coreference Capabilities of Machine Translation Models
Asaf Yehudai, Arie Cattan, Omri Abend, Gabriel Stanovsky
cs.CL, cs.AI, cs.LG
Machine translation (MT) requires a wide range of linguistic capabilities, which current end-to-end models are expected to learn implicitly by observing aligned sentences in bilingual corpora. In this work, we ask: \emph{How well do MT models learn coreference resolution from implicit signal?} To answer this question, ...
http://arxiv.org/pdf/2302.08464v1
2023-02-16T18:16:09
2023-02-16T18:16:09
{ "title": [ -0.011749268, -0.010475159, 0.0030059814, -0.048339844, -0.007549286, 0.009567261, -0.024978638, -0.0053253174, 0.008522034, 0.06112671, -0.02166748, 0.024124146, 0.04244995, 0.001461029, -0.058624268, 0.038238525, -0.0059661865, 0.0...
http://arxiv.org/abs/2302.10914v1
GLUECons: A Generic Benchmark for Learning Under Constraints
Hossein Rajaby Faghihi, Aliakbar Nafar, Chen Zheng, Roshanak Mirzaee, Yue Zhang, Andrzej Uszok, Alexander Wan, Tanawan Premsri, Dan Roth, Parisa Kordjamshidi
cs.LG, cs.AI, cs.CL
Recent research has shown that integrating domain knowledge into deep learning architectures is effective -- it helps reduce the amount of required data, improves the accuracy of the models' decisions, and improves the interpretability of models. However, the research community is missing a convened benchmark for syste...
http://arxiv.org/pdf/2302.10914v1
2023-02-16T16:45:36
2023-02-16T16:45:36
{ "title": [ -0.05404663, -0.0024280548, -0.05618286, -0.037200928, 0.0074501038, 0.005847931, -0.04272461, -0.0158844, 0.008148193, 0.0071487427, -0.037322998, 0.031188965, -0.03729248, -0.0061531067, -0.031234741, -0.01234436, 0.007865906, 0.03...
http://arxiv.org/abs/2302.09301v1
Exploring the Representation Manifolds of Stable Diffusion Through the Lens of Intrinsic Dimension
Henry Kvinge, Davis Brown, Charles Godfrey
cs.CL, cs.CV, cs.LG
Prompting has become an important mechanism by which users can more effectively interact with many flavors of foundation model. Indeed, the last several years have shown that well-honed prompts can sometimes unlock emergent capabilities within such models. While there has been a substantial amount of empirical explorat...
http://arxiv.org/pdf/2302.09301v1
2023-02-16T16:22:30
2023-02-16T16:22:30
{ "title": [ 0.01133728, 0.0046310425, -0.033569336, -0.04257202, 0.023529053, -0.019180298, 0.029449463, -0.012008667, -0.00579834, 0.029190063, -0.006980896, 0.0003077984, -0.007144928, -0.02218628, -0.030822754, -0.00013792515, 0.021255493, 0....
http://arxiv.org/abs/2302.10144v1
Stabilising and accelerating light gated recurrent units for automatic speech recognition
Adel Moumen, Titouan Parcollet
eess.AS, cs.CL, cs.LG, cs.SD
The light gated recurrent units (Li-GRU) is well-known for achieving impressive results in automatic speech recognition (ASR) tasks while being lighter and faster to train than a standard gated recurrent units (GRU). However, the unbounded nature of its rectified linear unit on the candidate recurrent gate induces an i...
http://arxiv.org/pdf/2302.10144v1
2023-02-16T16:18:58
2023-02-16T16:18:58
{ "title": [ -0.006603241, 0.0046195984, -0.062927246, -0.061065674, 0.018554688, -0.02418518, -0.012687683, 0.023330688, -0.014160156, 0.050231934, -0.024215698, 0.022766113, 0.029174805, -0.011077881, -0.0335083, 0.012130737, -0.0073623657, 0.0...
http://arxiv.org/abs/2302.08399v5
Large Language Models Fail on Trivial Alterations to Theory-of-Mind Tasks
Tomer Ullman
cs.AI, cs.CL, I.2.7
Intuitive psychology is a pillar of common-sense reasoning. The replication of this reasoning in machine intelligence is an important stepping-stone on the way to human-like artificial intelligence. Several recent tasks and benchmarks for examining this reasoning in Large-Large Models have focused in particular on beli...
http://arxiv.org/pdf/2302.08399v5
2023-03-14T13:47:26
2023-02-16T16:18:03
{ "title": [ -0.0066184998, -0.037963867, 0.006893158, 0.0096588135, -0.06021118, -0.011222839, 0.017059326, -0.014251709, 0.0025043488, 0.052764893, -0.030593872, -0.029418945, 0.038024902, -0.03439331, -0.02696228, -0.019744873, -0.0031318665, ...
http://arxiv.org/abs/2302.08387v1
LEALLA: Learning Lightweight Language-agnostic Sentence Embeddings with Knowledge Distillation
Zhuoyuan Mao, Tetsuji Nakagawa
cs.CL
Large-scale language-agnostic sentence embedding models such as LaBSE (Feng et al., 2022) obtain state-of-the-art performance for parallel sentence alignment. However, these large-scale models can suffer from inference speed and computation overhead. This study systematically explores learning language-agnostic sentenc...
http://arxiv.org/pdf/2302.08387v1
2023-02-16T16:05:34
2023-02-16T16:05:34
{ "title": [ -0.020935059, -0.019638062, -0.039001465, -0.048950195, -0.011779785, -0.0032901764, -0.05731201, 0.0042381287, 0.02810669, 0.016815186, -0.028152466, -0.0015382767, -0.0021839142, -0.004184723, -0.04638672, 0.0015773773, -0.0019817352, ...
http://arxiv.org/abs/2302.10314v1
Dynamic Named Entity Recognition
Tristan Luiggi, Laure Soulier, Vincent Guigue, Siwar Jendoubi, Aurélien Baelde
cs.CL, cs.AI
Named Entity Recognition (NER) is a challenging and widely studied task that involves detecting and typing entities in text. So far,NER still approaches entity typing as a task of classification into universal classes (e.g. date, person, or location). Recent advances innatural language processing focus on architectures...
http://arxiv.org/pdf/2302.10314v1
2023-02-16T15:50:02
2023-02-16T15:50:02
{ "title": [ -0.012664795, 0.005718231, -0.056640625, -0.08880615, 0.022750854, -0.009063721, -0.02911377, 0.012168884, -0.010673523, 0.028259277, -0.023117065, 0.0010957718, -0.00016975403, -0.016708374, -0.0211792, 0.000790596, 0.005859375, 0.0...
http://arxiv.org/abs/2302.08374v3
Efficiency 360: Efficient Vision Transformers
Badri N. Patro, Vijay Srinivas Agneeswaran
cs.CV, cs.AI, cs.CL, cs.LG
Transformers are widely used for solving tasks in natural language processing, computer vision, speech, and music domains. In this paper, we talk about the efficiency of transformers in terms of memory (the number of parameters), computation cost (number of floating points operations), and performance of models, includ...
http://arxiv.org/pdf/2302.08374v3
2023-02-23T19:36:20
2023-02-16T15:43:32
{ "title": [ 0.005203247, 0.0047454834, -0.033172607, -0.0546875, 0.013465881, -0.023376465, 0.0016489029, -0.029464722, 0.006580353, 0.029373169, -0.045166016, -0.012565613, 0.029342651, -0.032287598, -0.035705566, -0.00078868866, 0.023498535, 0...
http://arxiv.org/abs/2302.08362v2
Conversation Style Transfer using Few-Shot Learning
Shamik Roy, Raphael Shu, Nikolaos Pappas, Elman Mansimov, Yi Zhang, Saab Mansour, Dan Roth
cs.CL
Conventional text style transfer approaches focus on sentence-level style transfer without considering contextual information, and the style is described with attributes (e.g., formality). When applying style transfer in conversations such as task-oriented dialogues, existing approaches suffer from these limitations as...
http://arxiv.org/pdf/2302.08362v2
2023-09-21T23:14:02
2023-02-16T15:27:00
{ "title": [ -0.013786316, -0.013252258, -0.055541992, -0.029647827, -0.006336212, -0.011650085, -0.036254883, 0.014045715, -0.00000333786, 0.036224365, 0.007762909, 0.0082473755, -0.003047943, -0.027313232, -0.027236938, 0.00642395, 0.009017944, ...
http://arxiv.org/abs/2302.09302v1
Bridge the Gap between Language models and Tabular Understanding
Nuo Chen, Linjun Shou, Ming Gong, Jian Pei, Chenyu You, Jianhui Chang, Daxin Jiang, Jia Li
cs.CL
Table pretrain-then-finetune paradigm has been proposed and employed at a rapid pace after the success of pre-training in the natural language domain. Despite the promising findings in tabular pre-trained language models (TPLMs), there is an input gap between pre-training and fine-tuning phases. For instance, TPLMs joi...
http://arxiv.org/pdf/2302.09302v1
2023-02-16T15:16:55
2023-02-16T15:16:55
{ "title": [ 0.0018005371, -0.01600647, -0.023147583, -0.06488037, -0.057800293, 0.0025520325, -0.09954834, -0.004924774, -0.00774765, 0.038604736, -0.045135498, -0.046722412, 0.01939392, -0.02671814, -0.011619568, -0.046325684, 0.008628845, -0.0...
http://arxiv.org/abs/2302.08351v2
A Survey on Event-based News Narrative Extraction
Brian Keith Norambuena, Tanushree Mitra, Chris North
cs.CL, cs.IR
Narratives are fundamental to our understanding of the world, providing us with a natural structure for knowledge representation over time. Computational narrative extraction is a subfield of artificial intelligence that makes heavy use of information retrieval and natural language processing techniques. Despite the im...
http://arxiv.org/pdf/2302.08351v2
2023-03-10T20:39:14
2023-02-16T15:11:53
{ "title": [ -0.02571106, -0.002084732, -0.036071777, -0.076293945, 0.0146102905, 0.013298035, -0.036499023, 0.0069084167, -0.014076233, 0.033569336, -0.0053901672, 0.013679504, -0.010864258, -0.0013837814, -0.04840088, -0.00031995773, -0.034851074, ...
http://arxiv.org/abs/2302.08343v1
Cluster-based Deep Ensemble Learning for Emotion Classification in Internet Memes
Xiaoyu Guo, Jing Ma, Arkaitz Zubiaga
cs.CV, cs.CL
Memes have gained popularity as a means to share visual ideas through the Internet and social media by mixing text, images and videos, often for humorous purposes. Research enabling automated analysis of memes has gained attention in recent years, including among others the task of classifying the emotion expressed in ...
http://arxiv.org/pdf/2302.08343v1
2023-02-16T15:01:07
2023-02-16T15:01:07
{ "title": [ 0.029052734, 0.013641357, -0.029937744, -0.024780273, 0.02406311, -0.02961731, 0.019378662, -0.018127441, 0.03918457, 0.031341553, -0.023834229, 0.0076675415, 0.014335632, -0.022964478, -0.050872803, 0.024002075, -0.0087509155, 0.014...
http://arxiv.org/abs/2302.08326v1
NUAA-QMUL-AIIT at Memotion 3: Multi-modal Fusion with Squeeze-and-Excitation for Internet Meme Emotion Analysis
Xiaoyu Guo, Jing Ma, Arkaitz Zubiaga
cs.CL
This paper describes the participation of our NUAA-QMUL-AIIT team in the Memotion 3 shared task on meme emotion analysis. We propose a novel multi-modal fusion method, Squeeze-and-Excitation Fusion (SEFusion), and embed it into our system for emotion classification in memes. SEFusion is a simple fusion method that empl...
http://arxiv.org/pdf/2302.08326v1
2023-02-16T14:29:58
2023-02-16T14:29:58
{ "title": [ 0.026504517, 0.016830444, -0.008621216, 0.007675171, -0.0026054382, -0.021636963, 0.0107040405, -0.05456543, 0.0736084, 0.051483154, -0.021896362, 0.0018024445, 0.052124023, -0.039245605, -0.042877197, 0.037139893, -0.006210327, 0.00...
http://arxiv.org/abs/2302.08268v1
Retrieval-augmented Image Captioning
Rita Ramos, Desmond Elliott, Bruno Martins
cs.CV, cs.CL
Inspired by retrieval-augmented language generation and pretrained Vision and Language (V&L) encoders, we present a new approach to image captioning that generates sentences given the input image and a set of captions retrieved from a datastore, as opposed to the image alone. The encoder in our model jointly processes ...
http://arxiv.org/pdf/2302.08268v1
2023-02-16T12:54:13
2023-02-16T12:54:13
{ "title": [ 0.016647339, 0.009902954, -0.0031471252, -0.029891968, 0.000110685825, -0.016220093, 0.010513306, -0.030059814, 0.013404846, 0.06488037, -0.027175903, 0.020584106, 0.0115737915, 0.0018320084, -0.048706055, 0.011383057, -0.0046043396, ...
http://arxiv.org/abs/2302.08258v1
Tragic and Comical Networks. Clustering Dramatic Genres According to Structural Properties
Szemes Botond, Vida Bence
cs.CL
There is a growing tradition in the joint field of network studies and drama history that produces interpretations from the character networks of the plays.The potential of such an interpretation is that the diagrams provide a different representation of the relationships between characters as compared to reading the t...
http://arxiv.org/pdf/2302.08258v1
2023-02-16T12:36:16
2023-02-16T12:36:16
{ "title": [ 0.0012264252, -0.0041542053, -0.05935669, -0.025894165, -0.013687134, -0.031066895, -0.022247314, 0.010910034, -0.030792236, 0.012901306, 0.0039978027, 0.02003479, 0.024047852, -0.051239014, 0.038116455, -0.002401352, -0.0057907104, ...
http://arxiv.org/abs/2302.08229v1
Improving Spoken Language Identification with Map-Mix
Shangeth Rajaa, Kriti Anandan, Swaraj Dalmia, Tarun Gupta, Eng Siong Chng
cs.LG, cs.CL
The pre-trained multi-lingual XLSR model generalizes well for language identification after fine-tuning on unseen languages. However, the performance significantly degrades when the languages are not very distinct from each other, for example, in the case of dialects. Low resource dialect classification remains a chall...
http://arxiv.org/pdf/2302.08229v1
2023-02-16T11:27:46
2023-02-16T11:27:46
{ "title": [ -0.0050201416, 0.00982666, -0.023773193, -0.028335571, 0.004589081, -0.0013599396, -0.037322998, -0.025817871, -0.030807495, 0.03451538, 0.002904892, -0.0035247803, -0.008422852, -0.032073975, -0.030593872, -0.01586914, -0.016967773, ...
http://arxiv.org/abs/2302.08220v2
Dialogue State Distillation Network with Inter-slot Contrastive Learning for Dialogue State Tracking
Jing Xu, Dandan Song, Chong Liu, Siu Cheung Hui, Fei Li, Qiang Ju, Xiaonan He, Jian Xie
cs.CL
In task-oriented dialogue systems, Dialogue State Tracking (DST) aims to extract users' intentions from the dialogue history. Currently, most existing approaches suffer from error propagation and are unable to dynamically select relevant information when utilizing previous dialogue states. Moreover, the relations betwe...
http://arxiv.org/pdf/2302.08220v2
2023-03-07T07:00:35
2023-02-16T11:05:24
{ "title": [ 0.02331543, -0.006187439, -0.021743774, -0.064575195, -0.00642395, -0.013023376, -0.03942871, -0.014007568, 0.033081055, 0.003540039, -0.032104492, -0.00014555454, -0.030456543, -0.032806396, -0.041656494, -0.0178833, -0.022216797, 0...
http://arxiv.org/abs/2302.08215v2
Aligning Language Models with Preferences through f-divergence Minimization
Dongyoung Go, Tomasz Korbak, Germán Kruszewski, Jos Rozen, Nahyeon Ryu, Marc Dymetman
cs.CL, cs.LG, stat.ML
Aligning language models with preferences can be posed as approximating a target distribution representing some desired behavior. Existing approaches differ both in the functional form of the target distribution and the algorithm used to approximate it. For instance, Reinforcement Learning from Human Feedback (RLHF) co...
http://arxiv.org/pdf/2302.08215v2
2023-06-06T13:09:21
2023-02-16T10:59:39
{ "title": [ -0.025238037, 0.0036678314, -0.05532837, -0.037506104, -0.0059547424, -0.0037994385, -0.05053711, -0.002128601, -0.0045928955, 0.029052734, -0.022964478, -0.005191803, 0.004722595, -0.0078048706, -0.04815674, -0.0056381226, -0.014312744,...
http://arxiv.org/abs/2302.08198v1
Un mod{è}le de base de connaissances terminologiques
Patrick Séguéla, Nathalie Aussenac-Gilles
cs.AI, cs.CL
In the present paper, we argue that Terminological Knowledge Bases (TKB) are all the more useful for addressing various needs as they do not fulfill formal criteria. Moreover, they intend to clarify the terminology of a given domain by illustrating term uses in various contexts. Thus we designed a TKB structure includi...
http://arxiv.org/pdf/2302.08198v1
2023-02-16T10:28:23
2023-02-16T10:28:23
{ "title": [ 0.003118515, -0.0060768127, -0.029632568, -0.061340332, -0.016784668, 0.010787964, -0.02708435, 0.013183594, 0.011192322, 0.02734375, 0.012741089, 0.019180298, 0.016647339, -0.062805176, 0.028289795, 0.023376465, -0.012023926, 0.0230...
http://arxiv.org/abs/2302.08150v2
Reanalyzing L2 Preposition Learning with Bayesian Mixed Effects and a Pretrained Language Model
Jakob Prange, Man Ho Ivy Wong
cs.CL
We use both Bayesian and neural models to dissect a data set of Chinese learners' pre- and post-interventional responses to two tests measuring their understanding of English prepositions. The results mostly replicate previous findings from frequentist analyses and newly reveal crucial interactions between student abil...
http://arxiv.org/pdf/2302.08150v2
2023-05-23T08:00:08
2023-02-16T08:54:05
{ "title": [ 0.00459671, -0.0023765564, -0.00033783913, -0.029464722, -0.028930664, -0.0088272095, -0.028793335, -0.0041046143, -0.017929077, 0.020401001, -0.019760132, 0.015838623, -0.014381409, -0.03326416, -0.031036377, -0.017654419, -0.020080566,...
http://arxiv.org/abs/2302.08148v1
Empirical Investigation of Neural Symbolic Reasoning Strategies
Yoichi Aoki, Keito Kudo, Tatsuki Kuribayashi, Ana Brassard, Masashi Yoshikawa, Keisuke Sakaguchi, Kentaro Inui
cs.AI, cs.CL
Neural reasoning accuracy improves when generating intermediate reasoning steps. However, the source of this improvement is yet unclear. Here, we investigate and factorize the benefit of generating intermediate steps for symbolic reasoning. Specifically, we decompose the reasoning strategy w.r.t. step granularity and c...
http://arxiv.org/pdf/2302.08148v1
2023-02-16T08:49:47
2023-02-16T08:49:47
{ "title": [ -0.011741638, -0.006427765, -0.019378662, -0.0029296875, -0.010971069, -0.026412964, 0.012954712, -0.019699097, -0.026550293, 0.040893555, 0.0026016235, -0.007820129, 0.0013446808, -0.037475586, -0.0385437, 0.0037403107, -0.000010550022,...
http://arxiv.org/abs/2302.08146v1
CluCDD:Contrastive Dialogue Disentanglement via Clustering
Jingsheng Gao, Zeyu Li, Suncheng Xiang, Ting Liu, Yuzhuo Fu
cs.CL
A huge number of multi-participant dialogues happen online every day, which leads to difficulty in understanding the nature of dialogue dynamics for both humans and machines. Dialogue disentanglement aims at separating an entangled dialogue into detached sessions, thus increasing the readability of long disordered dial...
http://arxiv.org/pdf/2302.08146v1
2023-02-16T08:47:51
2023-02-16T08:47:51
{ "title": [ 0.022460938, -0.00944519, -0.045898438, -0.064941406, 0.009033203, -0.021255493, -0.020721436, -0.0030670166, 0.01574707, 0.024887085, 0.0016460419, 0.010437012, 0.0022850037, -0.024978638, -0.027770996, 0.007106781, -0.018753052, 0....
http://arxiv.org/abs/2302.08143v2
Learning to Initialize: Can Meta Learning Improve Cross-task Generalization in Prompt Tuning?
Chengwei Qin, Shafiq Joty, Qian Li, Ruochen Zhao
cs.CL, cs.AI
Prompt tuning (PT) which only tunes the embeddings of an additional sequence of tokens per task, keeping the pre-trained language model (PLM) frozen, has shown remarkable performance in few-shot learning. Despite this, PT has been shown to rely heavily on good initialization of the prompt embeddings. In this work, we s...
http://arxiv.org/pdf/2302.08143v2
2023-05-20T23:48:41
2023-02-16T08:37:22
{ "title": [ -0.030654907, -0.0058174133, -0.011581421, -0.026748657, -0.032958984, 0.0070266724, -0.059692383, -0.029769897, -0.018188477, 0.035949707, -0.01826477, 0.02748108, -0.010314941, -0.019058228, -0.024780273, -0.014198303, -0.0033550262, ...
http://arxiv.org/abs/2302.08102v1
Prompt Tuning of Deep Neural Networks for Speaker-adaptive Visual Speech Recognition
Minsu Kim, Hyung-Il Kim, Yong Man Ro
cs.CL, cs.AI, cs.CV, cs.SD, eess.AS, eess.IV
Visual Speech Recognition (VSR) aims to infer speech into text depending on lip movements alone. As it focuses on visual information to model the speech, its performance is inherently sensitive to personal lip appearances and movements, and this makes the VSR models show degraded performance when they are applied to un...
http://arxiv.org/pdf/2302.08102v1
2023-02-16T06:01:31
2023-02-16T06:01:31
{ "title": [ -0.009811401, 0.009414673, -0.04840088, -0.055145264, -0.0035629272, -0.006187439, -0.045135498, 0.0037727356, 0.00023555756, 0.017440796, -0.031829834, 0.0061454773, -0.016815186, -0.016342163, -0.05154419, -0.015098572, 0.014312744, ...
http://arxiv.org/abs/2302.08095v1
PAAPLoss: A Phonetic-Aligned Acoustic Parameter Loss for Speech Enhancement
Muqiao Yang, Joseph Konan, David Bick, Yunyang Zeng, Shuo Han, Anurag Kumar, Shinji Watanabe, Bhiksha Raj
cs.SD, cs.CL, eess.AS
Despite rapid advancement in recent years, current speech enhancement models often produce speech that differs in perceptual quality from real clean speech. We propose a learning objective that formalizes differences in perceptual quality, by using domain knowledge of acoustic-phonetics. We identify temporal acoustic p...
http://arxiv.org/pdf/2302.08095v1
2023-02-16T05:17:06
2023-02-16T05:17:06
{ "title": [ -0.025680542, 0.0005555153, -0.04660034, -0.046722412, 0.0093688965, -0.0010700226, -0.011672974, 0.010238647, 0.0132369995, 0.05038452, 0.002544403, -0.0049819946, 0.05682373, -0.04232788, -0.030960083, 0.024841309, -0.012702942, 0....
http://arxiv.org/abs/2302.08092v2
Product Question Answering in E-Commerce: A Survey
Yang Deng, Wenxuan Zhang, Qian Yu, Wai Lam
cs.CL, cs.IR
Product question answering (PQA), aiming to automatically provide instant responses to customer's questions in E-Commerce platforms, has drawn increasing attention in recent years. Compared with typical QA problems, PQA exhibits unique challenges such as the subjectivity and reliability of user-generated contents in E-...
http://arxiv.org/pdf/2302.08092v2
2023-05-03T07:29:41
2023-02-16T05:09:41
{ "title": [ -0.037841797, -0.015960693, -0.034698486, -0.051940918, -0.00022864342, 0.0031414032, -0.036193848, 0.011199951, 0.00504303, 0.032104492, -0.023025513, -0.009147644, 0.005874634, -0.00573349, -0.059020996, -0.012207031, 0.0031261444, ...
http://arxiv.org/abs/2302.08091v1
Do We Still Need Clinical Language Models?
Eric Lehman, Evan Hernandez, Diwakar Mahajan, Jonas Wulff, Micah J. Smith, Zachary Ziegler, Daniel Nadler, Peter Szolovits, Alistair Johnson, Emily Alsentzer
cs.CL
Although recent advances in scaling large language models (LLMs) have resulted in improvements on many NLP tasks, it remains unclear whether these models trained primarily with general web text are the right tool in highly specialized, safety critical domains such as clinical text. Recent results have suggested that LL...
http://arxiv.org/pdf/2302.08091v1
2023-02-16T05:08:34
2023-02-16T05:08:34
{ "title": [ -0.03439331, -0.012916565, -0.025619507, -0.056915283, -0.039367676, 0.013549805, -0.08087158, 0.00022685528, -0.01474762, 0.03717041, 0.011154175, -0.02822876, 0.0035533905, -0.043914795, -0.02557373, 0.020751953, 0.0032405853, 0.02...
http://arxiv.org/abs/2302.08088v1
TAPLoss: A Temporal Acoustic Parameter Loss for Speech Enhancement
Yunyang Zeng, Joseph Konan, Shuo Han, David Bick, Muqiao Yang, Anurag Kumar, Shinji Watanabe, Bhiksha Raj
cs.CL, cs.SD, eess.AS
Speech enhancement models have greatly progressed in recent years, but still show limits in perceptual quality of their speech outputs. We propose an objective for perceptual quality based on temporal acoustic parameters. These are fundamental speech features that play an essential role in various applications, includi...
http://arxiv.org/pdf/2302.08088v1
2023-02-16T04:57:11
2023-02-16T04:57:11
{ "title": [ -0.0074691772, -0.0029945374, -0.057434082, -0.058746338, 0.014640808, -0.0060653687, -0.010765076, 0.0070228577, -0.001200676, 0.052459717, 0.008003235, -0.008201599, 0.034210205, -0.037109375, -0.053100586, 0.01411438, -0.009765625, ...
http://arxiv.org/abs/2302.08084v1
Learning Multi-Object Positional Relationships via Emergent Communication
Yicheng Feng, Boshi An, Zongqing Lu
cs.LG, cs.AI, cs.CL
The study of emergent communication has been dedicated to interactive artificial intelligence. While existing work focuses on communication about single objects or complex image scenes, we argue that communicating relationships between multiple objects is important in more realistic tasks, but understudied. In this pap...
http://arxiv.org/pdf/2302.08084v1
2023-02-16T04:44:53
2023-02-16T04:44:53
{ "title": [ -0.012306213, -0.0051231384, -0.038604736, -0.05029297, 0.005596161, -0.013198853, -0.015823364, -0.011978149, 0.029708862, 0.01625061, -0.026321411, 0.0016756058, -0.009117126, -0.006324768, -0.035308838, -0.006542206, -0.0069465637, ...