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/2310.20274v1
Extracting Entities of Interest from Comparative Product Reviews
Jatin Arora, Sumit Agrawal, Pawan Goyal, Sayan Pathak
cs.IR, cs.CL, cs.LG, I.2.7; H.3.3
This paper presents a deep learning based approach to extract product comparison information out of user reviews on various e-commerce websites. Any comparative product review has three major entities of information: the names of the products being compared, the user opinion (predicate) and the feature or aspect under ...
http://arxiv.org/pdf/2310.20274v1
2023-10-31T08:43:11
2023-10-31T08:43:11
{ "title": [ -0.024978638, -0.006565094, -0.04119873, -0.0104522705, -0.0041007996, -0.025756836, -0.008049011, -0.029968262, -0.010032654, 0.037750244, -0.020843506, 0.021194458, -0.016860962, 0.009117126, -0.05593872, -0.0034809113, -0.004383087, ...
http://arxiv.org/abs/2310.20260v1
Learning to Play Chess from Textbooks (LEAP): a Corpus for Evaluating Chess Moves based on Sentiment Analysis
Haifa Alrdahi, Riza Batista-Navarro
cs.CL
Learning chess strategies has been investigated widely, with most studies focussing on learning from previous games using search algorithms. Chess textbooks encapsulate grandmaster knowledge, explain playing strategies and require a smaller search space compared to traditional chess agents. This paper examines chess te...
http://arxiv.org/pdf/2310.20260v1
2023-10-31T08:26:02
2023-10-31T08:26:02
{ "title": [ 0.026016235, 0.015594482, -0.0079422, 0.012626648, 0.010017395, -0.030181885, 0.021697998, -0.040405273, 0.017105103, 0.034118652, -0.059387207, 0.0045318604, 0.02418518, -0.006450653, -0.026626587, -0.022766113, 0.006134033, 0.00014...
http://arxiv.org/abs/2310.20256v2
PsyCoT: Psychological Questionnaire as Powerful Chain-of-Thought for Personality Detection
Tao Yang, Tianyuan Shi, Fanqi Wan, Xiaojun Quan, Qifan Wang, Bingzhe Wu, Jiaxiang Wu
cs.CL
Recent advances in large language models (LLMs), such as ChatGPT, have showcased remarkable zero-shot performance across various NLP tasks. However, the potential of LLMs in personality detection, which involves identifying an individual's personality from their written texts, remains largely unexplored. Drawing inspir...
http://arxiv.org/pdf/2310.20256v2
2023-11-05T03:19:18
2023-10-31T08:23:33
{ "title": [ -0.010627747, -0.001086235, -0.03463745, -0.018859863, 0.01512146, -0.016433716, 0.012634277, -0.019454956, 0.0032176971, 0.01928711, -0.018051147, -0.006084442, 0.0013952255, -0.04031372, -0.032318115, 0.009101868, -0.0025119781, 0....
http://arxiv.org/abs/2310.20246v3
Breaking Language Barriers in Multilingual Mathematical Reasoning: Insights and Observations
Nuo Chen, Zinan Zheng, Ning Wu, Ming Gong, Yangqiu Song, Dongmei Zhang, Jia Li
cs.CL, cs.AI
Existing research predominantly focuses on developing powerful language learning models (LLMs) for mathematical reasoning within monolingual languages, with few explorations in preserving efficacy in a multilingual context. To bridge this gap, this paper pioneers exploring and training powerful Multilingual Math Reason...
http://arxiv.org/pdf/2310.20246v3
2023-11-07T12:13:02
2023-10-31T08:09:20
{ "title": [ -0.027877808, -0.010444641, -0.023757935, -0.049072266, -0.00031113625, -0.0015535355, -0.028289795, -0.0032901764, 0.009170532, 0.017959595, -0.0048942566, -0.009468079, 0.00030899048, -0.030776978, -0.019348145, 0.0037670135, -0.010513...
http://arxiv.org/abs/2310.20236v1
Dynamically Updating Event Representations for Temporal Relation Classification with Multi-category Learning
Fei Cheng, Masayuki Asahara, Ichiro Kobayashi, Sadao Kurohashi
cs.CL
Temporal relation classification is a pair-wise task for identifying the relation of a temporal link (TLINK) between two mentions, i.e. event, time, and document creation time (DCT). It leads to two crucial limits: 1) Two TLINKs involving a common mention do not share information. 2) Existing models with independent cl...
http://arxiv.org/pdf/2310.20236v1
2023-10-31T07:41:24
2023-10-31T07:41:24
{ "title": [ 0.03845215, 0.0051841736, -0.049987793, -0.11102295, 0.029968262, -0.022964478, -0.045898438, 0.01184082, 0.019592285, 0.021499634, 0.0024280548, 0.0036010742, -0.014076233, -0.01763916, -0.043762207, 0.0030212402, 0.00066423416, 0.0...
http://arxiv.org/abs/2310.20216v1
Does GPT-4 Pass the Turing Test?
Cameron Jones, Benjamin Bergen
cs.AI, cs.CL
We evaluated GPT-4 in a public online Turing Test. The best-performing GPT-4 prompt passed in 41% of games, outperforming baselines set by ELIZA (27%) and GPT-3.5 (14%), but falling short of chance and the baseline set by human participants (63%). Participants' decisions were based mainly on linguistic style (35%) and ...
http://arxiv.org/pdf/2310.20216v1
2023-10-31T06:27:52
2023-10-31T06:27:52
{ "title": [ 0.0047454834, 0.016647339, -0.0059165955, 0.009346008, -0.01335144, -0.007259369, 0.004180908, -0.05493164, -0.018753052, 0.029144287, -0.015060425, -0.018859863, -0.013343811, -0.025054932, -0.019195557, -0.012268066, -0.012290955, ...
http://arxiv.org/abs/2310.20204v1
General-Purpose Retrieval-Enhanced Medical Prediction Model Using Near-Infinite History
Junu Kim, Chaeeun Shim, Bosco Seong Kyu Yang, Chami Im, Sung Yoon Lim, Han-Gil Jeong, Edward Choi
cs.LG, cs.CL
Developing clinical prediction models (e.g., mortality prediction) based on electronic health records (EHRs) typically relies on expert opinion for feature selection and adjusting observation window size. This burdens experts and creates a bottleneck in the development process. We propose Retrieval-Enhanced Medical pre...
http://arxiv.org/pdf/2310.20204v1
2023-10-31T06:04:18
2023-10-31T06:04:18
{ "title": [ 0.033569336, 0.014137268, -0.026397705, -0.09020996, 0.03741455, 0.004211426, -0.026733398, -0.014183044, 0.024490356, 0.039398193, 0.0038776398, 0.0009965897, -0.005519867, -0.028396606, -0.061706543, 0.022888184, 0.006275177, 0.005...
http://arxiv.org/abs/2310.20201v1
Video-Helpful Multimodal Machine Translation
Yihang Li, Shuichiro Shimizu, Chenhui Chu, Sadao Kurohashi, Wei Li
cs.CL
Existing multimodal machine translation (MMT) datasets consist of images and video captions or instructional video subtitles, which rarely contain linguistic ambiguity, making visual information ineffective in generating appropriate translations. Recent work has constructed an ambiguous subtitles dataset to alleviate t...
http://arxiv.org/pdf/2310.20201v1
2023-10-31T05:51:56
2023-10-31T05:51:56
{ "title": [ -0.02319336, 0.0029582977, -0.0340271, -0.050231934, -0.019302368, -0.010261536, -0.040222168, -0.0025157928, -0.016799927, 0.059631348, -0.017791748, 0.0079956055, 0.036315918, -0.022247314, -0.02116394, 0.011932373, 0.007762909, 0....
http://arxiv.org/abs/2310.20195v2
Generating Continuations in Multilingual Idiomatic Contexts
Rhitabrat Pokharel, Ameeta Agrawal
cs.CL, cs.AI
The ability to process idiomatic or literal multiword expressions is a crucial aspect of understanding and generating any language. The task of generating contextually relevant continuations for narratives containing idiomatic (or literal) expressions can allow us to test the ability of generative language models (LMs)...
http://arxiv.org/pdf/2310.20195v2
2023-11-04T04:15:18
2023-10-31T05:40:33
{ "title": [ -0.0067253113, -0.026901245, 0.006088257, -0.0793457, -0.0068359375, 0.03164673, -0.06378174, -0.002532959, 0.0036640167, 0.060058594, -0.003364563, 0.01737976, 0.021316528, -0.007972717, -0.033935547, 0.035003662, -0.026184082, 0.03...
http://arxiv.org/abs/2310.20170v1
DIVKNOWQA: Assessing the Reasoning Ability of LLMs via Open-Domain Question Answering over Knowledge Base and Text
Wenting Zhao, Ye Liu, Tong Niu, Yao Wan, Philip S. Yu, Shafiq Joty, Yingbo Zhou, Semih Yavuz
cs.CL
Large Language Models (LLMs) have exhibited impressive generation capabilities, but they suffer from hallucinations when solely relying on their internal knowledge, especially when answering questions that require less commonly known information. Retrieval-augmented LLMs have emerged as a potential solution to ground L...
http://arxiv.org/pdf/2310.20170v1
2023-10-31T04:37:57
2023-10-31T04:37:57
{ "title": [ -0.056365967, -0.004989624, -0.047454834, -0.03237915, -0.024429321, 0.008552551, -0.052612305, -0.009414673, 0.028808594, -0.009361267, -0.041931152, 0.004589081, -0.0048599243, -0.0362854, -0.014472961, -0.019470215, -0.014228821, ...
http://arxiv.org/abs/2310.20158v1
GAR-meets-RAG Paradigm for Zero-Shot Information Retrieval
Daman Arora, Anush Kini, Sayak Ray Chowdhury, Nagarajan Natarajan, Gaurav Sinha, Amit Sharma
cs.CL
Given a query and a document corpus, the information retrieval (IR) task is to output a ranked list of relevant documents. Combining large language models (LLMs) with embedding-based retrieval models, recent work shows promising results on the zero-shot retrieval problem, i.e., no access to labeled data from the target...
http://arxiv.org/pdf/2310.20158v1
2023-10-31T03:52:08
2023-10-31T03:52:08
{ "title": [ 0.005065918, -0.005214691, -0.044525146, -0.036376953, 0.004886627, -0.027496338, 0.013618469, 0.003314972, 0.011169434, 0.045410156, -0.0077819824, 0.0050964355, -0.009155273, -0.018585205, -0.058288574, 0.008811951, 0.003004074, 0....
http://arxiv.org/abs/2310.20153v1
Interactive Multi-fidelity Learning for Cost-effective Adaptation of Language Model with Sparse Human Supervision
Jiaxin Zhang, Zhuohang Li, Kamalika Das, Sricharan Kumar
cs.CL, cs.AI
Large language models (LLMs) have demonstrated remarkable capabilities in various tasks. However, their suitability for domain-specific tasks, is limited due to their immense scale at deployment, susceptibility to misinformation, and more importantly, high data annotation costs. We propose a novel Interactive Multi-Fid...
http://arxiv.org/pdf/2310.20153v1
2023-10-31T03:39:23
2023-10-31T03:39:23
{ "title": [ -0.009239197, -0.010948181, -0.029708862, -0.056365967, -0.028244019, -0.006752014, -0.06726074, 0.008987427, -0.01007843, 0.02947998, 0.0045318604, 0.00005376339, -0.00573349, -0.024963379, -0.039886475, -0.006832123, 0.00579834, 0....
http://arxiv.org/abs/2310.20151v1
Multi-Agent Consensus Seeking via Large Language Models
Huaben Chen, Wenkang Ji, Lufeng Xu, Shiyu Zhao
cs.CL, cs.RO, cs.SY, eess.SY
Multi-agent systems driven by large language models (LLMs) have shown promising abilities for solving complex tasks in a collaborative manner. This work considers a fundamental problem in multi-agent collaboration: consensus seeking. When multiple agents work together, we are interested in how they can reach a consensu...
http://arxiv.org/pdf/2310.20151v1
2023-10-31T03:37:11
2023-10-31T03:37:11
{ "title": [ -0.04586792, 0.0065612793, -0.04385376, -0.02418518, 0.0062332153, -0.004108429, -0.017181396, -0.006904602, 0.010322571, 0.036987305, -0.013999939, -0.0035381317, -0.011306763, 0.0015306473, -0.05255127, -0.002796173, -0.01878357, 0...
http://arxiv.org/abs/2310.20150v1
Unlearn What You Want to Forget: Efficient Unlearning for LLMs
Jiaao Chen, Diyi Yang
cs.CL, cs.AI
Large language models (LLMs) have achieved significant progress from pre-training on and memorizing a wide range of textual data, however, this process might suffer from privacy issues and violations of data protection regulations. As a result, the ability to easily remove data related to individual users from such mod...
http://arxiv.org/pdf/2310.20150v1
2023-10-31T03:35:59
2023-10-31T03:35:59
{ "title": [ -0.03137207, -0.011528015, -0.058563232, -0.044952393, -0.01991272, -0.017196655, -0.04421997, 0.0033988953, -0.00674057, 0.027618408, -0.028305054, 0.013046265, 0.010643005, -0.034118652, -0.022369385, -0.017150879, 0.012062073, 0.0...
http://arxiv.org/abs/2310.20144v1
EELBERT: Tiny Models through Dynamic Embeddings
Gabrielle Cohn, Rishika Agarwal, Deepanshu Gupta, Siddharth Patwardhan
cs.CL, cs.AI, cs.LG, 68T07, I.2.7; I.2.6
We introduce EELBERT, an approach for compression of transformer-based models (e.g., BERT), with minimal impact on the accuracy of downstream tasks. This is achieved by replacing the input embedding layer of the model with dynamic, i.e. on-the-fly, embedding computations. Since the input embedding layer accounts for a ...
http://arxiv.org/pdf/2310.20144v1
2023-10-31T03:28:08
2023-10-31T03:28:08
{ "title": [ -0.024368286, -0.012138367, -0.032470703, -0.048675537, -0.0039863586, 0.013549805, -0.035064697, -0.024246216, 0.009735107, 0.027816772, -0.023788452, 0.005332947, 0.0061454773, -0.0008044243, -0.03665161, 0.012046814, 0.004310608, ...
http://arxiv.org/abs/2310.20138v1
DEPN: Detecting and Editing Privacy Neurons in Pretrained Language Models
Xinwei Wu, Junzhuo Li, Minghui Xu, Weilong Dong, Shuangzhi Wu, Chao Bian, Deyi Xiong
cs.CR, cs.CL
Large language models pretrained on a huge amount of data capture rich knowledge and information in the training data. The ability of data memorization and regurgitation in pretrained language models, revealed in previous studies, brings the risk of data leakage. In order to effectively reduce these risks, we propose a...
http://arxiv.org/pdf/2310.20138v1
2023-10-31T03:09:36
2023-10-31T03:09:36
{ "title": [ -0.009162903, -0.017578125, -0.029800415, -0.02532959, -0.004016876, -0.017059326, -0.015731812, -0.008003235, -0.019805908, 0.016326904, 0.0018568039, -0.0062065125, 0.024108887, -0.035980225, -0.016784668, 0.014205933, -0.021072388, ...
http://arxiv.org/abs/2310.20127v1
Improving Prompt Tuning with Learned Prompting Layers
Wei Zhu, Ming Tan
cs.CL
Prompt tuning prepends a soft prompt to the input embeddings or hidden states and only optimizes the prompt to adapt pretrained models (PTMs) to downstream tasks. The previous work manually selects prompt layers which are far from optimal and failed to exploit the potential of prompt tuning. In this work, we propose a ...
http://arxiv.org/pdf/2310.20127v1
2023-10-31T02:07:51
2023-10-31T02:07:51
{ "title": [ -0.011138916, -0.012107849, -0.022781372, -0.051635742, -0.024932861, -0.0038890839, -0.05532837, 0.0033950806, -0.017974854, 0.020401001, -0.013244629, -0.000975132, -0.019454956, -0.03265381, -0.018447876, -0.0022583008, 0.0129852295, ...
http://arxiv.org/abs/2310.20121v1
Ling-CL: Understanding NLP Models through Linguistic Curricula
Mohamed Elgaar, Hadi Amiri
cs.CL
We employ a characterization of linguistic complexity from psycholinguistic and language acquisition research to develop data-driven curricula to understand the underlying linguistic knowledge that models learn to address NLP tasks. The novelty of our approach is in the development of linguistic curricula derived from ...
http://arxiv.org/pdf/2310.20121v1
2023-10-31T01:44:33
2023-10-31T01:44:33
{ "title": [ -0.043670654, -0.031829834, -0.026260376, -0.039764404, -0.017425537, 0.007194519, -0.040802002, -0.0021629333, 0.0041275024, 0.013641357, -0.027160645, 0.006011963, -0.003168106, -0.017532349, -0.026824951, 0.0035266876, -0.02583313, ...
http://arxiv.org/abs/2310.20111v1
Making Large Language Models Better Data Creators
Dong-Ho Lee, Jay Pujara, Mohit Sewak, Ryen W. White, Sujay Kumar Jauhar
cs.CL
Although large language models (LLMs) have advanced the state-of-the-art in NLP significantly, deploying them for downstream applications is still challenging due to cost, responsiveness, control, or concerns around privacy and security. As such, trainable models are still the preferred option in some cases. However, t...
http://arxiv.org/pdf/2310.20111v1
2023-10-31T01:08:34
2023-10-31T01:08:34
{ "title": [ -0.018814087, -0.020339966, -0.0362854, -0.087402344, -0.015457153, 0.009231567, -0.07080078, 0.022705078, -0.0012102127, 0.030090332, -0.026260376, -0.0058403015, -0.0126571655, -0.012069702, -0.034118652, 0.010940552, 0.0007648468, ...
http://arxiv.org/abs/2310.20105v1
Efficient Classification of Student Help Requests in Programming Courses Using Large Language Models
Jaromir Savelka, Paul Denny, Mark Liffiton, Brad Sheese
cs.CY, cs.AI, cs.CL
The accurate classification of student help requests with respect to the type of help being sought can enable the tailoring of effective responses. Automatically classifying such requests is non-trivial, but large language models (LLMs) appear to offer an accessible, cost-effective solution. This study evaluates the pe...
http://arxiv.org/pdf/2310.20105v1
2023-10-31T00:56:33
2023-10-31T00:56:33
{ "title": [ -0.0074882507, -0.016021729, -0.017959595, -0.040985107, 0.0020866394, -0.01838684, -0.010604858, 0.0036582947, -0.0020160675, 0.023635864, -0.042266846, 0.008361816, 0.00012362003, -0.007980347, -0.035583496, -0.025527954, -0.020812988,...
http://arxiv.org/abs/2310.20093v1
Evaluating Neural Language Models as Cognitive Models of Language Acquisition
Héctor Javier Vázquez Martínez, Annika Lea Heuser, Charles Yang, Jordan Kodner
cs.CL, cs.AI
The success of neural language models (LMs) on many technological tasks has brought about their potential relevance as scientific theories of language despite some clear differences between LM training and child language acquisition. In this paper we argue that some of the most prominent benchmarks for evaluating the s...
http://arxiv.org/pdf/2310.20093v1
2023-10-31T00:16:17
2023-10-31T00:16:17
{ "title": [ -0.03656006, -0.0033721924, -0.02078247, -0.011680603, -0.015960693, -0.004371643, -0.015823364, -0.008155823, -0.0284729, 0.027572632, -0.0043525696, -0.0051879883, -0.011878967, -0.03201294, -0.047058105, 0.0007367134, -0.02180481, ...
http://arxiv.org/abs/2310.20089v1
Keyword-optimized Template Insertion for Clinical Information Extraction via Prompt-based Learning
Eugenia Alleva, Isotta Landi, Leslee J Shaw, Erwin Böttinger, Thomas J Fuchs, Ipek Ensari
cs.CL
Clinical note classification is a common clinical NLP task. However, annotated data-sets are scarse. Prompt-based learning has recently emerged as an effective method to adapt pre-trained models for text classification using only few training examples. A critical component of prompt design is the definition of the temp...
http://arxiv.org/pdf/2310.20089v1
2023-10-31T00:07:11
2023-10-31T00:07:11
{ "title": [ -0.021087646, 0.0056610107, -0.04055786, -0.06652832, 0.0014047623, 0.015319824, -0.054260254, -0.025985718, -0.008865356, 0.02619934, -0.019866943, 0.019821167, -0.018753052, -0.02053833, -0.028625488, -0.015853882, -0.0026378632, -...
http://arxiv.org/abs/2310.20081v1
Integrating Summarization and Retrieval for Enhanced Personalization via Large Language Models
Chris Richardson, Yao Zhang, Kellen Gillespie, Sudipta Kar, Arshdeep Singh, Zeynab Raeesy, Omar Zia Khan, Abhinav Sethy
cs.CL, cs.AI, cs.IR, I.2.7; H.3.3
Personalization, the ability to tailor a system to individual users, is an essential factor in user experience with natural language processing (NLP) systems. With the emergence of Large Language Models (LLMs), a key question is how to leverage these models to better personalize user experiences. To personalize a langu...
http://arxiv.org/pdf/2310.20081v1
2023-10-30T23:40:41
2023-10-30T23:40:41
{ "title": [ 0.013130188, -0.0026893616, -0.026657104, -0.09185791, 0.0075187683, 0.014701843, -0.052124023, 0.0008254051, -0.004169464, 0.057159424, 0.0065841675, 0.009262085, 0.015510559, -0.0050239563, -0.047058105, 0.01687622, -0.0054244995, ...
http://arxiv.org/abs/2310.20077v1
Partial Tensorized Transformers for Natural Language Processing
Subhadra Vadlamannati, Ryan Solgi
cs.CL, cs.LG
The transformer architecture has revolutionized Natural Language Processing (NLP) and other machine-learning tasks, due to its unprecedented accuracy. However, their extensive memory and parameter requirements often hinder their practical applications. In this work, we study the effect of tensor-train decomposition to ...
http://arxiv.org/pdf/2310.20077v1
2023-10-30T23:19:06
2023-10-30T23:19:06
{ "title": [ -0.0026779175, -0.007865906, -0.0022583008, -0.072509766, 0.009529114, 0.024810791, -0.04397583, -0.012397766, -0.0029582977, 0.04006958, -0.018814087, 0.008544922, 0.0184021, -0.014259338, -0.040039062, 0.021408081, -0.009887695, 0....
http://arxiv.org/abs/2310.20072v1
Automatic Evaluation of Generative Models with Instruction Tuning
Shuhaib Mehri, Vered Shwartz
cs.CL, cs.LG
Automatic evaluation of natural language generation has long been an elusive goal in NLP.A recent paradigm fine-tunes pre-trained language models to emulate human judgements for a particular task and evaluation criterion. Inspired by the generalization ability of instruction-tuned models, we propose a learned metric ba...
http://arxiv.org/pdf/2310.20072v1
2023-10-30T23:00:52
2023-10-30T23:00:52
{ "title": [ -0.047698975, -0.014663696, -0.014915466, -0.047851562, -0.010269165, 0.013694763, -0.059020996, -0.019012451, 0.012397766, 0.018920898, -0.030090332, 0.0066947937, -0.037628174, -0.017990112, -0.04953003, -0.0036201477, -0.010437012, ...
http://arxiv.org/abs/2310.20046v1
Which Examples to Annotate for In-Context Learning? Towards Effective and Efficient Selection
Costas Mavromatis, Balasubramaniam Srinivasan, Zhengyuan Shen, Jiani Zhang, Huzefa Rangwala, Christos Faloutsos, George Karypis
cs.CL
Large Language Models (LLMs) can adapt to new tasks via in-context learning (ICL). ICL is efficient as it does not require any parameter updates to the trained LLM, but only few annotated examples as input for the LLM. In this work, we investigate an active learning approach for ICL, where there is a limited budget for...
http://arxiv.org/pdf/2310.20046v1
2023-10-30T22:03:55
2023-10-30T22:03:55
{ "title": [ 0.02558899, -0.002407074, -0.028656006, -0.034942627, 0.02545166, -0.025238037, 0.0053482056, -0.030014038, -0.006477356, 0.032562256, -0.02217102, 0.02192688, -0.017944336, -0.017349243, -0.04168701, -0.0059280396, -0.0057678223, 0....
http://arxiv.org/abs/2311.01469v1
Leveraging Language Models to Detect Greenwashing
Avalon Vinella, Margaret Capetz, Rebecca Pattichis, Christina Chance, Reshmi Ghosh
cs.CL, cs.LG
In recent years, climate change repercussions have increasingly captured public interest. Consequently, corporations are emphasizing their environmental efforts in sustainability reports to bolster their public image. Yet, the absence of stringent regulations in review of such reports allows potential greenwashing. In ...
http://arxiv.org/pdf/2311.01469v1
2023-10-30T21:41:49
2023-10-30T21:41:49
{ "title": [ 0.0047416687, 0.012992859, -0.046020508, -0.04550171, 0.0206604, -0.008132935, -0.02168274, -0.015777588, -0.03967285, 0.031051636, -0.03173828, 0.022903442, -0.03527832, -0.012779236, -0.039489746, -0.033325195, -0.024551392, 0.0482...
http://arxiv.org/abs/2310.20033v2
Synthetic Imitation Edit Feedback for Factual Alignment in Clinical Summarization
Prakamya Mishra, Zonghai Yao, Shuwei Chen, Beining Wang, Rohan Mittal, Hong Yu
cs.CL, cs.AI
Large Language Models (LLMs) like the GPT and LLaMA families have demonstrated exceptional capabilities in capturing and condensing critical contextual information and achieving state-of-the-art performance in the summarization task. However, community concerns about these models' hallucination issues continue to rise....
http://arxiv.org/pdf/2310.20033v2
2023-11-03T13:49:16
2023-10-30T21:33:22
{ "title": [ -0.012710571, -0.0110321045, -0.01826477, -0.06854248, -0.039367676, 0.013809204, -0.07849121, -0.012687683, -0.021652222, 0.06951904, 0.013343811, 0.003929138, 0.020523071, -0.011528015, -0.0446167, 0.0017051697, -0.004524231, 0.017...
http://arxiv.org/abs/2310.20003v1
Early Detection of Depression and Eating Disorders in Spanish: UNSL at MentalRiskES 2023
Horacio Thompson, Marcelo Errecalde
cs.CL
MentalRiskES is a novel challenge that proposes to solve problems related to early risk detection for the Spanish language. The objective is to detect, as soon as possible, Telegram users who show signs of mental disorders considering different tasks. Task 1 involved the users' detection of eating disorders, Task 2 foc...
http://arxiv.org/pdf/2310.20003v1
2023-10-30T20:38:31
2023-10-30T20:38:31
{ "title": [ -0.001332283, 0.011154175, 0.02204895, 0.0003848076, -0.027954102, 0.01739502, -0.036376953, -0.031921387, -0.008163452, 0.022918701, 0.037261963, -0.038146973, 0.019622803, -0.09112549, 0.017684937, -0.021408081, 0.023483276, 0.0301...
http://arxiv.org/abs/2310.19998v1
Generative retrieval-augmented ontologic graph and multi-agent strategies for interpretive large language model-based materials design
Markus J. Buehler
cs.CL, cond-mat.dis-nn, cond-mat.mes-hall, cond-mat.mtrl-sci, physics.app-ph
Transformer neural networks show promising capabilities, in particular for uses in materials analysis, design and manufacturing, including their capacity to work effectively with both human language, symbols, code, and numerical data. Here we explore the use of large language models (LLMs) as a tool that can support en...
http://arxiv.org/pdf/2310.19998v1
2023-10-30T20:31:50
2023-10-30T20:31:50
{ "title": [ -0.056152344, -0.026931763, -0.046936035, -0.054382324, -0.053131104, 0.010696411, -0.09375, 0.011962891, 0.002023697, 0.018081665, 0.004272461, 0.0039043427, -0.013259888, -0.027175903, -0.014465332, 0.00025820732, -0.018554688, 0.0...
http://arxiv.org/abs/2311.01605v1
Faithful and Robust Local Interpretability for Textual Predictions
Gianluigi Lopardo, Frederic Precioso, Damien Garreau
cs.CL, cs.LG, stat.ML
Interpretability is essential for machine learning models to be trusted and deployed in critical domains. However, existing methods for interpreting text models are often complex, lack solid mathematical foundations, and their performance is not guaranteed. In this paper, we propose FRED (Faithful and Robust Explainer ...
http://arxiv.org/pdf/2311.01605v1
2023-10-30T20:27:36
2023-10-30T20:27:36
{ "title": [ -0.023712158, 0.014785767, -0.019546509, -0.10046387, 0.011726379, 0.037200928, -0.09667969, -0.020050049, 0.03805542, 0.020141602, -0.013580322, -0.01084137, 0.0041656494, 0.0007634163, -0.061340332, 0.007637024, 0.00011330843, 0.03...
http://arxiv.org/abs/2310.19975v2
BioInstruct: Instruction Tuning of Large Language Models for Biomedical Natural Language Processing
Hieu Tran, Zhichao Yang, Zonghai Yao, Hong Yu
cs.CL, cs.AI
To enhance the performance of large language models (LLMs) in biomedical natural language processing (BioNLP) by introducing a domain-specific instruction dataset and examining its impact when combined with multi-task learning principles. We created the BioInstruct, comprising 25,005 instructions to instruction-tune LL...
http://arxiv.org/pdf/2310.19975v2
2023-11-06T15:05:34
2023-10-30T19:38:50
{ "title": [ -0.052124023, -0.028259277, -0.02961731, -0.04864502, -0.02961731, 0.013893127, -0.07861328, -0.00248909, 0.0036830902, 0.017959595, -0.030319214, -0.012664795, 0.0073547363, -0.025665283, -0.033599854, -0.0059776306, -0.0073547363, ...
http://arxiv.org/abs/2310.19970v1
Strategies to Harness the Transformers' Potential: UNSL at eRisk 2023
Horacio Thompson, Leticia Cagnina, Marcelo Errecalde
cs.CL
The CLEF eRisk Laboratory explores solutions to different tasks related to risk detection on the Internet. In the 2023 edition, Task 1 consisted of searching for symptoms of depression, the objective of which was to extract user writings according to their relevance to the BDI Questionnaire symptoms. Task 2 was related...
http://arxiv.org/pdf/2310.19970v1
2023-10-30T19:34:33
2023-10-30T19:34:33
{ "title": [ 0.017944336, -0.002313614, -0.019012451, -0.013122559, -0.013175964, -0.017974854, -0.0010910034, -0.015289307, -0.05886841, 0.034729004, 0.004348755, 0.008872986, 0.030944824, -0.067993164, 0.04928589, -0.025787354, 0.02507019, 0.02...
http://arxiv.org/abs/2311.01468v1
Remember what you did so you know what to do next
Manuel R. Ciosici, Alex Hedges, Yash Kankanampati, Justin Martin, Marjorie Freedman, Ralph Weischedel
cs.CL, cs.LG
We explore using a moderately sized large language model (GPT-J 6B parameters) to create a plan for a simulated robot to achieve 30 classes of goals in ScienceWorld, a text game simulator for elementary science experiments. Previously published empirical work claimed that large language models (LLMs) are a poor fit (Wa...
http://arxiv.org/pdf/2311.01468v1
2023-10-30T19:29:00
2023-10-30T19:29:00
{ "title": [ 0.03753662, 0.0001705885, -0.02897644, -0.101135254, 0.0036258698, -0.0028190613, -0.04800415, 0.03866577, 0.013458252, 0.043182373, 0.0015058517, 0.0036525726, -0.017532349, -0.015106201, -0.039978027, -0.010192871, 0.04171753, -0.0...
http://arxiv.org/abs/2310.19956v1
The Impact of Depth and Width on Transformer Language Model Generalization
Jackson Petty, Sjoerd van Steenkiste, Ishita Dasgupta, Fei Sha, Dan Garrette, Tal Linzen
cs.CL
To process novel sentences, language models (LMs) must generalize compositionally -- combine familiar elements in new ways. What aspects of a model's structure promote compositional generalization? Focusing on transformers, we test the hypothesis, motivated by recent theoretical and empirical work, that transformers ge...
http://arxiv.org/pdf/2310.19956v1
2023-10-30T19:10:06
2023-10-30T19:10:06
{ "title": [ -0.031204224, -0.0011367798, -0.04345703, -0.04837036, -0.002609253, -0.0057678223, -0.072509766, -0.001824379, -0.008033752, 0.042144775, -0.017242432, -0.003627777, 0.012870789, -0.016220093, -0.038085938, -0.0019283295, -0.005695343, ...
http://arxiv.org/abs/2310.19942v1
Split-NER: Named Entity Recognition via Two Question-Answering-based Classifications
Jatin Arora, Youngja Park
cs.CL, cs.IR, cs.LG, I.2.7; H.3.3
In this work, we address the NER problem by splitting it into two logical sub-tasks: (1) Span Detection which simply extracts entity mention spans irrespective of entity type; (2) Span Classification which classifies the spans into their entity types. Further, we formulate both sub-tasks as question-answering (QA) prob...
http://arxiv.org/pdf/2310.19942v1
2023-10-30T18:57:28
2023-10-30T18:57:28
{ "title": [ -0.021957397, -0.0035572052, -0.035125732, -0.07385254, 0.014442444, 0.008758545, -0.033050537, 0.0040130615, 0.011688232, -0.0062026978, -0.042266846, 0.024993896, -0.016647339, -0.007972717, -0.026367188, -0.011528015, -0.009666443, ...
http://arxiv.org/abs/2310.19923v1
Jina Embeddings 2: 8192-Token General-Purpose Text Embeddings for Long Documents
Michael Günther, Jackmin Ong, Isabelle Mohr, Alaeddine Abdessalem, Tanguy Abel, Mohammad Kalim Akram, Susana Guzman, Georgios Mastrapas, Saba Sturua, Bo Wang, Maximilian Werk, Nan Wang, Han Xiao
cs.CL, cs.AI, cs.LG, 68T50, I.2.7
Text embedding models have emerged as powerful tools for transforming sentences into fixed-sized feature vectors that encapsulate semantic information. While these models are essential for tasks like information retrieval, semantic clustering, and text re-ranking, most existing open-source models, especially those buil...
http://arxiv.org/pdf/2310.19923v1
2023-10-30T18:35:30
2023-10-30T18:35:30
{ "title": [ -0.03744507, -0.010673523, -0.066223145, -0.011016846, -0.030853271, -0.032165527, -0.039978027, 0.014022827, -0.020370483, 0.02458191, 0.0012464523, -0.013069153, 0.03074646, -0.05053711, 0.02545166, -0.019744873, 0.004119873, -0.01...
http://arxiv.org/abs/2310.19792v1
The Eval4NLP 2023 Shared Task on Prompting Large Language Models as Explainable Metrics
Christoph Leiter, Juri Opitz, Daniel Deutsch, Yang Gao, Rotem Dror, Steffen Eger
cs.CL
With an increasing number of parameters and pre-training data, generative large language models (LLMs) have shown remarkable capabilities to solve tasks with minimal or no task-related examples. Notably, LLMs have been successfully employed as evaluation metrics in text generation tasks. Within this context, we introdu...
http://arxiv.org/pdf/2310.19792v1
2023-10-30T17:55:08
2023-10-30T17:55:08
{ "title": [ -0.02709961, -0.018295288, -0.025634766, -0.05529785, -0.020523071, 0.008460999, -0.06500244, -0.009223938, -0.0154800415, -0.009750366, 0.012916565, -0.01625061, -0.0082092285, -0.057373047, 0.014846802, 0.0141067505, -0.0043640137, ...
http://arxiv.org/abs/2310.19791v1
LILO: Learning Interpretable Libraries by Compressing and Documenting Code
Gabriel Grand, Lionel Wong, Matthew Bowers, Theo X. Olausson, Muxin Liu, Joshua B. Tenenbaum, Jacob Andreas
cs.CL, cs.AI, cs.LG, cs.PL
While large language models (LLMs) now excel at code generation, a key aspect of software development is the art of refactoring: consolidating code into libraries of reusable and readable programs. In this paper, we introduce LILO, a neurosymbolic framework that iteratively synthesizes, compresses, and documents code t...
http://arxiv.org/pdf/2310.19791v1
2023-10-30T17:55:02
2023-10-30T17:55:02
{ "title": [ -0.06451416, -0.014816284, -0.058013916, -0.041168213, -0.022415161, 0.0077705383, -0.055389404, 0.01486969, 0.003124237, 0.0059280396, -0.019973755, 0.0037059784, -0.030670166, -0.02267456, -0.027740479, -0.026367188, 0.0039596558, ...
http://arxiv.org/abs/2310.19785v1
What's "up" with vision-language models? Investigating their struggle with spatial reasoning
Amita Kamath, Jack Hessel, Kai-Wei Chang
cs.CL, cs.CV, cs.LG
Recent vision-language (VL) models are powerful, but can they reliably distinguish "right" from "left"? We curate three new corpora to quantify model comprehension of such basic spatial relations. These tests isolate spatial reasoning more precisely than existing datasets like VQAv2, e.g., our What'sUp benchmark contai...
http://arxiv.org/pdf/2310.19785v1
2023-10-30T17:50:15
2023-10-30T17:50:15
{ "title": [ -0.036468506, -0.03152466, -0.02494812, -0.019424438, -0.02558899, -0.01335144, -0.016616821, -0.0070152283, 0.030593872, 0.026641846, -0.040405273, -0.018478394, 0.027999878, -0.014175415, -0.054840088, -0.0037994385, 0.0018606186, ...
http://arxiv.org/abs/2310.19750v1
Chain-of-Thought Embeddings for Stance Detection on Social Media
Joseph Gatto, Omar Sharif, Sarah Masud Preum
cs.CL
Stance detection on social media is challenging for Large Language Models (LLMs), as emerging slang and colloquial language in online conversations often contain deeply implicit stance labels. Chain-of-Thought (COT) prompting has recently been shown to improve performance on stance detection tasks -- alleviating some o...
http://arxiv.org/pdf/2310.19750v1
2023-10-30T17:18:10
2023-10-30T17:18:10
{ "title": [ 0.013336182, 0.008308411, -0.007873535, -0.045166016, 0.0028305054, -0.01210022, -0.023773193, -0.022079468, 0.030883789, 0.018661499, -0.05291748, 0.011177063, -0.008018494, -0.0017356873, -0.043182373, -0.007709503, -0.009597778, 0...
http://arxiv.org/abs/2310.19740v1
Collaborative Evaluation: Exploring the Synergy of Large Language Models and Humans for Open-ended Generation Evaluation
Qintong Li, Leyang Cui, Lingpeng Kong, Wei Bi
cs.CL
Humans are widely involved in the evaluation of open-ended natural language generation tasks (NLG) that demand creativity, as automatic metrics often exhibit weak correlations with human judgments. Large language models (LLMs) recently have emerged as a scalable and cost-effective alternative to human evaluations. Howe...
http://arxiv.org/pdf/2310.19740v1
2023-10-30T17:04:35
2023-10-30T17:04:35
{ "title": [ -0.027770996, 0.006454468, -0.037017822, -0.052642822, 0.0020751953, -0.010116577, -0.058776855, -0.015304565, 0.0022258759, 0.030532837, -0.00472641, -0.006866455, -0.024368286, -0.022705078, -0.05819702, -0.00869751, -0.043426514, ...
http://arxiv.org/abs/2310.19736v2
Evaluating Large Language Models: A Comprehensive Survey
Zishan Guo, Renren Jin, Chuang Liu, Yufei Huang, Dan Shi, Supryadi, Linhao Yu, Yan Liu, Jiaxuan Li, Bojian Xiong, Deyi Xiong
cs.CL, cs.AI
Large language models (LLMs) have demonstrated remarkable capabilities across a broad spectrum of tasks. They have attracted significant attention and been deployed in numerous downstream applications. Nevertheless, akin to a double-edged sword, LLMs also present potential risks. They could suffer from private data lea...
http://arxiv.org/pdf/2310.19736v2
2023-10-31T05:00:28
2023-10-30T17:00:52
{ "title": [ -0.03930664, -0.0134887695, -0.04284668, -0.06707764, 0.0087890625, 0.0135269165, -0.068725586, 0.01285553, 0.007762909, 0.031677246, -0.0104599, -0.012413025, 0.008293152, -0.016830444, -0.05847168, 0.011932373, -0.023376465, 0.0404...
http://arxiv.org/abs/2310.19727v1
Generating Medical Instructions with Conditional Transformer
Samuel Belkadi, Nicolo Micheletti, Lifeng Han, Warren Del-Pinto, Goran Nenadic
cs.CL, cs.AI, cs.LG
Access to real-world medical instructions is essential for medical research and healthcare quality improvement. However, access to real medical instructions is often limited due to the sensitive nature of the information expressed. Additionally, manually labelling these instructions for training and fine-tuning Natural...
http://arxiv.org/pdf/2310.19727v1
2023-10-30T16:53:11
2023-10-30T16:53:11
{ "title": [ -0.017333984, 0.026611328, -0.05432129, -0.074157715, 0.018737793, 0.018096924, -0.085876465, -0.013572693, -0.003200531, 0.04827881, -0.0149383545, -0.0020503998, 0.030792236, -0.032836914, -0.022033691, 0.011291504, 0.020858765, 0....
http://arxiv.org/abs/2310.19708v3
Combining Language Models For Specialized Domains: A Colorful Approach
Daniel Eitan, Menachem Pirchi, Neta Glazer, Shai Meital, Gil Ayach, Gidon Krendel, Aviv Shamsian, Aviv Navon, Gil Hetz, Joseph Keshet
cs.CL, cs.LG
General purpose language models (LMs) encounter difficulties when processing domain-specific jargon and terminology, which are frequently utilized in specialized fields such as medicine or industrial settings. Moreover, they often find it challenging to interpret mixed speech that blends general language with specializ...
http://arxiv.org/pdf/2310.19708v3
2023-11-01T07:55:28
2023-10-30T16:35:55
{ "title": [ -0.025131226, -0.0003182888, -0.029556274, -0.09680176, 0.009002686, 0.021102905, -0.09716797, 0.0011253357, -0.00844574, 0.043029785, 0.012535095, -0.0024375916, 0.011421204, -0.022201538, -0.043151855, 0.018997192, -0.017196655, 0....
http://arxiv.org/abs/2310.19698v1
When Do Prompting and Prefix-Tuning Work? A Theory of Capabilities and Limitations
Aleksandar Petrov, Philip H. S. Torr, Adel Bibi
cs.LG, cs.CL
Context-based fine-tuning methods, including prompting, in-context learning, soft prompting (also known as prompt tuning), and prefix-tuning, have gained popularity due to their ability to often match the performance of full fine-tuning with a fraction of the parameters. Despite their empirical successes, there is litt...
http://arxiv.org/pdf/2310.19698v1
2023-10-30T16:19:34
2023-10-30T16:19:34
{ "title": [ -0.026565552, -0.033721924, -0.0005097389, -0.047180176, -0.02330017, 0.014228821, -0.038604736, -0.009544373, -0.05340576, 0.013549805, 0.0007762909, 0.004173279, -0.03326416, -0.04626465, -0.019058228, -0.01966858, -0.00029540062, ...
http://arxiv.org/abs/2310.19687v1
Sentiment Analysis in Digital Spaces: An Overview of Reviews
Laura E. M. Ayravainen, Joanne Hinds, Brittany I. Davidson
cs.CY, cs.CL, I.2.7
Sentiment analysis (SA) is commonly applied to digital textual data, revealing insight into opinions and feelings. Many systematic reviews have summarized existing work, but often overlook discussions of validity and scientific practices. Here, we present an overview of reviews, synthesizing 38 systematic reviews, cont...
http://arxiv.org/pdf/2310.19687v1
2023-10-30T16:04:35
2023-10-30T16:04:35
{ "title": [ -0.009864807, -0.010864258, -0.012825012, -0.02166748, 0.004573822, -0.014808655, -0.0058784485, -0.027297974, 0.008743286, 0.039093018, -0.0395813, 0.0025615692, -0.0032653809, -0.0042419434, -0.06750488, -0.0017538071, -0.005847931, ...
http://arxiv.org/abs/2310.19680v2
Integrating Pre-trained Language Model into Neural Machine Translation
Soon-Jae Hwang, Chang-Sung Jeong
cs.CL, cs.AI, cs.LG
Neural Machine Translation (NMT) has become a significant technology in natural language processing through extensive research and development. However, the deficiency of high-quality bilingual language pair data still poses a major challenge to improving NMT performance. Recent studies are exploring the use of context...
http://arxiv.org/pdf/2310.19680v2
2023-11-13T16:12:02
2023-10-30T16:00:13
{ "title": [ -0.012817383, 0.0010509491, -0.012718201, -0.06008911, 0.00021493435, 0.009841919, -0.04421997, -0.0032176971, -0.0107803345, 0.04647827, 0.00076818466, -0.00047636032, 0.019180298, -0.02027893, -0.038116455, 0.02204895, 0.000033140182, ...
http://arxiv.org/abs/2310.19677v2
MoCa: Measuring Human-Language Model Alignment on Causal and Moral Judgment Tasks
Allen Nie, Yuhui Zhang, Atharva Amdekar, Chris Piech, Tatsunori Hashimoto, Tobias Gerstenberg
cs.CL
Human commonsense understanding of the physical and social world is organized around intuitive theories. These theories support making causal and moral judgments. When something bad happens, we naturally ask: who did what, and why? A rich literature in cognitive science has studied people's causal and moral intuitions....
http://arxiv.org/pdf/2310.19677v2
2023-10-31T06:36:13
2023-10-30T15:57:32
{ "title": [ -0.030731201, -0.009796143, -0.05883789, -0.027053833, -0.011528015, -0.024902344, -0.034240723, -0.0017375946, 0.0007252693, 0.027709961, -0.0018672943, -0.0044670105, 0.019714355, -0.025970459, -0.040283203, 0.0030784607, -0.022521973,...
http://arxiv.org/abs/2310.19671v2
Large Language Models: The Need for Nuance in Current Debates and a Pragmatic Perspective on Understanding
Bram M. A. van Dijk, Tom Kouwenhoven, Marco R. Spruit, Max J. van Duijn
cs.CL, cs.AI
Current Large Language Models (LLMs) are unparalleled in their ability to generate grammatically correct, fluent text. LLMs are appearing rapidly, and debates on LLM capacities have taken off, but reflection is lagging behind. Thus, in this position paper, we first zoom in on the debate and critically assess three poin...
http://arxiv.org/pdf/2310.19671v2
2023-10-31T08:17:29
2023-10-30T15:51:04
{ "title": [ -0.026504517, -0.013160706, -0.042388916, -0.0715332, 0.0047187805, -0.002817154, -0.05239868, 0.0066490173, 0.004005432, 0.018081665, -0.012542725, -0.019332886, 0.002462387, -0.032806396, -0.042907715, 0.00026798248, -0.012054443, ...
http://arxiv.org/abs/2310.19660v1
Interpretable-by-Design Text Classification with Iteratively Generated Concept Bottleneck
Josh Magnus Ludan, Qing Lyu, Yue Yang, Liam Dugan, Mark Yatskar, Chris Callison-Burch
cs.CL
Deep neural networks excel in text classification tasks, yet their application in high-stakes domains is hindered by their lack of interpretability. To address this, we propose Text Bottleneck Models (TBMs), an intrinsically interpretable text classification framework that offers both global and local explanations. Rat...
http://arxiv.org/pdf/2310.19660v1
2023-10-30T15:41:32
2023-10-30T15:41:32
{ "title": [ -0.035217285, 0.017593384, -0.04385376, -0.028518677, 0.00182724, -0.0107421875, -0.04260254, -0.021835327, 0.013145447, 0.010940552, -0.05053711, 0.03161621, -0.04043579, 0.01386261, -0.052734375, -0.037200928, -0.022247314, 0.06262...
http://arxiv.org/abs/2310.19658v1
Explaining Tree Model Decisions in Natural Language for Network Intrusion Detection
Noah Ziems, Gang Liu, John Flanagan, Meng Jiang
cs.CL, cs.AI
Network intrusion detection (NID) systems which leverage machine learning have been shown to have strong performance in practice when used to detect malicious network traffic. Decision trees in particular offer a strong balance between performance and simplicity, but require users of NID systems to have background know...
http://arxiv.org/pdf/2310.19658v1
2023-10-30T15:40:34
2023-10-30T15:40:34
{ "title": [ -0.04876709, -0.022262573, -0.023666382, -0.03866577, -0.03945923, 0.014587402, -0.05053711, 0.0071907043, -0.016708374, -0.0071754456, -0.055786133, 0.010169983, -0.03842163, -0.015205383, -0.014717102, -0.048187256, -0.025390625, 0...
http://arxiv.org/abs/2310.19651v1
Dynamics of Instruction Tuning: Each Ability of Large Language Models Has Its Own Growth Pace
Chiyu Song, Zhanchao Zhou, Jianhao Yan, Yuejiao Fei, Zhenzhong Lan, Yue Zhang
cs.CL
Instruction tuning is a burgeoning method to elicit the general intelligence of Large Language Models (LLMs). However, the creation of instruction data is still largely heuristic, leading to significant variation in quality and distribution across existing datasets. Experimental conclusions drawn from these datasets ar...
http://arxiv.org/pdf/2310.19651v1
2023-10-30T15:37:10
2023-10-30T15:37:10
{ "title": [ -0.059936523, -0.008415222, -0.06677246, -0.007663727, -0.03994751, -0.033203125, -0.046051025, 0.024124146, -0.0541687, 0.031829834, -0.04144287, 0.0362854, -0.024612427, -0.013137817, -0.0051994324, -0.037109375, 0.011878967, 0.033...
http://arxiv.org/abs/2310.19650v1
KeyGen2Vec: Learning Document Embedding via Multi-label Keyword Generation in Question-Answering
Iftitahu Ni'mah, Samaneh Khoshrou, Vlado Menkovski, Mykola Pechenizkiy
cs.CL
Representing documents into high dimensional embedding space while preserving the structural similarity between document sources has been an ultimate goal for many works on text representation learning. Current embedding models, however, mainly rely on the availability of label supervision to increase the expressivenes...
http://arxiv.org/pdf/2310.19650v1
2023-10-30T15:35:45
2023-10-30T15:35:45
{ "title": [ -0.019760132, -0.017791748, -0.023590088, -0.037628174, -0.018127441, 0.0010299683, -0.011375427, 0.00932312, 0.0033931732, 0.020065308, -0.017944336, 0.023986816, -0.0146484375, -0.0040664673, -0.038757324, -0.00027823448, -0.007762909,...
http://arxiv.org/abs/2310.19619v1
Towards A Holistic Landscape of Situated Theory of Mind in Large Language Models
Ziqiao Ma, Jacob Sansom, Run Peng, Joyce Chai
cs.CL, cs.AI
Large Language Models (LLMs) have generated considerable interest and debate regarding their potential emergence of Theory of Mind (ToM). Several recent inquiries reveal a lack of robust ToM in these models and pose a pressing demand to develop new benchmarks, as current ones primarily focus on different aspects of ToM...
http://arxiv.org/pdf/2310.19619v1
2023-10-30T15:12:09
2023-10-30T15:12:09
{ "title": [ -0.004863739, -0.026382446, -0.006965637, -0.057250977, -0.007194519, -0.0013179779, -0.018035889, 0.0038433075, -0.005970001, 0.024978638, -0.0132369995, -0.0038547516, 0.001876831, -0.021377563, -0.043304443, 0.0016517639, -0.03100586,...
http://arxiv.org/abs/2310.19596v2
LLMaAA: Making Large Language Models as Active Annotators
Ruoyu Zhang, Yanzeng Li, Yongliang Ma, Ming Zhou, Lei Zou
cs.CL, cs.AI
Prevalent supervised learning methods in natural language processing (NLP) are notoriously data-hungry, which demand large amounts of high-quality annotated data. In practice, acquiring such data is a costly endeavor. Recently, the superior few-shot performance of large language models (LLMs) has propelled the developm...
http://arxiv.org/pdf/2310.19596v2
2023-10-31T08:19:52
2023-10-30T14:54:15
{ "title": [ -0.02192688, -0.006046295, -0.025619507, -0.085510254, 0.0066375732, 0.015487671, -0.05645752, 0.008102417, 0.01473999, 0.03933716, -0.0013093948, -0.00056552887, 0.00497818, -0.01574707, -0.046203613, 0.025619507, -0.006263733, 0.03...
http://arxiv.org/abs/2310.19588v1
DPATD: Dual-Phase Audio Transformer for Denoising
Junhui Li, Pu Wang, Jialu Li, Xinzhe Wang, Youshan Zhang
cs.SD, cs.CL, eess.AS
Recent high-performance transformer-based speech enhancement models demonstrate that time domain methods could achieve similar performance as time-frequency domain methods. However, time-domain speech enhancement systems typically receive input audio sequences consisting of a large number of time steps, making it chall...
http://arxiv.org/pdf/2310.19588v1
2023-10-30T14:44:59
2023-10-30T14:44:59
{ "title": [ 0.007659912, 0.02178955, -0.049957275, -0.04840088, 0.010948181, -0.021575928, -0.0070533752, -0.025054932, -0.009811401, 0.052490234, -0.0063552856, -0.009391785, 0.06402588, -0.040161133, -0.02015686, 0.00573349, 0.0005326271, 0.02...
http://arxiv.org/abs/2310.19572v1
Improving Input-label Mapping with Demonstration Replay for In-context Learning
Zhuocheng Gong, Jiahao Liu, Qifan Wang, Jingang Wang, Xunliang Cai, Dongyan Zhao, Rui Yan
cs.CL
In-context learning (ICL) is an emerging capability of large autoregressive language models where a few input-label demonstrations are appended to the input to enhance the model's understanding of downstream NLP tasks, without directly adjusting the model parameters. The effectiveness of ICL can be attributed to the st...
http://arxiv.org/pdf/2310.19572v1
2023-10-30T14:29:41
2023-10-30T14:29:41
{ "title": [ 0.013069153, -0.009300232, 0.0028133392, -0.039367676, -0.011436462, -0.014251709, -0.008049011, -0.025344849, -0.017578125, 0.032073975, -0.01966858, 0.02835083, -0.010917664, -0.014854431, -0.014419556, -0.013366699, -0.020050049, ...
http://arxiv.org/abs/2310.19567v1
CreoleVal: Multilingual Multitask Benchmarks for Creoles
Heather Lent, Kushal Tatariya, Raj Dabre, Yiyi Chen, Marcell Fekete, Esther Ploeger, Li Zhou, Hans Erik Heje, Diptesh Kanojia, Paul Belony, Marcel Bollmann, Loïc Grobol, Miryam de Lhoneux, Daniel Hershcovich, Michel DeGraff, Anders Søgaard, Johannes Bjerva
cs.CL, cs.AI
Creoles represent an under-explored and marginalized group of languages, with few available resources for NLP research. While the genealogical ties between Creoles and other highly-resourced languages imply a significant potential for transfer learning, this potential is hampered due to this lack of annotated data. In ...
http://arxiv.org/pdf/2310.19567v1
2023-10-30T14:24:20
2023-10-30T14:24:20
{ "title": [ -0.044555664, -0.022155762, -0.025939941, -0.06262207, -0.009460449, 0.017501831, -0.06161499, 0.020950317, -0.013412476, 0.043884277, -0.018920898, 0.021774292, 0.005203247, -0.016159058, -0.03866577, 0.0073013306, -0.00907135, 0.04...
http://arxiv.org/abs/2310.19539v1
A Novel Representation to Improve Team Problem Solving in Real-Time
Alex Doboli
cs.CL
This paper proposes a novel representation to support computing metrics that help understanding and improving in real-time a team's behavior during problem solving in real-life. Even though teams are important in modern activities, there is little computing aid to improve their activity. The representation captures the...
http://arxiv.org/pdf/2310.19539v1
2023-10-30T13:46:24
2023-10-30T13:46:24
{ "title": [ -0.00491333, 0.0027542114, -0.014915466, -0.032043457, 0.017166138, -0.02128601, -0.000040769577, -0.018997192, 0.007972717, 0.031951904, -0.016723633, -0.0048980713, -0.025054932, -0.0055007935, -0.050689697, -0.016723633, 0.014640808, ...
http://arxiv.org/abs/2310.19531v3
InfoEntropy Loss to Mitigate Bias of Learning Difficulties for Generative Language Models
Zhenpeng Su, Xing Wu, Xue Bai, Zijia Lin, Hui Chen, Guiguang Ding, Wei Zhou, Songlin Hu
cs.CL
Generative language models are usually pretrained on large text corpus via predicting the next token (i.e., sub-word/word/phrase) given the previous ones. Recent works have demonstrated the impressive performance of large generative language models on downstream tasks. However, existing generative language models gener...
http://arxiv.org/pdf/2310.19531v3
2023-11-10T09:35:30
2023-10-30T13:33:21
{ "title": [ -0.020217896, -0.011940002, -0.044891357, -0.03152466, -0.0287323, -0.012710571, -0.038085938, -0.0034999847, 0.024337769, 0.029342651, -0.03475952, 0.018295288, 0.03466797, -0.017227173, -0.026473999, 0.02911377, -0.018203735, 0.037...
http://arxiv.org/abs/2310.19462v2
Constituency Parsing using LLMs
Xuefeng Bai, Jialong Wu, Yulong Chen, Zhongqing Wang, Yue Zhang
cs.CL
Constituency parsing is a fundamental yet unsolved natural language processing task. In this paper, we explore the potential of recent large language models (LLMs) that have exhibited remarkable performance across various domains and tasks to tackle this task. We employ three linearization strategies to transform outpu...
http://arxiv.org/pdf/2310.19462v2
2023-10-31T07:19:51
2023-10-30T11:39:11
{ "title": [ -0.01852417, -0.0019550323, -0.0143585205, -0.040527344, -0.041778564, 0.013755798, -0.078552246, -0.01423645, -0.02848816, 0.029037476, -0.038879395, 0.00032949448, -0.007575989, -0.035186768, -0.0052261353, -0.014007568, 0.00015926361,...
http://arxiv.org/abs/2311.01606v1
KG-FRUS: a Novel Graph-based Dataset of 127 Years of US Diplomatic Relations
Gökberk Özsoy, Luis Salamanca, Matthew Connelly, Raymond Hicks, Fernando Pérez-Cruz
cs.CL
In the current paper, we present the KG-FRUS dataset, comprised of more than 300,000 US government diplomatic documents encoded in a Knowledge Graph (KG). We leverage the data of the Foreign Relations of the United States (FRUS) (available as XML files) to extract information about the documents and the individuals and...
http://arxiv.org/pdf/2311.01606v1
2023-10-30T10:53:02
2023-10-30T10:53:02
{ "title": [ 0.019729614, -0.009490967, -0.025802612, -0.039764404, 0.0011262894, -0.027435303, -0.010948181, -0.006843567, 0.0049552917, 0.011108398, -0.03488159, 0.009086609, -0.02796936, -0.024963379, -0.035369873, -0.015853882, -0.00283432, 0...
http://arxiv.org/abs/2310.19420v1
Mean BERTs make erratic language teachers: the effectiveness of latent bootstrapping in low-resource settings
David Samuel
cs.CL, cs.LG
This paper explores the use of latent bootstrapping, an alternative self-supervision technique, for pretraining language models. Unlike the typical practice of using self-supervision on discrete subwords, latent bootstrapping leverages contextualized embeddings for a richer supervision signal. We conduct experiments to...
http://arxiv.org/pdf/2310.19420v1
2023-10-30T10:31:32
2023-10-30T10:31:32
{ "title": [ -0.031021118, -0.00995636, -0.017120361, -0.001074791, -0.029632568, -0.015792847, -0.024429321, -0.02079773, -0.048736572, 0.0020561218, -0.007030487, 0.03616333, -0.01474762, -0.04296875, 0.016952515, -0.009689331, -0.02923584, 0.0...
http://arxiv.org/abs/2310.19403v1
A Lightweight Method to Generate Unanswerable Questions in English
Vagrant Gautam, Miaoran Zhang, Dietrich Klakow
cs.CL, I.2.7
If a question cannot be answered with the available information, robust systems for question answering (QA) should know _not_ to answer. One way to build QA models that do this is with additional training data comprised of unanswerable questions, created either by employing annotators or through automated methods for u...
http://arxiv.org/pdf/2310.19403v1
2023-10-30T10:14:52
2023-10-30T10:14:52
{ "title": [ -0.0073509216, -0.013633728, -0.0289917, -0.022842407, -0.03945923, -0.011123657, -0.014541626, -0.018615723, 0.0030670166, 0.018997192, 0.0013275146, -0.00983429, -0.011543274, -0.042663574, -0.027755737, 0.005214691, 0.0132751465, ...
http://arxiv.org/abs/2310.19349v1
Japanese SimCSE Technical Report
Hayato Tsukagoshi, Ryohei Sasano, Koichi Takeda
cs.CL
We report the development of Japanese SimCSE, Japanese sentence embedding models fine-tuned with SimCSE. Since there is a lack of sentence embedding models for Japanese that can be used as a baseline in sentence embedding research, we conducted extensive experiments on Japanese sentence embeddings involving 24 pre-trai...
http://arxiv.org/pdf/2310.19349v1
2023-10-30T08:43:26
2023-10-30T08:43:26
{ "title": [ 0.00995636, -0.011650085, 0.012214661, -0.031921387, -0.0259552, 0.011039734, -0.046051025, -0.025726318, -0.03765869, 0.06121826, 0.017150879, -0.041107178, 0.03942871, -0.041503906, -0.008163452, -0.009513855, 0.007167816, 0.037811...
http://arxiv.org/abs/2310.19347v3
Improving Factual Consistency of Text Summarization by Adversarially Decoupling Comprehension and Embellishment Abilities of LLMs
Huawen Feng, Yan Fan, Xiong Liu, Ting-En Lin, Zekun Yao, Yuchuan Wu, Fei Huang, Yongbin Li, Qianli Ma
cs.CL, cs.AI
Despite the recent progress in text summarization made by large language models (LLMs), they often generate summaries that are factually inconsistent with original articles, known as "hallucinations" in text generation. Unlike previous small models (e.g., BART, T5), current LLMs make fewer silly mistakes but more sophi...
http://arxiv.org/pdf/2310.19347v3
2023-11-14T06:55:56
2023-10-30T08:40:16
{ "title": [ -0.017822266, 0.014198303, -0.021148682, -0.01751709, -0.031433105, 0.004119873, -0.048706055, -0.035583496, -0.025222778, 0.05203247, -0.003376007, 0.02607727, 0.014709473, -0.013587952, -0.037963867, -0.007949829, -0.021484375, 0.0...
http://arxiv.org/abs/2310.19345v1
Test Suites Task: Evaluation of Gender Fairness in MT with MuST-SHE and INES
Beatrice Savoldi, Marco Gaido, Matteo Negri, Luisa Bentivogli
cs.CL
As part of the WMT-2023 "Test suites" shared task, in this paper we summarize the results of two test suites evaluations: MuST-SHE-WMT23 and INES. By focusing on the en-de and de-en language pairs, we rely on these newly created test suites to investigate systems' ability to translate feminine and masculine gender and ...
http://arxiv.org/pdf/2310.19345v1
2023-10-30T08:36:30
2023-10-30T08:36:30
{ "title": [ -0.04397583, -0.004760742, -0.07922363, -0.02217102, -0.008773804, -0.022964478, -0.009384155, 0.042877197, 0.009086609, 0.05178833, -0.034973145, -0.0022583008, 0.053466797, -0.0154800415, -0.022888184, -0.015838623, -0.03692627, 0....
http://arxiv.org/abs/2310.19341v1
Skywork: A More Open Bilingual Foundation Model
Tianwen Wei, Liang Zhao, Lichang Zhang, Bo Zhu, Lijie Wang, Haihua Yang, Biye Li, Cheng Cheng, Weiwei Lü, Rui Hu, Chenxia Li, Liu Yang, Xilin Luo, Xuejie Wu, Lunan Liu, Wenjun Cheng, Peng Cheng, Jianhao Zhang, Xiaoyu Zhang, Lei Lin, Xiaokun Wang, Yutuan Ma, Chuanhai Dong, Yanqi Sun, Yifu Chen, Yongyi Peng, Xiaojuan Lia...
cs.CL, cs.AI
In this technical report, we present Skywork-13B, a family of large language models (LLMs) trained on a corpus of over 3.2 trillion tokens drawn from both English and Chinese texts. This bilingual foundation model is the most extensively trained and openly published LLMs of comparable size to date. We introduce a two-s...
http://arxiv.org/pdf/2310.19341v1
2023-10-30T08:31:47
2023-10-30T08:31:47
{ "title": [ -0.030395508, -0.019699097, -0.05706787, -0.04763794, -0.031143188, -0.005554199, -0.045440674, 0.009796143, -0.034210205, 0.032165527, -0.0026474, -0.017959595, 0.015159607, -0.052459717, -0.0037708282, -0.03086853, -0.007572174, 0....
http://arxiv.org/abs/2310.19301v1
ROME: Evaluating Pre-trained Vision-Language Models on Reasoning beyond Visual Common Sense
Kankan Zhou, Eason Lai, Wei Bin Au Yeong, Kyriakos Mouratidis, Jing Jiang
cs.CL, cs.AI, cs.CV
Humans possess a strong capability for reasoning beyond common sense. For example, given an unconventional image of a goldfish laying on the table next to an empty fishbowl, a human would effortlessly determine that the fish is not inside the fishbowl. The case, however, may be different for a vision-language model, wh...
http://arxiv.org/pdf/2310.19301v1
2023-10-30T06:35:37
2023-10-30T06:35:37
{ "title": [ -0.017425537, -0.03149414, -0.035980225, -0.03945923, -0.02407837, -0.021194458, -0.040802002, -0.00983429, 0.016235352, 0.009544373, -0.022903442, -0.022918701, 0.002313614, -0.02545166, -0.046051025, -0.019500732, 0.0126571655, 0.0...
http://arxiv.org/abs/2310.19292v1
Fusing Temporal Graphs into Transformers for Time-Sensitive Question Answering
Xin Su, Phillip Howard, Nagib Hakim, Steven Bethard
cs.CL
Answering time-sensitive questions from long documents requires temporal reasoning over the times in questions and documents. An important open question is whether large language models can perform such reasoning solely using a provided text document, or whether they can benefit from additional temporal information ext...
http://arxiv.org/pdf/2310.19292v1
2023-10-30T06:12:50
2023-10-30T06:12:50
{ "title": [ 0.019302368, -0.0069885254, -0.026489258, -0.084350586, 0.003036499, -0.008399963, -0.043762207, 0.015159607, 0.01864624, 0.025878906, -0.03225708, 0.00042319298, -0.014793396, -0.002954483, -0.07366943, -0.014251709, -0.00032949448, ...
http://arxiv.org/abs/2310.19271v1
Learning to love diligent trolls: Accounting for rater effects in the dialogue safety task
Michael John Ilagan
cs.CL
Chatbots have the risk of generating offensive utterances, which must be avoided. Post-deployment, one way for a chatbot to continuously improve is to source utterance/label pairs from feedback by live users. However, among users are trolls, who provide training examples with incorrect labels. To de-troll training data...
http://arxiv.org/pdf/2310.19271v1
2023-10-30T05:08:23
2023-10-30T05:08:23
{ "title": [ 0.006034851, 0.012016296, -0.039611816, -0.046417236, 0.0028896332, -0.029281616, -0.0121536255, -0.0043907166, -0.004928589, 0.008605957, -0.046875, 0.014183044, -0.008125305, -0.016098022, -0.025405884, -0.018127441, -0.027359009, ...
http://arxiv.org/abs/2310.19268v1
Moral Judgments in Narratives on Reddit: Investigating Moral Sparks via Social Commonsense and Linguistic Signals
Ruijie Xi, Munindar P. Singh
cs.SI, cs.CL, cs.CY
Given the increasing realism of social interactions online, social media offers an unprecedented avenue to evaluate real-life moral scenarios. We examine posts from Reddit, where authors and commenters share their moral judgments on who is blameworthy. We employ computational techniques to investigate factors influenci...
http://arxiv.org/pdf/2310.19268v1
2023-10-30T05:03:26
2023-10-30T05:03:26
{ "title": [ 0.02116394, 0.0099487305, -0.00034356117, -0.0043640137, -0.004814148, -0.031311035, 0.030151367, -0.041107178, -0.0023956299, 0.03744507, -0.026535034, 0.025878906, 0.00066947937, 0.0008993149, -0.036254883, 0.005466461, -0.023284912, ...
http://arxiv.org/abs/2310.19267v1
Overview of the CLAIMSCAN-2023: Uncovering Truth in Social Media through Claim Detection and Identification of Claim Spans
Megha Sundriyal, Md Shad Akhtar, Tanmoy Chakraborty
cs.CL
A significant increase in content creation and information exchange has been made possible by the quick development of online social media platforms, which has been very advantageous. However, these platforms have also become a haven for those who disseminate false information, propaganda, and fake news. Claims are ess...
http://arxiv.org/pdf/2310.19267v1
2023-10-30T04:57:41
2023-10-30T04:57:41
{ "title": [ -0.009963989, -0.013404846, -0.049560547, -0.0039520264, -0.022476196, -0.031280518, -0.0023040771, 0.010948181, -0.05731201, 0.020263672, -0.061401367, 0.0435791, 0.023361206, -0.031188965, 0.055389404, -0.03439331, -0.002111435, 0....
http://arxiv.org/abs/2310.19240v1
M4LE: A Multi-Ability Multi-Range Multi-Task Multi-Domain Long-Context Evaluation Benchmark for Large Language Models
Wai-Chung Kwan, Xingshan Zeng, Yufei Wang, Yusen Sun, Liangyou Li, Lifeng Shang, Qun Liu, Kam-Fai Wong
cs.CL
Managing long sequences has become an important and necessary feature for large language models (LLMs). However, it is still an open question of how to comprehensively and systematically evaluate the long-sequence capability of LLMs. One of the reasons is that conventional and widely-used benchmarks mainly consist of s...
http://arxiv.org/pdf/2310.19240v1
2023-10-30T03:11:30
2023-10-30T03:11:30
{ "title": [ -0.027282715, 0.00074768066, -0.05331421, -0.07269287, 0.0069618225, -0.012832642, -0.0569458, 0.01979065, 0.00018751621, 0.04989624, 0.015701294, 0.01473999, 0.015563965, -0.024810791, -0.043762207, 0.024765015, -0.03555298, 0.03976...
http://arxiv.org/abs/2310.19233v3
Building Real-World Meeting Summarization Systems using Large Language Models: A Practical Perspective
Md Tahmid Rahman Laskar, Xue-Yong Fu, Cheng Chen, Shashi Bhushan TN
cs.CL
This paper studies how to effectively build meeting summarization systems for real-world usage using large language models (LLMs). For this purpose, we conduct an extensive evaluation and comparison of various closed-source and open-source LLMs, namely, GPT-4, GPT- 3.5, PaLM-2, and LLaMA-2. Our findings reveal that mos...
http://arxiv.org/pdf/2310.19233v3
2023-11-08T02:23:28
2023-10-30T02:25:21
{ "title": [ -0.019424438, -0.0049972534, -0.055480957, -0.105163574, -0.000046551228, 0.015022278, -0.1003418, 0.00894165, 0.018310547, 0.050476074, -0.008659363, -0.0011901855, 0.028900146, -0.004714966, -0.06021118, 0.009086609, -0.00018370152, ...
http://arxiv.org/abs/2310.19232v1
Adapter Pruning using Tropical Characterization
Rishabh Bhardwaj, Tushar Vaidya, Soujanya Poria
cs.CL
Adapters are widely popular parameter-efficient transfer learning approaches in natural language processing that insert trainable modules in between layers of a pre-trained language model. Apart from several heuristics, however, there has been a lack of studies analyzing the optimal number of adapter parameters needed ...
http://arxiv.org/pdf/2310.19232v1
2023-10-30T02:20:44
2023-10-30T02:20:44
{ "title": [ 0.0054016113, 0.014587402, -0.032836914, -0.02520752, -0.042541504, -0.012893677, -0.0435791, -0.0071411133, -0.0069465637, 0.015525818, 0.00116539, -0.021377563, 0.007545471, -0.068237305, -0.001004219, -0.016464233, 0.027694702, 0....
http://arxiv.org/abs/2310.19212v1
EHRTutor: Enhancing Patient Understanding of Discharge Instructions
Zihao Zhang, Zonghai Yao, Huixue Zhou, Feiyun ouyang, Hong Yu
cs.CL, cs.AI
Large language models have shown success as a tutor in education in various fields. Educating patients about their clinical visits plays a pivotal role in patients' adherence to their treatment plans post-discharge. This paper presents EHRTutor, an innovative multi-component framework leveraging the Large Language Mode...
http://arxiv.org/pdf/2310.19212v1
2023-10-30T00:46:03
2023-10-30T00:46:03
{ "title": [ 0.0027503967, 0.006416321, -0.051452637, -0.09338379, 0.008522034, -0.0007724762, -0.06939697, -0.010635376, -0.0027618408, 0.035614014, -0.021636963, 0.00166893, 0.022369385, -0.023620605, -0.02911377, 0.017196655, 0.007408142, 0.02...
http://arxiv.org/abs/2310.19208v1
LitCab: Lightweight Calibration of Language Models on Outputs of Varied Lengths
Xin Liu, Muhammad Khalifa, Lu Wang
cs.CL
A model is considered well-calibrated when its probability estimate aligns with the actual likelihood of the output being correct. Calibrating language models (LMs) is crucial, as it plays a vital role in detecting and mitigating hallucinations, a common issue of LMs, as well as building more trustworthy models. Yet, p...
http://arxiv.org/pdf/2310.19208v1
2023-10-30T00:30:34
2023-10-30T00:30:34
{ "title": [ -0.06414795, -0.026275635, -0.04647827, -0.052642822, -0.026306152, 0.009880066, -0.092041016, 0.0037117004, -0.004508972, 0.023635864, -0.014678955, 0.0065612793, 0.0040893555, -0.02067566, -0.045898438, 0.016860962, -0.0035743713, ...
http://arxiv.org/abs/2310.19181v1
From Chatbots to PhishBots? -- Preventing Phishing scams created using ChatGPT, Google Bard and Claude
Sayak Saha Roy, Poojitha Thota, Krishna Vamsi Naragam, Shirin Nilizadeh
cs.CR, cs.CL
The advanced capabilities of Large Language Models (LLMs) have made them invaluable across various applications, from conversational agents and content creation to data analysis, research, and innovation. However, their effectiveness and accessibility also render them susceptible to abuse for generating malicious conte...
http://arxiv.org/pdf/2310.19181v1
2023-10-29T22:52:40
2023-10-29T22:52:40
{ "title": [ 0.01914978, -0.0035915375, -0.01713562, -0.0015277863, -0.012969971, -0.025924683, -0.004886627, -0.023361206, -0.051727295, 0.037017822, 0.0037670135, 0.034057617, 0.013282776, -0.037597656, 0.030227661, -0.01687622, -0.02734375, 0....
http://arxiv.org/abs/2310.19177v1
Robustifying Language Models with Test-Time Adaptation
Noah Thomas McDermott, Junfeng Yang, Chengzhi Mao
cs.CL, cs.LG
Large-scale language models achieved state-of-the-art performance over a number of language tasks. However, they fail on adversarial language examples, which are sentences optimized to fool the language models but with similar semantic meanings for humans. While prior work focuses on making the language model robust at...
http://arxiv.org/pdf/2310.19177v1
2023-10-29T22:37:54
2023-10-29T22:37:54
{ "title": [ -0.025665283, -0.014801025, -0.04309082, -0.076416016, 0.00022113323, -0.0010166168, -0.07299805, 0.0054626465, 0.0049438477, 0.035614014, -0.020690918, 0.007045746, -0.008415222, -0.0030784607, -0.06222534, 0.0003619194, -0.013442993, ...
http://arxiv.org/abs/2310.19156v1
Poisoning Retrieval Corpora by Injecting Adversarial Passages
Zexuan Zhong, Ziqing Huang, Alexander Wettig, Danqi Chen
cs.CL, cs.IR
Dense retrievers have achieved state-of-the-art performance in various information retrieval tasks, but to what extent can they be safely deployed in real-world applications? In this work, we propose a novel attack for dense retrieval systems in which a malicious user generates a small number of adversarial passages by...
http://arxiv.org/pdf/2310.19156v1
2023-10-29T21:13:31
2023-10-29T21:13:31
{ "title": [ 0.02696228, -0.0032405853, -0.00092077255, -0.03918457, 0.0030288696, -0.0045394897, 0.022109985, -0.0335083, -0.021896362, 0.03579712, -0.018066406, 0.03326416, -0.004776001, -0.0014839172, -0.029907227, -0.003452301, -0.01058197, 0...
http://arxiv.org/abs/2310.19152v2
BERT Lost Patience Won't Be Robust to Adversarial Slowdown
Zachary Coalson, Gabriel Ritter, Rakesh Bobba, Sanghyun Hong
cs.LG, cs.CL, cs.CR
In this paper, we systematically evaluate the robustness of multi-exit language models against adversarial slowdown. To audit their robustness, we design a slowdown attack that generates natural adversarial text bypassing early-exit points. We use the resulting WAFFLE attack as a vehicle to conduct a comprehensive eval...
http://arxiv.org/pdf/2310.19152v2
2023-10-31T04:19:59
2023-10-29T21:06:34
{ "title": [ 0.0368042, -0.026397705, 0.038970947, 0.0024871826, -0.03540039, 0.014053345, 0.041778564, -0.014274597, -0.035369873, 0.043518066, -0.031707764, 0.013801575, -0.039611816, -0.020065308, -0.018173218, -0.04800415, 0.013168335, 0.0118...
http://arxiv.org/abs/2310.19145v1
Learning to Follow Object-Centric Image Editing Instructions Faithfully
Tuhin Chakrabarty, Kanishk Singh, Arkadiy Saakyan, Smaranda Muresan
cs.CL, cs.CV
Natural language instructions are a powerful interface for editing the outputs of text-to-image diffusion models. However, several challenges need to be addressed: 1) underspecification (the need to model the implicit meaning of instructions) 2) grounding (the need to localize where the edit has to be performed), 3) fa...
http://arxiv.org/pdf/2310.19145v1
2023-10-29T20:39:11
2023-10-29T20:39:11
{ "title": [ 0.010559082, 0.0068626404, -0.023513794, -0.034973145, -0.018203735, -0.025253296, -0.023086548, -0.015014648, -0.0011539459, 0.0395813, -0.028259277, 0.008148193, 0.020477295, -0.019866943, -0.015487671, 0.00844574, 0.017074585, 0.0...
http://arxiv.org/abs/2310.19130v1
Women Wearing Lipstick: Measuring the Bias Between an Object and Its Related Gender
Ahmed Sabir, Lluís Padró
cs.CL, cs.CV
In this paper, we investigate the impact of objects on gender bias in image captioning systems. Our results show that only gender-specific objects have a strong gender bias (e.g., women-lipstick). In addition, we propose a visual semantic-based gender score that measures the degree of bias and can be used as a plug-in ...
http://arxiv.org/pdf/2310.19130v1
2023-10-29T19:39:03
2023-10-29T19:39:03
{ "title": [ -0.03250122, -0.000029861927, -0.07757568, 0.018814087, -0.015319824, -0.04562378, 0.009666443, -0.017547607, -0.02545166, 0.0005245209, -0.044525146, 0.0035076141, -0.0042037964, -0.037231445, 0.015335083, -0.047851562, 0.0033950806, ...
http://arxiv.org/abs/2310.19127v1
Unified Representation for Non-compositional and Compositional Expressions
Ziheng Zeng, Suma Bhat
cs.CL
Accurate processing of non-compositional language relies on generating good representations for such expressions. In this work, we study the representation of language non-compositionality by proposing a language model, PIER, that builds on BART and can create semantically meaningful and contextually appropriate repres...
http://arxiv.org/pdf/2310.19127v1
2023-10-29T19:28:22
2023-10-29T19:28:22
{ "title": [ 0.030410767, 0.007598877, -0.021392822, -0.09234619, 0.0034332275, -0.009262085, -0.083862305, -0.0044555664, 0.0026779175, 0.046813965, 0.0064430237, -0.014320374, 0.023239136, -0.0413208, -0.013793945, 0.020889282, -0.006526947, 0....
http://arxiv.org/abs/2310.19123v1
Three Dogmas, a Puzzle and its Solution
Elnaserledinellah Mahmood Abdelwahab
cs.CL, cs.AI
Modern Logics, as formulated notably by Frege, Russell and Tarski involved basic assumptions about Natural Languages in general and Indo-European Languages in particular, which are contested by Linguists. Based upon those assumptions, formal Languages were designed to overcome what Logicians claimed to be 'defects' of ...
http://arxiv.org/pdf/2310.19123v1
2023-10-29T19:20:38
2023-10-29T19:20:38
{ "title": [ 0.047973633, 0.0039520264, -0.05505371, -0.015838623, -0.008163452, -0.052764893, 0.011985779, -0.0025901794, -0.030517578, 0.041107178, -0.033416748, -0.007621765, -0.006389618, -0.024368286, -0.01373291, -0.04458618, 0.02897644, 0....
http://arxiv.org/abs/2310.19106v1
PACuna: Automated Fine-Tuning of Language Models for Particle Accelerators
Antonin Sulc, Raimund Kammering, Annika Eichler, Tim Wilksen
cs.CL
Navigating the landscape of particle accelerators has become increasingly challenging with recent surges in contributions. These intricate devices challenge comprehension, even within individual facilities. To address this, we introduce PACuna, a fine-tuned language model refined through publicly available accelerator ...
http://arxiv.org/pdf/2310.19106v1
2023-10-29T18:43:19
2023-10-29T18:43:19
{ "title": [ -0.061401367, -0.016784668, -0.047332764, -0.040161133, -0.007411957, 0.0050354004, -0.05593872, -0.0010108948, 0.0034179688, 0.038085938, 0.023391724, -0.022201538, 0.0017557144, -0.037719727, -0.053894043, 0.018157959, -0.021881104, ...
http://arxiv.org/abs/2310.19089v1
Pushdown Layers: Encoding Recursive Structure in Transformer Language Models
Shikhar Murty, Pratyusha Sharma, Jacob Andreas, Christopher D. Manning
cs.CL
Recursion is a prominent feature of human language, and fundamentally challenging for self-attention due to the lack of an explicit recursive-state tracking mechanism. Consequently, Transformer language models poorly capture long-tail recursive structure and exhibit sample-inefficient syntactic generalization. This wor...
http://arxiv.org/pdf/2310.19089v1
2023-10-29T17:27:18
2023-10-29T17:27:18
{ "title": [ -0.034118652, -0.0042533875, -0.026397705, -0.07696533, 0.009857178, 0.024337769, -0.07678223, -0.0032749176, -0.00056648254, 0.03353882, 0.004310608, -0.016418457, 0.037322998, -0.0362854, -0.02973938, 0.018753052, -0.012519836, 0.0...
http://arxiv.org/abs/2310.19084v1
Roles of Scaling and Instruction Tuning in Language Perception: Model vs. Human Attention
Changjiang Gao, Shujian Huang, Jixing Li, Jiajun Chen
cs.CL, cs.AI
Recent large language models (LLMs) have revealed strong abilities to understand natural language. Since most of them share the same basic structure, i.e. the transformer block, possible contributors to their success in the training process are scaling and instruction tuning. However, how these factors affect the model...
http://arxiv.org/pdf/2310.19084v1
2023-10-29T17:16:40
2023-10-29T17:16:40
{ "title": [ -0.020095825, -0.0056037903, -0.025009155, -0.023330688, -0.025817871, -0.018447876, -0.024841309, -0.012008667, -0.019683838, 0.021575928, -0.034362793, -0.004272461, 0.0004749298, -0.02168274, -0.03729248, -0.023025513, -0.01499939, ...
http://arxiv.org/abs/2310.19060v1
TESTA: Temporal-Spatial Token Aggregation for Long-form Video-Language Understanding
Shuhuai Ren, Sishuo Chen, Shicheng Li, Xu Sun, Lu Hou
cs.CV, cs.AI, cs.CL
Large-scale video-language pre-training has made remarkable strides in advancing video-language understanding tasks. However, the heavy computational burden of video encoding remains a formidable efficiency bottleneck, particularly for long-form videos. These videos contain massive visual tokens due to their inherent 3...
http://arxiv.org/pdf/2310.19060v1
2023-10-29T16:25:32
2023-10-29T16:25:32
{ "title": [ 0.0012378693, -0.015670776, -0.0423584, -0.07122803, 0.0070228577, -0.019607544, -0.02772522, 0.01247406, 0.01739502, 0.037261963, -0.043823242, -0.010444641, 0.019134521, 0.0006556511, -0.058929443, -0.017837524, -0.0023765564, 0.02...
http://arxiv.org/abs/2310.19056v2
MILL: Mutual Verification with Large Language Models for Zero-Shot Query Expansion
Pengyue Jia, Yiding Liu, Xiangyu Zhao, Xiaopeng Li, Changying Hao, Shuaiqiang Wang, Dawei Yin
cs.IR, cs.AI, cs.CL
Query expansion is a commonly-used technique in many search systems to better represent users' information needs with additional query terms. Existing studies for this task usually propose to expand a query with retrieved or generated contextual documents. However, both types of methods have clear limitations. For retr...
http://arxiv.org/pdf/2310.19056v2
2023-11-13T15:09:01
2023-10-29T16:04:10
{ "title": [ -0.0005788803, -0.011062622, -0.061645508, -0.10003662, 0.03237915, -0.0010976791, -0.054656982, 0.029022217, -0.005958557, 0.035705566, -0.014709473, 0.011474609, 0.009216309, -0.010513306, -0.046081543, 0.007575989, -0.005039215, 0...
http://arxiv.org/abs/2310.19055v1
A Survey on Recent Named Entity Recognition and Relation Classification Methods with Focus on Few-Shot Learning Approaches
Sakher Alqaaidi, Elika Bozorgi
cs.CL
Named entity recognition and relation classification are key stages for extracting information from unstructured text. Several natural language processing applications utilize the two tasks, such as information retrieval, knowledge graph construction and completion, question answering and other domain-specific applicat...
http://arxiv.org/pdf/2310.19055v1
2023-10-29T16:02:46
2023-10-29T16:02:46
{ "title": [ -0.013534546, -0.018554688, -0.05166626, -0.047546387, 0.0047569275, -0.008369446, -0.039733887, 0.016708374, 0.012702942, 0.00086688995, -0.03189087, 0.011955261, -0.0032100677, -0.017318726, -0.020874023, -0.015640259, 0.000259161, ...
http://arxiv.org/abs/2310.19034v1
ArBanking77: Intent Detection Neural Model and a New Dataset in Modern and Dialectical Arabic
Mustafa Jarrar, Ahmet Birim, Mohammed Khalilia, Mustafa Erden, Sana Ghanem
cs.CL
This paper presents the ArBanking77, a large Arabic dataset for intent detection in the banking domain. Our dataset was arabized and localized from the original English Banking77 dataset, which consists of 13,083 queries to ArBanking77 dataset with 31,404 queries in both Modern Standard Arabic (MSA) and Palestinian dia...
http://arxiv.org/pdf/2310.19034v1
2023-10-29T14:46:11
2023-10-29T14:46:11
{ "title": [ 0.0012369156, 0.009033203, -0.038757324, -0.025985718, 0.017654419, -0.031677246, -0.0046653748, -0.016540527, -0.0027751923, 0.016647339, 0.016555786, -0.011558533, -0.0059928894, -0.05178833, -0.027297974, 0.012573242, -0.0046653748, ...
http://arxiv.org/abs/2310.19029v1
SALMA: Arabic Sense-Annotated Corpus and WSD Benchmarks
Mustafa Jarrar, Sanad Malaysha, Tymaa Hammouda, Mohammed Khalilia
cs.CL
SALMA, the first Arabic sense-annotated corpus, consists of ~34K tokens, which are all sense-annotated. The corpus is annotated using two different sense inventories simultaneously (Modern and Ghani). SALMA novelty lies in how tokens and senses are associated. Instead of linking a token to only one intended sense, SALM...
http://arxiv.org/pdf/2310.19029v1
2023-10-29T14:36:37
2023-10-29T14:36:37
{ "title": [ -0.031982422, -0.015655518, -0.049621582, -0.062683105, 0.0011711121, -0.004486084, -0.041015625, 0.006717682, 0.025009155, 0.042144775, -0.0031661987, -0.0037574768, -0.0016908646, -0.02217102, -0.061584473, 0.011024475, 0.012649536, ...
http://arxiv.org/abs/2310.19019v2
TeacherLM: Teaching to Fish Rather Than Giving the Fish, Language Modeling Likewise
Nan He, Hanyu Lai, Chenyang Zhao, Zirui Cheng, Junting Pan, Ruoyu Qin, Ruofan Lu, Rui Lu, Yunchen Zhang, Gangming Zhao, Zhaohui Hou, Zhiyuan Huang, Shaoqing Lu, Ding Liang, Mingjie Zhan
cs.CL, cs.AI
Large Language Models (LLMs) exhibit impressive reasoning and data augmentation capabilities in various NLP tasks. However, what about small models? In this work, we propose TeacherLM-7.1B, capable of annotating relevant fundamentals, chain of thought, and common mistakes for most NLP samples, which makes annotation mo...
http://arxiv.org/pdf/2310.19019v2
2023-10-31T06:49:13
2023-10-29T14:16:54
{ "title": [ 0.025741577, -0.011146545, -0.0024814606, 0.016418457, -0.01676941, -0.017181396, 0.025756836, -0.028198242, -0.01838684, 0.012237549, -0.02947998, 0.012619019, 0.0017738342, -0.038879395, 0.018539429, -0.03488159, -0.018753052, 0.05...
http://arxiv.org/abs/2310.18992v1
Bipartite Graph Pre-training for Unsupervised Extractive Summarization with Graph Convolutional Auto-Encoders
Qianren Mao, Shaobo Zhao, Jiarui Li, Xiaolei Gu, Shizhu He, Bo Li, Jianxin Li
cs.CL, cs.AI
Pre-trained sentence representations are crucial for identifying significant sentences in unsupervised document extractive summarization. However, the traditional two-step paradigm of pre-training and sentence-ranking, creates a gap due to differing optimization objectives. To address this issue, we argue that utilizin...
http://arxiv.org/pdf/2310.18992v1
2023-10-29T12:27:18
2023-10-29T12:27:18
{ "title": [ -0.023910522, -0.004863739, -0.04598999, -0.07751465, -0.025115967, 0.0037193298, -0.08502197, 0.0014400482, 0.009567261, 0.037506104, -0.010810852, 0.027145386, 0.01247406, -0.017730713, -0.017181396, 0.018615723, 0.0013713837, 0.03...
http://arxiv.org/abs/2310.18974v1
EtiCor: Corpus for Analyzing LLMs for Etiquettes
Ashutosh Dwivedi, Pradhyumna Lavania, Ashutosh Modi
cs.CL, cs.AI, cs.LG
Etiquettes are an essential ingredient of day-to-day interactions among people. Moreover, etiquettes are region-specific, and etiquettes in one region might contradict those in other regions. In this paper, we propose EtiCor, an Etiquettes Corpus, having texts about social norms from five different regions across the g...
http://arxiv.org/pdf/2310.18974v1
2023-10-29T10:47:23
2023-10-29T10:47:23
{ "title": [ -0.056610107, -0.01927185, -0.05307007, -0.019821167, -0.03857422, -0.008865356, -0.054473877, 0.007686615, -0.0047836304, 0.027252197, -0.06298828, 0.0126571655, 0.04925537, -0.02407837, 0.009605408, -0.012741089, -0.012886047, 0.05...
http://arxiv.org/abs/2310.18964v1
LLMs and Finetuning: Benchmarking cross-domain performance for hate speech detection
Ahmad Nasir, Aadish Sharma, Kokil Jaidka
cs.CL
This paper compares different pre-trained and fine-tuned large language models (LLMs) for hate speech detection. Our research underscores challenges in LLMs' cross-domain validity and overfitting risks. Through evaluations, we highlight the need for fine-tuned models that grasp the nuances of hate speech through greate...
http://arxiv.org/pdf/2310.18964v1
2023-10-29T10:07:32
2023-10-29T10:07:32
{ "title": [ 0.023651123, 0.00843811, -0.034210205, -0.050842285, 0.027145386, -0.038391113, 0.007843018, -0.017349243, -0.016830444, 0.03074646, -0.026748657, 0.035888672, 0.0033798218, -0.0044898987, -0.028015137, -0.0008254051, -0.025772095, 0...
http://arxiv.org/abs/2310.18956v1
End-to-End Autoregressive Retrieval via Bootstrapping for Smart Reply Systems
Benjamin Towle, Ke Zhou
cs.CL, cs.AI, cs.LG
Reply suggestion systems represent a staple component of many instant messaging and email systems. However, the requirement to produce sets of replies, rather than individual replies, makes the task poorly suited for out-of-the-box retrieval architectures, which only consider individual message-reply similarity. As a r...
http://arxiv.org/pdf/2310.18956v1
2023-10-29T09:56:17
2023-10-29T09:56:17
{ "title": [ 0.0001386404, 0.011352539, -0.034179688, -0.07080078, 0.020980835, 0.005859375, -0.027053833, -0.0014925003, 0.034088135, 0.049926758, -0.015022278, -0.004058838, 0.0016107559, -0.0093307495, -0.09320068, 0.008346558, 0.014770508, -0...