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/2301.11357v1
Multimodal Event Transformer for Image-guided Story Ending Generation
Yucheng Zhou, Guodong Long
cs.CV, cs.CL
Image-guided story ending generation (IgSEG) is to generate a story ending based on given story plots and ending image. Existing methods focus on cross-modal feature fusion but overlook reasoning and mining implicit information from story plots and ending image. To tackle this drawback, we propose a multimodal event tr...
http://arxiv.org/pdf/2301.11357v1
2023-01-26T19:10:07
2023-01-26T19:10:07
{ "title": [ -0.0032901764, 0.00008779764, -0.031341553, -0.06781006, -0.013389587, 0.009208679, -0.051818848, -0.0055656433, 0.00932312, 0.035888672, 0.028045654, 0.0018157959, 0.030670166, -0.034454346, -0.031555176, 0.022384644, -0.003932953, ...
http://arxiv.org/abs/2301.11312v1
LoRaLay: A Multilingual and Multimodal Dataset for Long Range and Layout-Aware Summarization
Laura Nguyen, Thomas Scialom, Benjamin Piwowarski, Jacopo Staiano
cs.CL
Text Summarization is a popular task and an active area of research for the Natural Language Processing community. By definition, it requires to account for long input texts, a characteristic which poses computational challenges for neural models. Moreover, real-world documents come in a variety of complex, visually-ri...
http://arxiv.org/pdf/2301.11312v1
2023-01-26T18:50:54
2023-01-26T18:50:54
{ "title": [ -0.019592285, -0.00055646896, -0.04901123, -0.09008789, -0.013191223, 0.0074424744, -0.08654785, 0.0004413128, 0.007106781, 0.054260254, 0.00007611513, 0.034118652, -0.0005412102, 0.0018148422, -0.033935547, 0.019744873, 0.00415802, ...
http://arxiv.org/abs/2301.11309v2
SemSup-XC: Semantic Supervision for Zero and Few-shot Extreme Classification
Pranjal Aggarwal, Ameet Deshpande, Karthik Narasimhan
cs.CL
Extreme classification (XC) involves predicting over large numbers of classes (thousands to millions), with real-world applications like news article classification and e-commerce product tagging. The zero-shot version of this task requires generalization to novel classes without additional supervision. In this paper, ...
http://arxiv.org/pdf/2301.11309v2
2023-06-22T06:56:24
2023-01-26T18:49:02
{ "title": [ 0.008163452, -0.0138168335, -0.033203125, -0.064331055, 0.004295349, -0.006793976, -0.012626648, -0.0021381378, 0.024047852, 0.013626099, -0.038360596, 0.02279663, -0.0014009476, -0.018997192, -0.019317627, -0.0037879944, -0.01776123, ...
http://arxiv.org/abs/2301.11305v2
DetectGPT: Zero-Shot Machine-Generated Text Detection using Probability Curvature
Eric Mitchell, Yoonho Lee, Alexander Khazatsky, Christopher D. Manning, Chelsea Finn
cs.CL, cs.AI
The increasing fluency and widespread usage of large language models (LLMs) highlight the desirability of corresponding tools aiding detection of LLM-generated text. In this paper, we identify a property of the structure of an LLM's probability function that is useful for such detection. Specifically, we demonstrate th...
http://arxiv.org/pdf/2301.11305v2
2023-07-23T04:18:36
2023-01-26T18:44:06
{ "title": [ -0.01600647, -0.004837036, -0.027618408, -0.033966064, 0.0006122589, -0.008605957, -0.010345459, -0.01687622, -0.00048184395, 0.03543091, -0.023162842, 0.016326904, 0.011787415, -0.012077332, -0.037872314, 0.010482788, -0.014656067, ...
http://arxiv.org/abs/2301.11293v1
Understanding Finetuning for Factual Knowledge Extraction from Language Models
Mehran Kazemi, Sid Mittal, Deepak Ramachandran
cs.CL, cs.LG
Language models (LMs) pretrained on large corpora of text from the web have been observed to contain large amounts of various types of knowledge about the world. This observation has led to a new and exciting paradigm in knowledge graph construction where, instead of manual curation or text mining, one extracts knowled...
http://arxiv.org/pdf/2301.11293v1
2023-01-26T18:29:50
2023-01-26T18:29:50
{ "title": [ -0.011962891, -0.012680054, -0.048583984, -0.084350586, -0.0002090931, -0.008583069, -0.04425049, 0.017868042, 0.005432129, 0.04333496, -0.026809692, 0.0082473755, 0.0022907257, 0.0005893707, -0.06689453, 0.010597229, -0.021987915, 0...
http://arxiv.org/abs/2301.11259v5
Domain-Agnostic Molecular Generation with Self-feedback
Yin Fang, Ningyu Zhang, Zhuo Chen, Lingbing Guo, Xiaohui Fan, Huajun Chen
cs.LG, cs.AI, cs.CE, cs.CL
The generation of molecules with desired properties has gained tremendous popularity, revolutionizing the way scientists design molecular structures and providing valuable support for chemical and drug design. However, despite the potential of language models in molecule generation, they face numerous challenges such a...
http://arxiv.org/pdf/2301.11259v5
2023-10-02T15:17:34
2023-01-26T17:52:56
{ "title": [ -0.04949951, -0.017074585, -0.046966553, -0.07446289, 0.000998497, 0.012763977, -0.04650879, 0.038970947, -0.0021915436, 0.035949707, 0.023483276, 0.02355957, -0.0017156601, -0.02217102, -0.0335083, 0.022155762, 0.012107849, 0.019775...
http://arxiv.org/abs/2301.11219v2
Characterizing the Entities in Harmful Memes: Who is the Hero, the Villain, the Victim?
Shivam Sharma, Atharva Kulkarni, Tharun Suresh, Himanshi Mathur, Preslav Nakov, Md. Shad Akhtar, Tanmoy Chakraborty
cs.CL, cs.CY
Memes can sway people's opinions over social media as they combine visual and textual information in an easy-to-consume manner. Since memes instantly turn viral, it becomes crucial to infer their intent and potentially associated harmfulness to take timely measures as needed. A common problem associated with meme compr...
http://arxiv.org/pdf/2301.11219v2
2023-04-10T20:07:38
2023-01-26T16:55:15
{ "title": [ 0.046813965, 0.011039734, -0.029708862, -0.026473999, 0.026397705, -0.03036499, 0.039276123, -0.033203125, -0.035125732, 0.034484863, 0.022232056, 0.036590576, 0.01864624, -0.0418396, -0.014678955, 0.011360168, -0.014595032, 0.049102...
http://arxiv.org/abs/2301.11174v1
Semi-Supervised Image Captioning by Adversarially Propagating Labeled Data
Dong-Jin Kim, Tae-Hyun Oh, Jinsoo Choi, In So Kweon
cs.CV, cs.AI, cs.CL, cs.LG
We present a novel data-efficient semi-supervised framework to improve the generalization of image captioning models. Constructing a large-scale labeled image captioning dataset is an expensive task in terms of labor, time, and cost. In contrast to manually annotating all the training samples, separately collecting uni...
http://arxiv.org/pdf/2301.11174v1
2023-01-26T15:25:43
2023-01-26T15:25:43
{ "title": [ 0.006111145, 0.016189575, -0.01260376, -0.035614014, 0.0087890625, -0.007003784, -0.00969696, -0.032806396, -0.019546509, 0.033294678, -0.014419556, 0.023254395, -0.021636963, -0.010528564, -0.030258179, -0.02003479, -0.012634277, 0....
http://arxiv.org/abs/2301.11125v1
A benchmark for toxic comment classification on Civil Comments dataset
Corentin Duchene, Henri Jamet, Pierre Guillaume, Reda Dehak
cs.CL, eess.AS
Toxic comment detection on social media has proven to be essential for content moderation. This paper compares a wide set of different models on a highly skewed multi-label hate speech dataset. We consider inference time and several metrics to measure performance and bias in our comparison. We show that all BERTs have ...
http://arxiv.org/pdf/2301.11125v1
2023-01-26T14:25:09
2023-01-26T14:25:09
{ "title": [ 0.015808105, 0.007659912, -0.03466797, -0.075927734, 0.030700684, -0.016525269, -0.024627686, -0.005214691, -0.005153656, 0.029022217, -0.036224365, 0.029815674, -0.029067993, 0.0062675476, -0.051239014, -0.012924194, -0.01008606, 0....
http://arxiv.org/abs/2301.11030v2
Paraphrase Acquisition from Image Captions
Marcel Gohsen, Matthias Hagen, Martin Potthast, Benno Stein
cs.CL
We propose to use image captions from the Web as a previously underutilized resource for paraphrases (i.e., texts with the same "message") and to create and analyze a corresponding dataset. When an image is reused on the Web, an original caption is often assigned. We hypothesize that different captions for the same ima...
http://arxiv.org/pdf/2301.11030v2
2023-02-15T15:32:26
2023-01-26T10:54:51
{ "title": [ 0.0011987686, -0.0039787292, -0.007587433, -0.050567627, -0.013023376, -0.0047073364, -0.017211914, -0.016159058, 0.00028586388, 0.06036377, -0.031234741, 0.01197052, 0.010314941, 0.0029582977, -0.05319214, 0.011276245, -0.01121521, ...
http://arxiv.org/abs/2301.11004v5
NLP as a Lens for Causal Analysis and Perception Mining to Infer Mental Health on Social Media
Muskan Garg, Chandni Saxena, Usman Naseem, Bonnie J Dorr
cs.CL
Interactions among humans on social media often convey intentions behind their actions, yielding a psychological language resource for Mental Health Analysis (MHA) of online users. The success of Computational Intelligence Techniques (CIT) for inferring mental illness from such social media resources points to NLP as a...
http://arxiv.org/pdf/2301.11004v5
2023-08-23T01:15:04
2023-01-26T09:26:01
{ "title": [ 0.010017395, 0.0011758804, -0.0357666, -0.06390381, 0.013221741, -0.006336212, -0.020996094, 0.008476257, -0.017913818, 0.008529663, -0.039764404, 0.025802612, 0.0068969727, -0.021591187, -0.0029525757, -0.026733398, -0.018951416, 0....
http://arxiv.org/abs/2301.10988v1
Neural Dynamic Focused Topic Model
Kostadin Cvejoski, Ramsés J. Sánchez, César Ojeda
cs.CL, cs.LG
Topic models and all their variants analyse text by learning meaningful representations through word co-occurrences. As pointed out by Williamson et al. (2010), such models implicitly assume that the probability of a topic to be active and its proportion within each document are positively correlated. This correlation ...
http://arxiv.org/pdf/2301.10988v1
2023-01-26T08:37:34
2023-01-26T08:37:34
{ "title": [ -0.0047416687, 0.021133423, -0.024002075, -0.047973633, 0.0021305084, -0.0065307617, -0.019500732, -0.024414062, 0.014976501, 0.03173828, 0.014282227, 0.0076675415, -0.013023376, -0.030593872, -0.029388428, 0.028213501, 0.0032482147, ...
http://arxiv.org/abs/2301.10951v1
Cross Modal Global Local Representation Learning from Radiology Reports and X-Ray Chest Images
Nathan Hadjiyski, Ali Vosoughi, Axel Wismueller
cs.CV, cs.CL, I.2.7
Deep learning models can be applied successfully in real-work problems; however, training most of these models requires massive data. Recent methods use language and vision, but unfortunately, they rely on datasets that are not usually publicly available. Here we pave the way for further research in the multimodal lang...
http://arxiv.org/pdf/2301.10951v1
2023-01-26T06:02:28
2023-01-26T06:02:28
{ "title": [ -0.0014123917, 0.022613525, -0.054351807, -0.04827881, 0.015914917, -0.02104187, -0.035949707, -0.01448822, 0.023544312, 0.018707275, -0.045013428, 0.013420105, -0.012611389, -0.016357422, -0.028930664, 0.007106781, 0.020309448, 0.02...
http://arxiv.org/abs/2301.13853v1
LAGAN: Deep Semi-Supervised Linguistic-Anthropology Classification with Conditional Generative Adversarial Neural Network
Rossi Kamal, Zuzana Kubincova
cs.CL, cs.CY
Education is a right of all, however, every individual is different than others. Teachers in post-communism era discover inherent individualism to equally train all towards job market of fourth industrial revolution. We can consider scenario of ethnic minority education in academic practices. Ethnic minority group has ...
http://arxiv.org/pdf/2301.13853v1
2023-01-26T05:25:34
2023-01-26T05:25:34
{ "title": [ -0.017410278, -0.0009007454, -0.02607727, -0.044433594, -0.010902405, 0.0058021545, -0.045562744, -0.015609741, -0.0004954338, 0.023620605, -0.0033512115, 0.009811401, 0.013267517, -0.028381348, -0.021087646, 0.001326561, -0.008964539, ...
http://arxiv.org/abs/2301.10939v1
Affective Faces for Goal-Driven Dyadic Communication
Scott Geng, Revant Teotia, Purva Tendulkar, Sachit Menon, Carl Vondrick
cs.CV, cs.CL, cs.LG
We introduce a video framework for modeling the association between verbal and non-verbal communication during dyadic conversation. Given the input speech of a speaker, our approach retrieves a video of a listener, who has facial expressions that would be socially appropriate given the context. Our approach further all...
http://arxiv.org/pdf/2301.10939v1
2023-01-26T05:00:09
2023-01-26T05:00:09
{ "title": [ -0.020355225, 0.0015535355, -0.043914795, -0.02571106, -0.011352539, -0.022064209, -0.021896362, -0.025360107, 0.0043678284, 0.010383606, -0.00843811, -0.010543823, 0.0051193237, -0.03842163, -0.020080566, -0.004211426, 0.010482788, ...
http://arxiv.org/abs/2301.10915v2
Parameter-Efficient Low-Resource Dialogue State Tracking by Prompt Tuning
Mingyu Derek Ma, Jiun-Yu Kao, Shuyang Gao, Arpit Gupta, Di Jin, Tagyoung Chung, Nanyun Peng
cs.CL, cs.AI
Dialogue state tracking (DST) is an important step in dialogue management to keep track of users' beliefs. Existing works fine-tune all language model (LM) parameters to tackle the DST task, which requires significant data and computing resources for training and hosting. The cost grows exponentially in the real-world ...
http://arxiv.org/pdf/2301.10915v2
2023-05-30T00:23:15
2023-01-26T03:01:59
{ "title": [ 0.0038719177, -0.026504517, -0.026428223, -0.101379395, -0.012672424, 0.009918213, -0.06451416, 0.0072364807, -0.004459381, 0.0069351196, 0.007118225, 0.0035896301, -0.016036987, -0.039367676, -0.015014648, -0.007587433, -0.011665344, ...
http://arxiv.org/abs/2301.10896v3
Causal Reasoning of Entities and Events in Procedural Texts
Li Zhang, Hainiu Xu, Yue Yang, Shuyan Zhou, Weiqiu You, Manni Arora, Chris Callison-Burch
cs.CL
Entities and events are crucial to natural language reasoning and common in procedural texts. Existing work has focused either exclusively on entity state tracking (e.g., whether a pan is hot) or on event reasoning (e.g., whether one would burn themselves by touching the pan), while these two tasks are often causally r...
http://arxiv.org/pdf/2301.10896v3
2023-02-16T13:56:22
2023-01-26T01:43:17
{ "title": [ -0.018707275, -0.018508911, -0.04324341, -0.0947876, -0.010391235, 0.006801605, -0.055847168, 0.032928467, -0.0063438416, 0.01348114, -0.0043945312, 0.016525269, 0.011985779, -0.025634766, -0.006225586, 0.00024044514, -0.01586914, 0....
http://arxiv.org/abs/2301.10887v1
Improving Text-based Early Prediction by Distillation from Privileged Time-Series Text
Jinghui Liu, Daniel Capurro, Anthony Nguyen, Karin Verspoor
cs.CL
Modeling text-based time-series to make prediction about a future event or outcome is an important task with a wide range of applications. The standard approach is to train and test the model using the same input window, but this approach neglects the data collected in longer input windows between the prediction time a...
http://arxiv.org/pdf/2301.10887v1
2023-01-26T01:07:02
2023-01-26T01:07:02
{ "title": [ -0.005290985, -0.0010051727, -0.03289795, -0.07775879, 0.006801605, 0.00042891502, -0.053222656, -0.017547607, 0.020370483, 0.045684814, -0.045806885, 0.0028095245, -0.0052948, -0.0099487305, -0.066467285, -0.011985779, 0.0029411316, ...
http://arxiv.org/abs/2301.10884v2
Break It Down: Evidence for Structural Compositionality in Neural Networks
Michael A. Lepori, Thomas Serre, Ellie Pavlick
cs.CL, cs.AI
Though modern neural networks have achieved impressive performance in both vision and language tasks, we know little about the functions that they implement. One possibility is that neural networks implicitly break down complex tasks into subroutines, implement modular solutions to these subroutines, and compose them i...
http://arxiv.org/pdf/2301.10884v2
2023-11-06T19:25:02
2023-01-26T00:53:11
{ "title": [ -0.0010547638, 0.004383087, -0.026306152, -0.05657959, 0.02508545, -0.004737854, -0.018569946, -0.024215698, -0.009384155, 0.027816772, 0.0053100586, 0.003320694, -0.026062012, -0.023971558, -0.049316406, 0.0035743713, -0.008300781, ...
http://arxiv.org/abs/2301.10871v3
Qualitative Analysis of a Graph Transformer Approach to Addressing Hate Speech: Adapting to Dynamically Changing Content
Liam Hebert, Hong Yi Chen, Robin Cohen, Lukasz Golab
cs.LG, cs.CL, cs.SI
Our work advances an approach for predicting hate speech in social media, drawing out the critical need to consider the discussions that follow a post to successfully detect when hateful discourse may arise. Using graph transformer networks, coupled with modelling attention and BERT-level natural language processing, o...
http://arxiv.org/pdf/2301.10871v3
2023-05-01T02:53:18
2023-01-25T23:32:32
{ "title": [ 0.055755615, 0.01878357, -0.04019165, -0.023284912, 0.019378662, -0.06021118, 0.01687622, -0.01574707, -0.027267456, 0.035583496, -0.009498596, 0.005214691, 0.032928467, -0.040985107, -0.011428833, -0.024353027, 0.011169434, 0.038818...
http://arxiv.org/abs/2301.10856v2
Partial Mobilization: Tracking Multilingual Information Flows Amongst Russian Media Outlets and Telegram
Hans W. A. Hanley, Zakir Durumeric
cs.CY, cs.CL, cs.LG, cs.SI
In response to disinformation and propaganda from Russian online media following the Russian invasion of Ukraine, Russian outlets including Russia Today and Sputnik News were banned throughout Europe. To maintain viewership, many of these Russian outlets began to heavily promote their content on messaging services like...
http://arxiv.org/pdf/2301.10856v2
2023-05-16T16:19:35
2023-01-25T22:27:40
{ "title": [ 0.021392822, -0.012435913, -0.00014150143, -0.029327393, -0.015159607, -0.011795044, 0.00019919872, -0.013198853, -0.043823242, 0.047729492, 0.017456055, 0.040161133, 0.034973145, -0.05331421, 0.03692627, 0.010887146, 0.0043258667, 0...
http://arxiv.org/abs/2301.10810v1
On the inconsistency of separable losses for structured prediction
Caio Corro
cs.LG, cs.CL, stat.ML
In this paper, we prove that separable negative log-likelihood losses for structured prediction are not necessarily Bayes consistent, or, in other words, minimizing these losses may not result in a model that predicts the most probable structure in the data distribution for a given input. This fact opens the question o...
http://arxiv.org/pdf/2301.10810v1
2023-01-25T20:02:07
2023-01-25T20:02:07
{ "title": [ -0.025314331, -0.005012512, -0.021469116, -0.071777344, 0.0021038055, 0.01739502, -0.050231934, -0.0090789795, 0.0017375946, 0.015167236, -0.027832031, 0.015975952, -0.027023315, -0.015220642, -0.0284729, -0.008178711, 0.009429932, 0...
http://arxiv.org/abs/2301.10799v2
Towards a Unified Model for Generating Answers and Explanations in Visual Question Answering
Chenxi Whitehouse, Tillman Weyde, Pranava Madhyastha
cs.CL
The field of visual question answering (VQA) has recently seen a surge in research focused on providing explanations for predicted answers. However, current systems mostly rely on separate models to predict answers and generate explanations, leading to less grounded and frequently inconsistent results. To address this,...
http://arxiv.org/pdf/2301.10799v2
2023-02-13T14:02:12
2023-01-25T19:29:19
{ "title": [ -0.000828743, -0.004310608, -0.0001937151, -0.0491333, -0.016189575, 0.007259369, -0.025878906, -0.006641388, 0.023880005, 0.025558472, -0.018356323, 0.0029964447, -0.009414673, -0.0024528503, -0.06707764, -0.00062704086, 0.010757446, ...
http://arxiv.org/abs/2301.10761v4
Fillers in Spoken Language Understanding: Computational and Psycholinguistic Perspectives
Tanvi Dinkar, Chloé Clavel, Ioana Vasilescu
cs.CL, cs.HC
Disfluencies (i.e. interruptions in the regular flow of speech), are ubiquitous to spoken discourse. Fillers ("uh", "um") are disfluencies that occur the most frequently compared to other kinds of disfluencies. Yet, to the best of our knowledge, there isn't a resource that brings together the research perspectives infl...
http://arxiv.org/pdf/2301.10761v4
2023-03-24T15:35:49
2023-01-25T18:55:05
{ "title": [ -0.005958557, -0.011131287, 0.0016479492, -0.06237793, -0.008460999, 0.003162384, -0.04763794, -0.0063323975, -0.00068998337, 0.013465881, -0.03427124, -0.0088272095, -0.00074386597, -0.023391724, -0.042785645, -0.024246216, -0.026809692...
http://arxiv.org/abs/2301.12974v1
Improving the Inference of Topic Models via Infinite Latent State Replications
Daniel Rugeles, Zhen Hai, Juan Felipe Carmona, Manoranjan Dash, Gao Cong
cs.CL, cs.AI, cs.LG, math.ST, stat.TH, G.3; I.2.6
In text mining, topic models are a type of probabilistic generative models for inferring latent semantic topics from text corpus. One of the most popular inference approaches to topic models is perhaps collapsed Gibbs sampling (CGS), which typically samples one single topic label for each observed document-word pair. I...
http://arxiv.org/pdf/2301.12974v1
2023-01-25T17:07:25
2023-01-25T17:07:25
{ "title": [ -0.012634277, 0.0024929047, -0.033111572, -0.05709839, 0.010231018, 0.00097608566, -0.024520874, -0.027374268, 0.022399902, 0.019607544, -0.03616333, 0.040161133, -0.033996582, -0.0051994324, -0.032348633, 0.0049934387, 0.0008125305, ...
http://arxiv.org/abs/2301.10684v1
Consistency is Key: Disentangling Label Variation in Natural Language Processing with Intra-Annotator Agreement
Gavin Abercrombie, Verena Rieser, Dirk Hovy
cs.CL
We commonly use agreement measures to assess the utility of judgements made by human annotators in Natural Language Processing (NLP) tasks. While inter-annotator agreement is frequently used as an indication of label reliability by measuring consistency between annotators, we argue for the additional use of intra-annot...
http://arxiv.org/pdf/2301.10684v1
2023-01-25T16:38:11
2023-01-25T16:38:11
{ "title": [ -0.025177002, 0.0071144104, -0.05114746, -0.043060303, 0.010482788, -0.022018433, -0.028121948, -0.008972168, -0.0048065186, 0.018951416, -0.022506714, 0.014465332, -0.0069885254, -0.014884949, -0.030014038, -0.0049324036, -0.024749756, ...
http://arxiv.org/abs/2301.13820v1
Explaining Large Language Model-Based Neural Semantic Parsers (Student Abstract)
Daking Rai, Yilun Zhou, Bailin Wang, Ziyu Yao
cs.CL, cs.AI, 68T50, I.2.7
While large language models (LLMs) have demonstrated strong capability in structured prediction tasks such as semantic parsing, few amounts of research have explored the underlying mechanisms of their success. Our work studies different methods for explaining an LLM-based semantic parser and qualitatively discusses the...
http://arxiv.org/pdf/2301.13820v1
2023-01-25T16:12:43
2023-01-25T16:12:43
{ "title": [ -0.0317688, 0.011276245, -0.027282715, -0.06323242, -0.010360718, 0.0018863678, -0.057617188, 0.0030174255, -0.02784729, 0.016708374, -0.024017334, -0.020874023, -0.007160187, -0.018463135, -0.010276794, -0.01927185, -0.012161255, 0....
http://arxiv.org/abs/2301.10656v1
Probing Taxonomic and Thematic Embeddings for Taxonomic Information
Filip Klubička, John D. Kelleher
cs.CL, cs.AI, cs.LG, 68T30
Modelling taxonomic and thematic relatedness is important for building AI with comprehensive natural language understanding. The goal of this paper is to learn more about how taxonomic information is structurally encoded in embeddings. To do this, we design a new hypernym-hyponym probing task and perform a comparative ...
http://arxiv.org/pdf/2301.10656v1
2023-01-25T15:59:26
2023-01-25T15:59:26
{ "title": [ -0.013259888, -0.01474762, -0.039916992, -0.052459717, 0.00605011, -0.018753052, -0.0046844482, -0.0090408325, 0.016159058, 0.008781433, -0.035095215, -0.0029315948, -0.0054893494, -0.027008057, -0.02583313, 0.003189087, 0.0034008026, ...
http://arxiv.org/abs/2301.10606v1
A Holistic Cascade System, benchmark, and Human Evaluation Protocol for Expressive Speech-to-Speech Translation
Wen-Chin Huang, Benjamin Peloquin, Justine Kao, Changhan Wang, Hongyu Gong, Elizabeth Salesky, Yossi Adi, Ann Lee, Peng-Jen Chen
cs.CL, cs.SD, eess.AS
Expressive speech-to-speech translation (S2ST) aims to transfer prosodic attributes of source speech to target speech while maintaining translation accuracy. Existing research in expressive S2ST is limited, typically focusing on a single expressivity aspect at a time. Likewise, this research area lacks standard evaluat...
http://arxiv.org/pdf/2301.10606v1
2023-01-25T14:27:00
2023-01-25T14:27:00
{ "title": [ -0.03225708, 0.0074691772, -0.026260376, -0.06915283, -0.00178051, 0.010848999, -0.06530762, 0.014144897, 0.0020503998, 0.03704834, -0.011894226, 0.008598328, 0.02658081, -0.023971558, -0.056793213, 0.007385254, -0.00008970499, 0.022...
http://arxiv.org/abs/2301.11333v1
Knowing About Knowing: An Illusion of Human Competence Can Hinder Appropriate Reliance on AI Systems
Gaole He, Lucie Kuiper, Ujwal Gadiraju
cs.HC, cs.AI, cs.CL
The dazzling promises of AI systems to augment humans in various tasks hinge on whether humans can appropriately rely on them. Recent research has shown that appropriate reliance is the key to achieving complementary team performance in AI-assisted decision making. This paper addresses an under-explored problem of whet...
http://arxiv.org/pdf/2301.11333v1
2023-01-25T14:26:10
2023-01-25T14:26:10
{ "title": [ 0.0022392273, 0.012290955, -0.03012085, 0.004638672, 0.0030651093, -0.05230713, 0.03286743, -0.039733887, -0.018676758, 0.024108887, -0.06567383, -0.0042037964, 0.012382507, -0.0206604, -0.023727417, -0.04748535, 0.009925842, 0.01541...
http://arxiv.org/abs/2301.10604v1
Automated multilingual detection of Pro-Kremlin propaganda in newspapers and Telegram posts
Veronika Solopova, Oana-Iuliana Popescu, Christoph Benzmüller, Tim Landgraf
cs.CL, cs.LG
The full-scale conflict between the Russian Federation and Ukraine generated an unprecedented amount of news articles and social media data reflecting opposing ideologies and narratives. These polarized campaigns have led to mutual accusations of misinformation and fake news, shaping an atmosphere of confusion and mist...
http://arxiv.org/pdf/2301.10604v1
2023-01-25T14:25:37
2023-01-25T14:25:37
{ "title": [ 0.03414917, 0.022460938, -0.011367798, -0.030563354, 0.032409668, -0.020401001, 0.024230957, -0.03289795, -0.04638672, 0.042388916, -0.008224487, 0.023406982, 0.019241333, -0.023956299, -0.01209259, -0.0021915436, -0.012382507, 0.049...
http://arxiv.org/abs/2301.10595v1
Distilling Text into Circuits
Vincent Wang-Mascianica, Jonathon Liu, Bob Coecke
cs.CL, cs.AI, cs.LO, math.CT
This paper concerns the structure of meanings within natural language. Earlier, a framework named DisCoCirc was sketched that (1) is compositional and distributional (a.k.a. vectorial); (2) applies to general text; (3) captures linguistic `connections' between meanings (cf. grammar) (4) updates word meanings as text pr...
http://arxiv.org/pdf/2301.10595v1
2023-01-25T13:56:34
2023-01-25T13:56:34
{ "title": [ 0.00028014183, -0.021636963, -0.0181427, -0.05203247, -0.044311523, 0.0051078796, -0.05227661, 0.011192322, 0.013496399, 0.060913086, -0.05014038, -0.0070228577, 0.019836426, -0.0028209686, -0.014381409, -0.005077362, 0.014373779, 0....
http://arxiv.org/abs/2301.10577v1
ARDIAS: AI-Enhanced Research Management, Discovery, and Advisory System
Debayan Banerjee, Seid Muhie Yimam, Sushil Awale, Chris Biemann
cs.CL, cs.AI, cs.IR, cs.LG
In this work, we present ARDIAS, a web-based application that aims to provide researchers with a full suite of discovery and collaboration tools. ARDIAS currently allows searching for authors and articles by name and gaining insights into the research topics of a particular researcher. With the aid of AI-based tools, A...
http://arxiv.org/pdf/2301.10577v1
2023-01-25T13:30:10
2023-01-25T13:30:10
{ "title": [ -0.043060303, -0.026565552, -0.036132812, -0.018569946, -0.022766113, 0.0014324188, -0.011047363, -0.016860962, -0.0036201477, 0.026809692, -0.020065308, -0.0044898987, 0.0066184998, -0.022781372, -0.044677734, -0.011528015, -0.002952575...
http://arxiv.org/abs/2301.10576v1
A Study on FGSM Adversarial Training for Neural Retrieval
Simon Lupart, Stéphane Clinchant
cs.IR, cs.CL
Neural retrieval models have acquired significant effectiveness gains over the last few years compared to term-based methods. Nevertheless, those models may be brittle when faced to typos, distribution shifts or vulnerable to malicious attacks. For instance, several recent papers demonstrated that such variations sever...
http://arxiv.org/pdf/2301.10576v1
2023-01-25T13:28:54
2023-01-25T13:28:54
{ "title": [ 0.013160706, 0.020019531, -0.01525116, -0.012931824, 0.020477295, -0.011856079, 0.015045166, -0.030914307, -0.01586914, 0.048217773, 0.009635925, 0.00083065033, -0.00818634, -0.02268982, -0.044769287, -0.0013103485, 0.004459381, -0.0...
http://arxiv.org/abs/2301.10546v1
Backward Compatibility During Data Updates by Weight Interpolation
Raphael Schumann, Elman Mansimov, Yi-An Lai, Nikolaos Pappas, Xibin Gao, Yi Zhang
cs.LG, cs.CL
Backward compatibility of model predictions is a desired property when updating a machine learning driven application. It allows to seamlessly improve the underlying model without introducing regression bugs. In classification tasks these bugs occur in the form of negative flips. This means an instance that was correct...
http://arxiv.org/pdf/2301.10546v1
2023-01-25T12:23:10
2023-01-25T12:23:10
{ "title": [ -0.020202637, 0.022994995, -0.07849121, -0.085876465, 0.016143799, -0.010360718, -0.059387207, 0.006126404, 0.014022827, 0.052093506, -0.002090454, -0.040283203, 0.011459351, -0.036865234, -0.043884277, -0.008117676, 0.04232788, 0.01...
http://arxiv.org/abs/2301.10527v1
Cross-lingual Argument Mining in the Medical Domain
Anar Yeginbergenova, Rodrigo Agerri
cs.CL
Nowadays the medical domain is receiving more and more attention in applications involving Artificial Intelligence. Clinicians have to deal with an enormous amount of unstructured textual data to make a conclusion about patients' health in their everyday life. Argument mining helps to provide a structure to such data b...
http://arxiv.org/pdf/2301.10527v1
2023-01-25T11:21:12
2023-01-25T11:21:12
{ "title": [ 0.021728516, -0.012779236, -0.009010315, -0.08258057, 0.017791748, 0.0118637085, -0.048095703, -0.015823364, -0.03161621, 0.024459839, 0.012893677, -0.00013744831, -0.01977539, -0.029342651, -0.029312134, -0.00459671, -0.008094788, 0...
http://arxiv.org/abs/2301.10521v2
ExaRanker: Explanation-Augmented Neural Ranker
Fernando Ferraretto, Thiago Laitz, Roberto Lotufo, Rodrigo Nogueira
cs.CL, cs.AI, cs.IR
Recent work has shown that inducing a large language model (LLM) to generate explanations prior to outputting an answer is an effective strategy to improve performance on a wide range of reasoning tasks. In this work, we show that neural rankers also benefit from explanations. We use LLMs such as GPT-3.5 to augment ret...
http://arxiv.org/pdf/2301.10521v2
2023-06-03T17:11:28
2023-01-25T11:03:04
{ "title": [ -0.004924774, -0.0064926147, -0.016433716, -0.03086853, -0.0007958412, -0.007896423, -0.010047913, -0.034484863, 0.0095825195, -0.0013542175, -0.002204895, -0.009429932, -0.014602661, -0.036895752, -0.017089844, 0.0011501312, -0.01313018...
http://arxiv.org/abs/2301.10517v1
Multi-Tenant Optimization For Few-Shot Task-Oriented FAQ Retrieval
Asha Vishwanathan, Rajeev Unnikrishnan Warrier, Gautham Vadakkekara Suresh, Chandra Shekhar Kandpal
cs.CL
Business-specific Frequently Asked Questions (FAQ) retrieval in task-oriented dialog systems poses unique challenges vis-\`a-vis community based FAQs. Each FAQ question represents an intent which is usually an umbrella term for many related user queries. We evaluate performance for such Business FAQs both with standard...
http://arxiv.org/pdf/2301.10517v1
2023-01-25T10:55:45
2023-01-25T10:55:45
{ "title": [ 0.011886597, -0.02166748, -0.02947998, -0.045928955, -0.023162842, -0.017044067, -0.0044403076, 0.03475952, 0.015167236, 0.060150146, -0.0026550293, -0.010375977, 0.023345947, -0.019515991, -0.05947876, -0.018203735, 0.017227173, -0....
http://arxiv.org/abs/2301.10483v1
SWING: Balancing Coverage and Faithfulness for Dialogue Summarization
Kung-Hsiang Huang, Siffi Singh, Xiaofei Ma, Wei Xiao, Feng Nan, Nicholas Dingwall, William Yang Wang, Kathleen McKeown
cs.CL
Missing information is a common issue of dialogue summarization where some information in the reference summaries is not covered in the generated summaries. To address this issue, we propose to utilize natural language inference (NLI) models to improve coverage while avoiding introducing factual inconsistencies. Specif...
http://arxiv.org/pdf/2301.10483v1
2023-01-25T09:33:11
2023-01-25T09:33:11
{ "title": [ -0.015701294, -0.016830444, -0.03366089, -0.097473145, -0.012763977, 0.016204834, -0.0703125, 0.01864624, -0.0019893646, 0.047821045, 0.0037975311, 0.021881104, -0.0107803345, -0.0012149811, -0.06201172, 0.013687134, -0.01158905, 0.0...
http://arxiv.org/abs/2301.10481v2
FewShotTextGCN: K-hop neighborhood regularization for few-shot learning on graphs
Niels van der Heijden, Ekaterina Shutova, Helen Yannakoudakis
cs.CL, cs.LG
We present FewShotTextGCN, a novel method designed to effectively utilize the properties of word-document graphs for improved learning in low-resource settings. We introduce K-hop Neighbourhood Regularization, a regularizer for heterogeneous graphs, and show that it stabilizes and improves learning when only a few trai...
http://arxiv.org/pdf/2301.10481v2
2023-02-06T15:53:12
2023-01-25T09:30:32
{ "title": [ -0.034698486, -0.02947998, -0.042388916, -0.028366089, -0.021972656, 0.01776123, -0.04119873, 0.019470215, 0.0040359497, 0.0074310303, -0.031082153, 0.037506104, -0.011100769, -0.020462036, 0.0052223206, -0.0033740997, 0.013015747, 0...
http://arxiv.org/abs/2301.10472v2
XLM-V: Overcoming the Vocabulary Bottleneck in Multilingual Masked Language Models
Davis Liang, Hila Gonen, Yuning Mao, Rui Hou, Naman Goyal, Marjan Ghazvininejad, Luke Zettlemoyer, Madian Khabsa
cs.CL, cs.LG
Large multilingual language models typically rely on a single vocabulary shared across 100+ languages. As these models have increased in parameter count and depth, vocabulary size has remained largely unchanged. This \textit{vocabulary bottleneck} limits the representational capabilities of multilingual models like XLM...
http://arxiv.org/pdf/2301.10472v2
2023-10-13T18:36:06
2023-01-25T09:15:17
{ "title": [ -0.05831909, -0.0050201416, -0.061187744, -0.04309082, -0.0025424957, 0.007987976, -0.079589844, 0.0058555603, -0.0048179626, 0.0423584, -0.0007123947, 0.0074424744, 0.041168213, -0.029876709, -0.021148682, 0.029251099, -0.02468872, ...
http://arxiv.org/abs/2301.10458v1
Improved Stock Price Movement Classification Using News Articles Based on Embeddings and Label Smoothing
Luis Villamil, Ryan Bausback, Shaeke Salman, Ting L. Liu, Conrad Horn, Xiuwen Liu
cs.LG, cs.AI, cs.CL, cs.NE
Stock price movement prediction is a challenging and essential problem in finance. While it is well established in modern behavioral finance that the share prices of related stocks often move after the release of news via reactions and overreactions of investors, how to capture the relationships between price movements...
http://arxiv.org/pdf/2301.10458v1
2023-01-25T08:33:45
2023-01-25T08:33:45
{ "title": [ -0.019104004, 0.019744873, -0.0579834, -0.012702942, 0.027389526, -0.027511597, 0.016479492, -0.025009155, -0.004673004, 0.021057129, -0.015899658, -0.0031433105, -0.025299072, -0.011878967, -0.041259766, -0.026031494, -0.009101868, ...
http://arxiv.org/abs/2301.10451v1
Knowledge-augmented Graph Neural Networks with Concept-aware Attention for Adverse Drug Event Detection
Shaoxiong Ji, Ya Gao, Pekka Marttinen
cs.CL
Adverse drug events (ADEs) are an important aspect of drug safety. Various texts such as biomedical literature, drug reviews, and user posts on social media and medical forums contain a wealth of information about ADEs. Recent studies have applied word embedding and deep learning -based natural language processing to a...
http://arxiv.org/pdf/2301.10451v1
2023-01-25T08:01:45
2023-01-25T08:01:45
{ "title": [ 0.012496948, -0.0042915344, -0.03955078, -0.07739258, -0.0053749084, -0.01335144, -0.060394287, 0.0143966675, 0.0138549805, 0.02130127, -0.0064086914, 0.016708374, -0.003156662, -0.031585693, -0.023040771, 0.01474762, -0.0051841736, ...
http://arxiv.org/abs/2301.10448v2
Pre-computed memory or on-the-fly encoding? A hybrid approach to retrieval augmentation makes the most of your compute
Michiel de Jong, Yury Zemlyanskiy, Nicholas FitzGerald, Joshua Ainslie, Sumit Sanghai, Fei Sha, William Cohen
cs.CL, cs.AI, cs.LG
Retrieval-augmented language models such as Fusion-in-Decoder are powerful, setting the state of the art on a variety of knowledge-intensive tasks. However, they are also expensive, due to the need to encode a large number of retrieved passages. Some work avoids this cost by pre-encoding a text corpus into a memory and...
http://arxiv.org/pdf/2301.10448v2
2023-06-02T23:13:05
2023-01-25T07:55:45
{ "title": [ 0.049560547, -0.021133423, -0.006412506, -0.041900635, -0.01776123, -0.015151978, 0.016433716, 0.02330017, -0.015159607, 0.009155273, -0.030426025, 0.018234253, -0.015434265, -0.026657104, 0.015365601, -0.04333496, 0.018081665, -0.01...
http://arxiv.org/abs/2301.11353v2
Using novel data and ensemble models to improve automated labeling of Sustainable Development Goals
Dirk U. Wulff, Dominik S. Meier, Rui Mata
cs.CL
A number of labeling systems based on text have been proposed to help monitor work on the United Nations (UN) Sustainable Development Goals (SDGs). Here, we present a systematic comparison of systems using a variety of text sources and show that systems differ considerably in their specificity (i.e., true-positive rate...
http://arxiv.org/pdf/2301.11353v2
2023-02-01T12:44:44
2023-01-25T07:44:46
{ "title": [ -0.023101807, 0.016494751, -0.046417236, -0.04055786, 0.010101318, -0.012145996, -0.03805542, -0.013137817, -0.017807007, 0.016815186, -0.020462036, 0.014015198, -0.038909912, -0.019744873, -0.038238525, -0.005584717, -0.005744934, 0...
http://arxiv.org/abs/2301.10444v1
An Experimental Study on Pretraining Transformers from Scratch for IR
Carlos Lassance, Hervé Déjean, Stéphane Clinchant
cs.IR, cs.CL
Finetuning Pretrained Language Models (PLM) for IR has been de facto the standard practice since their breakthrough effectiveness few years ago. But, is this approach well understood? In this paper, we study the impact of the pretraining collection on the final IR effectiveness. In particular, we challenge the current ...
http://arxiv.org/pdf/2301.10444v1
2023-01-25T07:43:05
2023-01-25T07:43:05
{ "title": [ 0.0552063, 0.015083313, -0.004878998, -0.058685303, 0.006515503, -0.016571045, -0.00819397, -0.05606079, -0.04925537, 0.026870728, -0.025344849, 0.028121948, -0.015274048, -0.03250122, 0.014076233, -0.028900146, 0.02973938, 0.0361633...
http://arxiv.org/abs/2301.10439v2
ViDeBERTa: A powerful pre-trained language model for Vietnamese
Cong Dao Tran, Nhut Huy Pham, Anh Nguyen, Truong Son Hy, Tu Vu
cs.CL, cs.LG
This paper presents ViDeBERTa, a new pre-trained monolingual language model for Vietnamese, with three versions - ViDeBERTa_xsmall, ViDeBERTa_base, and ViDeBERTa_large, which are pre-trained on a large-scale corpus of high-quality and diverse Vietnamese texts using DeBERTa architecture. Although many successful pre-tra...
http://arxiv.org/pdf/2301.10439v2
2023-02-10T15:55:58
2023-01-25T07:26:54
{ "title": [ -0.015357971, -0.025222778, -0.026977539, -0.060028076, -0.017974854, 0.0011606216, -0.045654297, 0.0107803345, -0.014343262, 0.032714844, -0.015106201, -0.008270264, 0.009162903, -0.028549194, -0.033416748, -0.0037670135, -0.0146484375,...
http://arxiv.org/abs/2301.10412v1
BDMMT: Backdoor Sample Detection for Language Models through Model Mutation Testing
Jiali Wei, Ming Fan, Wenjing Jiao, Wuxia Jin, Ting Liu
cs.CL, cs.AI, cs.CR
Deep neural networks (DNNs) and natural language processing (NLP) systems have developed rapidly and have been widely used in various real-world fields. However, they have been shown to be vulnerable to backdoor attacks. Specifically, the adversary injects a backdoor into the model during the training phase, so that in...
http://arxiv.org/pdf/2301.10412v1
2023-01-25T05:24:46
2023-01-25T05:24:46
{ "title": [ -0.03137207, -0.011741638, -0.0131073, -0.05178833, -0.019485474, 0.015472412, -0.07116699, -0.018493652, -0.03137207, 0.02734375, -0.008560181, 0.02708435, 0.003967285, -0.025756836, -0.014205933, -0.0010614395, -0.01914978, 0.03942...
http://arxiv.org/abs/2301.10410v5
One Model for All Domains: Collaborative Domain-Prefix Tuning for Cross-Domain NER
Xiang Chen, Lei Li, Shuofei Qiao, Ningyu Zhang, Chuanqi Tan, Yong Jiang, Fei Huang, Huajun Chen
cs.CL, cs.AI, cs.DB, cs.IR, cs.LG
Cross-domain NER is a challenging task to address the low-resource problem in practical scenarios. Previous typical solutions mainly obtain a NER model by pre-trained language models (PLMs) with data from a rich-resource domain and adapt it to the target domain. Owing to the mismatch issue among entity types in differe...
http://arxiv.org/pdf/2301.10410v5
2023-09-18T16:51:00
2023-01-25T05:16:43
{ "title": [ -0.014297485, -0.013664246, -0.014884949, -0.07891846, -0.010025024, 0.009712219, -0.05419922, 0.0015792847, -0.036132812, 0.027435303, 0.021133423, 0.015388489, 0.008811951, -0.051513672, 0.016494751, 0.029708862, -0.02545166, 0.044...
http://arxiv.org/abs/2301.10405v6
Editing Language Model-based Knowledge Graph Embeddings
Siyuan Cheng, Ningyu Zhang, Bozhong Tian, Xi Chen, Qingbing Liu, Huajun Chen
cs.CL, cs.AI, cs.DB, cs.IR, cs.LG
Recently decades have witnessed the empirical success of framing Knowledge Graph (KG) embeddings via language models. However, language model-based KG embeddings are usually deployed as static artifacts, making them difficult to modify post-deployment without re-training after deployment. To address this issue, we prop...
http://arxiv.org/pdf/2301.10405v6
2023-11-15T14:00:17
2023-01-25T04:45:06
{ "title": [ -0.045043945, -0.019317627, -0.055541992, -0.0703125, -0.01612854, 0.008979797, -0.07269287, 0.028305054, 0.0039901733, 0.029251099, -0.0067749023, -0.012077332, 0.018081665, -0.023132324, -0.026443481, 0.008903503, -0.0059318542, 0....
http://arxiv.org/abs/2301.10385v1
XNLI: Explaining and Diagnosing NLI-based Visual Data Analysis
Yingchaojie Feng, Xingbo Wang, Bo Pan, Kam Kwai Wong, Yi Ren, Shi Liu, Zihan Yan, Yuxin Ma, Huamin Qu, Wei Chen
cs.HC, cs.CL
Natural language interfaces (NLIs) enable users to flexibly specify analytical intentions in data visualization. However, diagnosing the visualization results without understanding the underlying generation process is challenging. Our research explores how to provide explanations for NLIs to help users locate the probl...
http://arxiv.org/pdf/2301.10385v1
2023-01-25T02:24:16
2023-01-25T02:24:16
{ "title": [ -0.02670288, -0.01689148, -0.041137695, -0.033447266, -0.014137268, -0.0059165955, -0.035858154, -0.026031494, 0.004802704, 0.00068855286, -0.04748535, -0.0020256042, -0.031402588, -0.021148682, -0.040222168, -0.02079773, -0.00466156, ...
http://arxiv.org/abs/2301.10371v1
Weakly Supervised Headline Dependency Parsing
Adrian Benton, Tianze Shi, Ozan İrsoy, Igor Malioutov
cs.CL, cs.AI, cs.LG, I.2.7
English news headlines form a register with unique syntactic properties that have been documented in linguistics literature since the 1930s. However, headlines have received surprisingly little attention from the NLP syntactic parsing community. We aim to bridge this gap by providing the first news headline corpus of U...
http://arxiv.org/pdf/2301.10371v1
2023-01-25T01:00:16
2023-01-25T01:00:16
{ "title": [ -0.029403687, 0.005138397, -0.037017822, -0.09094238, -0.003227234, 0.016204834, -0.07977295, 0.0050697327, -0.020767212, 0.020355225, -0.03173828, 0.022018433, -0.0368042, 0.005558014, -0.040008545, -0.021774292, -0.015716553, 0.034...
http://arxiv.org/abs/2301.10368v1
Language Model Detoxification in Dialogue with Contextualized Stance Control
Jing Qian, Xifeng Yan
cs.CL, cs.LG
To reduce the toxic degeneration in a pretrained Language Model (LM), previous work on Language Model detoxification has focused on reducing the toxicity of the generation itself (self-toxicity) without consideration of the context. As a result, a type of implicit offensive language where the generations support the of...
http://arxiv.org/pdf/2301.10368v1
2023-01-25T00:47:28
2023-01-25T00:47:28
{ "title": [ -0.001627922, 0.010856628, -0.040771484, -0.06500244, 0.00035333633, -0.012588501, -0.04800415, -0.000047266483, 0.017578125, 0.044555664, -0.00315094, -0.022918701, 0.014129639, -0.028945923, -0.060333252, 0.0015907288, -0.0013742447, ...
http://arxiv.org/abs/2301.11322v1
Semi-Automated Construction of Food Composition Knowledge Base
Jason Youn, Fangzhou Li, Ilias Tagkopoulos
cs.CL, cs.AI
A food composition knowledge base, which stores the essential phyto-, micro-, and macro-nutrients of foods is useful for both research and industrial applications. Although many existing knowledge bases attempt to curate such information, they are often limited by time-consuming manual curation processes. Outside of th...
http://arxiv.org/pdf/2301.11322v1
2023-01-24T22:08:49
2023-01-24T22:08:49
{ "title": [ -0.029037476, 0.0023536682, -0.060394287, -0.08673096, 0.0031337738, 0.00068330765, -0.064086914, 0.014015198, 0.011909485, 0.031585693, -0.008140564, 0.00484848, -0.0011129379, -0.027770996, -0.02720642, 0.0023231506, 0.0016860962, ...
http://arxiv.org/abs/2301.10309v1
Interactive-Chain-Prompting: Ambiguity Resolution for Crosslingual Conditional Generation with Interaction
Jonathan Pilault, Xavier Garcia, Arthur Bražinskas, Orhan Firat
cs.LG, cs.AI, cs.CL
Crosslingual conditional generation (e.g., machine translation) has long enjoyed the benefits of scaling. Nonetheless, there are still issues that scale alone may not overcome. A source query in one language, for instance, may yield several translation options in another language without any extra context. Only one tra...
http://arxiv.org/pdf/2301.10309v1
2023-01-24T21:08:13
2023-01-24T21:08:13
{ "title": [ -0.03567505, -0.0008511543, -0.047668457, -0.10119629, -0.004886627, 0.028762817, -0.097961426, 0.007587433, 0.026473999, 0.0067443848, -0.031433105, -0.0135650635, 0.005619049, -0.034606934, -0.032470703, -0.013008118, -0.012504578, ...
http://arxiv.org/abs/2301.10297v1
Large language models can segment narrative events similarly to humans
Sebastian Michelmann, Manoj Kumar, Kenneth A. Norman, Mariya Toneva
cs.CL, q-bio.NC
Humans perceive discrete events such as "restaurant visits" and "train rides" in their continuous experience. One important prerequisite for studying human event perception is the ability of researchers to quantify when one event ends and another begins. Typically, this information is derived by aggregating behavioral ...
http://arxiv.org/pdf/2301.10297v1
2023-01-24T20:34:37
2023-01-24T20:34:37
{ "title": [ 0.001876831, -0.0046958923, -0.021942139, -0.046691895, -0.06329346, -0.016326904, -0.06262207, 0.020965576, -0.035736084, 0.05090332, -0.023284912, 0.00472641, 0.04296875, -0.026992798, 0.033111572, 0.0067596436, -0.008529663, 0.013...
http://arxiv.org/abs/2301.10283v1
Audience-Centric Natural Language Generation via Style Infusion
Samraj Moorjani, Adit Krishnan, Hari Sundaram, Ewa Maslowska, Aravind Sankar
cs.CL, cs.LG
Adopting contextually appropriate, audience-tailored linguistic styles is critical to the success of user-centric language generation systems (e.g., chatbots, computer-aided writing, dialog systems). While existing approaches demonstrate textual style transfer with large volumes of parallel or non-parallel data, we arg...
http://arxiv.org/pdf/2301.10283v1
2023-01-24T19:57:50
2023-01-24T19:57:50
{ "title": [ -0.024658203, -0.024291992, -0.04498291, -0.05834961, -0.01725769, 0.0029563904, -0.0473938, 0.02406311, -0.008926392, 0.046569824, -0.0072631836, 0.01927185, 0.005329132, -0.00340271, -0.03363037, 0.010986328, -0.004611969, 0.022171...
http://arxiv.org/abs/2301.13819v2
Causal-Discovery Performance of ChatGPT in the context of Neuropathic Pain Diagnosis
Ruibo Tu, Chao Ma, Cheng Zhang
cs.CL, cs.LG
ChatGPT has demonstrated exceptional proficiency in natural language conversation, e.g., it can answer a wide range of questions while no previous large language models can. Thus, we would like to push its limit and explore its ability to answer causal discovery questions by using a medical benchmark (Tu et al. 2019) i...
http://arxiv.org/pdf/2301.13819v2
2023-02-06T12:03:38
2023-01-24T19:23:38
{ "title": [ 0.0056877136, -0.008972168, -0.019058228, -0.041168213, 0.00020587444, 0.0032157898, -0.018981934, -0.012901306, -0.016098022, -0.004863739, -0.021575928, 0.018478394, -0.017593384, -0.056549072, 0.018997192, -0.008956909, -0.028579712, ...
http://arxiv.org/abs/2301.10226v3
A Watermark for Large Language Models
John Kirchenbauer, Jonas Geiping, Yuxin Wen, Jonathan Katz, Ian Miers, Tom Goldstein
cs.LG, cs.CL, cs.CR
Potential harms of large language models can be mitigated by watermarking model output, i.e., embedding signals into generated text that are invisible to humans but algorithmically detectable from a short span of tokens. We propose a watermarking framework for proprietary language models. The watermark can be embedded ...
http://arxiv.org/pdf/2301.10226v3
2023-06-06T17:50:01
2023-01-24T18:52:59
{ "title": [ -0.01991272, 0.0012722015, -0.04248047, -0.06921387, 0.016220093, 0.00409317, -0.05996704, -0.0039787292, 0.00013649464, 0.036865234, -0.01524353, -0.0017108917, 0.001821518, -0.016220093, -0.038482666, 0.010421753, -0.012321472, 0.0...
http://arxiv.org/abs/2301.10186v2
ViHOS: Hate Speech Spans Detection for Vietnamese
Phu Gia Hoang, Canh Duc Luu, Khanh Quoc Tran, Kiet Van Nguyen, Ngan Luu-Thuy Nguyen
cs.CL
The rise in hateful and offensive language directed at other users is one of the adverse side effects of the increased use of social networking platforms. This could make it difficult for human moderators to review tagged comments filtered by classification systems. To help address this issue, we present the ViHOS (Vie...
http://arxiv.org/pdf/2301.10186v2
2023-01-26T08:58:01
2023-01-24T17:53:21
{ "title": [ 0.0154953, -0.008003235, -0.02822876, -0.035369873, 0.011329651, -0.029708862, 0.013572693, 0.00039482117, -0.020401001, 0.03540039, -0.046325684, 0.0033817291, 0.014160156, -0.012535095, -0.041900635, -0.03201294, -0.029800415, 0.06...
http://arxiv.org/abs/2301.10140v1
The Semantic Scholar Open Data Platform
Rodney Kinney, Chloe Anastasiades, Russell Authur, Iz Beltagy, Jonathan Bragg, Alexandra Buraczynski, Isabel Cachola, Stefan Candra, Yoganand Chandrasekhar, Arman Cohan, Miles Crawford, Doug Downey, Jason Dunkelberger, Oren Etzioni, Rob Evans, Sergey Feldman, Joseph Gorney, David Graham, Fangzhou Hu, Regan Huff, Daniel...
cs.DL, cs.CL
The volume of scientific output is creating an urgent need for automated tools to help scientists keep up with developments in their field. Semantic Scholar (S2) is an open data platform and website aimed at accelerating science by helping scholars discover and understand scientific literature. We combine public and pr...
http://arxiv.org/pdf/2301.10140v1
2023-01-24T17:13:08
2023-01-24T17:13:08
{ "title": [ -0.004814148, 0.013717651, -0.062561035, -0.08325195, 0.016586304, -0.013748169, -0.034240723, -0.016220093, -0.007663727, 0.043029785, -0.033721924, 0.0019054413, -0.0014715195, -0.025375366, -0.025268555, -0.002696991, 0.014511108, ...
http://arxiv.org/abs/2301.10095v2
Large Language Models as Fiduciaries: A Case Study Toward Robustly Communicating With Artificial Intelligence Through Legal Standards
John J. Nay
cs.CL, cs.AI, cs.CY
Artificial Intelligence (AI) is taking on increasingly autonomous roles, e.g., browsing the web as a research assistant and managing money. But specifying goals and restrictions for AI behavior is difficult. Similar to how parties to a legal contract cannot foresee every potential "if-then" contingency of their future ...
http://arxiv.org/pdf/2301.10095v2
2023-01-30T18:25:18
2023-01-24T16:03:20
{ "title": [ -0.055419922, -0.013801575, -0.05493164, -0.049072266, -0.0017633438, -0.0011091232, -0.04107666, 0.014801025, 0.013053894, 0.028411865, -0.03829956, -0.004634857, 0.020385742, -0.011779785, -0.047851562, -0.005329132, -0.021362305, ...
http://arxiv.org/abs/2301.10075v3
Gender Neutralization for an Inclusive Machine Translation: from Theoretical Foundations to Open Challenges
Andrea Piergentili, Dennis Fucci, Beatrice Savoldi, Luisa Bentivogli, Matteo Negri
cs.CL
Gender inclusivity in language technologies has become a prominent research topic. In this study, we explore gender-neutral translation (GNT) as a form of gender inclusivity and a goal to be achieved by machine translation (MT) models, which have been found to perpetuate gender bias and discrimination. Specifically, we...
http://arxiv.org/pdf/2301.10075v3
2023-07-04T09:01:41
2023-01-24T15:26:36
{ "title": [ -0.05569458, -0.0026168823, -0.034576416, -0.033081055, 0.0013799667, 0.0023899078, -0.019104004, 0.005207062, 0.0011882782, 0.072509766, -0.0071372986, 0.005142212, 0.05038452, -0.00052690506, -0.066467285, 0.023040771, -0.025131226, ...
http://arxiv.org/abs/2301.09992v1
Multitask Instruction-based Prompting for Fallacy Recognition
Tariq Alhindi, Tuhin Chakrabarty, Elena Musi, Smaranda Muresan
cs.CL
Fallacies are used as seemingly valid arguments to support a position and persuade the audience about its validity. Recognizing fallacies is an intrinsically difficult task both for humans and machines. Moreover, a big challenge for computational models lies in the fact that fallacies are formulated differently across ...
http://arxiv.org/pdf/2301.09992v1
2023-01-24T13:39:23
2023-01-24T13:39:23
{ "title": [ 0.022109985, 0.0035381317, -0.031799316, -0.0619812, -0.002861023, -0.029953003, -0.030807495, -0.002319336, -0.0068244934, 0.02330017, -0.033966064, 0.022247314, -0.011802673, -0.010559082, -0.013412476, -0.0154953, -0.009613037, 0....
http://arxiv.org/abs/2302.04662v2
Generating High-Precision Feedback for Programming Syntax Errors using Large Language Models
Tung Phung, José Cambronero, Sumit Gulwani, Tobias Kohn, Rupak Majumdar, Adish Singla, Gustavo Soares
cs.PL, cs.AI, cs.CL
Large language models (LLMs), such as Codex, hold great promise in enhancing programming education by automatically generating feedback for students. We investigate using LLMs to generate feedback for fixing syntax errors in Python programs, a key scenario in introductory programming. More concretely, given a student's...
http://arxiv.org/pdf/2302.04662v2
2023-04-28T11:33:44
2023-01-24T13:00:25
{ "title": [ -0.026748657, -0.03012085, -0.028686523, -0.07086182, -0.019546509, 0.00007075071, -0.07171631, 0.01802063, 0.025665283, 0.026473999, -0.037384033, -0.0028305054, 0.0031337738, -0.010192871, -0.056152344, -0.005859375, -0.0055274963, ...
http://arxiv.org/abs/2301.09919v2
Opportunities and Challenges in Neural Dialog Tutoring
Jakub Macina, Nico Daheim, Lingzhi Wang, Tanmay Sinha, Manu Kapur, Iryna Gurevych, Mrinmaya Sachan
cs.CL, cs.AI
Designing dialog tutors has been challenging as it involves modeling the diverse and complex pedagogical strategies employed by human tutors. Although there have been significant recent advances in neural conversational systems using large language models (LLMs) and growth in available dialog corpora, dialog tutoring h...
http://arxiv.org/pdf/2301.09919v2
2023-03-27T19:13:35
2023-01-24T11:00:17
{ "title": [ 0.0030784607, -0.0069122314, -0.020706177, -0.037261963, -0.014312744, -0.01525116, -0.014045715, -0.006729126, -0.014297485, 0.041259766, -0.00003027916, -0.00067281723, -0.0013561249, -0.018325806, -0.04751587, 0.0031757355, -0.0196228...
http://arxiv.org/abs/2301.09912v1
Applications and Challenges of Sentiment Analysis in Real-life Scenarios
Diptesh Kanojia, Aditya Joshi
cs.CL
Sentiment analysis has benefited from the availability of lexicons and benchmark datasets created over decades of research. However, its applications to the real world are a driving force for research in SA. This chapter describes some of these applications and related challenges in real-life scenarios. In this chapter...
http://arxiv.org/pdf/2301.09912v1
2023-01-24T10:49:21
2023-01-24T10:49:21
{ "title": [ 0.012138367, -0.013671875, -0.019958496, -0.030563354, 0.015319824, -0.020950317, 0.0071258545, -0.020614624, 0.0051116943, 0.04119873, -0.02961731, -0.010223389, -0.009094238, -0.0103302, -0.05947876, 0.0013542175, -0.0019140244, 0....
http://arxiv.org/abs/2301.09911v1
Conclusion-based Counter-Argument Generation
Milad Alshomary, Henning Wachsmuth
cs.CL
In real-world debates, the most common way to counter an argument is to reason against its main point, that is, its conclusion. Existing work on the automatic generation of natural language counter-arguments does not address the relation to the conclusion, possibly because many arguments leave their conclusion implicit...
http://arxiv.org/pdf/2301.09911v1
2023-01-24T10:49:01
2023-01-24T10:49:01
{ "title": [ 0.029296875, -0.007972717, -0.00032949448, -0.07104492, -0.0079422, 0.0058517456, -0.02355957, -0.016204834, -0.014801025, 0.02609253, 0.0029582977, 0.0025157928, -0.042144775, -0.02607727, -0.041809082, -0.005493164, -0.006729126, 0...
http://arxiv.org/abs/2301.09908v1
Cross-lingual German Biomedical Information Extraction: from Zero-shot to Human-in-the-Loop
Siting Liang, Mareike Hartmann, Daniel Sonntag
cs.HC, cs.CL
This paper presents our project proposal for extracting biomedical information from German clinical narratives with limited amounts of annotations. We first describe the applied strategies in transfer learning and active learning for solving our problem. After that, we discuss the design of the user interface for both ...
http://arxiv.org/pdf/2301.09908v1
2023-01-24T10:35:28
2023-01-24T10:35:28
{ "title": [ -0.034454346, -0.006713867, -0.037872314, -0.04550171, -0.010154724, 0.012084961, -0.037597656, -0.0050697327, 0.0063323975, 0.034729004, 0.011169434, -0.01058197, 0.019546509, -0.035308838, -0.041778564, 0.022094727, -0.0016670227, ...
http://arxiv.org/abs/2301.09896v1
Automated Identification of Disaster News For Crisis Management Using Machine Learning
Lord Christian Carl H. Regacho, Ai Matsushita, Angie M. Ceniza-Canillo
cs.CL
A lot of news sources picked up on Typhoon Rai (also known locally as Typhoon Odette), along with fake news outlets. The study honed in on the issue, to create a model that can identify between legitimate and illegitimate news articles. With this in mind, we chose the following machine learning algorithms in our develo...
http://arxiv.org/pdf/2301.09896v1
2023-01-24T10:13:00
2023-01-24T10:13:00
{ "title": [ -0.026824951, 0.018966675, -0.0513916, -0.037078857, 0.011131287, -0.02217102, -0.021621704, -0.017730713, -0.031402588, 0.029449463, -0.034973145, 0.011642456, -0.0011043549, -0.027404785, -0.022338867, -0.02204895, -0.010406494, 0....
http://arxiv.org/abs/2301.09820v1
A Stability Analysis of Fine-Tuning a Pre-Trained Model
Zihao Fu, Anthony Man-Cho So, Nigel Collier
cs.LG, cs.AI, cs.CL
Fine-tuning a pre-trained model (such as BERT, ALBERT, RoBERTa, T5, GPT, etc.) has proven to be one of the most promising paradigms in recent NLP research. However, numerous recent works indicate that fine-tuning suffers from the instability problem, i.e., tuning the same model under the same setting results in signifi...
http://arxiv.org/pdf/2301.09820v1
2023-01-24T05:11:17
2023-01-24T05:11:17
{ "title": [ 0.04244995, 0.011253357, 0.015472412, -0.039794922, -0.0048217773, -0.014549255, -0.019821167, -0.045013428, -0.011131287, 0.050109863, -0.037628174, 0.01828003, -0.0073394775, -0.0052108765, -0.03451538, -0.003730774, 0.018600464, 0...
http://arxiv.org/abs/2301.09809v2
Low-Resource Compositional Semantic Parsing with Concept Pretraining
Subendhu Rongali, Mukund Sridhar, Haidar Khan, Konstantine Arkoudas, Wael Hamza, Andrew McCallum
cs.CL
Semantic parsing plays a key role in digital voice assistants such as Alexa, Siri, and Google Assistant by mapping natural language to structured meaning representations. When we want to improve the capabilities of a voice assistant by adding a new domain, the underlying semantic parsing model needs to be retrained usi...
http://arxiv.org/pdf/2301.09809v2
2023-01-30T20:49:49
2023-01-24T04:27:27
{ "title": [ 0.0010957718, -0.015914917, -0.02986145, -0.10626221, 0.000025629997, 0.0032253265, -0.08300781, 0.017852783, 0.0043563843, 0.0340271, -0.013183594, 0.022918701, -0.020843506, -0.0066337585, -0.043304443, -0.0035648346, -0.016998291, ...
http://arxiv.org/abs/2301.10416v2
AI vs. Human -- Differentiation Analysis of Scientific Content Generation
Yongqiang Ma, Jiawei Liu, Fan Yi, Qikai Cheng, Yong Huang, Wei Lu, Xiaozhong Liu
cs.CL, cs.AI
Recent neural language models have taken a significant step forward in producing remarkably controllable, fluent, and grammatical text. Although studies have found that AI-generated text is not distinguishable from human-written text for crowd-sourcing workers, there still exist errors in AI-generated text which are ev...
http://arxiv.org/pdf/2301.10416v2
2023-02-12T11:58:22
2023-01-24T04:23:20
{ "title": [ -0.038238525, -0.000013291836, -0.049560547, -0.021514893, -0.0069503784, -0.026229858, 0.0002977848, -0.01423645, -0.030303955, 0.033416748, -0.016738892, 0.016998291, -0.004962921, -0.02671814, -0.01966858, -0.006164551, -0.008857727, ...
http://arxiv.org/abs/2301.09790v3
The Next Chapter: A Study of Large Language Models in Storytelling
Zhuohan Xie, Trevor Cohn, Jey Han Lau
cs.CL
To enhance the quality of generated stories, recent story generation models have been investigating the utilization of higher-level attributes like plots or commonsense knowledge. The application of prompt-based learning with large language models (LLMs), exemplified by GPT-3, has exhibited remarkable performance in di...
http://arxiv.org/pdf/2301.09790v3
2023-07-24T10:03:01
2023-01-24T02:44:02
{ "title": [ -0.037200928, -0.015914917, -0.05517578, -0.062164307, -0.012710571, -0.016784668, -0.05267334, 0.022888184, -0.032287598, 0.036193848, -0.01146698, 0.003227234, 0.018508911, -0.029296875, -0.0010032654, -0.0038604736, 0.002691269, 0...
http://arxiv.org/abs/2301.09785v1
Transformer-Patcher: One Mistake worth One Neuron
Zeyu Huang, Yikang Shen, Xiaofeng Zhang, Jie Zhou, Wenge Rong, Zhang Xiong
cs.CL, cs.AI
Large Transformer-based Pretrained Language Models (PLMs) dominate almost all Natural Language Processing (NLP) tasks. Nevertheless, they still make mistakes from time to time. For a model deployed in an industrial environment, fixing these mistakes quickly and robustly is vital to improve user experiences. Previous wo...
http://arxiv.org/pdf/2301.09785v1
2023-01-24T02:12:42
2023-01-24T02:12:42
{ "title": [ 0.02696228, 0.0013523102, -0.027511597, -0.0069122314, -0.027496338, -0.030792236, -0.025634766, -0.045074463, -0.035339355, 0.022064209, -0.026184082, -0.010932922, 0.020751953, -0.051239014, 0.01197052, -0.027236938, -0.005104065, ...
http://arxiv.org/abs/2301.09767v3
Truveta Mapper: A Zero-shot Ontology Alignment Framework
Mariyam Amir, Murchana Baruah, Mahsa Eslamialishah, Sina Ehsani, Alireza Bahramali, Sadra Naddaf-Sh, Saman Zarandioon
cs.LG, cs.AI, cs.CL
In this paper, a new perspective is suggested for unsupervised Ontology Matching (OM) or Ontology Alignment (OA) by treating it as a translation task. Ontologies are represented as graphs, and the translation is performed from a node in the source ontology graph to a path in the target ontology graph. The proposed fram...
http://arxiv.org/pdf/2301.09767v3
2023-08-22T00:22:42
2023-01-24T00:32:56
{ "title": [ -0.02067566, -0.008911133, -0.0592041, -0.07928467, 0.0035037994, -0.001991272, -0.0491333, -0.007701874, 0.03048706, 0.019760132, 0.01134491, -0.03302002, 0.0112838745, -0.032226562, -0.032226562, 0.011192322, 0.002922058, 0.0152587...
http://arxiv.org/abs/2301.09759v1
Topic Ontologies for Arguments
Yamen Ajjour, Johannes Kiesel, Benno Stein, Martin Potthast
cs.CL
Many computational argumentation tasks, like stance classification, are topic-dependent: the effectiveness of approaches to these tasks significantly depends on whether the approaches were trained on arguments from the same topics as those they are tested on. So, which are these topics that researchers train approaches...
http://arxiv.org/pdf/2301.09759v1
2023-01-23T23:43:24
2023-01-23T23:43:24
{ "title": [ 0.021255493, -0.008018494, -0.02609253, -0.09350586, -0.013420105, 0.00037884712, -0.04006958, 0.0030078888, -0.020095825, 0.02015686, 0.0015468597, 0.02708435, -0.028717041, -0.038757324, 0.015586853, 0.022476196, 0.011436462, 0.017...
http://arxiv.org/abs/2301.09715v2
PrimeQA: The Prime Repository for State-of-the-Art Multilingual Question Answering Research and Development
Avirup Sil, Jaydeep Sen, Bhavani Iyer, Martin Franz, Kshitij Fadnis, Mihaela Bornea, Sara Rosenthal, Scott McCarley, Rong Zhang, Vishwajeet Kumar, Yulong Li, Md Arafat Sultan, Riyaz Bhat, Radu Florian, Salim Roukos
cs.CL, cs.IR, cs.LG
The field of Question Answering (QA) has made remarkable progress in recent years, thanks to the advent of large pre-trained language models, newer realistic benchmark datasets with leaderboards, and novel algorithms for key components such as retrievers and readers. In this paper, we introduce PRIMEQA: a one-stop and ...
http://arxiv.org/pdf/2301.09715v2
2023-01-25T15:48:03
2023-01-23T20:43:26
{ "title": [ -0.015853882, -0.026657104, -0.027420044, -0.069885254, 0.011711121, 0.013038635, -0.038360596, 0.0039482117, 0.018218994, 0.02911377, -0.003458023, 0.0009937286, 0.0075035095, -0.014572144, -0.05102539, 0.022781372, -0.0014925003, 0...
http://arxiv.org/abs/2301.09685v2
Noisy Parallel Data Alignment
Ruoyu Xie, Antonios Anastasopoulos
cs.CL, cs.LG
An ongoing challenge in current natural language processing is how its major advancements tend to disproportionately favor resource-rich languages, leaving a significant number of under-resourced languages behind. Due to the lack of resources required to train and evaluate models, most modern language technologies are ...
http://arxiv.org/pdf/2301.09685v2
2023-02-10T17:21:23
2023-01-23T19:26:34
{ "title": [ -0.05480957, 0.013130188, -0.05053711, -0.03366089, -0.022781372, -0.00095939636, -0.040405273, -0.0051078796, 0.027008057, 0.06088257, -0.033843994, 0.0030021667, 0.011146545, -0.021514893, -0.026382446, 0.02420044, 0.024414062, 0.0...
http://arxiv.org/abs/2301.09656v3
Selective Explanations: Leveraging Human Input to Align Explainable AI
Vivian Lai, Yiming Zhang, Chacha Chen, Q. Vera Liao, Chenhao Tan
cs.AI, cs.CL, cs.HC, cs.LG
While a vast collection of explainable AI (XAI) algorithms have been developed in recent years, they are often criticized for significant gaps with how humans produce and consume explanations. As a result, current XAI techniques are often found to be hard to use and lack effectiveness. In this work, we attempt to close...
http://arxiv.org/pdf/2301.09656v3
2023-08-07T17:40:40
2023-01-23T19:00:02
{ "title": [ -0.021240234, -0.009567261, -0.01461792, -0.059143066, -0.009315491, 0.014511108, -0.033996582, -0.019363403, -0.017501831, 0.00075769424, -0.01977539, 0.01977539, -0.03338623, -0.00869751, -0.020706177, -0.01448822, -0.0039634705, 0...
http://arxiv.org/abs/2301.09626v1
Efficient Language Model Training through Cross-Lingual and Progressive Transfer Learning
Malte Ostendorff, Georg Rehm
cs.CL, cs.AI
Most Transformer language models are primarily pretrained on English text, limiting their use for other languages. As the model sizes grow, the performance gap between English and other languages with fewer compute and data resources increases even further. Consequently, more resource-efficient training methods are nee...
http://arxiv.org/pdf/2301.09626v1
2023-01-23T18:56:12
2023-01-23T18:56:12
{ "title": [ -0.020431519, -0.0036411285, -0.03668213, -0.07055664, 0.0038795471, 0.0032577515, -0.05831909, -0.0068206787, -0.0006375313, 0.024887085, 0.007785797, 0.009269714, 0.0084991455, -0.033355713, -0.01737976, 0.022323608, -0.013000488, ...
http://arxiv.org/abs/2301.09445v1
The Energy Worker Profiler from Technologies to Skills to Realize Energy Efficiency in Manufacturing
Silvia Fareri, Riccardo Apreda, Valentina Mulas, Ruben Alonso
cs.CL, cs.CY
In recent years, the manufacturing sector has been responsible for nearly 55 percent of total energy consumption, inducing a major impact on the global ecosystem. Although stricter regulations, restrictions on heavy manufacturing and technological advances are increasing its sustainability, zero-emission and fuel-effic...
http://arxiv.org/pdf/2301.09445v1
2023-01-23T14:08:34
2023-01-23T14:08:34
{ "title": [ -0.0006008148, 0.020523071, -0.055389404, -0.010482788, 0.031097412, -0.04272461, 0.04815674, -0.0031032562, -0.06958008, 0.016983032, -0.0079193115, 0.032165527, 0.0034770966, -0.038970947, 0.048309326, -0.015808105, 0.015899658, 0....
http://arxiv.org/abs/2301.09412v1
Deep Learning Mental Health Dialogue System
Lennart Brocki, George C. Dyer, Anna Gładka, Neo Christopher Chung
cs.CL, cs.AI, cs.CY, cs.HC
Mental health counseling remains a major challenge in modern society due to cost, stigma, fear, and unavailability. We posit that generative artificial intelligence (AI) models designed for mental health counseling could help improve outcomes by lowering barriers to access. To this end, we have developed a deep learnin...
http://arxiv.org/pdf/2301.09412v1
2023-01-23T13:10:23
2023-01-23T13:10:23
{ "title": [ 0.007293701, 0.011253357, -0.04425049, -0.06829834, 0.020874023, -0.01638794, -0.032470703, 0.000044941902, 0.00027275085, 0.037719727, -0.012084961, -0.024887085, 0.00447464, -0.044525146, -0.037353516, -0.0062217712, 0.01109314, 0....
http://arxiv.org/abs/2301.09361v1
SMDDH: Singleton Mention detection using Deep Learning in Hindi Text
Kusum Lata, Pardeep Singh, Kamlesh Dutta
cs.CL, cs.AI, cs.LG
Mention detection is an important component of coreference resolution system, where mentions such as name, nominal, and pronominals are identified. These mentions can be purely coreferential mentions or singleton mentions (non-coreferential mentions). Coreferential mentions are those mentions in a text that refer to th...
http://arxiv.org/pdf/2301.09361v1
2023-01-23T10:58:18
2023-01-23T10:58:18
{ "title": [ -0.008010864, -0.008102417, -0.04736328, -0.068237305, 0.0014743805, -0.004310608, -0.021469116, 0.015213013, -0.012542725, 0.039245605, -0.03314209, 0.046325684, 0.012519836, -0.008010864, -0.033233643, -0.0042800903, -0.02229309, 0...
http://arxiv.org/abs/2301.09350v2
Large-scale investigation of weakly-supervised deep learning for the fine-grained semantic indexing of biomedical literature
Anastasios Nentidis, Thomas Chatzopoulos, Anastasia Krithara, Grigorios Tsoumakas, Georgios Paliouras
cs.CL, cs.DL, cs.LG
Objective: Semantic indexing of biomedical literature is usually done at the level of MeSH descriptors with several related but distinct biomedical concepts often grouped together and treated as a single topic. This study proposes a new method for the automated refinement of subject annotations at the level of MeSH con...
http://arxiv.org/pdf/2301.09350v2
2023-10-05T14:17:39
2023-01-23T10:33:22
{ "title": [ -0.01411438, 0.013404846, -0.04333496, -0.046722412, 0.009063721, -0.0038928986, -0.017807007, -0.00806427, 0.008880615, 0.03152466, -0.00868988, 0.0031490326, 0.010498047, -0.021011353, -0.04714966, -0.0035362244, -0.00019204617, 0....
http://arxiv.org/abs/2301.10165v1
Lexi: Self-Supervised Learning of the UI Language
Pratyay Banerjee, Shweti Mahajan, Kushal Arora, Chitta Baral, Oriana Riva
cs.CL, cs.AI
Humans can learn to operate the user interface (UI) of an application by reading an instruction manual or how-to guide. Along with text, these resources include visual content such as UI screenshots and images of application icons referenced in the text. We explore how to leverage this data to learn generic visio-lingu...
http://arxiv.org/pdf/2301.10165v1
2023-01-23T09:05:49
2023-01-23T09:05:49
{ "title": [ -0.03353882, -0.026275635, -0.04119873, -0.051330566, -0.005882263, 0.0021457672, -0.055847168, 0.0010080338, 0.012207031, 0.02659607, -0.014823914, 0.005054474, -0.007423401, -0.014144897, -0.049621582, 0.0030574799, -0.008605957, 0...
http://arxiv.org/abs/2301.09295v1
Sensemaking About Contraceptive Methods Across Online Platforms
LeAnn McDowall, Maria Antoniak, David Mimno
cs.CL
Selecting a birth control method is a complex healthcare decision. While birth control methods provide important benefits, they can also cause unpredictable side effects and be stigmatized, leading many people to seek additional information online, where they can find reviews, advice, hypotheses, and experiences of oth...
http://arxiv.org/pdf/2301.09295v1
2023-01-23T06:44:38
2023-01-23T06:44:38
{ "title": [ 0.01927185, -0.0060768127, -0.024642944, -0.027770996, -0.019805908, -0.02029419, -0.003332138, 0.023544312, -0.018585205, 0.030944824, -0.004650116, -0.00091791153, 0.032806396, -0.042541504, 0.018234253, 0.007648468, 0.02368164, 0....
http://arxiv.org/abs/2301.09279v2
StockEmotions: Discover Investor Emotions for Financial Sentiment Analysis and Multivariate Time Series
Jean Lee, Hoyoul Luis Youn, Josiah Poon, Soyeon Caren Han
cs.CL, cs.AI, cs.LG
There has been growing interest in applying NLP techniques in the financial domain, however, resources are extremely limited. This paper introduces StockEmotions, a new dataset for detecting emotions in the stock market that consists of 10,000 English comments collected from StockTwits, a financial social media platfor...
http://arxiv.org/pdf/2301.09279v2
2023-02-06T10:42:47
2023-01-23T05:32:42
{ "title": [ 0.0060157776, -0.006095886, -0.047210693, -0.037353516, 0.024307251, -0.029296875, 0.020324707, -0.012763977, 0.01776123, 0.036499023, -0.045562744, 0.0008711815, 0.023620605, -0.008956909, -0.052856445, -0.0021915436, 0.0018777847, ...
http://arxiv.org/abs/2301.09244v2
Efficient Encoders for Streaming Sequence Tagging
Ayush Kaushal, Aditya Gupta, Shyam Upadhyay, Manaal Faruqui
cs.CL, cs.AI
A naive application of state-of-the-art bidirectional encoders for streaming sequence tagging would require encoding each token from scratch for each new token in an incremental streaming input (like transcribed speech). The lack of re-usability of previous computation leads to a higher number of Floating Point Operati...
http://arxiv.org/pdf/2301.09244v2
2023-03-16T08:09:49
2023-01-23T02:20:39
{ "title": [ -0.011726379, 0.0084991455, -0.04397583, -0.0892334, 0.030319214, -0.0007100105, -0.03338623, 0.014755249, 0.007663727, 0.046844482, -0.035339355, 0.008956909, -0.0015134811, 0.00031757355, -0.04586792, 0.012359619, 0.008773804, 0.03...
http://arxiv.org/abs/2301.09237v1
Semantic-aware Contrastive Learning for Electroencephalography-to-Text Generation with Curriculum Learning
Xiachong Feng, Xiaocheng Feng, Bing Qin
cs.HC, cs.CL
Electroencephalography-to-Text generation (EEG-to-Text), which aims to directly generate natural text from EEG signals has drawn increasing attention in recent years due to the enormous potential for Brain-computer interfaces (BCIs). However, the remarkable discrepancy between the subject-dependent EEG representation a...
http://arxiv.org/pdf/2301.09237v1
2023-01-23T00:54:48
2023-01-23T00:54:48
{ "title": [ -0.013160706, -0.00957489, 0.0051231384, 0.015045166, -0.029663086, -0.015541077, 0.008758545, -0.051116943, 0.0059509277, 0.035614014, -0.004852295, 0.022750854, 0.030654907, -0.04058838, -0.01512146, 0.007484436, -0.004837036, 0.03...
http://arxiv.org/abs/2301.09211v1
An Empirical Study of Metrics to Measure Representational Harms in Pre-Trained Language Models
Saghar Hosseini, Hamid Palangi, Ahmed Hassan Awadallah
cs.CL, cs.AI, I.2.7
Large-scale Pre-Trained Language Models (PTLMs) capture knowledge from massive human-written data which contains latent societal biases and toxic contents. In this paper, we leverage the primary task of PTLMs, i.e., language modeling, and propose a new metric to quantify manifested implicit representational harms in PT...
http://arxiv.org/pdf/2301.09211v1
2023-01-22T21:47:26
2023-01-22T21:47:26
{ "title": [ -0.000910759, -0.00060892105, -0.021850586, -0.046142578, 0.0032463074, -0.0047912598, -0.03656006, -0.021560669, -0.0076904297, 0.04348755, -0.013946533, 0.016464233, 0.04107666, -0.017623901, -0.018936157, 0.012901306, -0.008354187, ...
http://arxiv.org/abs/2301.09209v3
Summarize the Past to Predict the Future: Natural Language Descriptions of Context Boost Multimodal Object Interaction
Razvan-George Pasca, Alexey Gavryushin, Yen-Ling Kuo, Luc Van Gool, Otmar Hilliges, Xi Wang
cs.CV, cs.CL
We study object interaction anticipation in egocentric videos. This task requires an understanding of the spatiotemporal context formed by past actions on objects, coined action context. We propose TransFusion, a multimodal transformer-based architecture. It exploits the representational power of language by summarisin...
http://arxiv.org/pdf/2301.09209v3
2023-06-23T10:43:14
2023-01-22T21:30:12
{ "title": [ 0.028900146, -0.004688263, -0.056365967, -0.082336426, 0.0020275116, -0.034057617, -0.053649902, 0.0119018555, -0.019851685, 0.037597656, -0.019363403, 0.026351929, 0.037231445, -0.022109985, -0.0051956177, -0.007888794, 0.016616821, ...
http://arxiv.org/abs/2301.09175v1
Ensemble Transfer Learning for Multilingual Coreference Resolution
Tuan Manh Lai, Heng Ji
cs.CL
Entity coreference resolution is an important research problem with many applications, including information extraction and question answering. Coreference resolution for English has been studied extensively. However, there is relatively little work for other languages. A problem that frequently occurs when working wit...
http://arxiv.org/pdf/2301.09175v1
2023-01-22T18:22:55
2023-01-22T18:22:55
{ "title": [ 0.008270264, 0.008201599, -0.025314331, -0.07055664, 0.009590149, -0.003238678, -0.032348633, 0.010612488, -0.0048675537, 0.061553955, -0.001698494, 0.023635864, 0.028259277, -0.004436493, -0.0435791, 0.03564453, -0.013183594, 0.0428...
http://arxiv.org/abs/2302.09304v1
Interpretability in Activation Space Analysis of Transformers: A Focused Survey
Soniya Vijayakumar
cs.CL
The field of natural language processing has reached breakthroughs with the advent of transformers. They have remained state-of-the-art since then, and there also has been much research in analyzing, interpreting, and evaluating the attention layers and the underlying embedding space. In addition to the self-attention ...
http://arxiv.org/pdf/2302.09304v1
2023-01-22T18:03:20
2023-01-22T18:03:20
{ "title": [ 0.029846191, 0.037353516, -0.016296387, -0.044799805, 0.041107178, 0.0020256042, -0.020935059, -0.052490234, -0.00831604, 0.013870239, -0.010810852, 0.005191803, -0.012062073, -0.030395508, -0.029693604, -0.01373291, -0.006198883, 0....
http://arxiv.org/abs/2301.09169v1
Representing Interlingual Meaning in Lexical Databases
Fausto Giunchiglia, Gabor Bella, Nandu Chandran Nair, Yang Chi, Hao Xu
cs.CL, cs.AI, I.2.4; I.2.7
In today's multilingual lexical databases, the majority of the world's languages are under-represented. Beyond a mere issue of resource incompleteness, we show that existing lexical databases have structural limitations that result in a reduced expressivity on culturally-specific words and in mapping them across langua...
http://arxiv.org/pdf/2301.09169v1
2023-01-22T17:41:29
2023-01-22T17:41:29
{ "title": [ -0.025375366, -0.005104065, -0.030349731, -0.05871582, 0.0099105835, 0.015327454, -0.045654297, -0.005592346, 0.024276733, 0.04095459, -0.031555176, -0.01121521, -0.0014076233, -0.0017623901, -0.066589355, -0.005405426, -0.017044067, ...
http://arxiv.org/abs/2301.09112v2
Differentially Private Natural Language Models: Recent Advances and Future Directions
Lijie Hu, Ivan Habernal, Lei Shen, Di Wang
cs.CL, cs.AI, cs.LG
Recent developments in deep learning have led to great success in various natural language processing (NLP) tasks. However, these applications may involve data that contain sensitive information. Therefore, how to achieve good performance while also protecting the privacy of sensitive data is a crucial challenge in NLP...
http://arxiv.org/pdf/2301.09112v2
2023-10-23T13:28:35
2023-01-22T12:29:03
{ "title": [ -0.03338623, -0.010154724, -0.042053223, -0.06652832, 0.0079193115, 0.017227173, -0.058685303, 0.014892578, -0.02130127, 0.014060974, -0.015625, 0.0034160614, 0.0077400208, -0.021270752, -0.014373779, -0.007572174, -0.022949219, 0.05...
http://arxiv.org/abs/2301.09099v2
Unsupervised Data Selection for TTS: Using Arabic Broadcast News as a Case Study
Massa Baali, Tomoki Hayashi, Hamdy Mubarak, Soumi Maiti, Shinji Watanabe, Wassim El-Hajj, Ahmed Ali
cs.CL, cs.SD, eess.AS
Several high-resource Text to Speech (TTS) systems currently produce natural, well-established human-like speech. In contrast, low-resource languages, including Arabic, have very limited TTS systems due to the lack of resources. We propose a fully unsupervised method for building TTS, including automatic data selection...
http://arxiv.org/pdf/2301.09099v2
2023-01-26T07:51:54
2023-01-22T10:41:58
{ "title": [ -0.008544922, 0.00082826614, -0.0440979, -0.046447754, 0.00844574, -0.017715454, -0.019256592, -0.0013952255, -0.017227173, 0.028717041, -0.028671265, 0.0024280548, -0.018310547, -0.02897644, -0.03744507, -0.0112838745, 0.0010528564, ...
http://arxiv.org/abs/2301.09640v2
Weakly-Supervised Questions for Zero-Shot Relation Extraction
Saeed Najafi, Alona Fyshe
cs.CL, cs.AI, cs.LG
Zero-Shot Relation Extraction (ZRE) is the task of Relation Extraction where the training and test sets have no shared relation types. This very challenging domain is a good test of a model's ability to generalize. Previous approaches to ZRE reframed relation extraction as Question Answering (QA), allowing for the use ...
http://arxiv.org/pdf/2301.09640v2
2023-02-09T00:03:17
2023-01-21T22:18:24
{ "title": [ 0.0011262894, -0.003566742, -0.037231445, -0.079833984, 0.0067214966, 0.006160736, -0.029876709, 0.022750854, 0.03677368, 0.004852295, -0.008613586, 0.008666992, -0.02607727, -0.0101623535, -0.04159546, 0.004486084, 0.004207611, 0.03...
http://arxiv.org/abs/2301.09017v2
Transfer Knowledge from Natural Language to Electrocardiography: Can We Detect Cardiovascular Disease Through Language Models?
Jielin Qiu, William Han, Jiacheng Zhu, Mengdi Xu, Michael Rosenberg, Emerson Liu, Douglas Weber, Ding Zhao
cs.CL
Recent advancements in Large Language Models (LLMs) have drawn increasing attention since the learned embeddings pretrained on large-scale datasets have shown powerful ability in various downstream applications. However, whether the learned knowledge by LLMs can be transferred to clinical cardiology remains unknown. In...
http://arxiv.org/pdf/2301.09017v2
2023-02-04T17:29:01
2023-01-21T21:58:00
{ "title": [ -0.0033950806, -0.0026626587, -0.013145447, -0.016906738, -0.013381958, -0.0043525696, -0.015205383, -0.026794434, 0.0020713806, 0.034851074, -0.012435913, -0.0066108704, 0.045898438, -0.05102539, -0.025512695, 0.011329651, 0.0063476562,...
http://arxiv.org/abs/2301.09008v3
Poor Man's Quality Estimation: Predicting Reference-Based MT Metrics Without the Reference
Vilém Zouhar, Shehzaad Dhuliawala, Wangchunshu Zhou, Nico Daheim, Tom Kocmi, Yuchen Eleanor Jiang, Mrinmaya Sachan
cs.CL
Machine translation quality estimation (QE) predicts human judgements of a translation hypothesis without seeing the reference. State-of-the-art QE systems based on pretrained language models have been achieving remarkable correlations with human judgements yet they are computationally heavy and require human annotatio...
http://arxiv.org/pdf/2301.09008v3
2023-04-25T13:10:59
2023-01-21T21:02:16
{ "title": [ -0.046295166, -0.0012273788, -0.01600647, -0.011741638, -0.016082764, 0.004005432, -0.024795532, -0.044311523, -0.012382507, 0.03881836, -0.015312195, -0.025939941, 0.012878418, -0.025009155, -0.04458618, -0.009590149, -0.015686035, ...