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/2311.02216v1
Exploring the Numerical Reasoning Capabilities of Language Models: A Comprehensive Analysis on Tabular Data
Mubashara Akhtar, Abhilash Shankarampeta, Vivek Gupta, Arpit Patil, Oana Cocarascu, Elena Simperl
cs.CL, cs.LG
Numbers are crucial for various real-world domains such as finance, economics, and science. Thus, understanding and reasoning with numbers are essential skills for language models to solve different tasks. While different numerical benchmarks have been introduced in recent years, they are limited to specific numerical ...
http://arxiv.org/pdf/2311.02216v1
2023-11-03T20:05:30
2023-11-03T20:05:30
{ "title": [ -0.029891968, -0.015617371, -0.033721924, -0.0602417, -0.01802063, 0.0006890297, -0.06414795, 0.008377075, -0.0007996559, 0.030761719, -0.010818481, -0.020080566, 0.0113220215, -0.022766113, -0.0473938, -0.006679535, -0.0178833, 0.03...
http://arxiv.org/abs/2311.04929v1
An Interdisciplinary Outlook on Large Language Models for Scientific Research
James Boyko, Joseph Cohen, Nathan Fox, Maria Han Veiga, Jennifer I-Hsiu Li, Jing Liu, Bernardo Modenesi, Andreas H. Rauch, Kenneth N. Reid, Soumi Tribedi, Anastasia Visheratina, Xin Xie
cs.CL, cs.AI, cs.DL, cs.LG
In this paper, we describe the capabilities and constraints of Large Language Models (LLMs) within disparate academic disciplines, aiming to delineate their strengths and limitations with precision. We examine how LLMs augment scientific inquiry, offering concrete examples such as accelerating literature review by summ...
http://arxiv.org/pdf/2311.04929v1
2023-11-03T19:41:09
2023-11-03T19:41:09
{ "title": [ -0.046295166, -0.011169434, -0.053894043, -0.055114746, -0.0069503784, -0.00040102005, -0.061553955, 0.0030136108, -0.0071525574, 0.040527344, 0.003446579, -0.010093689, -0.00013804436, -0.027633667, -0.04724121, 0.022781372, -0.00830841...
http://arxiv.org/abs/2311.02205v2
An Introduction to Natural Language Processing Techniques and Framework for Clinical Implementation in Radiation Oncology
Reza Khanmohammadi, Mohammad M. Ghassemi, Kyle Verdecchia, Ahmed I. Ghanem, Luo Bing, Indrin J. Chetty, Hassan Bagher-Ebadian, Farzan Siddiqui, Mohamed Elshaikh, Benjamin Movsas, Kundan Thind
cs.CL
Natural Language Processing (NLP) is a key technique for developing Medical Artificial Intelligence (AI) systems that leverage Electronic Health Record (EHR) data to build diagnostic and prognostic models. NLP enables the conversion of unstructured clinical text into structured data that can be fed into AI algorithms. ...
http://arxiv.org/pdf/2311.02205v2
2023-11-08T11:51:16
2023-11-03T19:32:35
{ "title": [ 0.015808105, -0.0061187744, -0.016479492, -0.049713135, -0.0018444061, -0.0060195923, -0.014801025, 0.009902954, -0.044433594, 0.030899048, -0.02116394, 0.001663208, 0.011451721, -0.023010254, -0.0055999756, -0.014678955, -0.001162529, ...
http://arxiv.org/abs/2311.02192v1
Automating Governing Knowledge Commons and Contextual Integrity (GKC-CI) Privacy Policy Annotations with Large Language Models
Jake Chanenson, Madison Pickering, Noah Apthorpe
cs.CY, cs.CL, cs.LG
Identifying contextual integrity (CI) and governing knowledge commons (GKC) parameters in privacy policy texts can facilitate normative privacy analysis. However, GKC-CI annotation has heretofore required manual or crowdsourced effort. This paper demonstrates that high-accuracy GKC-CI parameter annotation of privacy po...
http://arxiv.org/pdf/2311.02192v1
2023-11-03T18:49:05
2023-11-03T18:49:05
{ "title": [ -0.014556885, -0.003074646, -0.0725708, -0.068847656, -0.018936157, -0.01878357, -0.074157715, 0.026367188, -0.03665161, 0.04699707, -0.03842163, -0.006046295, 0.043884277, -0.03488159, 0.0022640228, -0.026794434, -0.020370483, 0.046...
http://arxiv.org/abs/2311.04928v1
Leveraging Large Language Models for Collective Decision-Making
Marios Papachristou, Longqi Yang, Chin-Chia Hsu
cs.CL, cs.AI, cs.HC, cs.SI
In various work contexts, such as meeting scheduling, collaborating, and project planning, collective decision-making is essential but often challenging due to diverse individual preferences, varying work focuses, and power dynamics among members. To address this, we propose a system leveraging Large Language Models (L...
http://arxiv.org/pdf/2311.04928v1
2023-11-03T18:27:21
2023-11-03T18:27:21
{ "title": [ -0.019348145, -0.009552002, -0.052764893, -0.061798096, 0.010566711, 0.0019893646, -0.04248047, 0.010154724, 0.0061950684, 0.033599854, -0.031402588, -0.0057296753, 0.001824379, 0.004299164, -0.04510498, -0.005241394, -0.010726929, 0...
http://arxiv.org/abs/2311.02069v1
Grounded Intuition of GPT-Vision's Abilities with Scientific Images
Alyssa Hwang, Andrew Head, Chris Callison-Burch
cs.CL
GPT-Vision has impressed us on a range of vision-language tasks, but it comes with the familiar new challenge: we have little idea of its capabilities and limitations. In our study, we formalize a process that many have instinctively been trying already to develop "grounded intuition" of this new model. Inspired by the...
http://arxiv.org/pdf/2311.02069v1
2023-11-03T17:53:43
2023-11-03T17:53:43
{ "title": [ 0.03302002, 0.025985718, -0.011077881, 0.019241333, -0.022033691, -0.042297363, 0.025238037, -0.054534912, 0.009422302, 0.030319214, -0.045562744, -0.008468628, 0.02671814, -0.03012085, -0.016983032, -0.0007185936, 0.020858765, 0.036...
http://arxiv.org/abs/2311.02049v1
Post Turing: Mapping the landscape of LLM Evaluation
Alexey Tikhonov, Ivan P. Yamshchikov
cs.CL, cs.AI, 68T50, I.2.7
In the rapidly evolving landscape of Large Language Models (LLMs), introduction of well-defined and standardized evaluation methodologies remains a crucial challenge. This paper traces the historical trajectory of LLM evaluations, from the foundational questions posed by Alan Turing to the modern era of AI research. We...
http://arxiv.org/pdf/2311.02049v1
2023-11-03T17:24:50
2023-11-03T17:24:50
{ "title": [ -0.05545044, -0.0040016174, -0.040008545, -0.026641846, -0.03829956, 0.00466156, -0.04574585, -0.009124756, -0.024169922, 0.00069761276, -0.039489746, 0.018936157, -0.034973145, -0.028808594, -0.017578125, -0.018127441, -0.04119873, ...
http://arxiv.org/abs/2311.04927v1
Contextualizing the Limits of Model & Evaluation Dataset Curation on Semantic Similarity Classification Tasks
Daniel Theron
cs.CL, cs.LG
This paper demonstrates how the limitations of pre-trained models and open evaluation datasets factor into assessing the performance of binary semantic similarity classification tasks. As (1) end-user-facing documentation around the curation of these datasets and pre-trained model training regimes is often not easily a...
http://arxiv.org/pdf/2311.04927v1
2023-11-03T17:12:07
2023-11-03T17:12:07
{ "title": [ -0.014808655, 0.00667572, -0.052337646, -0.038269043, 0.018676758, -0.019515991, -0.010566711, -0.019424438, 0.004798889, 0.033355713, -0.052612305, 0.021057129, -0.0027694702, 0.0070877075, -0.05114746, -0.014328003, -0.009277344, 0...
http://arxiv.org/abs/2311.01623v1
VQPy: An Object-Oriented Approach to Modern Video Analytics
Shan Yu, Zhenting Zhu, Yu Chen, Hanchen Xu, Pengzhan Zhao, Yang Wang, Arthi Padmanabhan, Hugo Latapie, Harry Xu
cs.CV, cs.AI, cs.CL, cs.LG
Video analytics is widely used in contemporary systems and services. At the forefront of video analytics are video queries that users develop to find objects of particular interest. Building upon the insight that video objects (e.g., human, animals, cars, etc.), the center of video analytics, are similar in spirit to o...
http://arxiv.org/pdf/2311.01623v1
2023-11-03T16:58:10
2023-11-03T16:58:10
{ "title": [ 0.00007176399, -0.0014514923, -0.05206299, -0.07397461, 0.039367676, -0.019638062, -0.00044584274, -0.0020656586, -0.0047187805, 0.034698486, -0.032409668, -0.0060577393, -0.014526367, -0.008903503, -0.061645508, -0.016937256, 0.01239013...
http://arxiv.org/abs/2311.02025v1
Vicinal Risk Minimization for Few-Shot Cross-lingual Transfer in Abusive Language Detection
Gretel Liz De la Peña Sarracén, Paolo Rosso, Robert Litschko, Goran Glavaš, Simone Paolo Ponzetto
cs.CL
Cross-lingual transfer learning from high-resource to medium and low-resource languages has shown encouraging results. However, the scarcity of resources in target languages remains a challenge. In this work, we resort to data augmentation and continual pre-training for domain adaptation to improve cross-lingual abusiv...
http://arxiv.org/pdf/2311.02025v1
2023-11-03T16:51:07
2023-11-03T16:51:07
{ "title": [ 0.027648926, 0.0022258759, -0.027862549, -0.027069092, 0.010650635, -0.030456543, -0.0019292831, -0.024780273, -0.0074920654, 0.025436401, -0.020767212, 0.026443481, 0.020568848, -0.02305603, -0.0013504028, -0.009780884, -0.007621765, ...
http://arxiv.org/abs/2311.01981v1
ProSG: Using Prompt Synthetic Gradients to Alleviate Prompt Forgetting of RNN-like Language Models
Haotian Luo, Kunming Wu, Cheng Dai, Sixian Ding, Xinhao Chen
cs.CL
RNN-like language models are getting renewed attention from NLP researchers in recent years and several models have made significant progress, which demonstrates performance comparable to traditional transformers. However, due to the recurrent nature of RNNs, this kind of language model can only store information in a ...
http://arxiv.org/pdf/2311.01981v1
2023-11-03T15:34:02
2023-11-03T15:34:02
{ "title": [ -0.031311035, 0.011451721, -0.041046143, -0.0569458, -0.014007568, 0.0015792847, -0.06567383, 0.00010305643, 0.00091552734, 0.029846191, -0.012046814, -0.016525269, 0.022262573, -0.033111572, -0.029541016, 0.009506226, -0.0022354126, ...
http://arxiv.org/abs/2311.01967v1
The language of prompting: What linguistic properties make a prompt successful?
Alina Leidinger, Robert van Rooij, Ekaterina Shutova
cs.CL, cs.AI, cs.LG
The latest generation of LLMs can be prompted to achieve impressive zero-shot or few-shot performance in many NLP tasks. However, since performance is highly sensitive to the choice of prompts, considerable effort has been devoted to crowd-sourcing prompts or designing methods for prompt optimisation. Yet, we still lac...
http://arxiv.org/pdf/2311.01967v1
2023-11-03T15:03:36
2023-11-03T15:03:36
{ "title": [ -0.016342163, -0.010520935, -0.034942627, -0.031951904, -0.008392334, -0.019165039, -0.038726807, -0.025161743, -0.0234375, 0.013725281, -0.0065612793, -0.0045394897, 0.0050239563, -0.03164673, -0.017654419, -0.00095176697, 0.008560181, ...
http://arxiv.org/abs/2311.01964v1
Don't Make Your LLM an Evaluation Benchmark Cheater
Kun Zhou, Yutao Zhu, Zhipeng Chen, Wentong Chen, Wayne Xin Zhao, Xu Chen, Yankai Lin, Ji-Rong Wen, Jiawei Han
cs.CL, cs.AI
Large language models~(LLMs) have greatly advanced the frontiers of artificial intelligence, attaining remarkable improvement in model capacity. To assess the model performance, a typical approach is to construct evaluation benchmarks for measuring the ability level of LLMs in different aspects. Despite that a number o...
http://arxiv.org/pdf/2311.01964v1
2023-11-03T14:59:54
2023-11-03T14:59:54
{ "title": [ -0.0026245117, -0.002216339, 0.002445221, 0.0018386841, -0.056274414, -0.0072784424, -0.017425537, -0.042236328, -0.053588867, 0.00982666, 0.011924744, -0.003993988, -0.06210327, -0.041137695, 0.011222839, -0.0435791, -0.036834717, 0...
http://arxiv.org/abs/2311.01955v1
Too Much Information: Keeping Training Simple for BabyLMs
Lukas Edman, Lisa Bylinina
cs.CL
This paper details the work of the University of Groningen for the BabyLM Challenge. We follow the idea that, like babies, language models should be introduced to simpler concepts first and build off of that knowledge to understand more complex concepts. We examine this strategy of simple-then-complex through a variety...
http://arxiv.org/pdf/2311.01955v1
2023-11-03T14:50:00
2023-11-03T14:50:00
{ "title": [ 0.019485474, -0.007286072, -0.007926941, -0.04458618, -0.025604248, -0.0049705505, 0.010223389, 0.0035934448, -0.074523926, 0.039611816, -0.0102005005, 0.0035305023, -0.0126953125, -0.05618286, 0.017608643, -0.030944824, 0.030288696, ...
http://arxiv.org/abs/2311.04926v1
More Robots are Coming: Large Multimodal Models (ChatGPT) can Solve Visually Diverse Images of Parsons Problems
Irene Hou, Owen Man, Sophie Mettille, Sebastian Gutierrez, Kenneth Angelikas, Stephen MacNeil
cs.CL, cs.AI
The advent of large language models is reshaping computing education. Recent research has demonstrated that these models can produce better explanations than students, answer multiple-choice questions at or above the class average, and generate code that can pass automated tests in introductory courses. These capabilit...
http://arxiv.org/pdf/2311.04926v1
2023-11-03T14:47:17
2023-11-03T14:47:17
{ "title": [ 0.043823242, -0.01033783, 0.00907135, -0.009185791, -0.027496338, -0.023529053, -0.00573349, -0.035339355, -0.017715454, 0.01953125, -0.03652954, 0.033966064, -0.0158844, 0.0022125244, -0.004447937, -0.03930664, -0.011100769, 0.04495...
http://arxiv.org/abs/2311.01949v1
Hint-enhanced In-Context Learning wakes Large Language Models up for knowledge-intensive tasks
Yifan Wang, Qingyan Guo, Xinzhe Ni, Chufan Shi, Lemao Liu, Haiyun Jiang, Yujiu Yang
cs.CL
In-context learning (ICL) ability has emerged with the increasing scale of large language models (LLMs), enabling them to learn input-label mappings from demonstrations and perform well on downstream tasks. However, under the standard ICL setting, LLMs may sometimes neglect query-related information in demonstrations, ...
http://arxiv.org/pdf/2311.01949v1
2023-11-03T14:39:20
2023-11-03T14:39:20
{ "title": [ -0.0010919571, 0.0048599243, -0.046661377, -0.059020996, -0.0004222393, -0.031677246, -0.026657104, 0.011482239, -0.00541687, 0.026901245, -0.04058838, 0.012748718, 0.011352539, -0.013458252, -0.0023784637, -0.00308609, 0.010787964, ...
http://arxiv.org/abs/2311.01928v1
Constructing Temporal Dynamic Knowledge Graphs from Interactive Text-based Games
Keunwoo Peter Yu
cs.CL
In natural language processing, interactive text-based games serve as a test bed for interactive AI systems. Prior work has proposed to play text-based games by acting based on discrete knowledge graphs constructed by the Discrete Graph Updater (DGU) to represent the game state from the natural language description. Wh...
http://arxiv.org/pdf/2311.01928v1
2023-11-03T14:09:31
2023-11-03T14:09:31
{ "title": [ -0.015419006, -0.014785767, -0.057861328, -0.08062744, 0.0030117035, -0.013748169, -0.039245605, 0.014030457, 0.0015487671, 0.02520752, -0.013542175, 0.012969971, -0.05419922, -0.0047721863, -0.062561035, -0.028900146, 0.012428284, 0...
http://arxiv.org/abs/2311.01927v1
GateLoop: Fully Data-Controlled Linear Recurrence for Sequence Modeling
Tobias Katsch
cs.LG, cs.AI, cs.CL, cs.DS
Linear Recurrence has proven to be a powerful tool for modeling long sequences efficiently. In this work, we show that existing models fail to take full advantage of its potential. Motivated by this finding, we develop GateLoop, a foundational sequence model that generalizes linear recurrent models such as S4, S5, LRU ...
http://arxiv.org/pdf/2311.01927v1
2023-11-03T14:08:39
2023-11-03T14:08:39
{ "title": [ -0.03137207, -0.0069847107, -0.06451416, -0.08453369, 0.009155273, -0.00003886223, -0.07006836, 0.008056641, 0.011657715, 0.033233643, -0.009895325, 0.010353088, 0.020065308, -0.01953125, -0.042114258, 0.028274536, -0.006214142, 0.03...
http://arxiv.org/abs/2311.04925v1
Investigating Deep-Learning NLP for Automating the Extraction of Oncology Efficacy Endpoints from Scientific Literature
Aline Gendrin-Brokmann, Eden Harrison, Julianne Noveras, Leonidas Souliotis, Harris Vince, Ines Smit, Francisco Costa, David Milward, Sashka Dimitrievska, Paul Metcalfe, Emilie Louvet
cs.CL, cs.AI
Benchmarking drug efficacy is a critical step in clinical trial design and planning. The challenge is that much of the data on efficacy endpoints is stored in scientific papers in free text form, so extraction of such data is currently a largely manual task. Our objective is to automate this task as much as possible. I...
http://arxiv.org/pdf/2311.04925v1
2023-11-03T14:01:54
2023-11-03T14:01:54
{ "title": [ -0.015556335, 0.010047913, -0.02571106, -0.031921387, -0.006046295, -0.011558533, 0.0047340393, -0.028182983, -0.014144897, 0.03387451, -0.0033740997, 0.013435364, -0.009590149, -0.015594482, -0.04147339, 0.005962372, -0.02897644, 0....
http://arxiv.org/abs/2311.01918v1
Large Language Models Illuminate a Progressive Pathway to Artificial Healthcare Assistant: A Review
Mingze Yuan, Peng Bao, Jiajia Yuan, Yunhao Shen, Zifan Chen, Yi Xie, Jie Zhao, Yang Chen, Li Zhang, Lin Shen, Bin Dong
cs.CL, cs.AI, cs.LG
With the rapid development of artificial intelligence, large language models (LLMs) have shown promising capabilities in mimicking human-level language comprehension and reasoning. This has sparked significant interest in applying LLMs to enhance various aspects of healthcare, ranging from medical education to clinical...
http://arxiv.org/pdf/2311.01918v1
2023-11-03T13:51:36
2023-11-03T13:51:36
{ "title": [ -0.006965637, -0.008506775, -0.03665161, -0.07543945, 0.014854431, -0.0046157837, -0.04522705, 0.007911682, -0.015960693, 0.038238525, -0.0013608932, 0.004219055, 0.017837524, -0.02078247, -0.024353027, 0.013999939, -0.0010509491, 0....
http://arxiv.org/abs/2311.01907v1
BoschAI @ PLABA 2023: Leveraging Edit Operations in End-to-End Neural Sentence Simplification
Valentin Knappich, Simon Razniewski, Annemarie Friedrich
cs.CL
Automatic simplification can help laypeople to comprehend complex scientific text. Language models are frequently applied to this task by translating from complex to simple language. In this paper, we describe our system based on Llama 2, which ranked first in the PLABA shared task addressing the simplification of biom...
http://arxiv.org/pdf/2311.01907v1
2023-11-03T13:34:08
2023-11-03T13:34:08
{ "title": [ -0.018737793, -0.011177063, -0.021469116, -0.052825928, -0.017578125, 0.01625061, -0.05117798, -0.0058135986, -0.005317688, 0.048095703, 0.025100708, 0.0076904297, 0.01889038, -0.02267456, -0.04348755, 0.02796936, -0.006008148, 0.018...
http://arxiv.org/abs/2311.01882v1
Indicative Summarization of Long Discussions
Shahbaz Syed, Dominik Schwabe, Khalid Al-Khatib, Martin Potthast
cs.CL
Online forums encourage the exchange and discussion of different stances on many topics. Not only do they provide an opportunity to present one's own arguments, but may also gather a broad cross-section of others' arguments. However, the resulting long discussions are difficult to overview. This paper presents a novel ...
http://arxiv.org/pdf/2311.01882v1
2023-11-03T12:44:59
2023-11-03T12:44:59
{ "title": [ 0.054626465, 0.0038375854, 0.011291504, -0.10247803, -0.04623413, 0.0015439987, -0.055389404, -0.0036182404, -0.037719727, 0.023986816, 0.0037269592, -0.0020656586, -0.008857727, -0.05038452, 0.029266357, -0.0085372925, 0.02482605, 0...
http://arxiv.org/abs/2311.01876v1
Sentiment Analysis through LLM Negotiations
Xiaofei Sun, Xiaoya Li, Shengyu Zhang, Shuhe Wang, Fei Wu, Jiwei Li, Tianwei Zhang, Guoyin Wang
cs.CL
A standard paradigm for sentiment analysis is to rely on a singular LLM and makes the decision in a single round under the framework of in-context learning. This framework suffers the key disadvantage that the single-turn output generated by a single LLM might not deliver the perfect decision, just as humans sometimes ...
http://arxiv.org/pdf/2311.01876v1
2023-11-03T12:35:29
2023-11-03T12:35:29
{ "title": [ -0.008270264, -0.008224487, -0.025009155, -0.004131317, -0.06274414, -0.011703491, -0.032928467, -0.011955261, -0.018875122, 0.026626587, -0.051452637, -0.0024433136, -0.0063209534, -0.03555298, -0.018630981, -0.039001465, -0.009246826, ...
http://arxiv.org/abs/2311.01873v1
Efficient Black-Box Adversarial Attacks on Neural Text Detectors
Vitalii Fishchuk, Daniel Braun
cs.CL
Neural text detectors are models trained to detect whether a given text was generated by a language model or written by a human. In this paper, we investigate three simple and resource-efficient strategies (parameter tweaking, prompt engineering, and character-level mutations) to alter texts generated by GPT-3.5 that a...
http://arxiv.org/pdf/2311.01873v1
2023-11-03T12:29:32
2023-11-03T12:29:32
{ "title": [ -0.010986328, -0.0039024353, -0.020324707, -0.033477783, 0.020996094, -0.013641357, 0.008369446, -0.019882202, -0.025665283, 0.037963867, -0.0116119385, 0.020843506, -0.01096344, -0.002090454, -0.04675293, -0.008460999, -0.013389587, ...
http://arxiv.org/abs/2311.01870v1
Multi-EuP: The Multilingual European Parliament Dataset for Analysis of Bias in Information Retrieval
Jinrui Yang, Timothy Baldwin, Trevor Cohn
cs.CL, cs.AI, cs.IR
We present Multi-EuP, a new multilingual benchmark dataset, comprising 22K multi-lingual documents collected from the European Parliament, spanning 24 languages. This dataset is designed to investigate fairness in a multilingual information retrieval (IR) context to analyze both language and demographic bias in a ranki...
http://arxiv.org/pdf/2311.01870v1
2023-11-03T12:29:11
2023-11-03T12:29:11
{ "title": [ 0.0032444, 0.0067443848, -0.056884766, -0.012580872, 0.025726318, -0.032073975, -0.0018129349, -0.038482666, 0.023727417, 0.037384033, -0.04486084, 0.011329651, 0.040527344, -0.004562378, -0.037200928, 0.009803772, -0.018371582, 0.05...
http://arxiv.org/abs/2311.01866v1
Towards Concept-Aware Large Language Models
Chen Shani, Jilles Vreeken, Dafna Shahaf
cs.CL, cs.AI
Concepts play a pivotal role in various human cognitive functions, including learning, reasoning and communication. However, there is very little work on endowing machines with the ability to form and reason with concepts. In particular, state-of-the-art large language models (LLMs) work at the level of tokens, not con...
http://arxiv.org/pdf/2311.01866v1
2023-11-03T12:19:22
2023-11-03T12:19:22
{ "title": [ -0.01474762, -0.017700195, -0.0390625, -0.0769043, 0.0042152405, 0.0074005127, -0.05783081, 0.021697998, 0.03439331, 0.031982422, -0.01361084, -0.017791748, 0.019104004, -0.013008118, -0.051727295, 0.020980835, -0.017166138, 0.029495...
http://arxiv.org/abs/2311.01864v1
SortNet: Learning To Rank By a Neural-Based Sorting Algorithm
Leonardo Rigutini, Tiziano Papini, Marco Maggini, Franco Scarselli
cs.LG, cs.AI, cs.CL, cs.IR
The problem of relevance ranking consists of sorting a set of objects with respect to a given criterion. Since users may prefer different relevance criteria, the ranking algorithms should be adaptable to the user needs. Two main approaches exist in literature for the task of learning to rank: 1) a score function, learn...
http://arxiv.org/pdf/2311.01864v1
2023-11-03T12:14:26
2023-11-03T12:14:26
{ "title": [ 0.012069702, 0.005153656, -0.024536133, -0.023391724, 0.033294678, -0.020431519, 0.027435303, -0.031066895, 0.024475098, 0.0309906, -0.007259369, -0.007419586, -0.0064888, -0.019515991, -0.054534912, 0.004711151, -0.008850098, 0.0088...
http://arxiv.org/abs/2311.01862v1
$R^3$-NL2GQL: A Hybrid Models Approach for for Accuracy Enhancing and Hallucinations Mitigation
Yuhang Zhou, He Yu, Siyu Tian, Dan Chen, Liuzhi Zhou, Xinlin Yu, Chuanjun Ji, Sen Liu, Guangnan Ye, Hongfeng Chai
cs.CL, cs.DB
While current NL2SQL tasks constructed using Foundation Models have achieved commendable results, their direct application to Natural Language to Graph Query Language (NL2GQL) tasks poses challenges due to the significant differences between GQL and SQL expressions, as well as the numerous types of GQL. Our extensive e...
http://arxiv.org/pdf/2311.01862v1
2023-11-03T12:11:12
2023-11-03T12:11:12
{ "title": [ -0.027572632, 0.0024738312, -0.033935547, -0.041137695, -0.008659363, 0.0067481995, -0.044677734, -0.024917603, -0.015571594, 0.04852295, 0.006412506, -0.011482239, 0.034942627, -0.04168701, -0.025299072, 0.0095825195, -0.0026893616, ...
http://arxiv.org/abs/2311.01860v1
FAME: Flexible, Scalable Analogy Mappings Engine
Shahar Jacob, Chen Shani, Dafna Shahaf
cs.CL, cs.AI
Analogy is one of the core capacities of human cognition; when faced with new situations, we often transfer prior experience from other domains. Most work on computational analogy relies heavily on complex, manually crafted input. In this work, we relax the input requirements, requiring only names of entities to be map...
http://arxiv.org/pdf/2311.01860v1
2023-11-03T12:08:02
2023-11-03T12:08:02
{ "title": [ -0.0065994263, -0.027435303, -0.014411926, -0.066711426, 0.010353088, 0.016296387, -0.013748169, -0.02784729, 0.031097412, 0.03677368, 0.0020484924, -0.013717651, 0.030166626, -0.03543091, -0.0491333, 0.034301758, -0.0013160706, 0.03...
http://arxiv.org/abs/2311.01825v2
Large Language Models to the Rescue: Reducing the Complexity in Scientific Workflow Development Using ChatGPT
Mario Sänger, Ninon De Mecquenem, Katarzyna Ewa Lewińska, Vasilis Bountris, Fabian Lehmann, Ulf Leser, Thomas Kosch
cs.DC, cs.CL, cs.HC
Scientific workflow systems are increasingly popular for expressing and executing complex data analysis pipelines over large datasets, as they offer reproducibility, dependability, and scalability of analyses by automatic parallelization on large compute clusters. However, implementing workflows is difficult due to the...
http://arxiv.org/pdf/2311.01825v2
2023-11-06T11:43:33
2023-11-03T10:28:53
{ "title": [ -0.054748535, -0.022277832, -0.048095703, -0.044525146, -0.009025574, 0.012275696, -0.060455322, 0.01259613, 0.006839752, 0.038024902, 0.012779236, 0.0008249283, 0.0042800903, -0.028259277, -0.033691406, 0.019454956, -0.018569946, 0....
http://arxiv.org/abs/2311.01820v1
Minimalist Grammar: Construction without Overgeneration
Isidor Konrad Maier, Johannes Kuhn, Jesse Beisegel, Markus Huber-Liebl, Matthias Wolff
cs.CL
In this paper we give instructions on how to write a minimalist grammar (MG). In order to present the instructions as an algorithm, we use a variant of context free grammars (CFG) as an input format. We can exclude overgeneration, if the CFG has no recursion, i.e. no non-terminal can (indirectly) derive to a right-hand...
http://arxiv.org/pdf/2311.01820v1
2023-11-03T10:02:00
2023-11-03T10:02:00
{ "title": [ -0.016983032, -0.009185791, -0.04083252, -0.076049805, 0.011116028, 0.016662598, -0.039154053, 0.00818634, -0.020629883, 0.027664185, 0.006916046, 0.011009216, -0.04522705, -0.036376953, -0.045074463, 0.000007927418, -0.010276794, 0....
http://arxiv.org/abs/2311.01817v1
Mitigating Framing Bias with Polarity Minimization Loss
Yejin Bang, Nayeon Lee, Pascale Fung
cs.CL
Framing bias plays a significant role in exacerbating political polarization by distorting the perception of actual events. Media outlets with divergent political stances often use polarized language in their reporting of the same event. We propose a new loss function that encourages the model to minimize the polarity ...
http://arxiv.org/pdf/2311.01817v1
2023-11-03T09:50:23
2023-11-03T09:50:23
{ "title": [ 0.012504578, 0.017364502, -0.019119263, -0.02949524, 0.0034275055, -0.008148193, -0.021514893, -0.06695557, -0.018753052, 0.048797607, -0.0076942444, 0.0033607483, 0.02772522, -0.016494751, -0.020446777, 0.005706787, -0.0059928894, 0...
http://arxiv.org/abs/2311.04924v1
Tuning-less Object Naming with a Foundation Model
Andrej Lucny, Pavel Petrovic
cs.CL, cs.AI
We implement a real-time object naming system that enables learning a set of named entities never seen. Our approach employs an existing foundation model that we consider ready to see anything before starting. It turns seen images into relatively small feature vectors that we associate with index to a gradually built v...
http://arxiv.org/pdf/2311.04924v1
2023-11-03T09:11:49
2023-11-03T09:11:49
{ "title": [ -0.022628784, -0.011657715, -0.061798096, -0.06512451, -0.025970459, -0.014198303, -0.06488037, 0.0035514832, -0.00064611435, 0.03945923, -0.0050239563, 0.0041618347, 0.02267456, -0.027236938, 0.006996155, 0.01360321, 0.0031852722, 0...
http://arxiv.org/abs/2311.01792v1
AFPQ: Asymmetric Floating Point Quantization for LLMs
Yijia Zhang, Sicheng Zhang, Shijie Cao, Dayou Du, Jianyu Wei, Ting Cao, Ningyi Xu
cs.CL, cs.AI
Large language models (LLMs) show great performance in various tasks, but face deployment challenges from limited memory capacity and bandwidth. Low-bit weight quantization can save memory and accelerate inference. Although floating-point (FP) formats show good performance in LLM quantization, they tend to perform poor...
http://arxiv.org/pdf/2311.01792v1
2023-11-03T09:07:09
2023-11-03T09:07:09
{ "title": [ -0.06829834, 0.008171082, -0.046661377, -0.014854431, -0.0051460266, -0.0035171509, -0.025756836, -0.009643555, -0.021728516, 0.056518555, -0.008773804, 0.009773254, 0.009536743, -0.021697998, -0.037506104, 0.012710571, 0.007858276, ...
http://arxiv.org/abs/2311.04923v1
Is one brick enough to break the wall of spoken dialogue state tracking?
Lucas Druart, Valentin Vielzeuf, Yannick Estève
cs.CL, cs.AI, eess.SP
In Task-Oriented Dialogue (TOD) systems, correctly updating the system's understanding of the user's needs (a.k.a dialogue state tracking) is key to a smooth interaction. Traditionally, TOD systems perform this update in three steps: transcription of the user's utterance, semantic extraction of the key concepts, and co...
http://arxiv.org/pdf/2311.04923v1
2023-11-03T08:59:51
2023-11-03T08:59:51
{ "title": [ 0.029174805, -0.03363037, -0.04269409, -0.052368164, -0.04486084, -0.031707764, -0.033569336, 0.02142334, -0.059661865, 0.0017023087, 0.0013685226, 0.01473999, -0.0076828003, -0.05545044, 0.047851562, -0.034820557, -0.00774765, 0.012...
http://arxiv.org/abs/2311.01786v1
TCM-GPT: Efficient Pre-training of Large Language Models for Domain Adaptation in Traditional Chinese Medicine
Guoxing Yang, Jianyu Shi, Zan Wang, Xiaohong Liu, Guangyu Wang
cs.CL, cs.AI
Pre-training and fine-tuning have emerged as a promising paradigm across various natural language processing (NLP) tasks. The effectiveness of pretrained large language models (LLM) has witnessed further enhancement, holding potential for applications in the field of medicine, particularly in the context of Traditional...
http://arxiv.org/pdf/2311.01786v1
2023-11-03T08:54:50
2023-11-03T08:54:50
{ "title": [ -0.016967773, -0.011123657, -0.037384033, -0.05831909, -0.005832672, -0.0010118484, -0.058807373, 0.00011378527, 0.004016876, 0.021591187, 0.020385742, -0.016220093, 0.019897461, -0.05596924, -0.01927185, 0.01335144, -0.0013456345, 0...
http://arxiv.org/abs/2311.04922v1
Are cascade dialogue state tracking models speaking out of turn in spoken dialogues?
Lucas Druart, Léo Jacqmin, Benoît Favre, Lina Maria Rojas-Barahona, Valentin Vielzeuf
cs.CL, cs.AI, eess.SP
In Task-Oriented Dialogue (TOD) systems, correctly updating the system's understanding of the user's needs is key to a smooth interaction. Traditionally TOD systems are composed of several modules that interact with one another. While each of these components is the focus of active research communities, their behavior ...
http://arxiv.org/pdf/2311.04922v1
2023-11-03T08:45:22
2023-11-03T08:45:22
{ "title": [ 0.010917664, 0.0006995201, -0.019500732, -0.042236328, -0.025131226, -0.018829346, -0.034210205, 0.0022468567, 0.009864807, 0.0033435822, -0.032196045, -0.019515991, 0.00029611588, -0.044067383, -0.0084991455, -0.0501709, -0.02532959, ...
http://arxiv.org/abs/2311.01775v1
UP4LS: User Profile Constructed by Multiple Attributes for Enhancing Linguistic Steganalysis
Yihao Wang, Ruiqi Song, Ru Zhang, Jianyi Liu
cs.CL
Linguistic steganalysis (LS) tasks aim to effectively detect stegos generated by linguistic steganography. Existing LS methods overlook the distinctive user characteristics, leading to weak performance in social networks. The limited occurrence of stegos further complicates detection. In this paper, we propose the UP4L...
http://arxiv.org/pdf/2311.01775v1
2023-11-03T08:20:48
2023-11-03T08:20:48
{ "title": [ -0.0096206665, 0.00806427, -0.03050232, -0.03933716, 0.0030879974, -0.0012645721, -0.022857666, -0.023803711, -0.032287598, 0.042175293, 0.05267334, 0.010887146, 0.03173828, -0.049194336, -0.022460938, 0.037078857, -0.032714844, 0.03...
http://arxiv.org/abs/2311.01767v2
PPTC Benchmark: Evaluating Large Language Models for PowerPoint Task Completion
Yiduo Guo, Zekai Zhang, Yaobo Liang, Dongyan Zhao, Nan Duan
cs.CL
Recent evaluations of Large Language Models (LLMs) have centered around testing their zero-shot/few-shot capabilities for basic natural language tasks and their ability to translate instructions into tool APIs. However, the evaluation of LLMs utilizing complex tools to finish multi-turn, multi-modal instructions in a c...
http://arxiv.org/pdf/2311.01767v2
2023-11-07T10:13:34
2023-11-03T08:06:35
{ "title": [ -0.013069153, -0.0011672974, -0.040802002, -0.06903076, 0.0052948, -0.0012426376, -0.05621338, -0.0033035278, -0.0025672913, 0.048095703, -0.0035877228, 0.010368347, 0.023757935, -0.004863739, -0.046020508, 0.0022335052, -0.011413574, ...
http://arxiv.org/abs/2311.01766v2
Support or Refute: Analyzing the Stance of Evidence to Detect Out-of-Context Mis- and Disinformation
Xin Yuan, Jie Guo, Weidong Qiu, Zheng Huang, Shujun Li
cs.CL, cs.CV
Mis- and disinformation online have become a major societal problem as major sources of online harms of different kinds. One common form of mis- and disinformation is out-of-context (OOC) information, where different pieces of information are falsely associated, e.g., a real image combined with a false textual caption ...
http://arxiv.org/pdf/2311.01766v2
2023-11-06T16:33:10
2023-11-03T08:05:54
{ "title": [ 0.012931824, -0.0012750626, -0.05038452, -0.039367676, 0.00983429, -0.041412354, 0.0018234253, -0.016784668, -0.03250122, 0.01713562, -0.048583984, 0.012634277, -0.0062065125, -0.029571533, -0.0063667297, -0.044525146, -0.010429382, ...
http://arxiv.org/abs/2311.05640v1
FinGPT: Large Generative Models for a Small Language
Risto Luukkonen, Ville Komulainen, Jouni Luoma, Anni Eskelinen, Jenna Kanerva, Hanna-Mari Kupari, Filip Ginter, Veronika Laippala, Niklas Muennighoff, Aleksandra Piktus, Thomas Wang, Nouamane Tazi, Teven Le Scao, Thomas Wolf, Osma Suominen, Samuli Sairanen, Mikko Merioksa, Jyrki Heinonen, Aija Vahtola, Samuel Antao, Sa...
cs.CL
Large language models (LLMs) excel in many tasks in NLP and beyond, but most open models have very limited coverage of smaller languages and LLM work tends to focus on languages where nearly unlimited data is available for pretraining. In this work, we study the challenges of creating LLMs for Finnish, a language spoke...
http://arxiv.org/pdf/2311.05640v1
2023-11-03T08:05:04
2023-11-03T08:05:04
{ "title": [ -0.016647339, -0.025985718, -0.03970337, -0.07952881, 0.014984131, 0.011421204, -0.058532715, 0.014167786, 0.008811951, 0.019348145, -0.006225586, 0.01486969, 0.0039100647, -0.025634766, -0.029922485, 0.020492554, -0.018005371, 0.043...
http://arxiv.org/abs/2311.01757v1
Indo LEGO-ABSA: A Multitask Generative Aspect Based Sentiment Analysis for Indonesian Language
Randy Zakya Suchrady, Ayu Purwarianti
cs.CL, cs.AI, I.2.7
Aspect-based sentiment analysis is a method in natural language processing aimed at identifying and understanding sentiments related to specific aspects of an entity. Aspects are words or phrases that represent an aspect or attribute of a particular entity. Previous research has utilized generative pre-trained language...
http://arxiv.org/pdf/2311.01757v1
2023-11-03T07:28:12
2023-11-03T07:28:12
{ "title": [ 0.014411926, -0.035369873, -0.020767212, -0.06427002, -0.0181427, 0.0037670135, -0.052825928, 0.02154541, 0.0345459, 0.04876709, 0.0071983337, -0.0013742447, -0.0073547363, -0.01486969, -0.08807373, 0.011543274, -0.021133423, 0.03640...
http://arxiv.org/abs/2311.01751v1
EmojiLM: Modeling the New Emoji Language
Letian Peng, Zilong Wang, Hang Liu, Zihan Wang, Jingbo Shang
cs.CL
With the rapid development of the internet, online social media welcomes people with different backgrounds through its diverse content. The increasing usage of emoji becomes a noticeable trend thanks to emoji's rich information beyond cultural or linguistic borders. However, the current study on emojis is limited to si...
http://arxiv.org/pdf/2311.01751v1
2023-11-03T07:06:51
2023-11-03T07:06:51
{ "title": [ -0.016784668, -0.0071258545, -0.02154541, -0.037719727, 0.004180908, 0.0079422, -0.02670288, -0.0115356445, 0.0010385513, 0.05041504, -0.00067710876, -0.0041046143, 0.032928467, -0.020507812, -0.043914795, 0.016357422, -0.0074005127, ...
http://arxiv.org/abs/2311.01740v1
SAC$^3$: Reliable Hallucination Detection in Black-Box Language Models via Semantic-aware Cross-check Consistency
Jiaxin Zhang, Zhuohang Li, Kamalika Das, Bradley A. Malin, Sricharan Kumar
cs.CL
Hallucination detection is a critical step toward understanding the trustworthiness of modern language models (LMs). To achieve this goal, we re-examine existing detection approaches based on the self-consistency of LMs and uncover two types of hallucinations resulting from 1) question-level and 2) model-level, which c...
http://arxiv.org/pdf/2311.01740v1
2023-11-03T06:32:43
2023-11-03T06:32:43
{ "title": [ 0.006164551, -0.0017080307, -0.020767212, -0.058502197, -0.0129776, -0.004878998, -0.051208496, -0.025344849, -0.0001206398, 0.043823242, -0.017211914, 0.019454956, 0.03866577, -0.012390137, -0.047698975, 0.0124435425, -0.02999878, 0...
http://arxiv.org/abs/2311.01732v2
Proto-lm: A Prototypical Network-Based Framework for Built-in Interpretability in Large Language Models
Sean Xie, Soroush Vosoughi, Saeed Hassanpour
cs.CL
Large Language Models (LLMs) have significantly advanced the field of Natural Language Processing (NLP), but their lack of interpretability has been a major concern. Current methods for interpreting LLMs are post hoc, applied after inference time, and have limitations such as their focus on low-level features and lack ...
http://arxiv.org/pdf/2311.01732v2
2023-11-12T04:28:43
2023-11-03T05:55:32
{ "title": [ -0.043121338, -0.0032310486, -0.022842407, -0.06982422, -0.019119263, 0.031036377, -0.092163086, -0.0023899078, 0.0074272156, 0.010093689, -0.00868988, -0.008216858, 0.0058631897, -0.026062012, -0.036132812, 0.00869751, -0.010002136, ...
http://arxiv.org/abs/2311.01713v1
An Empirical Study of Benchmarking Chinese Aspect Sentiment Quad Prediction
Junxian Zhou, Haiqin Yang, Ye Junpeng, Yuxuan He, Hao Mou
cs.CL, cs.AI
Aspect sentiment quad prediction (ASQP) is a critical subtask of aspect-level sentiment analysis. Current ASQP datasets are characterized by their small size and low quadruple density, which hinders technical development. To expand capacity, we construct two large Chinese ASQP datasets crawled from multiple online plat...
http://arxiv.org/pdf/2311.01713v1
2023-11-03T05:00:44
2023-11-03T05:00:44
{ "title": [ -0.008377075, -0.004688263, -0.025390625, -0.060516357, 0.0011634827, -0.001414299, -0.027618408, 0.007095337, 0.01977539, 0.06768799, 0.0065612793, 0.00014328957, 0.0034065247, -0.016906738, -0.09033203, 0.018005371, -0.0029945374, ...
http://arxiv.org/abs/2311.01712v1
A New Korean Text Classification Benchmark for Recognizing the Political Intents in Online Newspapers
Beomjune Kim, Eunsun Lee, Dongbin Na
cs.CL
Many users reading online articles in various magazines may suffer considerable difficulty in distinguishing the implicit intents in texts. In this work, we focus on automatically recognizing the political intents of a given online newspaper by understanding the context of the text. To solve this task, we present a nov...
http://arxiv.org/pdf/2311.01712v1
2023-11-03T04:59:55
2023-11-03T04:59:55
{ "title": [ 0.010040283, 0.0077552795, -0.014945984, -0.034210205, 0.0055274963, -0.020889282, -0.014907837, -0.0335083, -0.011474609, 0.023132324, -0.010025024, 0.024780273, -0.009063721, -0.009170532, -0.04800415, -0.015106201, -0.01651001, 0....
http://arxiv.org/abs/2311.01689v1
Data-Free Distillation of Language Model by Text-to-Text Transfer
Zheyuan Bai, Xinduo Liu, Hailin Hu, Tianyu Guo, Qinghua Zhang, Yunhe Wang
cs.CL, cs.AI
Data-Free Knowledge Distillation (DFKD) plays a vital role in compressing the model when original training data is unavailable. Previous works for DFKD in NLP mainly focus on distilling encoder-only structures like BERT on classification tasks, which overlook the notable progress of generative language modeling. In thi...
http://arxiv.org/pdf/2311.01689v1
2023-11-03T03:31:47
2023-11-03T03:31:47
{ "title": [ -0.009742737, -0.012298584, -0.03640747, -0.07739258, -0.0030784607, 0.006378174, -0.05859375, 0.0027046204, 0.02835083, 0.041870117, -0.004535675, -0.011207581, 0.02532959, -0.020233154, -0.057739258, 0.01689148, -0.0022640228, 0.02...
http://arxiv.org/abs/2311.01684v1
CASE: Commonsense-Augmented Score with an Expanded Answer Space
Wenkai Chen, Sahithya Ravi, Vered Shwartz
cs.CL
LLMs have demonstrated impressive zero-shot performance on NLP tasks thanks to the knowledge they acquired in their training. In multiple-choice QA tasks, the LM probabilities are used as an imperfect measure of the plausibility of each answer choice. One of the major limitations of the basic score is that it treats al...
http://arxiv.org/pdf/2311.01684v1
2023-11-03T03:15:26
2023-11-03T03:15:26
{ "title": [ 0.05355835, -0.022735596, -0.023651123, -0.0054969788, -0.029220581, -0.032287598, 0.01576233, -0.021011353, 0.023086548, 0.0054740906, -0.052215576, -0.014099121, -0.011268616, -0.031951904, 0.0056533813, -0.042663574, 0.0074806213, ...
http://arxiv.org/abs/2311.01677v1
DialogBench: Evaluating LLMs as Human-like Dialogue Systems
Jiao Ou, Junda Lu, Che Liu, Yihong Tang, Fuzheng Zhang, Di Zhang, Zhongyuan Wang, Kun Gai
cs.CL, cs.AI
Large language models (LLMs) have achieved remarkable breakthroughs in new dialogue capabilities, refreshing human's impressions on dialogue systems. The long-standing goal of dialogue systems is to be human-like enough to establish long-term connections with users by satisfying the need for communication, affection an...
http://arxiv.org/pdf/2311.01677v1
2023-11-03T02:59:56
2023-11-03T02:59:56
{ "title": [ -0.06500244, -0.009803772, -0.044799805, -0.022018433, -0.022720337, -0.0066490173, -0.054840088, -0.0066566467, -0.0023002625, 0.03363037, -0.042114258, 0.02192688, -0.024337769, -0.0015411377, -0.053131104, -0.015808105, -0.024871826, ...
http://arxiv.org/abs/2311.01666v1
Plot Retrieval as an Assessment of Abstract Semantic Association
Shicheng Xu, Liang Pang, Jiangnan Li, Mo Yu, Fandong Meng, Huawei Shen, Xueqi Cheng, Jie Zhou
cs.IR, cs.CL
Retrieving relevant plots from the book for a query is a critical task, which can improve the reading experience and efficiency of readers. Readers usually only give an abstract and vague description as the query based on their own understanding, summaries, or speculations of the plot, which requires the retrieval mode...
http://arxiv.org/pdf/2311.01666v1
2023-11-03T02:02:43
2023-11-03T02:02:43
{ "title": [ 0.013572693, 0.017822266, -0.02734375, -0.03201294, 0.021835327, -0.0178833, 0.013038635, -0.028198242, 0.022369385, 0.038970947, -0.00046634674, -0.0016441345, -0.009414673, -0.023880005, -0.062347412, 0.020736694, -0.0037269592, 0....
http://arxiv.org/abs/2311.01650v1
MARRS: Multimodal Reference Resolution System
Halim Cagri Ates, Shruti Bhargava, Site Li, Jiarui Lu, Siddhardha Maddula, Joel Ruben Antony Moniz, Anil Kumar Nalamalapu, Roman Hoang Nguyen, Melis Ozyildirim, Alkesh Patel, Dhivya Piraviperumal, Vincent Renkens, Ankit Samal, Thy Tran, Bo-Hsiang Tseng, Hong Yu, Yuan Zhang, Rong Zou
cs.CL, cs.AI, cs.LG
Successfully handling context is essential for any dialog understanding task. This context maybe be conversational (relying on previous user queries or system responses), visual (relying on what the user sees, for example, on their screen), or background (based on signals such as a ringing alarm or playing music). In t...
http://arxiv.org/pdf/2311.01650v1
2023-11-03T00:48:42
2023-11-03T00:48:42
{ "title": [ -0.035369873, -0.003944397, -0.06359863, -0.07513428, 0.012626648, -0.0026283264, -0.06317139, -0.016036987, 0.002363205, 0.04827881, 0.016860962, -0.021774292, 0.046661377, -0.051727295, -0.03213501, 0.024261475, -0.002729416, 0.022...
http://arxiv.org/abs/2311.04921v1
Successor Features for Efficient Multisubject Controlled Text Generation
Meng Cao, Mehdi Fatemi, Jackie Chi Kit Cheung, Samira Shabanian
cs.CL, cs.AI
While large language models (LLMs) have achieved impressive performance in generating fluent and realistic text, controlling the generated text so that it exhibits properties such as safety, factuality, and non-toxicity remains challenging. % such as DExperts, GeDi, and rectification Existing decoding-based methods are...
http://arxiv.org/pdf/2311.04921v1
2023-11-03T00:17:08
2023-11-03T00:17:08
{ "title": [ -0.020080566, -0.016098022, -0.036621094, -0.060028076, -0.016189575, 0.005584717, -0.053955078, 0.0019569397, 0.00026726723, 0.03186035, -0.020095825, 0.018005371, 0.007698059, -0.02166748, -0.014976501, 0.012466431, -0.014404297, 0...
http://arxiv.org/abs/2311.01620v1
ACQUIRED: A Dataset for Answering Counterfactual Questions In Real-Life Videos
Te-Lin Wu, Zi-Yi Dou, Qingyuan Hu, Yu Hou, Nischal Reddy Chandra, Marjorie Freedman, Ralph M. Weischedel, Nanyun Peng
cs.CV, cs.CL
Multimodal counterfactual reasoning is a vital yet challenging ability for AI systems. It involves predicting the outcomes of hypothetical circumstances based on vision and language inputs, which enables AI models to learn from failures and explore hypothetical scenarios. Despite its importance, there are only a few da...
http://arxiv.org/pdf/2311.01620v1
2023-11-02T22:17:03
2023-11-02T22:17:03
{ "title": [ 0.014595032, 0.000044584274, -0.043792725, -0.028427124, 0.0025539398, -0.040863037, -0.0012540817, -0.02519226, 0.0011320114, 0.005241394, -0.04736328, 0.004108429, -0.031173706, -0.016540527, -0.036132812, -0.030456543, 0.0090408325, ...
http://arxiv.org/abs/2311.01615v1
FLAP: Fast Language-Audio Pre-training
Ching-Feng Yeh, Po-Yao Huang, Vasu Sharma, Shang-Wen Li, Gargi Gosh
cs.SD, cs.CL, eess.AS
We propose Fast Language-Audio Pre-training (FLAP), a self-supervised approach that efficiently and effectively learns aligned audio and language representations through masking, contrastive learning and reconstruction. For efficiency, FLAP randomly drops audio spectrogram tokens, focusing solely on the remaining ones ...
http://arxiv.org/pdf/2311.01615v1
2023-11-02T21:58:50
2023-11-02T21:58:50
{ "title": [ -0.012802124, -0.025436401, -0.042419434, -0.04736328, -0.05114746, -0.0008587837, -0.07312012, 0.009719849, -0.046661377, 0.022766113, -0.017410278, 0.010299683, 0.034240723, -0.05847168, 0.0597229, -0.010345459, 0.0007119179, 0.025...
http://arxiv.org/abs/2311.01580v1
MetaReVision: Meta-Learning with Retrieval for Visually Grounded Compositional Concept Acquisition
Guangyue Xu, Parisa Kordjamshidi, Joyce Chai
cs.CL, cs.AI
Humans have the ability to learn novel compositional concepts by recalling and generalizing primitive concepts acquired from past experiences. Inspired by this observation, in this paper, we propose MetaReVision, a retrieval-enhanced meta-learning model to address the visually grounded compositional concept learning pr...
http://arxiv.org/pdf/2311.01580v1
2023-11-02T20:19:58
2023-11-02T20:19:58
{ "title": [ -0.005783081, -0.018997192, -0.031188965, -0.049072266, -0.013290405, -0.008216858, -0.044830322, -0.018554688, 0.038360596, 0.03945923, -0.026031494, 0.016555786, -0.0046920776, -0.007030487, -0.07196045, -0.0014476776, -0.016036987, ...
http://arxiv.org/abs/2311.01571v1
Preserving the knowledge of long clinical texts using aggregated ensembles of large language models
Mohammad Junayed Hasan, Suhra Noor, Mohammad Ashrafuzzaman Khan
cs.CL, cs.AI, I.2.7
Clinical texts, such as admission notes, discharge summaries, and progress notes, contain rich and valuable information that can be used for various clinical outcome prediction tasks. However, applying large language models, such as BERT-based models, to clinical texts poses two major challenges: the limitation of inpu...
http://arxiv.org/pdf/2311.01571v1
2023-11-02T19:50:02
2023-11-02T19:50:02
{ "title": [ -0.03741455, -0.013534546, -0.055725098, -0.079589844, -0.019241333, 0.011909485, -0.07800293, 0.021194458, 0.018356323, 0.046813965, -0.016708374, -0.004131317, 0.0149002075, -0.009498596, -0.059661865, 0.0014400482, -0.00012362003, ...
http://arxiv.org/abs/2311.01555v1
Instruction Distillation Makes Large Language Models Efficient Zero-shot Rankers
Weiwei Sun, Zheng Chen, Xinyu Ma, Lingyong Yan, Shuaiqiang Wang, Pengjie Ren, Zhumin Chen, Dawei Yin, Zhaochun Ren
cs.IR, cs.CL
Recent studies have demonstrated the great potential of Large Language Models (LLMs) serving as zero-shot relevance rankers. The typical approach involves making comparisons between pairs or lists of documents. Although effective, these listwise and pairwise methods are not efficient and also heavily rely on intricate ...
http://arxiv.org/pdf/2311.01555v1
2023-11-02T19:16:21
2023-11-02T19:16:21
{ "title": [ 0.013389587, -0.011497498, -0.025146484, -0.055480957, -0.022384644, -0.023773193, -0.052703857, 0.001996994, 0.021942139, 0.030014038, -0.032287598, -0.0061798096, 0.007003784, -0.02861023, -0.02949524, -0.0067253113, 0.0033416748, ...
http://arxiv.org/abs/2311.01544v2
Divergent Token Metrics: Measuring degradation to prune away LLM components -- and optimize quantization
Björn Deiseroth, Max Meuer, Nikolas Gritsch, Constantin Eichenberg, Patrick Schramowski, Matthias Aßenmacher, Kristian Kersting
cs.CL, cs.LG
Large Language Models (LLMs) have reshaped natural language processing with their impressive capabilities. Their ever-increasing size, however, raised concerns about their effective deployment and the need for LLM compressions. This study introduces the Divergent Token metrics (DTMs), a novel approach for assessing com...
http://arxiv.org/pdf/2311.01544v2
2023-11-13T15:33:35
2023-11-02T18:55:53
{ "title": [ -0.05807495, -0.0056915283, -0.062347412, -0.029342651, -0.036102295, -0.009132385, -0.06451416, 0.010536194, -0.018951416, 0.016540527, -0.03173828, 0.0035800934, -0.0045814514, -0.032073975, -0.0042877197, -0.039031982, -0.001490593, ...
http://arxiv.org/abs/2311.01460v1
Implicit Chain of Thought Reasoning via Knowledge Distillation
Yuntian Deng, Kiran Prasad, Roland Fernandez, Paul Smolensky, Vishrav Chaudhary, Stuart Shieber
cs.CL, cs.AI, cs.LG
To augment language models with the ability to reason, researchers usually prompt or finetune them to produce chain of thought reasoning steps before producing the final answer. However, although people use natural language to reason effectively, it may be that LMs could reason more effectively with some intermediate c...
http://arxiv.org/pdf/2311.01460v1
2023-11-02T17:59:49
2023-11-02T17:59:49
{ "title": [ 0.00012767315, -0.003917694, -0.046936035, -0.070251465, 0.013763428, -0.0113220215, -0.036743164, -0.0005354881, 0.047729492, 0.0002462864, -0.058441162, -0.03503418, -0.02999878, -0.021499634, -0.038482666, -0.034088135, 0.002418518, ...
http://arxiv.org/abs/2311.01449v1
TopicGPT: A Prompt-based Topic Modeling Framework
Chau Minh Pham, Alexander Hoyle, Simeng Sun, Mohit Iyyer
cs.CL
Topic modeling is a well-established technique for exploring text corpora. Conventional topic models (e.g., LDA) represent topics as bags of words that often require "reading the tea leaves" to interpret; additionally, they offer users minimal semantic control over topics. To tackle these issues, we introduce TopicGPT,...
http://arxiv.org/pdf/2311.01449v1
2023-11-02T17:57:10
2023-11-02T17:57:10
{ "title": [ -0.012565613, 0.0023765564, -0.041625977, -0.045806885, -0.00019574165, 0.006351471, -0.045562744, -0.043945312, 0.03050232, 0.01399231, -0.018722534, 0.031829834, -0.0124053955, -0.022506714, -0.024841309, 0.0206604, -0.0046463013, ...
http://arxiv.org/abs/2311.01398v1
Server-side Rescoring of Spoken Entity-centric Knowledge Queries for Virtual Assistants
Youyuan Zhang, Sashank Gondala, Thiago Fraga-Silva, Christophe Van Gysel
cs.CL, cs.SD, eess.AS
On-device Virtual Assistants (VAs) powered by Automatic Speech Recognition (ASR) require effective knowledge integration for the challenging entity-rich query recognition. In this paper, we conduct an empirical study of modeling strategies for server-side rescoring of spoken information domain queries using various cat...
http://arxiv.org/pdf/2311.01398v1
2023-11-02T17:07:23
2023-11-02T17:07:23
{ "title": [ -0.021972656, -0.031402588, -0.04849243, -0.07940674, 0.00072574615, 0.0028686523, -0.029022217, 0.055389404, 0.028701782, 0.08319092, -0.029800415, -0.01637268, 0.061523438, 0.009353638, -0.09362793, 0.020446777, 0.0036621094, 0.017...
http://arxiv.org/abs/2311.01386v1
Can Language Models Be Tricked by Language Illusions? Easier with Syntax, Harder with Semantics
Yuhan Zhang, Edward Gibson, Forrest Davis
cs.CL
Language models (LMs) have been argued to overlap substantially with human beings in grammaticality judgment tasks. But when humans systematically make errors in language processing, should we expect LMs to behave like cognitive models of language and mimic human behavior? We answer this question by investigating LMs' ...
http://arxiv.org/pdf/2311.01386v1
2023-11-02T16:44:24
2023-11-02T16:44:24
{ "title": [ -0.037628174, 0.0030918121, -0.035827637, -0.0335083, -0.01335144, -0.016906738, -0.02029419, -0.015838623, -0.0066184998, 0.020202637, -0.028015137, 0.009529114, 0.004337311, -0.021911621, -0.028762817, -0.006412506, -0.014083862, 0...
http://arxiv.org/abs/2311.04228v1
Graph Neural Networks for Topological Feature Extraction in ECG Classification
Kamyar Zeinalipour, Marco Gori
eess.SP, cs.CL, cs.LG
The electrocardiogram (ECG) is a dependable instrument for assessing the function of the cardiovascular system. There has recently been much emphasis on precisely classifying ECGs. While ECG situations have numerous similarities, little attention has been paid to categorizing ECGs using graph neural networks. In this s...
http://arxiv.org/pdf/2311.04228v1
2023-11-02T16:14:34
2023-11-02T16:14:34
{ "title": [ 0.007358551, 0.0022563934, -0.0012731552, -0.01725769, 0.009735107, -0.0053482056, 0.005355835, -0.030258179, -0.020950317, 0.058776855, 0.009712219, -0.010131836, 0.023208618, -0.032928467, -0.049468994, 0.010421753, 0.026031494, 0....
http://arxiv.org/abs/2311.01361v1
GPT-4V(ision) as a Generalist Evaluator for Vision-Language Tasks
Xinlu Zhang, Yujie Lu, Weizhi Wang, An Yan, Jun Yan, Lianke Qin, Heng Wang, Xifeng Yan, William Yang Wang, Linda Ruth Petzold
cs.CV, cs.CL
Automatically evaluating vision-language tasks is challenging, especially when it comes to reflecting human judgments due to limitations in accounting for fine-grained details. Although GPT-4V has shown promising results in various multi-modal tasks, leveraging GPT-4V as a generalist evaluator for these tasks has not y...
http://arxiv.org/pdf/2311.01361v1
2023-11-02T16:11:09
2023-11-02T16:11:09
{ "title": [ -0.0014886856, -0.0030994415, -0.02279663, -0.037719727, 0.0036811829, -0.012939453, -0.01448822, -0.016113281, 0.0046081543, 0.017181396, -0.01600647, -0.009048462, -0.007797241, -0.027511597, -0.04547119, -0.0047683716, -0.006706238, ...
http://arxiv.org/abs/2311.01326v1
Better Together: Enhancing Generative Knowledge Graph Completion with Language Models and Neighborhood Information
Alla Chepurova, Aydar Bulatov, Yuri Kuratov, Mikhail Burtsev
cs.CL, cs.AI
Real-world Knowledge Graphs (KGs) often suffer from incompleteness, which limits their potential performance. Knowledge Graph Completion (KGC) techniques aim to address this issue. However, traditional KGC methods are computationally intensive and impractical for large-scale KGs, necessitating the learning of dense nod...
http://arxiv.org/pdf/2311.01326v1
2023-11-02T15:38:39
2023-11-02T15:38:39
{ "title": [ -0.029556274, -0.024551392, -0.056488037, -0.10180664, -0.013366699, 0.021774292, -0.08898926, 0.029769897, 0.0067443848, 0.021911621, -0.040100098, 0.034729004, -0.0036067963, -0.0012378693, -0.021377563, 0.004550934, -0.0041275024, ...
http://arxiv.org/abs/2311.01487v1
What Makes for Good Visual Instructions? Synthesizing Complex Visual Reasoning Instructions for Visual Instruction Tuning
Yifan Du, Hangyu Guo, Kun Zhou, Wayne Xin Zhao, Jinpeng Wang, Chuyuan Wang, Mingchen Cai, Ruihua Song, Ji-Rong Wen
cs.CV, cs.CL
Visual instruction tuning is an essential approach to improving the zero-shot generalization capability of Multi-modal Large Language Models (MLLMs). A surge of visual instruction datasets with various focuses and characteristics have been proposed recently, enabling MLLMs to achieve surprising results on evaluation be...
http://arxiv.org/pdf/2311.01487v1
2023-11-02T15:36:12
2023-11-02T15:36:12
{ "title": [ -0.03074646, -0.019500732, -0.01134491, -0.04119873, -0.03439331, -0.0042152405, -0.047332764, -0.010848999, -0.00025606155, 0.032348633, -0.0473938, 0.0149002075, -0.0026054382, 0.0035152435, -0.036376953, -0.015640259, 0.018829346, ...
http://arxiv.org/abs/2311.01307v1
The Effect of Scaling, Retrieval Augmentation and Form on the Factual Consistency of Language Models
Lovisa Hagström, Denitsa Saynova, Tobias Norlund, Moa Johansson, Richard Johansson
cs.CL
Large Language Models (LLMs) make natural interfaces to factual knowledge, but their usefulness is limited by their tendency to deliver inconsistent answers to semantically equivalent questions. For example, a model might predict both "Anne Redpath passed away in Edinburgh." and "Anne Redpath's life ended in London." I...
http://arxiv.org/pdf/2311.01307v1
2023-11-02T15:20:11
2023-11-02T15:20:11
{ "title": [ -0.01335907, -0.0019721985, -0.05117798, -0.05105591, 0.0010623932, -0.017501831, -0.044433594, -0.006465912, 0.0065956116, 0.047332764, -0.011917114, -0.009788513, 0.024978638, -0.016616821, -0.06304932, 0.014930725, -0.014335632, 0...
http://arxiv.org/abs/2311.01305v3
AWEQ: Post-Training Quantization with Activation-Weight Equalization for Large Language Models
Baisong Li, Xingwang Wang, Haixiao Xu
cs.LG, cs.AI, cs.CL
Large language models(LLMs) exhibit excellent performance across a variety of tasks, but they come with significant computational and storage costs. Quantizing these models is an effective way to alleviate this issue. However, existing methods struggle to strike a balance between model accuracy and hardware efficiency....
http://arxiv.org/pdf/2311.01305v3
2023-11-12T07:54:09
2023-11-02T15:18:22
{ "title": [ -0.029785156, 0.0053634644, -0.050994873, -0.051208496, 0.019546509, -0.004508972, -0.05319214, -0.011398315, 0.005207062, 0.05895996, 0.010818481, -0.0043525696, 0.02909851, -0.03086853, -0.04699707, 0.030319214, -0.0074043274, 0.04...
http://arxiv.org/abs/2311.01282v3
FlashDecoding++: Faster Large Language Model Inference on GPUs
Ke Hong, Guohao Dai, Jiaming Xu, Qiuli Mao, Xiuhong Li, Jun Liu, Kangdi Chen, Yuhan Dong, Yu Wang
cs.LG, cs.CL
As the Large Language Model (LLM) becomes increasingly important in various domains. However, the following challenges still remain unsolved in accelerating LLM inference: (1) Synchronized partial softmax update. The softmax operation requires a synchronized update operation among each partial softmax result, leading t...
http://arxiv.org/pdf/2311.01282v3
2023-11-10T01:43:51
2023-11-02T14:57:03
{ "title": [ -0.034301758, -0.023605347, -0.03970337, -0.06524658, -0.00006699562, -0.0020828247, -0.059173584, 0.0074539185, -0.0004169941, 0.029342651, -0.014846802, 0.011047363, -0.0048599243, -0.020477295, -0.025482178, 0.013092041, -0.0012102127...
http://arxiv.org/abs/2311.01273v1
Finding Common Ground: Annotating and Predicting Common Ground in Spoken Conversations
Magdalena Markowska, Mohammad Taghizadeh, Adil Soubki, Seyed Abolghasem Mirroshandel, Owen Rambow
cs.CL
When we communicate with other humans, we do not simply generate a sequence of words. Rather, we use our cognitive state (beliefs, desires, intentions) and our model of the audience's cognitive state to create utterances that affect the audience's cognitive state in the intended manner. An important part of cognitive s...
http://arxiv.org/pdf/2311.01273v1
2023-11-02T14:37:28
2023-11-02T14:37:28
{ "title": [ -0.010482788, -0.00019729137, -0.041992188, -0.078430176, 0.012046814, -0.0023212433, -0.048339844, 0.0054626465, 0.0049934387, 0.046783447, -0.0045928955, 0.013801575, -0.010803223, -0.0045776367, -0.07232666, -0.009628296, -0.01638794,...
http://arxiv.org/abs/2311.01270v1
People Make Better Edits: Measuring the Efficacy of LLM-Generated Counterfactually Augmented Data for Harmful Language Detection
Indira Sen, Dennis Assenmacher, Mattia Samory, Isabelle Augenstein, Wil van der Aalst, Claudia Wagne
cs.CL, cs.CY
NLP models are used in a variety of critical social computing tasks, such as detecting sexist, racist, or otherwise hateful content. Therefore, it is imperative that these models are robust to spurious features. Past work has attempted to tackle such spurious features using training data augmentation, including Counter...
http://arxiv.org/pdf/2311.01270v1
2023-11-02T14:31:25
2023-11-02T14:31:25
{ "title": [ 0.008155823, 0.003030777, -0.03717041, -0.019927979, -0.0043792725, -0.025039673, -0.018600464, -0.02671814, -0.010398865, 0.015037537, -0.03503418, 0.027893066, -0.019943237, -0.015304565, -0.035461426, -0.026290894, -0.034210205, 0...
http://arxiv.org/abs/2311.01256v1
An energy-based comparative analysis of common approaches to text classification in the Legal domain
Sinan Gultekin, Achille Globo, Andrea Zugarini, Marco Ernandes, Leonardo Rigutini
cs.CL, cs.AI, cs.LG, cs.PF
Most Machine Learning research evaluates the best solutions in terms of performance. However, in the race for the best performing model, many important aspects are often overlooked when, on the contrary, they should be carefully considered. In fact, sometimes the gaps in performance between different approaches are neg...
http://arxiv.org/pdf/2311.01256v1
2023-11-02T14:16:48
2023-11-02T14:16:48
{ "title": [ -0.014457703, 0.012268066, -0.033813477, -0.03024292, 0.013198853, -0.020523071, -0.01146698, -0.017425537, 0.0014715195, 0.026351929, -0.047210693, 0.04232788, 0.0000192523, -0.009521484, -0.034210205, -0.008522034, -0.02230835, 0.0...
http://arxiv.org/abs/2311.04919v1
The Impact of Preference Agreement in Reinforcement Learning from Human Feedback: A Case Study in Summarization
Sian Gooding, Hassan Mansoor
cs.CL, cs.AI, cs.HC
Reinforcement Learning from Human Feedback (RLHF) can be used to capture complex and nuanced properties of text generation quality. As a result, the task of text summarization has been identified as a good candidate for this process. In this paper, we explore how preference agreement impacts the efficacy of RLHF for su...
http://arxiv.org/pdf/2311.04919v1
2023-11-02T13:21:23
2023-11-02T13:21:23
{ "title": [ 0.011764526, 0.017913818, -0.02961731, -0.04458618, -0.0006914139, -0.031677246, -0.018737793, -0.0236969, -0.024902344, 0.038604736, -0.03277588, 0.014305115, -0.032470703, 0.0099487305, -0.05368042, -0.030426025, -0.012908936, 0.03...
http://arxiv.org/abs/2311.01200v1
A Study of Continual Learning Under Language Shift
Evangelia Gogoulou, Timothée Lesort, Magnus Boman, Joakim Nivre
cs.CL, cs.LG
The recent increase in data and model scale for language model pre-training has led to huge training costs. In scenarios where new data become available over time, updating a model instead of fully retraining it would therefore provide significant gains. In this paper, we study the benefits and downsides of updating a ...
http://arxiv.org/pdf/2311.01200v1
2023-11-02T12:54:50
2023-11-02T12:54:50
{ "title": [ -0.002500534, 0.009635925, -0.044158936, -0.0102005005, -0.009422302, -0.053375244, -0.004257202, 0.0025177002, -0.031707764, 0.03793335, -0.035491943, 0.007167816, -0.00021076202, -0.026306152, 0.0020923615, -0.021255493, 0.023406982, ...
http://arxiv.org/abs/2311.01173v1
CRUSH4SQL: Collective Retrieval Using Schema Hallucination For Text2SQL
Mayank Kothyari, Dhruva Dhingra, Sunita Sarawagi, Soumen Chakrabarti
cs.CL
Existing Text-to-SQL generators require the entire schema to be encoded with the user text. This is expensive or impractical for large databases with tens of thousands of columns. Standard dense retrieval techniques are inadequate for schema subsetting of a large structured database, where the correct semantics of retr...
http://arxiv.org/pdf/2311.01173v1
2023-11-02T12:13:52
2023-11-02T12:13:52
{ "title": [ -0.024353027, -0.011878967, -0.05908203, -0.036254883, 0.0023384094, -0.005844116, 0.011756897, 0.014305115, 0.031311035, 0.053100586, -0.033416748, -0.0051879883, 0.034606934, -0.008308411, -0.05758667, 0.007789612, -0.0043029785, 0...
http://arxiv.org/abs/2311.01166v1
Generative Input: Towards Next-Generation Input Methods Paradigm
Keyu Ding, Yongcan Wang, Zihang Xu, Zhenzhen Jia, Shijin Wang, Cong Liu, Enhong Chen
cs.CL, cs.AI
Since the release of ChatGPT, generative models have achieved tremendous success and become the de facto approach for various NLP tasks. However, its application in the field of input methods remains under-explored. Many neural network approaches have been applied to the construction of Chinese input method engines(IME...
http://arxiv.org/pdf/2311.01166v1
2023-11-02T12:01:29
2023-11-02T12:01:29
{ "title": [ -0.006504059, -0.021987915, -0.031829834, -0.07183838, 0.008544922, 0.003276825, -0.042419434, -0.013511658, 0.00995636, 0.027359009, 0.0064048767, 0.0044403076, -0.023880005, -0.03579712, -0.048858643, 0.015396118, -0.02041626, 0.02...
http://arxiv.org/abs/2311.01161v1
Weakly Supervised Semantic Parsing with Execution-based Spurious Program Filtering
Kang-il Lee, Segwang Kim, Kyomin Jung
cs.CL, cs.AI
The problem of spurious programs is a longstanding challenge when training a semantic parser from weak supervision. To eliminate such programs that have wrong semantics but correct denotation, existing methods focus on exploiting similarities between examples based on domain-specific knowledge. In this paper, we propos...
http://arxiv.org/pdf/2311.01161v1
2023-11-02T11:45:40
2023-11-02T11:45:40
{ "title": [ -0.007045746, 0.009468079, -0.022827148, -0.06726074, 0.0064430237, -0.0028476715, -0.046783447, -0.005332947, -0.029830933, 0.016784668, -0.03225708, 0.015701294, -0.031280518, -0.010658264, -0.020614624, -0.03277588, -0.030090332, ...
http://arxiv.org/abs/2311.01153v1
ACES: Translation Accuracy Challenge Sets at WMT 2023
Chantal Amrhein, Nikita Moghe, Liane Guillou
cs.CL
We benchmark the performance of segmentlevel metrics submitted to WMT 2023 using the ACES Challenge Set (Amrhein et al., 2022). The challenge set consists of 36K examples representing challenges from 68 phenomena and covering 146 language pairs. The phenomena range from simple perturbations at the word/character level ...
http://arxiv.org/pdf/2311.01153v1
2023-11-02T11:29:09
2023-11-02T11:29:09
{ "title": [ -0.038238525, -0.008125305, -0.004558563, -0.012184143, -0.03894043, 0.039031982, -0.05596924, -0.012023926, -0.016052246, 0.05923462, 0.04849243, 0.0008149147, 0.059906006, -0.04434204, 0.01348114, 0.026260376, 0.036193848, -0.01100...
http://arxiv.org/abs/2311.01152v1
Predicting Question-Answering Performance of Large Language Models through Semantic Consistency
Ella Rabinovich, Samuel Ackerman, Orna Raz, Eitan Farchi, Ateret Anaby-Tavor
cs.CL
Semantic consistency of a language model is broadly defined as the model's ability to produce semantically-equivalent outputs, given semantically-equivalent inputs. We address the task of assessing question-answering (QA) semantic consistency of contemporary large language models (LLMs) by manually creating a benchmark...
http://arxiv.org/pdf/2311.01152v1
2023-11-02T11:27:21
2023-11-02T11:27:21
{ "title": [ -0.026428223, -0.0009803772, -0.044281006, -0.059020996, 0.0029144287, -0.0062789917, -0.041412354, 0.011489868, 0.0033512115, 0.04840088, -0.027770996, -0.0044517517, 0.024627686, -0.00198555, -0.06726074, 0.00187397, -0.008583069, ...
http://arxiv.org/abs/2311.01150v1
Revisiting the Knowledge Injection Frameworks
Peng Fu, Yiming Zhang, Haobo Wang, Weikang Qiu, Junbo Zhao
cs.CL, cs.AI
In recent years, large language models (LLMs), such as GPTs, have attained great impact worldwide. However, how to adapt these LLMs to better suit the vertical domain-specific tasks by utilizing external knowledge remains not completely solved. Indeed, there have emerged a few works on this line where most of them rely...
http://arxiv.org/pdf/2311.01150v1
2023-11-02T11:18:16
2023-11-02T11:18:16
{ "title": [ -0.03048706, -0.023895264, -0.04269409, -0.061584473, 0.0034637451, 0.0057678223, -0.025314331, 0.0014953613, -0.015068054, 0.022216797, -0.04171753, 0.0004761219, -0.036132812, -0.0076446533, -0.04748535, -0.01689148, -0.004928589, ...
http://arxiv.org/abs/2311.01149v2
ChineseWebText: Large-scale High-quality Chinese Web Text Extracted with Effective Evaluation Model
Jianghao Chen, Pu Jian, Tengxiao Xi, Dongyi Yi, Qianlong Du, Chenglin Ding, Guibo Zhu, Chengqing Zong, Jinqiao Wang, Jiajun Zhang
cs.CL
During the development of large language models (LLMs), the scale and quality of the pre-training data play a crucial role in shaping LLMs' capabilities. To accelerate the research of LLMs, several large-scale datasets, such as C4 [1], Pile [2], RefinedWeb [3] and WanJuan [4], have been released to the public. However,...
http://arxiv.org/pdf/2311.01149v2
2023-11-10T06:28:48
2023-11-02T11:13:51
{ "title": [ -0.019500732, 0.0058135986, -0.037750244, -0.033477783, 0.018600464, -0.025543213, -0.007724762, -0.020355225, 0.010269165, 0.02709961, -0.006958008, 0.01499939, 0.020370483, -0.03173828, -0.028640747, 0.015899658, -0.012313843, 0.03...
http://arxiv.org/abs/2311.04918v1
Low-Resource Named Entity Recognition: Can One-vs-All AUC Maximization Help?
Ngoc Dang Nguyen, Wei Tan, Lan Du, Wray Buntine, Richard Beare, Changyou Chen
cs.CL, cs.LG
Named entity recognition (NER), a task that identifies and categorizes named entities such as persons or organizations from text, is traditionally framed as a multi-class classification problem. However, this approach often overlooks the issues of imbalanced label distributions, particularly in low-resource settings, w...
http://arxiv.org/pdf/2311.04918v1
2023-11-02T10:14:52
2023-11-02T10:14:52
{ "title": [ -0.0262146, -0.0076065063, -0.04916382, -0.06402588, -0.0015621185, 0.00617218, -0.047851562, 0.016418457, -0.012382507, 0.040985107, -0.021499634, 0.033416748, -0.010566711, 0.0015592575, -0.035491943, 0.0033683777, -0.009597778, 0....
http://arxiv.org/abs/2311.01108v1
Noise-Robust Fine-Tuning of Pretrained Language Models via External Guidance
Song Wang, Zhen Tan, Ruocheng Guo, Jundong Li
cs.CL
Adopting a two-stage paradigm of pretraining followed by fine-tuning, Pretrained Language Models (PLMs) have achieved substantial advancements in the field of natural language processing. However, in real-world scenarios, data labels are often noisy due to the complex annotation process, making it essential to develop ...
http://arxiv.org/pdf/2311.01108v1
2023-11-02T09:20:38
2023-11-02T09:20:38
{ "title": [ -0.010192871, -0.004600525, -0.028152466, -0.073791504, -0.000042915344, -0.002506256, -0.060699463, 0.0032081604, -0.009750366, 0.038360596, -0.01739502, 0.022857666, 0.00983429, -0.014877319, -0.040618896, 0.012207031, 0.00073194504, ...
http://arxiv.org/abs/2311.04917v1
Adapting Fake News Detection to the Era of Large Language Models
Jinyan Su, Claire Cardie, Preslav Nakov
cs.CL, cs.AI
In the age of large language models (LLMs) and the widespread adoption of AI-driven content creation, the landscape of information dissemination has witnessed a paradigm shift. With the proliferation of both human-written and machine-generated real and fake news, robustly and effectively discerning the veracity of news...
http://arxiv.org/pdf/2311.04917v1
2023-11-02T08:39:45
2023-11-02T08:39:45
{ "title": [ -0.01007843, 0.0033435822, -0.027709961, -0.03555298, 0.0082473755, -0.0055503845, -0.011253357, -0.019302368, -0.018234253, 0.026382446, -0.028121948, 0.007091522, -0.0009646416, -0.005962372, -0.029922485, -0.0063667297, -0.037963867, ...
http://arxiv.org/abs/2311.01070v1
DistilWhisper: Efficient Distillation of Multi-task Speech Models via Language-Specific Experts
Thomas Palmeira Ferraz, Marcely Zanon Boito, Caroline Brun, Vassilina Nikoulina
cs.CL, cs.SD, eess.AS
Whisper is a multitask and multilingual speech model covering 99 languages. It yields commendable automatic speech recognition (ASR) results in a subset of its covered languages, but the model still under-performs on a non-negligible number of under-represented languages, a problem exacerbated in smaller model versions...
http://arxiv.org/pdf/2311.01070v1
2023-11-02T08:37:30
2023-11-02T08:37:30
{ "title": [ -0.0103302, 0.002412796, -0.07446289, -0.09844971, 0.002998352, -0.006752014, -0.10491943, 0.013061523, 0.011077881, 0.04006958, -0.005329132, -0.018600464, 0.041900635, -0.025344849, -0.03994751, 0.008514404, 0.0020961761, 0.0285797...
http://arxiv.org/abs/2311.01049v1
Multi-dimensional data refining strategy for effective fine-tuning LLMs
Thanh Nguyen Ngoc, Quang Nhat Tran, Arthur Tang, Bao Nguyen, Thuy Nguyen, Thanh Pham
cs.CL, cs.AI
Data is a cornerstone for fine-tuning large language models, yet acquiring suitable data remains challenging. Challenges encompassed data scarcity, linguistic diversity, and domain-specific content. This paper presents lessons learned while crawling and refining data tailored for fine-tuning Vietnamese language models....
http://arxiv.org/pdf/2311.01049v1
2023-11-02T07:50:43
2023-11-02T07:50:43
{ "title": [ -0.038482666, -0.006385803, -0.061035156, -0.078186035, -0.024520874, -0.010696411, -0.06994629, 0.025634766, -0.027862549, 0.030914307, -0.04257202, 0.014930725, -0.00077724457, -0.02041626, -0.020645142, -0.02104187, -0.004043579, ...
http://arxiv.org/abs/2311.01041v1
Learn to Refuse: Making Large Language Models More Controllable and Reliable through Knowledge Scope Limitation and Refusal Mechanism
Lang Cao
cs.CL, cs.AI
Large language models (LLMs) have demonstrated impressive language understanding and generation capabilities, enabling them to answer a wide range of questions across various domains. However, these models are not flawless and often produce responses that contain errors or misinformation. These inaccuracies, commonly r...
http://arxiv.org/pdf/2311.01041v1
2023-11-02T07:20:49
2023-11-02T07:20:49
{ "title": [ -0.021591187, -0.02859497, -0.040039062, -0.085754395, -0.016586304, 0.007045746, -0.06958008, 0.018569946, -0.0050201416, 0.022109985, -0.03741455, 0.0044517517, 0.0022010803, -0.004386902, -0.043762207, -0.007587433, -0.011238098, ...
http://arxiv.org/abs/2311.01036v1
ATHENA: Mathematical Reasoning with Thought Expansion
JB. Kim, Hazel Kim, Joonghyuk Hahn, Yo-Sub Han
cs.CL, cs.AI, I.2.7; I.2.3; F.4.1
Solving math word problems depends on how to articulate the problems, the lens through which models view human linguistic expressions. Real-world settings count on such a method even more due to the diverse practices of the same mathematical operations. Earlier works constrain available thinking processes by limited pr...
http://arxiv.org/pdf/2311.01036v1
2023-11-02T07:03:25
2023-11-02T07:03:25
{ "title": [ -0.021347046, -0.015808105, -0.04333496, -0.02330017, -0.012542725, -0.014427185, -0.012145996, -0.016937256, 0.0146102905, -0.0046653748, -0.03933716, -0.006225586, -0.007347107, -0.037506104, -0.020568848, -0.007911682, 0.0069007874, ...
http://arxiv.org/abs/2311.01030v1
Joint Learning of Local and Global Features for Aspect-based Sentiment Classification
Hao Niu, Yun Xiong, Xiaosu Wang, Philip S. Yu
cs.CL, cs.AI
Aspect-based sentiment classification (ASC) aims to judge the sentiment polarity conveyed by the given aspect term in a sentence. The sentiment polarity is not only determined by the local context but also related to the words far away from the given aspect term. Most recent efforts related to the attention-based model...
http://arxiv.org/pdf/2311.01030v1
2023-11-02T06:43:50
2023-11-02T06:43:50
{ "title": [ 0.0005044937, -0.001698494, -0.024337769, -0.066711426, 0.013793945, 0.0059928894, -0.033996582, 0.0018911362, 0.01991272, 0.045135498, -0.01222229, -0.021316528, 0.00032949448, -0.008598328, -0.07116699, 0.00009936094, -0.003917694, ...
http://arxiv.org/abs/2311.01012v2
COPAL-ID: Indonesian Language Reasoning with Local Culture and Nuances
Haryo Akbarianto Wibowo, Erland Hilman Fuadi, Made Nindyatama Nityasya, Radityo Eko Prasojo, Alham Fikri Aji
cs.CL, 68T50
We present publicly available COPAL-ID, a novel Indonesian language common sense reasoning dataset. Unlike the previous Indonesian COPA dataset (XCOPA-ID), COPAL-ID incorporates Indonesian local and cultural nuances, and therefore, provides a more natural portrayal of day-to-day causal reasoning within the Indonesian c...
http://arxiv.org/pdf/2311.01012v2
2023-11-13T18:19:44
2023-11-02T06:14:41
{ "title": [ -0.013946533, -0.036346436, -0.0062065125, -0.06896973, -0.022003174, 0.016159058, -0.061157227, 0.016342163, -0.007873535, 0.020828247, 0.0043907166, -0.012359619, 0.007843018, -0.032592773, -0.027359009, -0.0074653625, -0.018630981, ...
http://arxiv.org/abs/2311.00998v1
Replicable Benchmarking of Neural Machine Translation (NMT) on Low-Resource Local Languages in Indonesia
Lucky Susanto, Ryandito Diandaru, Adila Krisnadhi, Ayu Purwarianti, Derry Wijaya
cs.CL, cs.AI
Neural machine translation (NMT) for low-resource local languages in Indonesia faces significant challenges, including the need for a representative benchmark and limited data availability. This work addresses these challenges by comprehensively analyzing training NMT systems for four low-resource local languages in In...
http://arxiv.org/pdf/2311.00998v1
2023-11-02T05:27:48
2023-11-02T05:27:48
{ "title": [ -0.012771606, -0.013748169, 0.01713562, -0.029083252, -0.031677246, 0.013671875, -0.0524292, -0.021362305, -0.041625977, 0.050933838, 0.009429932, 0.022903442, -0.014450073, -0.011665344, -0.049987793, -0.0014972687, -0.0124053955, 0...
http://arxiv.org/abs/2311.04916v1
Explainable Identification of Hate Speech towards Islam using Graph Neural Networks
Azmine Toushik Wasi
cs.CL, cs.LG
Islamophobic language is a prevalent challenge on online social interaction platforms. Identifying and eliminating such hatred is a crucial step towards a future of harmony and peace. This study presents a novel paradigm for identifying and explaining hate speech towards Islam using graph neural networks. Utilizing the...
http://arxiv.org/pdf/2311.04916v1
2023-11-02T04:01:04
2023-11-02T04:01:04
{ "title": [ 0.014450073, 0.012939453, -0.023864746, -0.027435303, 0.014808655, -0.027893066, 0.012840271, -0.011016846, -0.036132812, 0.012275696, 0.0010881424, -0.0087509155, -0.026565552, -0.033569336, -0.039215088, -0.026794434, -0.0064697266, ...
http://arxiv.org/abs/2311.00967v1
Vision-Language Interpreter for Robot Task Planning
Keisuke Shirai, Cristian C. Beltran-Hernandez, Masashi Hamaya, Atsushi Hashimoto, Shohei Tanaka, Kento Kawaharazuka, Kazutoshi Tanaka, Yoshitaka Ushiku, Shinsuke Mori
cs.RO, cs.AI, cs.CL
Large language models (LLMs) are accelerating the development of language-guided robot planners. Meanwhile, symbolic planners offer the advantage of interpretability. This paper proposes a new task that bridges these two trends, namely, multimodal planning problem specification. The aim is to generate a problem descrip...
http://arxiv.org/pdf/2311.00967v1
2023-11-02T03:32:30
2023-11-02T03:32:30
{ "title": [ -0.047729492, -0.026763916, -0.04095459, -0.031036377, -0.021362305, -0.008728027, -0.028381348, 0.009529114, 0.00001859665, 0.016204834, -0.04714966, -0.011314392, -0.016784668, -0.006767273, -0.0385437, -0.029327393, 0.026489258, -...
http://arxiv.org/abs/2311.00958v1
IndoToD: A Multi-Domain Indonesian Benchmark For End-to-End Task-Oriented Dialogue Systems
Muhammad Dehan Al Kautsar, Rahmah Khoirussyifa' Nurdini, Samuel Cahyawijaya, Genta Indra Winata, Ayu Purwarianti
cs.CL, cs.AI
Task-oriented dialogue (ToD) systems have been mostly created for high-resource languages, such as English and Chinese. However, there is a need to develop ToD systems for other regional or local languages to broaden their ability to comprehend the dialogue contexts in various languages. This paper introduces IndoToD, ...
http://arxiv.org/pdf/2311.00958v1
2023-11-02T03:01:53
2023-11-02T03:01:53
{ "title": [ -0.013961792, -0.015182495, -0.027252197, -0.07757568, 0.00035858154, 0.0011405945, -0.05984497, 0.017288208, 0.0053863525, 0.033172607, 0.024154663, 0.007537842, -0.033050537, -0.028198242, -0.06262207, -0.0051078796, -0.02268982, 0...
http://arxiv.org/abs/2311.00953v1
Blending Reward Functions via Few Expert Demonstrations for Faithful and Accurate Knowledge-Grounded Dialogue Generation
Wanyu Du, Yangfeng Ji
cs.CL
The development of trustworthy conversational information-seeking systems relies on dialogue models that can generate faithful and accurate responses based on relevant knowledge texts. However, two main challenges hinder this task. Firstly, language models may generate hallucinations due to data biases present in their...
http://arxiv.org/pdf/2311.00953v1
2023-11-02T02:42:41
2023-11-02T02:42:41
{ "title": [ 0.009284973, -0.007217407, -0.032684326, -0.07318115, -0.0010051727, -0.01008606, -0.034179688, -0.0036201477, -0.015686035, 0.019470215, -0.0000011324883, 0.0021457672, -0.020980835, -0.02720642, -0.035888672, -0.017181396, -0.01084137,...
http://arxiv.org/abs/2311.00945v1
E3 TTS: Easy End-to-End Diffusion-based Text to Speech
Yuan Gao, Nobuyuki Morioka, Yu Zhang, Nanxin Chen
cs.SD, cs.CL, cs.LG, eess.AS
We propose Easy End-to-End Diffusion-based Text to Speech, a simple and efficient end-to-end text-to-speech model based on diffusion. E3 TTS directly takes plain text as input and generates an audio waveform through an iterative refinement process. Unlike many prior work, E3 TTS does not rely on any intermediate repres...
http://arxiv.org/pdf/2311.00945v1
2023-11-02T02:22:21
2023-11-02T02:22:21
{ "title": [ -0.017120361, -0.0054855347, -0.04272461, -0.06616211, 0.02709961, -0.0011873245, -0.012367249, 0.019439697, 0.0020656586, 0.04257202, -0.014984131, 0.009460449, 0.02659607, -0.025985718, -0.037475586, 0.012107849, -0.015640259, 0.03...
http://arxiv.org/abs/2311.04915v1
Chain of Empathy: Enhancing Empathetic Response of Large Language Models Based on Psychotherapy Models
Yoon Kyung Lee, Inju Lee, Minjung Shin, Seoyeon Bae, Sowon Hahn
cs.CL, cs.AI, cs.HC
We present a novel method, the Chain of Empathy (CoE) prompting, that utilizes insights from psychotherapy to induce Large Language Models (LLMs) to reason about human emotional states. This method is inspired by various psychotherapy approaches including Cognitive Behavioral Therapy (CBT), Dialectical Behavior Therapy...
http://arxiv.org/pdf/2311.04915v1
2023-11-02T02:21:39
2023-11-02T02:21:39
{ "title": [ -0.018951416, -0.0026340485, -0.031585693, -0.042663574, -0.006259918, -0.011894226, -0.025909424, -0.007610321, -0.0054397583, 0.021728516, 0.0017108917, 0.00036764145, 0.020523071, -0.04650879, -0.006504059, 0.0073661804, -0.018188477,...
http://arxiv.org/abs/2311.00915v1
Task-Agnostic Low-Rank Adapters for Unseen English Dialects
Zedian Xiao, William Held, Yanchen Liu, Diyi Yang
cs.CL
Large Language Models (LLMs) are trained on corpora disproportionally weighted in favor of Standard American English. As a result, speakers of other dialects experience significantly more failures when interacting with these technologies. In practice, these speakers often accommodate their speech to be better understoo...
http://arxiv.org/pdf/2311.00915v1
2023-11-02T01:17:29
2023-11-02T01:17:29
{ "title": [ -0.0102005005, 0.008239746, -0.029708862, -0.018707275, -0.011634827, -0.008804321, -0.03152466, -0.010986328, 0.016418457, 0.050628662, -0.0037975311, -0.022888184, 0.038604736, -0.023498535, -0.0423584, 0.02911377, 0.013496399, 0.0...
http://arxiv.org/abs/2311.00913v1
Self-Influence Guided Data Reweighting for Language Model Pre-training
Megh Thakkar, Tolga Bolukbasi, Sriram Ganapathy, Shikhar Vashishth, Sarath Chandar, Partha Talukdar
cs.CL
Language Models (LMs) pre-trained with self-supervision on large text corpora have become the default starting point for developing models for various NLP tasks. Once the pre-training corpus has been assembled, all data samples in the corpus are treated with equal importance during LM pre-training. However, due to vary...
http://arxiv.org/pdf/2311.00913v1
2023-11-02T01:00:46
2023-11-02T01:00:46
{ "title": [ -0.025817871, -0.0025043488, -0.044525146, -0.08331299, -0.023162842, 0.0062942505, -0.0770874, 0.004180908, -0.010955811, 0.040405273, -0.019119263, 0.009811401, 0.017105103, -0.021469116, -0.03375244, -0.0057029724, -0.0012397766, ...