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/2311.05085v1
Characterizing Large Language Models as Rationalizers of Knowledge-intensive Tasks
Aditi Mishra, Sajjadur Rahman, Hannah Kim, Kushan Mitra, Estevam Hruschka
cs.CL, cs.AI
Large language models (LLMs) are proficient at generating fluent text with minimal task-specific supervision. Yet, their ability to provide well-grounded rationalizations for knowledge-intensive tasks remains under-explored. Such tasks, like commonsense multiple-choice questions, require rationales based on world knowl...
http://arxiv.org/pdf/2311.05085v1
2023-11-09T01:04:44
2023-11-09T01:04:44
{ "title": [ -0.031341553, -0.016952515, -0.05340576, -0.06726074, -0.0037136078, -0.011169434, -0.059814453, 0.009086609, -0.0050582886, 0.035217285, -0.004852295, 0.0054626465, -0.0034942627, -0.014343262, -0.047851562, -0.0032291412, -0.007205963,...
http://arxiv.org/abs/2311.05657v1
Lumos: Learning Agents with Unified Data, Modular Design, and Open-Source LLMs
Da Yin, Faeze Brahman, Abhilasha Ravichander, Khyathi Chandu, Kai-Wei Chang, Yejin Choi, Bill Yuchen Lin
cs.AI, cs.CL, cs.LG
We introduce Lumos, a novel framework for training language agents that employs a unified data format and a modular architecture based on open-source large language models (LLMs). Lumos consists of three distinct modules: planning, grounding, and execution. The planning module breaks down a task into a series of high-l...
http://arxiv.org/pdf/2311.05657v1
2023-11-09T00:30:13
2023-11-09T00:30:13
{ "title": [ -0.04916382, -0.008720398, -0.058410645, -0.02633667, -0.026535034, -0.019851685, -0.03765869, -0.0038661957, -0.051361084, 0.034210205, -0.046783447, 0.0541687, 0.0052490234, -0.0043411255, 0.015151978, -0.01713562, -0.016677856, 0....
http://arxiv.org/abs/2311.05075v1
Mental Health Diagnosis in the Digital Age: Harnessing Sentiment Analysis on Social Media Platforms upon Ultra-Sparse Feature Content
Haijian Shao, Ming Zhu, Shengjie Zhai
cs.LG, cs.AI, cs.CL
Amid growing global mental health concerns, particularly among vulnerable groups, natural language processing offers a tremendous potential for early detection and intervention of people's mental disorders via analyzing their postings and discussions on social media platforms. However, ultra-sparse training data, often...
http://arxiv.org/pdf/2311.05075v1
2023-11-09T00:15:06
2023-11-09T00:15:06
{ "title": [ 0.012466431, -0.0019340515, -0.02684021, -0.04888916, 0.018875122, -0.01713562, -0.010871887, -0.0032291412, 0.010810852, 0.033691406, -0.026809692, 0.0034179688, 0.0031642914, -0.01927185, -0.049987793, -0.010826111, 0.015350342, 0....
http://arxiv.org/abs/2311.05074v1
A Framework to Assess (Dis)agreement Among Diverse Rater Groups
Vinodkumar Prabhakaran, Christopher Homan, Lora Aroyo, Alicia Parrish, Alex Taylor, Mark Díaz, Ding Wang
cs.CL, cs.AI
Recent advancements in conversational AI have created an urgent need for safety guardrails that prevent users from being exposed to offensive and dangerous content. Much of this work relies on human ratings and feedback, but does not account for the fact that perceptions of offense and safety are inherently subjective ...
http://arxiv.org/pdf/2311.05074v1
2023-11-09T00:12:21
2023-11-09T00:12:21
{ "title": [ 0.0019760132, 0.03717041, -0.074035645, -0.02444458, 0.011543274, -0.047698975, -0.016281128, -0.002357483, 0.014564514, -0.011398315, -0.039154053, -0.03982544, -0.039916992, -0.051086426, -0.025100708, -0.037994385, 0.007698059, 0....
http://arxiv.org/abs/2311.05051v1
Deep Learning Brasil at ABSAPT 2022: Portuguese Transformer Ensemble Approaches
Juliana Resplande Santanna Gomes, Eduardo Augusto Santos Garcia, Adalberto Ferreira Barbosa Junior, Ruan Chaves Rodrigues, Diogo Fernandes Costa Silva, Dyonnatan Ferreira Maia, Nádia Félix Felipe da Silva, Arlindo Rodrigues Galvão Filho, Anderson da Silva Soares
cs.CL
Aspect-based Sentiment Analysis (ABSA) is a task whose objective is to classify the individual sentiment polarity of all entities, called aspects, in a sentence. The task is composed of two subtasks: Aspect Term Extraction (ATE), identify all aspect terms in a sentence; and Sentiment Orientation Extraction (SOE), given...
http://arxiv.org/pdf/2311.05051v1
2023-11-08T22:58:34
2023-11-08T22:58:34
{ "title": [ -0.017913818, 0.021972656, -0.0680542, -0.040740967, 0.02558899, -0.015449524, -0.033325195, 0.00051021576, 0.0038852692, 0.041992188, -0.015014648, -0.009346008, 0.050445557, -0.045715332, -0.009254456, 0.008560181, 0.014022827, 0.0...
http://arxiv.org/abs/2311.05047v1
DeepLearningBrasil@LT-EDI-2023: Exploring Deep Learning Techniques for Detecting Depression in Social Media Text
Eduardo Garcia, Juliana Gomes, Adalberto Barbosa Júnior, Cardeque Borges, Nádia da Silva
cs.CL, cs.AI
In this paper, we delineate the strategy employed by our team, DeepLearningBrasil, which secured us the first place in the shared task DepSign-LT-EDI@RANLP-2023, achieving a 47.0% Macro F1-Score and a notable 2.4% advantage. The task was to classify social media texts into three distinct levels of depression - "not dep...
http://arxiv.org/pdf/2311.05047v1
2023-11-08T22:42:31
2023-11-08T22:42:31
{ "title": [ 0.019882202, -0.0021686554, -0.0063285828, -0.048187256, 0.008880615, -0.009666443, -0.013519287, -0.015419006, 0.00017166138, 0.03768921, -0.042266846, 0.016494751, 0.030471802, -0.0184021, -0.032348633, -0.0040016174, -0.011268616, ...
http://arxiv.org/abs/2311.05043v1
Zero-shot Translation of Attention Patterns in VQA Models to Natural Language
Leonard Salewski, A. Sophia Koepke, Hendrik P. A. Lensch, Zeynep Akata
cs.CV, cs.AI, cs.CL
Converting a model's internals to text can yield human-understandable insights about the model. Inspired by the recent success of training-free approaches for image captioning, we propose ZS-A2T, a zero-shot framework that translates the transformer attention of a given model into natural language without requiring any...
http://arxiv.org/pdf/2311.05043v1
2023-11-08T22:18:53
2023-11-08T22:18:53
{ "title": [ -0.012535095, 0.00024867058, -0.050445557, -0.03756714, -0.010978699, -0.025680542, -0.015899658, -0.00075387955, 0.026168823, 0.03668213, 0.0055885315, -0.04272461, 0.002538681, -0.03274536, -0.07733154, 0.003068924, -0.015182495, 0...
http://arxiv.org/abs/2311.07588v1
NLQxform: A Language Model-based Question to SPARQL Transformer
Ruijie Wang, Zhiruo Zhang, Luca Rossetto, Florian Ruosch, Abraham Bernstein
cs.CL, cs.AI
In recent years, scholarly data has grown dramatically in terms of both scale and complexity. It becomes increasingly challenging to retrieve information from scholarly knowledge graphs that include large-scale heterogeneous relationships, such as authorship, affiliation, and citation, between various types of entities...
http://arxiv.org/pdf/2311.07588v1
2023-11-08T21:41:45
2023-11-08T21:41:45
{ "title": [ -0.014923096, -0.0057868958, -0.053985596, -0.06149292, 0.00223732, 0.006668091, -0.043273926, 0.0063056946, 0.010894775, 0.0070648193, -0.027038574, 0.0010967255, -0.009750366, -0.027267456, -0.02861023, -0.0046081543, -0.011909485, ...
http://arxiv.org/abs/2311.05020v1
First Tragedy, then Parse: History Repeats Itself in the New Era of Large Language Models
Naomi Saphra, Eve Fleisig, Kyunghyun Cho, Adam Lopez
cs.CL
Many NLP researchers are experiencing an existential crisis triggered by the astonishing success of ChatGPT and other systems based on large language models (LLMs). After such a disruptive change to our understanding of the field, what is left to do? Taking a historical lens, we look for guidance from the first era of ...
http://arxiv.org/pdf/2311.05020v1
2023-11-08T21:13:38
2023-11-08T21:13:38
{ "title": [ -0.021392822, -0.01966858, -0.027664185, -0.087402344, -0.015975952, -0.005821228, -0.07055664, 0.024230957, -0.012199402, 0.018829346, -0.044799805, 0.03717041, 0.024490356, -0.008895874, -0.008934021, -0.0014305115, -0.010437012, 0...
http://arxiv.org/abs/2311.05014v1
Interpreting Pretrained Language Models via Concept Bottlenecks
Zhen Tan, Lu Cheng, Song Wang, Yuan Bo, Jundong Li, Huan Liu
cs.CL, cs.AI
Pretrained language models (PLMs) have made significant strides in various natural language processing tasks. However, the lack of interpretability due to their ``black-box'' nature poses challenges for responsible implementation. Although previous studies have attempted to improve interpretability by using, e.g., atte...
http://arxiv.org/pdf/2311.05014v1
2023-11-08T20:41:18
2023-11-08T20:41:18
{ "title": [ -0.026977539, -0.008865356, -0.026779175, -0.06707764, -0.013328552, 0.0022830963, -0.07159424, -0.011856079, 0.0015277863, 0.024520874, -0.018585205, -0.0028266907, -0.006679535, -0.011894226, -0.050048828, -0.004535675, -0.0036754608, ...
http://arxiv.org/abs/2311.07587v1
Frontier Language Models are not Robust to Adversarial Arithmetic, or "What do I need to say so you agree 2+2=5?
C. Daniel Freeman, Laura Culp, Aaron Parisi, Maxwell L Bileschi, Gamaleldin F Elsayed, Alex Rizkowsky, Isabelle Simpson, Alex Alemi, Azade Nova, Ben Adlam, Bernd Bohnet, Gaurav Mishra, Hanie Sedghi, Igor Mordatch, Izzeddin Gur, Jaehoon Lee, JD Co-Reyes, Jeffrey Pennington, Kelvin Xu, Kevin Swersky, Kshiteej Mahajan, Le...
cs.CL, cs.AI, cs.CY, cs.LG
We introduce and study the problem of adversarial arithmetic, which provides a simple yet challenging testbed for language model alignment. This problem is comprised of arithmetic questions posed in natural language, with an arbitrary adversarial string inserted before the question is complete. Even in the simple setti...
http://arxiv.org/pdf/2311.07587v1
2023-11-08T19:07:10
2023-11-08T19:07:10
{ "title": [ -0.019699097, 0.024993896, -0.029678345, 0.005355835, -0.00818634, 0.010253906, -0.04446411, -0.029449463, 0.012283325, 0.0034980774, -0.0012149811, 0.011772156, -0.026794434, -0.02923584, -0.00844574, -0.011650085, -0.011123657, 0.0...
http://arxiv.org/abs/2311.04978v1
On the steerability of large language models toward data-driven personas
Junyi Li, Ninareh Mehrabi, Charith Peris, Palash Goyal, Kai-Wei Chang, Aram Galstyan, Richard Zemel, Rahul Gupta
cs.CL
The recent surge in Large Language Model (LLM) related applications has led to a concurrent escalation in expectations for LLMs to accommodate a myriad of personas and encompass a broad spectrum of perspectives. An important first step towards addressing this demand is to align language models with specific personas, b...
http://arxiv.org/pdf/2311.04978v1
2023-11-08T19:01:13
2023-11-08T19:01:13
{ "title": [ -0.01914978, -0.01335144, -0.036193848, -0.07836914, -0.005428314, 0.011634827, -0.060760498, 0.019485474, -0.0011014938, 0.03302002, -0.02532959, 0.0026226044, 0.030639648, -0.010276794, -0.029022217, 0.012420654, -0.021560669, 0.05...
http://arxiv.org/abs/2311.04902v1
Beyond Size: How Gradients Shape Pruning Decisions in Large Language Models
Rocktim Jyoti Das, Liqun Ma, Zhiqiang Shen
cs.CL, cs.AI, cs.LG
Large Language Models (LLMs) with a billion or more parameters are prime targets for network pruning, which aims to reduce a portion of the network weights without compromising performance. Prior approaches such as Weights Magnitude, SparseGPT, and Wanda, either concentrated solely on weights or integrated weights with...
http://arxiv.org/pdf/2311.04902v1
2023-11-08T18:59:54
2023-11-08T18:59:54
{ "title": [ -0.026626587, -0.016983032, -0.034423828, -0.06866455, -0.00945282, 0.010314941, -0.06945801, 0.011886597, -0.0024051666, 0.017715454, -0.034698486, 0.015068054, 0.017044067, -0.013587952, -0.023803711, -0.0071525574, -0.012916565, 0...
http://arxiv.org/abs/2311.04900v1
How Abstract Is Linguistic Generalization in Large Language Models? Experiments with Argument Structure
Michael Wilson, Jackson Petty, Robert Frank
cs.CL
Language models are typically evaluated on their success at predicting the distribution of specific words in specific contexts. Yet linguistic knowledge also encodes relationships between contexts, allowing inferences between word distributions. We investigate the degree to which pre-trained Transformer-based large lan...
http://arxiv.org/pdf/2311.04900v1
2023-11-08T18:58:43
2023-11-08T18:58:43
{ "title": [ -0.016708374, -0.0060424805, -0.020492554, -0.075805664, -0.0036449432, 0.007774353, -0.056427002, 0.0022354126, -0.020111084, 0.0262146, -0.012283325, 0.028518677, -0.016921997, -0.0055999756, -0.034576416, 0.003665924, -0.019577026, ...
http://arxiv.org/abs/2311.04954v1
Prompt Sketching for Large Language Models
Luca Beurer-Kellner, Mark Niklas Müller, Marc Fischer, Martin Vechev
cs.CL, cs.AI
Many recent prompting strategies for large language models (LLMs) query the model multiple times sequentially -- first to produce intermediate results and then the final answer. However, using these methods, both decoder and model are unaware of potential follow-up prompts, leading to disconnected and undesirably wordy...
http://arxiv.org/pdf/2311.04954v1
2023-11-08T18:57:23
2023-11-08T18:57:23
{ "title": [ -0.039489746, -0.021896362, -0.033325195, -0.050598145, -0.02319336, 0.0063095093, -0.06726074, 0.0058288574, 0.0154953, 0.018814087, -0.04034424, -0.008766174, 0.0062332153, -0.012504578, -0.03289795, 0.0006213188, 0.006175995, 0.03...
http://arxiv.org/abs/2311.04897v1
Future Lens: Anticipating Subsequent Tokens from a Single Hidden State
Koyena Pal, Jiuding Sun, Andrew Yuan, Byron C. Wallace, David Bau
cs.CL, cs.LG
We conjecture that hidden state vectors corresponding to individual input tokens encode information sufficient to accurately predict several tokens ahead. More concretely, in this paper we ask: Given a hidden (internal) representation of a single token at position $t$ in an input, can we reliably anticipate the tokens ...
http://arxiv.org/pdf/2311.04897v1
2023-11-08T18:56:35
2023-11-08T18:56:35
{ "title": [ 0.011039734, -0.01197052, -0.006439209, -0.0725708, -0.0206604, 0.00031924248, -0.00630188, -0.0011768341, -0.008995056, 0.015670776, -0.020599365, -0.0022239685, 0.009666443, -0.034576416, -0.0118255615, -0.019485474, 0.00043725967, ...
http://arxiv.org/abs/2311.04892v1
Bias Runs Deep: Implicit Reasoning Biases in Persona-Assigned LLMs
Shashank Gupta, Vaishnavi Shrivastava, Ameet Deshpande, Ashwin Kalyan, Peter Clark, Ashish Sabharwal, Tushar Khot
cs.CL
Recent works have showcased the ability of large-scale language models (LLMs) to embody diverse personas in their responses, exemplified by prompts like 'You are Yoda. Explain the Theory of Relativity.' While this ability allows personalization of LLMs and enables human behavior simulation, its effect on LLMs' capabili...
http://arxiv.org/pdf/2311.04892v1
2023-11-08T18:52:17
2023-11-08T18:52:17
{ "title": [ -0.034057617, -0.0079422, -0.040405273, -0.021560669, -0.029312134, -0.016098022, -0.039276123, -0.013580322, 0.0029296875, 0.0031738281, -0.062042236, 0.009094238, -0.0044021606, -0.022735596, -0.0064888, -0.029541016, -0.04638672, ...
http://arxiv.org/abs/2311.04886v1
SEMQA: Semi-Extractive Multi-Source Question Answering
Tal Schuster, Adam D. Lelkes, Haitian Sun, Jai Gupta, Jonathan Berant, William W. Cohen, Donald Metzler
cs.CL, cs.AI, cs.LG
Recently proposed long-form question answering (QA) systems, supported by large language models (LLMs), have shown promising capabilities. Yet, attributing and verifying their generated abstractive answers can be difficult, and automatically evaluating their accuracy remains an ongoing challenge. In this work, we int...
http://arxiv.org/pdf/2311.04886v1
2023-11-08T18:46:32
2023-11-08T18:46:32
{ "title": [ -0.015991211, -0.00010919571, -0.03640747, -0.066589355, 0.013076782, 0.008811951, -0.042663574, -0.0036754608, 0.017501831, 0.023254395, -0.009262085, 0.008811951, -0.016723633, -0.006038666, -0.057495117, 0.006389618, -0.0061454773, ...
http://arxiv.org/abs/2311.04885v1
Profiling Irony & Stereotype: Exploring Sentiment, Topic, and Lexical Features
Tibor L. R. Krols, Marie Mortensen, Ninell Oldenburg
cs.CL
Social media has become a very popular source of information. With this popularity comes an interest in systems that can classify the information produced. This study tries to create such a system detecting irony in Twitter users. Recent work emphasize the importance of lexical features, sentiment features and the cont...
http://arxiv.org/pdf/2311.04885v1
2023-11-08T18:44:47
2023-11-08T18:44:47
{ "title": [ -0.0061187744, 0.00390625, -0.03086853, -0.025253296, 0.017944336, -0.011276245, -0.0019292831, -0.033050537, 0.0146102905, 0.018218994, -0.024261475, -0.014289856, 0.0025367737, -0.01889038, -0.04888916, -0.009864807, -0.02583313, 0...
http://arxiv.org/abs/2311.04879v2
LongQLoRA: Efficient and Effective Method to Extend Context Length of Large Language Models
Jianxin Yang
cs.CL, cs.AI
We present LongQLoRA, an efficient and effective method to extend context length of large language models with less training resources. LongQLoRA combines the advantages of Position Interpolation, QLoRA and Shift Short Attention of LongLoRA. With a single 32GB V100 GPU, LongQLoRA can extend the context length of LLaMA2...
http://arxiv.org/pdf/2311.04879v2
2023-11-09T05:21:28
2023-11-08T18:33:06
{ "title": [ -0.02923584, -0.019851685, -0.055511475, -0.080566406, 0.004966736, 0.0026226044, -0.06347656, 0.0061836243, 0.0009551048, 0.028518677, -0.0067481995, 0.01058197, 0.005142212, -0.021820068, -0.03277588, 0.025634766, -0.0087509155, 0....
http://arxiv.org/abs/2311.05772v1
ADaPT: As-Needed Decomposition and Planning with Language Models
Archiki Prasad, Alexander Koller, Mareike Hartmann, Peter Clark, Ashish Sabharwal, Mohit Bansal, Tushar Khot
cs.AI, cs.CL, cs.LG
Large Language Models (LLMs) are increasingly being used for interactive decision-making tasks requiring planning and adapting to the environment. Recent works employ LLMs-as-agents in broadly two ways: iteratively determining the next action (iterative executors) or generating plans and executing sub-tasks using LLMs ...
http://arxiv.org/pdf/2311.05772v1
2023-11-08T17:59:15
2023-11-08T17:59:15
{ "title": [ -0.048461914, -0.017196655, -0.03338623, -0.07788086, -0.0018472672, 0.033935547, -0.0715332, 0.009521484, -0.021606445, 0.019104004, -0.009803772, 0.0030975342, -0.006477356, -0.02142334, -0.013793945, -0.0012807846, 0.0035858154, 0...
http://arxiv.org/abs/2311.04850v2
Rethinking Benchmark and Contamination for Language Models with Rephrased Samples
Shuo Yang, Wei-Lin Chiang, Lianmin Zheng, Joseph E. Gonzalez, Ion Stoica
cs.CL, cs.AI
Large language models are increasingly trained on all the data ever produced by humans. Many have raised concerns about the trustworthiness of public benchmarks due to potential contamination in pre-training or fine-tuning datasets. While most data decontamination efforts apply string matching (e.g., n-gram overlap) to...
http://arxiv.org/pdf/2311.04850v2
2023-11-11T05:11:18
2023-11-08T17:35:20
{ "title": [ -0.030578613, -0.01626587, -0.026901245, -0.07989502, -0.0141067505, 0.019241333, -0.09301758, 0.0071754456, 0.00018501282, 0.04751587, -0.030792236, 0.01876831, -0.0005426407, 0.004207611, -0.07141113, 0.011642456, -0.021194458, 0.0...
http://arxiv.org/abs/2311.04823v1
Hierarchically Gated Recurrent Neural Network for Sequence Modeling
Zhen Qin, Songlin Yang, Yiran Zhong
cs.CL, cs.LG
Transformers have surpassed RNNs in popularity due to their superior abilities in parallel training and long-term dependency modeling. Recently, there has been a renewed interest in using linear RNNs for efficient sequence modeling. These linear RNNs often employ gating mechanisms in the output of the linear recurrence...
http://arxiv.org/pdf/2311.04823v1
2023-11-08T16:50:05
2023-11-08T16:50:05
{ "title": [ -0.016845703, 0.010055542, -0.05996704, -0.097839355, 0.017959595, -0.00041794777, -0.06689453, 0.01751709, 0.0060653687, 0.05331421, -0.001871109, 0.0066833496, 0.022644043, -0.016342163, -0.05130005, 0.030548096, -0.0060043335, 0.0...
http://arxiv.org/abs/2311.04816v1
MTGER: Multi-view Temporal Graph Enhanced Temporal Reasoning over Time-Involved Document
Zheng Chu, Zekun Wang, Jiafeng Liang, Ming Liu, Bing Qin
cs.CL, cs.AI
The facts and time in the document are intricately intertwined, making temporal reasoning over documents challenging. Previous work models time implicitly, making it difficult to handle such complex relationships. To address this issue, we propose MTGER, a novel Multi-view Temporal Graph Enhanced Temporal Reasoning fra...
http://arxiv.org/pdf/2311.04816v1
2023-11-08T16:41:37
2023-11-08T16:41:37
{ "title": [ 0.018936157, -0.013916016, -0.06933594, -0.11755371, 0.014373779, -0.017456055, -0.04925537, 0.037078857, 0.022460938, 0.014823914, -0.02154541, -0.0028095245, 0.011138916, -0.026809692, -0.021408081, -0.0073661804, -0.0055618286, 0....
http://arxiv.org/abs/2311.04799v1
DACBERT: Leveraging Dependency Agreement for Cost-Efficient Bert Pretraining
Martin Kuo, Jianyi Zhang, Yiran Chen
cs.CL, cs.AI
Building on the cost-efficient pretraining advancements brought about by Crammed BERT, we enhance its performance and interpretability further by introducing a novel pretrained model Dependency Agreement Crammed BERT (DACBERT) and its two-stage pretraining framework - Dependency Agreement Pretraining. This framework, g...
http://arxiv.org/pdf/2311.04799v1
2023-11-08T16:18:32
2023-11-08T16:18:32
{ "title": [ 0.004047394, -0.0345459, -0.047943115, -0.14050293, -0.006477356, 0.017089844, -0.09320068, 0.03353882, -0.022979736, 0.04248047, -0.007980347, 0.027023315, 0.028671265, -0.030014038, -0.006576538, 0.009780884, -0.004436493, 0.017822...
http://arxiv.org/abs/2311.04789v1
Determination of toxic comments and unintended model bias minimization using Deep learning approach
Md Azim Khan
cs.LG, cs.CL, cs.CY
Online conversations can be toxic and subjected to threats, abuse, or harassment. To identify toxic text comments, several deep learning and machine learning models have been proposed throughout the years. However, recent studies demonstrate that because of the imbalances in the training data, some models are more like...
http://arxiv.org/pdf/2311.04789v1
2023-11-08T16:10:28
2023-11-08T16:10:28
{ "title": [ 0.006511688, 0.008041382, -0.04510498, -0.060577393, 0.021774292, -0.014656067, -0.02986145, -0.027053833, -0.024917603, 0.030563354, -0.028411865, 0.009529114, -0.003074646, -0.015281677, -0.029083252, -0.022109985, -0.0043144226, 0...
http://arxiv.org/abs/2311.04742v1
Using large language models to study human memory for meaningful narratives
Antonios Georgiou Tankut Can, Mikhail Katkov, Misha Tsodyks
cs.CL, q-bio.NC
One of the most impressive achievements of the AI revolution is the development of large language models that can generate meaningful text and respond to instructions in plain English with no additional training necessary. Here we show that language models can be used as a scientific instrument for studying human memor...
http://arxiv.org/pdf/2311.04742v1
2023-11-08T15:11:57
2023-11-08T15:11:57
{ "title": [ -0.025909424, -0.011001587, -0.04046631, -0.057891846, -0.014312744, -0.014251709, -0.038513184, 0.0077323914, -0.0077400208, 0.037963867, -0.004322052, -0.0042266846, 0.009765625, -0.017349243, -0.042541504, 0.010795593, -0.009513855, ...
http://arxiv.org/abs/2311.04694v1
Evaluating Generative Ad Hoc Information Retrieval
Lukas Gienapp, Harrisen Scells, Niklas Deckers, Janek Bevendorff, Shuai Wang, Johannes Kiesel, Shahbaz Syed, Maik Fröbe, Guido Zuccon, Benno Stein, Matthias Hagen, Martin Potthast
cs.IR, cs.CL
Recent advances in large language models have enabled the development of viable generative information retrieval systems. A generative retrieval system returns a grounded generated text in response to an information need instead of the traditional document ranking. Quantifying the utility of these types of responses is...
http://arxiv.org/pdf/2311.04694v1
2023-11-08T14:05:00
2023-11-08T14:05:00
{ "title": [ 0.0015325546, -0.0052261353, -0.025497437, -0.031219482, 0.0023326874, -0.011207581, -0.0124435425, -0.024810791, 0.013374329, 0.03363037, -0.030166626, 0.008163452, -0.032165527, -0.0018424988, -0.06744385, -0.012718201, -0.010574341, ...
http://arxiv.org/abs/2311.04666v1
Pre-training LLMs using human-like development data corpus
Khushi Bhardwaj, Raj Sanjay Shah, Sashank Varma
cs.CL, cs.AI
Pre-trained Large Language Models (LLMs) have shown success in a diverse set of language inference and understanding tasks. The pre-training stage of LLMs looks at a large corpus of raw textual data. The BabyLM shared task compares LLM pre-training to human language acquisition, where the number of tokens seen by 13-ye...
http://arxiv.org/pdf/2311.04666v1
2023-11-08T13:13:23
2023-11-08T13:13:23
{ "title": [ -0.04977417, -0.008903503, -0.06161499, -0.023727417, -0.050933838, -0.007888794, -0.051483154, 0.009048462, -0.038116455, 0.035369873, -0.04309082, 0.029296875, 0.0043678284, -0.014137268, 0.021209717, -0.033569336, -0.0055160522, 0...
http://arxiv.org/abs/2311.04664v1
Speech language models lack important brain-relevant semantics
Subba Reddy Oota, Emin Çelik, Fatma Deniz, Mariya Toneva
cs.CL, cs.LG, eess.AS, q-bio.NC
Despite known differences between reading and listening in the brain, recent work has shown that text-based language models predict both text-evoked and speech-evoked brain activity to an impressive degree. This poses the question of what types of information language models truly predict in the brain. We investigate t...
http://arxiv.org/pdf/2311.04664v1
2023-11-08T13:11:48
2023-11-08T13:11:48
{ "title": [ -0.024551392, -0.0044059753, -0.013252258, 0.006298065, -0.05960083, -0.004627228, -0.03112793, -0.011672974, 0.003753662, 0.050109863, -0.015319824, -0.061828613, 0.013122559, -0.04031372, -0.03274536, -0.0016679764, 0.017044067, 0....
http://arxiv.org/abs/2311.04661v2
Massive Editing for Large Language Models via Meta Learning
Chenmien Tan, Ge Zhang, Jie Fu
cs.CL, cs.LG
While large language models (LLMs) have enabled learning knowledge from the pre-training corpora, the acquired knowledge may be fundamentally incorrect or outdated over time, which necessitates rectifying the knowledge of the language model (LM) after the training. A promising approach involves employing a hyper-networ...
http://arxiv.org/pdf/2311.04661v2
2023-11-09T11:07:15
2023-11-08T13:03:06
{ "title": [ -0.039093018, -0.016494751, -0.034454346, -0.060058594, -0.013450623, 0.023361206, -0.07574463, 0.0026988983, 0.008262634, 0.040222168, -0.0023326874, -0.008102417, 0.013160706, -0.018096924, -0.041870117, 0.016967773, -0.019836426, ...
http://arxiv.org/abs/2311.04948v1
Explained anomaly detection in text reviews: Can subjective scenarios be correctly evaluated?
David Novoa-Paradela, Oscar Fontenla-Romero, Bertha Guijarro-Berdiñas
cs.CL, cs.AI, cs.LG
This paper presents a pipeline to detect and explain anomalous reviews in online platforms. The pipeline is made up of three modules and allows the detection of reviews that do not generate value for users due to either worthless or malicious composition. The classifications are accompanied by a normality score and an ...
http://arxiv.org/pdf/2311.04948v1
2023-11-08T11:51:47
2023-11-08T11:51:47
{ "title": [ 0.009391785, -0.005710602, -0.0025634766, -0.015136719, -0.00440979, -0.010482788, 0.0113220215, -0.02809143, -0.020462036, 0.009895325, -0.044799805, 0.020141602, -0.0053520203, -0.0050468445, -0.035736084, -0.026687622, -0.037902832, ...
http://arxiv.org/abs/2311.04589v2
TEAL: Tokenize and Embed ALL for Multi-modal Large Language Models
Zhen Yang, Yingxue Zhang, Fandong Meng, Jie Zhou
cs.CL, cs.AI
Despite Multi-modal Large Language Models (MM-LLMs) have made exciting strides recently, they are still struggling to efficiently model the interactions among multi-modal inputs and the generation in non-textual modalities. In this work, we propose TEAL (Tokenize and Embed ALl)}, an approach to treat the input from any...
http://arxiv.org/pdf/2311.04589v2
2023-11-14T11:58:48
2023-11-08T10:34:16
{ "title": [ -0.026626587, -0.016921997, -0.05050659, -0.088378906, 0.0084991455, 0.010147095, -0.07116699, 0.0112838745, 0.0012893677, 0.029632568, -0.01763916, 0.009262085, 0.02583313, -0.014595032, -0.029571533, 0.013458252, -0.014228821, 0.04...
http://arxiv.org/abs/2311.04563v1
Investigating the Nature of Disagreements on Mid-Scale Ratings: A Case Study on the Abstractness-Concreteness Continuum
Urban Knupleš, Diego Frassinelli, Sabine Schulte im Walde
cs.CL
Humans tend to strongly agree on ratings on a scale for extreme cases (e.g., a CAT is judged as very concrete), but judgements on mid-scale words exhibit more disagreement. Yet, collected rating norms are heavily exploited across disciplines. Our study focuses on concreteness ratings and (i) implements correlations and...
http://arxiv.org/pdf/2311.04563v1
2023-11-08T09:52:58
2023-11-08T09:52:58
{ "title": [ 0.0018138885, 0.005744934, -0.04949951, -0.045013428, 0.0093307495, -0.03717041, 0.0066337585, -0.0073661804, -0.00088119507, 0.0031604767, -0.01663208, -0.03074646, -0.03491211, -0.044921875, -0.04043579, -0.056121826, -0.01852417, ...
http://arxiv.org/abs/2311.04554v1
Assessing Distractors in Multiple-Choice Tests
Vatsal Raina, Adian Liusie, Mark Gales
cs.CL
Multiple-choice tests are a common approach for assessing candidates' comprehension skills. Standard multiple-choice reading comprehension exams require candidates to select the correct answer option from a discrete set based on a question in relation to a contextual passage. For appropriate assessment, the distractor ...
http://arxiv.org/pdf/2311.04554v1
2023-11-08T09:37:09
2023-11-08T09:37:09
{ "title": [ 0.0067634583, 0.016479492, -0.006969452, -0.021652222, 0.003206253, -0.017669678, 0.019546509, -0.033843994, 0.0015201569, 0.02583313, -0.027252197, -0.015068054, -0.012619019, -0.027252197, -0.024765015, -0.00869751, -0.009208679, 0...
http://arxiv.org/abs/2311.04547v1
Large GPT-like Models are Bad Babies: A Closer Look at the Relationship between Linguistic Competence and Psycholinguistic Measures
Julius Steuer, Marius Mosbach, Dietrich Klakow
cs.CL
Research on the cognitive plausibility of language models (LMs) has so far mostly concentrated on modelling psycholinguistic response variables such as reading times, gaze durations and N400/P600 EEG signals, while mostly leaving out the dimension of what Mahowald et al. (2023) described as formal and functional lingui...
http://arxiv.org/pdf/2311.04547v1
2023-11-08T09:26:27
2023-11-08T09:26:27
{ "title": [ -0.018661499, 0.0082092285, -0.015487671, -0.002998352, -0.018188477, -0.015167236, -0.010192871, -0.027633667, -0.022994995, 0.016845703, -0.010345459, 0.004711151, -0.016113281, -0.041259766, -0.018692017, -0.008171082, -0.02835083, ...
http://arxiv.org/abs/2311.04535v1
RankAug: Augmented data ranking for text classification
Tiasa Singha Roy, Priyam Basu
cs.CL, cs.AI, cs.LG
Research on data generation and augmentation has been focused majorly on enhancing generation models, leaving a notable gap in the exploration and refinement of methods for evaluating synthetic data. There are several text similarity metrics within the context of generated data filtering which can impact the performanc...
http://arxiv.org/pdf/2311.04535v1
2023-11-08T08:47:49
2023-11-08T08:47:49
{ "title": [ -0.00084495544, -0.00030612946, -0.036132812, -0.043395996, 0.024459839, -0.013870239, 0.0020313263, -0.021942139, 0.011787415, 0.028137207, -0.029876709, 0.030776978, -0.023529053, 0.0023784637, -0.059814453, -0.00440979, -0.027038574, ...
http://arxiv.org/abs/2311.04534v1
Loss Masking Is Not Needed in Decoder-only Transformer for Discrete-token Based ASR
Qian Chen, Wen Wang, Qinglin Zhang, Siqi Zheng, Shiliang Zhang, Chong Deng, Yukun Ma, Hai Yu, Jiaqing Liu, Chong Zhang
cs.CL, cs.SD, eess.AS
Recently, unified speech-text models, such as SpeechGPT, VioLA, and AudioPaLM, have achieved remarkable performance on speech tasks. These models convert continuous speech signals into discrete tokens (speech discretization) and merge text and speech tokens into a shared vocabulary. Then they train a single decoder-onl...
http://arxiv.org/pdf/2311.04534v1
2023-11-08T08:45:14
2023-11-08T08:45:14
{ "title": [ -0.008384705, -0.00077056885, -0.043945312, -0.04663086, -0.016464233, -0.0036563873, -0.047912598, 0.029281616, -0.01838684, 0.043640137, -0.010543823, -0.022445679, 0.04925537, -0.04876709, 0.0060043335, -0.009208679, -0.011772156, ...
http://arxiv.org/abs/2311.04507v1
Conversation Understanding using Relational Temporal Graph Neural Networks with Auxiliary Cross-Modality Interaction
Cam-Van Thi Nguyen, Anh-Tuan Mai, The-Son Le, Hai-Dang Kieu, Duc-Trong Le
cs.CL, cs.MM
Emotion recognition is a crucial task for human conversation understanding. It becomes more challenging with the notion of multimodal data, e.g., language, voice, and facial expressions. As a typical solution, the global- and the local context information are exploited to predict the emotional label for every single se...
http://arxiv.org/pdf/2311.04507v1
2023-11-08T07:46:25
2023-11-08T07:46:25
{ "title": [ -0.012680054, -0.003993988, -0.044128418, -0.071777344, -0.0149002075, -0.0033359528, -0.074035645, 0.01335907, 0.0033283234, 0.023544312, -0.0076904297, -0.009689331, -0.004081726, -0.022567749, -0.047851562, -0.010360718, -0.009346008,...
http://arxiv.org/abs/2311.04498v3
NExT-Chat: An LMM for Chat, Detection and Segmentation
Ao Zhang, Wei Ji, Tat-Seng Chua
cs.CV, cs.AI, cs.CL
The development of large language models (LLMs) has greatly advanced the field of multimodal understanding, leading to the emergence of large multimodal models (LMMs). In order to enhance the level of visual comprehension, recent studies have equipped LMMs with region-level understanding capabilities by representing ob...
http://arxiv.org/pdf/2311.04498v3
2023-11-13T03:35:23
2023-11-08T07:15:05
{ "title": [ -0.018157959, 0.0063095093, -0.025222778, -0.051757812, 0.010513306, 0.008354187, -0.043182373, -0.026489258, -0.0067825317, 0.03439331, -0.04827881, 0.01398468, 0.0014762878, 0.0051460266, -0.043060303, -0.012504578, -0.0025501251, ...
http://arxiv.org/abs/2311.04495v1
Multi-label and Multi-target Sampling of Machine Annotation for Computational Stance Detection
Zhengyuan Liu, Hai Leong Chieu, Nancy F. Chen
cs.CL
Data collection from manual labeling provides domain-specific and task-aligned supervision for data-driven approaches, and a critical mass of well-annotated resources is required to achieve reasonable performance in natural language processing tasks. However, manual annotations are often challenging to scale up in term...
http://arxiv.org/pdf/2311.04495v1
2023-11-08T06:54:34
2023-11-08T06:54:34
{ "title": [ -0.021240234, -0.00044751167, -0.014465332, -0.035217285, -0.024139404, -0.014518738, -0.032989502, -0.015716553, 0.010757446, 0.049621582, -0.04321289, 0.0050735474, -0.0056533813, -0.0053901672, -0.056793213, -0.010749817, -0.001810073...
http://arxiv.org/abs/2311.04480v1
CLearViD: Curriculum Learning for Video Description
Cheng-Yu Chuang, Pooyan Fazli
cs.CV, cs.CL, cs.LG
Video description entails automatically generating coherent natural language sentences that narrate the content of a given video. We introduce CLearViD, a transformer-based model for video description generation that leverages curriculum learning to accomplish this task. In particular, we investigate two curriculum str...
http://arxiv.org/pdf/2311.04480v1
2023-11-08T06:20:32
2023-11-08T06:20:32
{ "title": [ -0.019165039, -0.008102417, -0.053985596, -0.04550171, 0.0037517548, -0.022827148, -0.017166138, -0.011489868, -0.022964478, 0.026153564, -0.040252686, 0.030395508, -0.021240234, -0.00605011, -0.013404846, -0.023605347, -0.011764526, ...
http://arxiv.org/abs/2311.04479v1
Twitter Sentiment Analysis of Covid Vacciness
Wenbo Zhu, Tiechuan Hu
cs.CL, cs.IR, cs.LG, cs.SI
In this paper, we look at a database of tweets sorted by various keywords that could indicate the users sentiment towards covid vaccines. With social media becoming such a prevalent source of opinion, sorting and ranking tweets that hold important information such as opinions on covid vaccines is of utmost importance. ...
http://arxiv.org/pdf/2311.04479v1
2023-11-08T06:16:04
2023-11-08T06:16:04
{ "title": [ 0.027999878, 0.0031547546, -0.036315918, -0.03555298, -0.017440796, -0.014602661, 0.00178051, 0.0016012192, -0.02470398, 0.046142578, -0.035827637, 0.025466919, 0.008415222, -0.02760315, 0.0064468384, -0.012748718, 0.0057868958, 0.05...
http://arxiv.org/abs/2311.04467v1
RDGCN: Reinforced Dependency Graph Convolutional Network for Aspect-based Sentiment Analysis
Xusheng Zhao, Hao Peng, Qiong Dai, Xu Bai, Huailiang Peng, Yanbing Liu, Qinglang Guo, Philip S. Yu
cs.CL, cs.AI
Aspect-based sentiment analysis (ABSA) is dedicated to forecasting the sentiment polarity of aspect terms within sentences. Employing graph neural networks to capture structural patterns from syntactic dependency parsing has been confirmed as an effective approach for boosting ABSA. In most works, the topology of depen...
http://arxiv.org/pdf/2311.04467v1
2023-11-08T05:37:49
2023-11-08T05:37:49
{ "title": [ 0.000767231, -0.022033691, -0.034851074, -0.08996582, -0.0004491806, 0.016921997, -0.04812622, 0.019195557, 0.01828003, 0.043121338, -0.0023860931, -0.0021953583, -0.0010070801, -0.010429382, -0.06726074, 0.009765625, -0.009475708, 0...
http://arxiv.org/abs/2311.04459v1
Improving Pacing in Long-Form Story Planning
Yichen Wang, Kevin Yang, Xiaoming Liu, Dan Klein
cs.CL, cs.AI
Existing LLM-based systems for writing long-form stories or story outlines frequently suffer from unnatural pacing, whether glossing over important events or over-elaborating on insignificant details, resulting in a jarring experience for the reader. We propose a CONCrete Outline ConTrol (CONCOCT) system to improve pac...
http://arxiv.org/pdf/2311.04459v1
2023-11-08T04:58:29
2023-11-08T04:58:29
{ "title": [ 0.00020086765, -0.00081157684, -0.039978027, -0.007949829, -0.03729248, -0.035461426, 0.0058059692, 0.009460449, -0.06585693, 0.024047852, -0.02494812, 0.012840271, -0.003921509, -0.03955078, 0.03842163, -0.01876831, 0.010116577, -0....
http://arxiv.org/abs/2311.04453v1
Lewis's Signaling Game as beta-VAE For Natural Word Lengths and Segments
Ryo Ueda, Tadahiro Taniguchi
cs.CL
As a sub-discipline of evolutionary and computational linguistics, emergent communication (EC) studies communication protocols, called emergent languages, arising in simulations where agents communicate. A key goal of EC is to give rise to languages that share statistical properties with natural languages. In this pape...
http://arxiv.org/pdf/2311.04453v1
2023-11-08T04:34:30
2023-11-08T04:34:30
{ "title": [ -0.043060303, -0.042510986, -0.035858154, -0.00046539307, -0.07537842, -0.022628784, -0.017837524, 0.01953125, -0.06121826, 0.029037476, 0.0035953522, 0.036499023, -0.029403687, -0.038757324, 0.01826477, -0.008430481, -0.0034675598, ...
http://arxiv.org/abs/2311.04449v1
Recursion in Recursion: Two-Level Nested Recursion for Length Generalization with Scalability
Jishnu Ray Chowdhury, Cornelia Caragea
cs.LG, cs.CL
Binary Balanced Tree RvNNs (BBT-RvNNs) enforce sequence composition according to a preset balanced binary tree structure. Thus, their non-linear recursion depth is just $\log_2 n$ ($n$ being the sequence length). Such logarithmic scaling makes BBT-RvNNs efficient and scalable on long sequence tasks such as Long Range A...
http://arxiv.org/pdf/2311.04449v1
2023-11-08T04:20:56
2023-11-08T04:20:56
{ "title": [ -0.018508911, -0.02432251, -0.044830322, -0.090270996, -0.0095825195, -0.00894928, -0.060516357, 0.028152466, -0.012542725, 0.027160645, -0.016479492, 0.029556274, -0.028915405, -0.008399963, -0.024505615, -0.000790596, 0.008644104, ...
http://arxiv.org/abs/2311.04939v1
LooGLE: Can Long-Context Language Models Understand Long Contexts?
Jiaqi Li, Mengmeng Wang, Zilong Zheng, Muhan Zhang
cs.CL, cs.AI
Large language models (LLMs), despite their impressive performance in various language tasks, are typically limited to processing texts within context-window size. This limitation has spurred significant research efforts to enhance LLMs' long-context understanding with high-quality long-sequence benchmarks. However, pr...
http://arxiv.org/pdf/2311.04939v1
2023-11-08T01:45:37
2023-11-08T01:45:37
{ "title": [ -0.009979248, 0.0001834631, -0.031280518, -0.082092285, 0.013175964, 0.0013971329, -0.05154419, -0.006126404, 0.004852295, 0.03111267, -0.031341553, 0.005268097, 0.014526367, -0.020202637, -0.030029297, 0.012786865, -0.022567749, 0.0...
http://arxiv.org/abs/2311.04420v1
Data Factors for Better Compositional Generalization
Xiang Zhou, Yichen Jiang, Mohit Bansal
cs.CL, cs.AI, cs.LG
Recent diagnostic datasets on compositional generalization, such as SCAN (Lake and Baroni, 2018) and COGS (Kim and Linzen, 2020), expose severe problems in models trained from scratch on these datasets. However, in contrast to this poor performance, state-of-the-art models trained on larger and more general datasets sh...
http://arxiv.org/pdf/2311.04420v1
2023-11-08T01:27:34
2023-11-08T01:27:34
{ "title": [ -0.009178162, -0.0060768127, -0.038024902, -0.09118652, -0.006427765, 0.0014343262, -0.074035645, 0.007881165, 0.0048065186, 0.028182983, -0.02229309, 0.009895325, -0.03314209, -0.024963379, -0.042022705, 0.001376152, -0.018676758, 0...
http://arxiv.org/abs/2311.04378v2
Watermarks in the Sand: Impossibility of Strong Watermarking for Generative Models
Hanlin Zhang, Benjamin L. Edelman, Danilo Francati, Daniele Venturi, Giuseppe Ateniese, Boaz Barak
cs.LG, cs.CL, cs.CR
Watermarking generative models consists of planting a statistical signal (watermark) in a model's output so that it can be later verified that the output was generated by the given model. A strong watermarking scheme satisfies the property that a computationally bounded attacker cannot erase the watermark without causi...
http://arxiv.org/pdf/2311.04378v2
2023-11-15T00:21:29
2023-11-07T22:52:54
{ "title": [ -0.004764557, 0.007522583, -0.032165527, -0.03253174, 0.0052452087, -0.0126953125, -0.021972656, -0.043060303, 0.00063943863, 0.016220093, -0.0037956238, 0.008529663, -0.0050582886, -0.04309082, -0.020401001, 0.0131073, -0.004611969, ...
http://arxiv.org/abs/2311.04368v1
Evaluating multiple large language models in pediatric ophthalmology
Jason Holmes, Rui Peng, Yiwei Li, Jinyu Hu, Zhengliang Liu, Zihao Wu, Huan Zhao, Xi Jiang, Wei Liu, Hong Wei, Jie Zou, Tianming Liu, Yi Shao
cs.CL
IMPORTANCE The response effectiveness of different large language models (LLMs) and various individuals, including medical students, graduate students, and practicing physicians, in pediatric ophthalmology consultations, has not been clearly established yet. OBJECTIVE Design a 100-question exam based on pediatric ophth...
http://arxiv.org/pdf/2311.04368v1
2023-11-07T22:23:51
2023-11-07T22:23:51
{ "title": [ -0.053649902, -0.0335083, -0.05496216, -0.068359375, -0.017669678, 0.011268616, -0.084472656, 0.020126343, -0.025680542, 0.039886475, -0.0061683655, 0.0068740845, -0.02267456, -0.017486572, -0.03189087, -0.011962891, 0.0017166138, 0....
http://arxiv.org/abs/2311.04364v1
Syntax-Guided Transformers: Elevating Compositional Generalization and Grounding in Multimodal Environments
Danial Kamali, Parisa Kordjamshidi
cs.CL, cs.RO
Compositional generalization, the ability of intelligent models to extrapolate understanding of components to novel compositions, is a fundamental yet challenging facet in AI research, especially within multimodal environments. In this work, we address this challenge by exploiting the syntactic structure of language to...
http://arxiv.org/pdf/2311.04364v1
2023-11-07T21:59:16
2023-11-07T21:59:16
{ "title": [ -0.0006146431, -0.0007863045, -0.015235901, -0.10723877, 0.0013875961, 0.024520874, -0.08685303, 0.01675415, 0.0011787415, 0.025344849, -0.024246216, 0.022872925, -0.0020771027, -0.017105103, -0.032104492, -0.01322937, -0.018203735, ...
http://arxiv.org/abs/2311.04354v1
Uncovering Causal Variables in Transformers using Circuit Probing
Michael A. Lepori, Thomas Serre, Ellie Pavlick
cs.CL
Neural network models have achieved high performance on a wide variety of complex tasks, but the algorithms that they implement are notoriously difficult to interpret. In order to understand these algorithms, it is often necessary to hypothesize intermediate variables involved in the network's computation. For example,...
http://arxiv.org/pdf/2311.04354v1
2023-11-07T21:27:17
2023-11-07T21:27:17
{ "title": [ 0.00012266636, -0.00049877167, -0.025054932, -0.066833496, 0.0055122375, -0.0046195984, -0.020248413, -0.010185242, -0.02947998, 0.008682251, -0.0066223145, -0.00067329407, 0.0014390945, -0.049468994, 0.011512756, -0.013580322, -0.012969...
http://arxiv.org/abs/2311.04348v1
Evaluating the Effectiveness of Retrieval-Augmented Large Language Models in Scientific Document Reasoning
Sai Munikoti, Anurag Acharya, Sridevi Wagle, Sameera Horawalavithana
cs.CL, cs.AI, I.2.7
Despite the dramatic progress in Large Language Model (LLM) development, LLMs often provide seemingly plausible but not factual information, often referred to as hallucinations. Retrieval-augmented LLMs provide a non-parametric approach to solve these issues by retrieving relevant information from external data sources...
http://arxiv.org/pdf/2311.04348v1
2023-11-07T21:09:57
2023-11-07T21:09:57
{ "title": [ -0.029449463, -0.019180298, -0.02281189, -0.047668457, -0.022415161, 0.018234253, -0.050933838, -0.0038032532, 0.019104004, 0.03338623, -0.0027828217, -0.02230835, 0.019134521, -0.018554688, -0.049346924, 0.0134887695, -0.015205383, ...
http://arxiv.org/abs/2311.04345v1
A Taxonomy of Rater Disagreements: Surveying Challenges & Opportunities from the Perspective of Annotating Online Toxicity
Wenbo Zhang, Hangzhi Guo, Ian D Kivlichan, Vinodkumar Prabhakaran, Davis Yadav, Amulya Yadav
cs.CL, cs.AI
Toxicity is an increasingly common and severe issue in online spaces. Consequently, a rich line of machine learning research over the past decade has focused on computationally detecting and mitigating online toxicity. These efforts crucially rely on human-annotated datasets that identify toxic content of various kinds...
http://arxiv.org/pdf/2311.04345v1
2023-11-07T21:00:51
2023-11-07T21:00:51
{ "title": [ 0.02532959, -0.011528015, 0.011665344, -0.052490234, -0.0044517517, -0.0011577606, 0.0055503845, -0.012130737, -0.0021591187, 0.0031356812, -0.017990112, 0.018096924, -0.03137207, -0.035339355, -0.01675415, -0.02810669, -0.007610321, ...
http://arxiv.org/abs/2311.04335v1
Sub-Sentence Encoder: Contrastive Learning of Propositional Semantic Representations
Sihao Chen, Hongming Zhang, Tong Chen, Ben Zhou, Wenhao Yu, Dian Yu, Baolin Peng, Hongwei Wang, Dan Roth, Dong Yu
cs.CL, cs.AI
We introduce sub-sentence encoder, a contrastively-learned contextual embedding model for fine-grained semantic representation of text. In contrast to the standard practice with sentence embeddings, where the meaning of an entire sequence of text is encoded into a fixed-length vector, the sub-sentence encoder learns to...
http://arxiv.org/pdf/2311.04335v1
2023-11-07T20:38:30
2023-11-07T20:38:30
{ "title": [ -0.008483887, 0.0013780594, -0.01576233, -0.042175293, -0.0010242462, -0.015670776, -0.018417358, -0.021133423, 0.023010254, 0.03201294, -0.036346436, 0.025115967, 0.00806427, -0.006542206, -0.036895752, 0.012298584, -0.0088272095, 0...
http://arxiv.org/abs/2311.04329v1
Formal Aspects of Language Modeling
Ryan Cotterell, Anej Svete, Clara Meister, Tianyu Liu, Li Du
cs.CL
Large language models have become one of the most commonly deployed NLP inventions. In the past half-decade, their integration into core natural language processing tools has dramatically increased the performance of such tools, and they have entered the public discourse surrounding artificial intelligence. Consequentl...
http://arxiv.org/pdf/2311.04329v1
2023-11-07T20:21:42
2023-11-07T20:21:42
{ "title": [ -0.030075073, -0.0060539246, -0.03768921, -0.09472656, 0.010658264, 0.020507812, -0.0736084, 0.01826477, -0.0010242462, 0.039733887, -0.0006480217, -0.0084991455, 0.004940033, -0.025222778, -0.05392456, 0.024627686, -0.01965332, 0.04...
http://arxiv.org/abs/2311.04936v1
A comparative analysis between Conformer-Transducer, Whisper, and wav2vec2 for improving the child speech recognition
Andrei Barcovschi, Rishabh Jain, Peter Corcoran
cs.CL, cs.AI, cs.SD, eess.AS
Automatic Speech Recognition (ASR) systems have progressed significantly in their performance on adult speech data; however, transcribing child speech remains challenging due to the acoustic differences in the characteristics of child and adult voices. This work aims to explore the potential of adapting state-of-the-ar...
http://arxiv.org/pdf/2311.04936v1
2023-11-07T19:32:48
2023-11-07T19:32:48
{ "title": [ 0.00069475174, -0.010307312, -0.052734375, -0.06628418, -0.016113281, -0.015045166, -0.046417236, 0.023254395, -0.04736328, 0.0637207, -0.025527954, 0.006095886, 0.018798828, -0.019866943, -0.028961182, -0.02810669, -0.016555786, 0.0...
http://arxiv.org/abs/2311.04292v1
Aspect-based Meeting Transcript Summarization: A Two-Stage Approach with Weak Supervision on Sentence Classification
Zhongfen Deng, Seunghyun Yoon, Trung Bui, Franck Dernoncourt, Quan Hung Tran, Shuaiqi Liu, Wenting Zhao, Tao Zhang, Yibo Wang, Philip S. Yu
cs.CL
Aspect-based meeting transcript summarization aims to produce multiple summaries, each focusing on one aspect of content in a meeting transcript. It is challenging as sentences related to different aspects can mingle together, and those relevant to a specific aspect can be scattered throughout the long transcript of a ...
http://arxiv.org/pdf/2311.04292v1
2023-11-07T19:06:31
2023-11-07T19:06:31
{ "title": [ -0.005794525, 0.007675171, -0.04284668, -0.108947754, 0.0006656647, 0.019683838, -0.08685303, 0.015068054, -0.0012359619, 0.057861328, 0.013175964, 0.0079193115, 0.0060768127, -0.0071258545, -0.062194824, 0.0048065186, -0.018859863, ...
http://arxiv.org/abs/2311.04284v1
CRAB: Assessing the Strength of Causal Relationships Between Real-world Events
Angelika Romanou, Syrielle Montariol, Debjit Paul, Leo Laugier, Karl Aberer, Antoine Bosselut
cs.CL, cs.AI
Understanding narratives requires reasoning about the cause-and-effect relationships between events mentioned in the text. While existing foundation models yield impressive results in many NLP tasks requiring reasoning, it is unclear whether they understand the complexity of the underlying network of causal relationshi...
http://arxiv.org/pdf/2311.04284v1
2023-11-07T19:00:44
2023-11-07T19:00:44
{ "title": [ 0.003484726, -0.01285553, -0.05404663, -0.059509277, 0.007133484, -0.023651123, -0.023986816, -0.00006234646, -0.010147095, -0.002298355, -0.020339966, 0.006465912, -0.0012750626, -0.043670654, -0.0011978149, -0.014633179, 0.0025539398, ...
http://arxiv.org/abs/2311.04205v1
Rephrase and Respond: Let Large Language Models Ask Better Questions for Themselves
Yihe Deng, Weitong Zhang, Zixiang Chen, Quanquan Gu
cs.CL, cs.AI, cs.LG
Misunderstandings arise not only in interpersonal communication but also between humans and Large Language Models (LLMs). Such discrepancies can make LLMs interpret seemingly unambiguous questions in unexpected ways, yielding incorrect responses. While it is widely acknowledged that the quality of a prompt, such as a q...
http://arxiv.org/pdf/2311.04205v1
2023-11-07T18:43:34
2023-11-07T18:43:34
{ "title": [ -0.02142334, -0.036376953, -0.0051956177, -0.045043945, -0.060668945, 0.011520386, -0.056152344, 0.0028591156, -0.01184082, 0.012046814, -0.015014648, -0.016937256, -0.020584106, -0.03475952, -0.015159607, -0.012275696, 0.008491516, ...
http://arxiv.org/abs/2311.04199v1
Exploring Recommendation Capabilities of GPT-4V(ision): A Preliminary Case Study
Peilin Zhou, Meng Cao, You-Liang Huang, Qichen Ye, Peiyan Zhang, Junling Liu, Yueqi Xie, Yining Hua, Jaeboum Kim
cs.IR, cs.CL
Large Multimodal Models (LMMs) have demonstrated impressive performance across various vision and language tasks, yet their potential applications in recommendation tasks with visual assistance remain unexplored. To bridge this gap, we present a preliminary case study investigating the recommendation capabilities of GP...
http://arxiv.org/pdf/2311.04199v1
2023-11-07T18:39:10
2023-11-07T18:39:10
{ "title": [ 0.008354187, 0.005970001, -0.015808105, -0.022888184, 0.007801056, -0.006828308, -0.0072631836, -0.0357666, -0.026870728, 0.016662598, -0.03161621, 0.011222839, -0.01777649, -0.018356323, -0.02458191, -0.01108551, -0.02633667, 0.0413...
http://arxiv.org/abs/2311.04196v1
JPAVE: A Generation and Classification-based Model for Joint Product Attribute Prediction and Value Extraction
Zhongfen Deng, Hao Peng, Tao Zhang, Shuaiqi Liu, Wenting Zhao, Yibo Wang, Philip S. Yu
cs.CL, cs.AI
Product attribute value extraction is an important task in e-Commerce which can help several downstream applications such as product search and recommendation. Most previous models handle this task using sequence labeling or question answering method which rely on the sequential position information of values in the pr...
http://arxiv.org/pdf/2311.04196v1
2023-11-07T18:36:16
2023-11-07T18:36:16
{ "title": [ -0.033203125, -0.0069885254, -0.056884766, -0.057922363, 0.014656067, -0.00014591217, -0.034820557, -0.009681702, -0.010513306, 0.024505615, 0.008460999, -0.007282257, -0.013893127, -0.02911377, -0.04824829, -0.0012302399, -0.03062439, ...
http://arxiv.org/abs/2311.04192v1
JaSPICE: Automatic Evaluation Metric Using Predicate-Argument Structures for Image Captioning Models
Yuiga Wada, Kanta Kaneda, Komei Sugiura
cs.CV, cs.CL
Image captioning studies heavily rely on automatic evaluation metrics such as BLEU and METEOR. However, such n-gram-based metrics have been shown to correlate poorly with human evaluation, leading to the proposal of alternative metrics such as SPICE for English; however, no equivalent metrics have been established for ...
http://arxiv.org/pdf/2311.04192v1
2023-11-07T18:33:34
2023-11-07T18:33:34
{ "title": [ 0.00012671947, 0.0013618469, -0.0024204254, -0.032470703, -0.0040740967, -0.0066375732, -0.014968872, -0.03729248, 0.0126953125, 0.043518066, -0.029251099, 0.011077881, -0.0007429123, -0.0067977905, -0.049560547, -0.009994507, -0.0115661...
http://arxiv.org/abs/2311.04189v1
SpaDeLeF: A Dataset for Hierarchical Classification of Lexical Functions for Collocations in Spanish
Yevhen Kostiuk, Grigori Sidorov, Olga Kolesnikova
cs.CL
In natural language processing (NLP), lexical function is a concept to unambiguously represent semantic and syntactic features of words and phrases in text first crafted in the Meaning-Text Theory. Hierarchical classification of lexical functions involves organizing these features into a tree-like hierarchy of categori...
http://arxiv.org/pdf/2311.04189v1
2023-11-07T18:32:34
2023-11-07T18:32:34
{ "title": [ -0.0068359375, 0.00019216537, -0.042755127, -0.06994629, 0.0006251335, 0.0044517517, -0.047424316, 0.012878418, 0.011764526, 0.035125732, -0.015853882, 0.026504517, 0.006855011, -0.016204834, -0.036010742, 0.009719849, -0.032470703, ...
http://arxiv.org/abs/2311.04934v1
Prompt Cache: Modular Attention Reuse for Low-Latency Inference
In Gim, Guojun Chen, Seung-seob Lee, Nikhil Sarda, Anurag Khandelwal, Lin Zhong
cs.CL, cs.AI
We present Prompt Cache, an approach for accelerating inference for large language models (LLM) by reusing attention states across different LLM prompts. Many input prompts have overlapping text segments, such as system messages, prompt templates, and documents provided for context. Our key insight is that by precomput...
http://arxiv.org/pdf/2311.04934v1
2023-11-07T18:17:05
2023-11-07T18:17:05
{ "title": [ -0.015052795, -0.007167816, -0.034698486, -0.08892822, 0.0047683716, -0.003370285, -0.041412354, 0.002872467, 0.023910522, 0.043914795, -0.03479004, -0.009963989, 0.01864624, -0.0013303757, -0.052612305, 0.006324768, 0.005947113, 0.0...
http://arxiv.org/abs/2311.04177v1
Enhancing LLM Intelligence with ARM-RAG: Auxiliary Rationale Memory for Retrieval Augmented Generation
Eric Melz
cs.CL, cs.AI
Large Language Models (LLMs) are smart but forgetful. Recent studies, (e.g., (Bubeck et al., 2023)) on modern LLMs have shown that they are capable of performing amazing tasks typically necessitating human-level intelligence. However, unlike humans, frozen LLMs do not improve over time; they neither acquire new knowled...
http://arxiv.org/pdf/2311.04177v1
2023-11-07T18:03:23
2023-11-07T18:03:23
{ "title": [ -0.026489258, -0.032226562, -0.048736572, -0.0070648193, -0.054718018, -0.01725769, -0.014312744, 0.008720398, 0.0036010742, 0.03894043, 0.00015938282, 0.014434814, 0.011077881, -0.034454346, -0.024337769, 0.018188477, -0.018707275, ...
http://arxiv.org/abs/2311.04166v1
Perturbed examples reveal invariances shared by language models
Ruchit Rawal, Mariya Toneva
cs.CL, cs.LG
An explosion of work in language is leading to ever-increasing numbers of available natural language processing models, with little understanding of how new models compare to better-understood models. One major reason for this difficulty is saturating benchmark datasets, which may not reflect well differences in model ...
http://arxiv.org/pdf/2311.04166v1
2023-11-07T17:48:35
2023-11-07T17:48:35
{ "title": [ 0.029891968, 0.0016202927, -0.016403198, -0.03277588, -0.034576416, -0.025665283, -0.041412354, -0.023086548, -0.010749817, 0.023452759, -0.02772522, 0.017715454, 0.010063171, -0.0345459, 0.004459381, -0.004348755, -0.0007481575, 0.0...
http://arxiv.org/abs/2311.04155v2
Black-Box Prompt Optimization: Aligning Large Language Models without Model Training
Jiale Cheng, Xiao Liu, Kehan Zheng, Pei Ke, Hongning Wang, Yuxiao Dong, Jie Tang, Minlie Huang
cs.CL
Large language models (LLMs) have shown impressive success in various applications. However, these models are often not well aligned with human intents, which calls for additional treatments on them, that is, the alignment problem. To make LLMs better follow user instructions, existing alignment methods mostly focus on...
http://arxiv.org/pdf/2311.04155v2
2023-11-08T04:21:41
2023-11-07T17:31:50
{ "title": [ -0.04095459, -0.006713867, -0.05960083, -0.06173706, -0.004360199, 0.00039076805, -0.08831787, -0.008132935, -0.0068740845, 0.03186035, -0.010688782, -0.0016546249, -0.0022029877, -0.011001587, -0.0413208, 0.007797241, -0.0007276535, ...
http://arxiv.org/abs/2311.04142v1
What is Lost in Knowledge Distillation?
Manas Mohanty, Tanya Roosta, Peyman Passban
cs.CL
Deep neural networks (DNNs) have improved NLP tasks significantly, but training and maintaining such networks could be costly. Model compression techniques, such as, knowledge distillation (KD), have been proposed to address the issue; however, the compression process could be lossy. Motivated by this, our work investi...
http://arxiv.org/pdf/2311.04142v1
2023-11-07T17:13:40
2023-11-07T17:13:40
{ "title": [ -0.03503418, -0.003824234, -0.056030273, 0.0075645447, -0.0284729, -0.016479492, -0.03152466, -0.011329651, 0.03213501, 0.032470703, -0.052459717, -0.058044434, 0.005214691, -0.013832092, -0.03793335, -0.053833008, 0.02519226, 0.0238...
http://arxiv.org/abs/2311.04139v1
Modelling Sentiment Analysis: LLMs and data augmentation techniques
Guillem Senabre Prades
cs.CL
This paper provides different approaches for a binary sentiment classification on a small training dataset. LLMs that provided state-of-the-art results in sentiment analysis and similar domains are being used, such as BERT, RoBERTa and XLNet.
http://arxiv.org/pdf/2311.04139v1
2023-11-07T17:12:39
2023-11-07T17:12:39
{ "title": [ -0.0131073, 0.0007362366, -0.04638672, -0.03286743, 0.0138168335, -0.010864258, -0.0087509155, -0.016616821, 0.013923645, 0.02583313, -0.038238525, -0.002943039, -0.021209717, -0.0105896, -0.06591797, -0.008735657, -0.012886047, 0.03...
http://arxiv.org/abs/2311.04131v1
Locating Cross-Task Sequence Continuation Circuits in Transformers
Michael Lan, Fazl Barez
cs.CL, cs.AI, cs.LG
While transformer models exhibit strong capabilities on linguistic tasks, their complex architectures make them difficult to interpret. Recent work has aimed to reverse engineer transformer models into human-readable representations called circuits that implement algorithmic functions. We extend this research by analyz...
http://arxiv.org/pdf/2311.04131v1
2023-11-07T16:58:51
2023-11-07T16:58:51
{ "title": [ 0.03378296, 0.017608643, -0.0048065186, -0.05230713, 0.019363403, -0.022766113, 0.0067825317, -0.041656494, -0.01838684, 0.059692383, -0.013465881, 0.014732361, 0.035858154, -0.010925293, -0.017425537, 0.024887085, -0.009254456, 0.05...
http://arxiv.org/abs/2311.04124v1
Unveiling Safety Vulnerabilities of Large Language Models
George Kour, Marcel Zalmanovici, Naama Zwerdling, Esther Goldbraich, Ora Nova Fandina, Ateret Anaby-Tavor, Orna Raz, Eitan Farchi
cs.CL, cs.AI, cs.LG, I.2.7
As large language models become more prevalent, their possible harmful or inappropriate responses are a cause for concern. This paper introduces a unique dataset containing adversarial examples in the form of questions, which we call AttaQ, designed to provoke such harmful or inappropriate responses. We assess the effi...
http://arxiv.org/pdf/2311.04124v1
2023-11-07T16:50:33
2023-11-07T16:50:33
{ "title": [ -0.015022278, -0.0052223206, -0.04824829, -0.090270996, 0.017822266, -0.0011043549, -0.048187256, 0.017593384, -0.01739502, 0.026367188, -0.008049011, 0.008018494, -0.009429932, -0.013656616, -0.0259552, 0.0059928894, -0.013053894, 0...
http://arxiv.org/abs/2311.04933v1
Evaluating Large Language Models in Ophthalmology
Jason Holmes, Shuyuan Ye, Yiwei Li, Shi-Nan Wu, Zhengliang Liu, Zihao Wu, Jinyu Hu, Huan Zhao, Xi Jiang, Wei Liu, Hong Wei, Jie Zou, Tianming Liu, Yi Shao
cs.CL, cs.AI
Purpose: The performance of three different large language models (LLMS) (GPT-3.5, GPT-4, and PaLM2) in answering ophthalmology professional questions was evaluated and compared with that of three different professional populations (medical undergraduates, medical masters, and attending physicians). Methods: A 100-item...
http://arxiv.org/pdf/2311.04933v1
2023-11-07T16:19:45
2023-11-07T16:19:45
{ "title": [ -0.05380249, -0.022781372, -0.04949951, -0.05633545, -0.011734009, 0.008705139, -0.07354736, 0.0071487427, -0.008026123, 0.03692627, -0.010017395, -0.012886047, -0.0128479, -0.019973755, -0.0519104, -0.006549835, 0.0134887695, 0.0090...
http://arxiv.org/abs/2311.04088v1
Personality Style Recognition via Machine Learning: Identifying Anaclitic and Introjective Personality Styles from Patients' Speech
Semere Kiros Bitew, Vincent Schelstraete, Klim Zaporojets, Kimberly Van Nieuwenhove, Reitske Meganck, Chris Develder
cs.CL
In disentangling the heterogeneity observed in psychopathology, personality of the patients is considered crucial. While it has been demonstrated that personality traits are reflected in the language used by a patient, we hypothesize that this enables automatic inference of the personality type directly from speech utt...
http://arxiv.org/pdf/2311.04088v1
2023-11-07T15:56:19
2023-11-07T15:56:19
{ "title": [ -0.006969452, 0.016159058, -0.048309326, -0.04168701, 0.0053596497, -0.025527954, -0.03643799, -0.015319824, -0.050048828, 0.060699463, 0.009300232, 0.0067367554, -0.019439697, -0.028381348, -0.040618896, -0.016357422, -0.021530151, ...
http://arxiv.org/abs/2311.04076v1
Do LLMs exhibit human-like response biases? A case study in survey design
Lindia Tjuatja, Valerie Chen, Sherry Tongshuang Wu, Ameet Talwalkar, Graham Neubig
cs.CL
As large language models (LLMs) become more capable, there is growing excitement about the possibility of using LLMs as proxies for humans in real-world tasks where subjective labels are desired, such as in surveys and opinion polling. One widely-cited barrier to the adoption of LLMs is their sensitivity to prompt word...
http://arxiv.org/pdf/2311.04076v1
2023-11-07T15:40:43
2023-11-07T15:40:43
{ "title": [ -0.04626465, 0.0059509277, -0.04071045, 0.036712646, -0.028198242, -0.03970337, 0.010368347, -0.04837036, -0.015007019, 0.0062942505, -0.058563232, 0.0340271, -0.059051514, -0.0004901886, -0.042785645, -0.03955078, -0.022827148, 0.02...
http://arxiv.org/abs/2311.04260v1
Fully Automated Task Management for Generation, Execution, and Evaluation: A Framework for Fetch-and-Carry Tasks with Natural Language Instructions in Continuous Space
Motonari Kambara, Komei Sugiura
cs.RO, cs.CL, cs.CV
This paper aims to develop a framework that enables a robot to execute tasks based on visual information, in response to natural language instructions for Fetch-and-Carry with Object Grounding (FCOG) tasks. Although there have been many frameworks, they usually rely on manually given instruction sentences. Therefore, e...
http://arxiv.org/pdf/2311.04260v1
2023-11-07T15:38:09
2023-11-07T15:38:09
{ "title": [ -0.00957489, -0.044952393, -0.028305054, -0.070007324, -0.036224365, -0.026992798, -0.06329346, 0.028640747, -0.009605408, 0.045410156, -0.004550934, -0.0005598068, 0.021118164, -0.016983032, -0.017166138, -0.004714966, 0.0112838745, ...
http://arxiv.org/abs/2311.04072v1
Beyond Imitation: Leveraging Fine-grained Quality Signals for Alignment
Geyang Guo, Ranchi Zhao, Tianyi Tang, Wayne Xin Zhao, Ji-Rong Wen
cs.CL
Alignment with human preference is a desired property of large language models (LLMs). Currently, the main alignment approach is based on reinforcement learning from human feedback (RLHF). Despite the effectiveness of RLHF, it is intricate to implement and train, thus recent studies explore how to develop alternative a...
http://arxiv.org/pdf/2311.04072v1
2023-11-07T15:36:40
2023-11-07T15:36:40
{ "title": [ -0.011398315, -0.0059013367, -0.01499176, -0.04083252, 0.0042762756, -0.0027236938, -0.031097412, -0.0357666, 0.000117242336, 0.04333496, -0.006526947, -0.0069732666, 0.011276245, -0.019332886, -0.053100586, 0.0022468567, 0.016052246, ...
http://arxiv.org/abs/2311.04064v1
Implementation and Comparison of Methods to Extract Reliability KPIs out of Textual Wind Turbine Maintenance Work Orders
Marc-Alexander Lutz, Bastian Schäfermeier, Rachael Sexton, Michael Sharp, Alden Dima, Stefan Faulstich, Jagan Mohini Aluri
cs.CL, cs.LG, I.2.7; I.7
Maintenance work orders are commonly used to document information about wind turbine operation and maintenance. This includes details about proactive and reactive wind turbine downtimes, such as preventative and corrective maintenance. However, the information contained in maintenance work orders is often unstructured ...
http://arxiv.org/pdf/2311.04064v1
2023-11-07T15:25:52
2023-11-07T15:25:52
{ "title": [ -0.0020122528, 0.00081157684, -0.04421997, -0.012763977, -0.0005068779, -0.030883789, 0.021194458, -0.0048065186, -0.03643799, 0.021240234, -0.057128906, 0.013679504, -0.009864807, -0.020507812, -0.0006532669, -0.048980713, -0.0016880035...
http://arxiv.org/abs/2311.04046v1
Reinforcement Learning Fine-tuning of Language Models is Biased Towards More Extractable Features
Diogo Cruz, Edoardo Pona, Alex Holness-Tofts, Elias Schmied, Víctor Abia Alonso, Charlie Griffin, Bogdan-Ionut Cirstea
cs.LG, cs.CL
Many capable large language models (LLMs) are developed via self-supervised pre-training followed by a reinforcement-learning fine-tuning phase, often based on human or AI feedback. During this stage, models may be guided by their inductive biases to rely on simpler features which may be easier to extract, at a cost to...
http://arxiv.org/pdf/2311.04046v1
2023-11-07T15:00:39
2023-11-07T15:00:39
{ "title": [ -0.00062942505, -0.0023670197, -0.024490356, -0.02659607, -0.021209717, -0.031829834, -0.012275696, -0.0026168823, -0.03781128, 0.026046753, -0.038116455, 0.024459839, -0.02659607, -0.012420654, -0.008201599, -0.024780273, 0.0001718998, ...
http://arxiv.org/abs/2311.04044v1
P-Bench: A Multi-level Privacy Evaluation Benchmark for Language Models
Haoran Li, Dadi Guo, Donghao Li, Wei Fan, Qi Hu, Xin Liu, Chunkit Chan, Duanyi Yao, Yangqiu Song
cs.CL, cs.CR
The rapid development of language models (LMs) brings unprecedented accessibility and usage for both models and users. On the one hand, powerful LMs, trained with massive textual data, achieve state-of-the-art performance over numerous downstream NLP tasks. On the other hand, more and more attention is paid to unrestri...
http://arxiv.org/pdf/2311.04044v1
2023-11-07T14:55:52
2023-11-07T14:55:52
{ "title": [ -0.04119873, -0.014335632, -0.050476074, -0.064086914, 0.0041122437, 0.00006991625, -0.04916382, 0.028289795, -0.008346558, 0.026367188, -0.015327454, 0.036376953, 0.012176514, -0.014190674, -0.025497437, 0.012298584, -0.025619507, 0...
http://arxiv.org/abs/2311.04257v2
mPLUG-Owl2: Revolutionizing Multi-modal Large Language Model with Modality Collaboration
Qinghao Ye, Haiyang Xu, Jiabo Ye, Ming Yan, Anwen Hu, Haowei Liu, Qi Qian, Ji Zhang, Fei Huang, Jingren Zhou
cs.CL, cs.CV
Multi-modal Large Language Models (MLLMs) have demonstrated impressive instruction abilities across various open-ended tasks. However, previous methods primarily focus on enhancing multi-modal capabilities. In this work, we introduce a versatile multi-modal large language model, mPLUG-Owl2, which effectively leverages ...
http://arxiv.org/pdf/2311.04257v2
2023-11-09T01:56:51
2023-11-07T14:21:29
{ "title": [ -0.025619507, -0.013687134, -0.05657959, -0.07635498, -0.0040359497, 0.0011320114, -0.06738281, 0.0079574585, -0.0032196045, 0.037017822, 0.004776001, 0.011703491, 0.01637268, -0.022567749, -0.02798462, 0.018447876, -0.020431519, 0.0...
http://arxiv.org/abs/2311.04020v1
Analyzing Film Adaptation through Narrative Alignment
Tanzir Pial, Shahreen Salim, Charuta Pethe, Allen Kim, Steven Skiena
cs.CL
Novels are often adapted into feature films, but the differences between the two media usually require dropping sections of the source text from the movie script. Here we study this screen adaptation process by constructing narrative alignments using the Smith-Waterman local alignment algorithm coupled with SBERT embed...
http://arxiv.org/pdf/2311.04020v1
2023-11-07T14:18:03
2023-11-07T14:18:03
{ "title": [ -0.02470398, -0.005420685, -0.046783447, 0.0000030398369, -0.03201294, -0.022247314, -0.0019245148, -0.011268616, -0.055755615, 0.04864502, 0.006187439, 0.008117676, -0.0070991516, -0.033325195, -0.02607727, -0.015136719, 0.00655365, ...
http://arxiv.org/abs/2311.06303v1
MatNexus: A Comprehensive Text Mining and Analysis Suite for Materials Discover
Lei Zhang, Markus Stricker
cond-mat.mtrl-sci, cs.CL, physics.chem-ph, H.4; H.5; I.5; I.7; J.2
MatNexus is a specialized software for the automated collection, processing, and analysis of text from scientific articles. Through an integrated suite of modules, the MatNexus facilitates the retrieval of scientific articles, processes textual data for insights, generates vector representations suitable for machine le...
http://arxiv.org/pdf/2311.06303v1
2023-11-07T14:14:36
2023-11-07T14:14:36
{ "title": [ -0.054840088, -0.033721924, -0.058135986, -0.04168701, -0.012420654, -0.0035305023, -0.03778076, 0.002128601, 0.0066452026, 0.016021729, -0.029830933, 0.014221191, -0.0096206665, -0.015686035, -0.030639648, 0.006504059, -0.023620605, ...
http://arxiv.org/abs/2311.03998v1
Exploring Jiu-Jitsu Argumentation for Writing Peer Review Rebuttals
Sukannya Purkayastha, Anne Lauscher, Iryna Gurevych
cs.CL
In many domains of argumentation, people's arguments are driven by so-called attitude roots, i.e., underlying beliefs and world views, and their corresponding attitude themes. Given the strength of these latent drivers of arguments, recent work in psychology suggests that instead of directly countering surface-level re...
http://arxiv.org/pdf/2311.03998v1
2023-11-07T13:54:01
2023-11-07T13:54:01
{ "title": [ 0.0036716461, -0.023712158, -0.011451721, -0.012367249, -0.041412354, 0.0002822876, -0.0023498535, -0.017715454, -0.0758667, 0.048828125, 0.03173828, -0.009483337, 0.0035381317, -0.04397583, -0.0032520294, -0.016906738, 0.011550903, ...
http://arxiv.org/abs/2311.03969v1
Factoring Hate Speech: A New Annotation Framework to Study Hate Speech in Social Media
Gal Ron, Effi Levi, Odelia Oshri, Shaul R. Shenhav
cs.CL
In this work we propose a novel annotation scheme which factors hate speech into five separate discursive categories. To evaluate our scheme, we construct a corpus of over 2.9M Twitter posts containing hateful expressions directed at Jews, and annotate a sample dataset of 1,050 tweets. We present a statistical analysis...
http://arxiv.org/pdf/2311.03969v1
2023-11-07T13:08:55
2023-11-07T13:08:55
{ "title": [ 0.027709961, -0.0006222725, -0.028335571, -0.06329346, 0.017807007, -0.020233154, -0.0006599426, 0.00027871132, -0.031051636, 0.04434204, -0.0027999878, 0.037719727, 0.007896423, -0.0128479, -0.04473877, 0.0044517517, -0.014579773, 0...
http://arxiv.org/abs/2311.03963v1
An Expectation-Realization Model for Metaphor Detection
Oseremen O. Uduehi, Razvan C. Bunescu
cs.CL, cs.AI
We propose a metaphor detection architecture that is structured around two main modules: an expectation component that estimates representations of literal word expectations given a context, and a realization component that computes representations of actual word meanings in context. The overall architecture is trained...
http://arxiv.org/pdf/2311.03963v1
2023-11-07T13:03:54
2023-11-07T13:03:54
{ "title": [ 0.0043258667, 0.012214661, -0.005584717, 0.007293701, 0.014717102, -0.015823364, 0.019592285, -0.049316406, 0.011413574, 0.03378296, -0.006038666, -0.02168274, 0.03866577, -0.01826477, -0.059326172, 0.004020691, -0.020858765, 0.05279...
http://arxiv.org/abs/2311.03952v1
An Analysis of Dialogue Repair in Voice Assistants
Matthew Galbraith
cs.CL, cs.HC, cs.RO
Spoken dialogue systems have transformed human-machine interaction by providing real-time responses to queries. However, misunderstandings between the user and system persist. This study explores the significance of interactional language in dialogue repair between virtual assistants and users by analyzing interactions...
http://arxiv.org/pdf/2311.03952v1
2023-11-07T12:50:11
2023-11-07T12:50:11
{ "title": [ 0.013366699, -0.0211792, -0.027145386, -0.076660156, -0.006374359, -0.010543823, -0.040008545, 0.029556274, -0.02557373, 0.055603027, -0.003156662, -0.002281189, 0.029159546, -0.018234253, -0.04434204, 0.010253906, -0.024597168, 0.02...
http://arxiv.org/abs/2311.03928v1
Improving Korean NLP Tasks with Linguistically Informed Subword Tokenization and Sub-character Decomposition
Taehee Jeon, Bongseok Yang, Changhwan Kim, Yoonseob Lim
cs.CL, 68T50
We introduce a morpheme-aware subword tokenization method that utilizes sub-character decomposition to address the challenges of applying Byte Pair Encoding (BPE) to Korean, a language characterized by its rich morphology and unique writing system. Our approach balances linguistic accuracy with computational efficiency...
http://arxiv.org/pdf/2311.03928v1
2023-11-07T12:08:21
2023-11-07T12:08:21
{ "title": [ -0.035614014, -0.011260986, -0.02130127, -0.039001465, -0.021408081, 0.0026893616, -0.050811768, -0.0059547424, -0.0055503845, 0.032196045, -0.024169922, 0.0259552, 0.029464722, -0.0045433044, -0.02029419, 0.008377075, -0.018707275, ...
http://arxiv.org/abs/2311.03896v1
iACOS: Advancing Implicit Sentiment Extraction with Informative and Adaptive Negative Examples
Xiancai Xu, Jia-Dong Zhang, Lei Xiong, Zhishang Liu
cs.CL
Aspect-based sentiment analysis (ABSA) have been extensively studied, but little light has been shed on the quadruple extraction consisting of four fundamental elements: aspects, categories, opinions and sentiments, especially with implicit aspects and opinions. In this paper, we propose a new method iACOS for extracti...
http://arxiv.org/pdf/2311.03896v1
2023-11-07T11:19:06
2023-11-07T11:19:06
{ "title": [ 0.017425537, -0.008583069, -0.0077056885, -0.053955078, 0.013656616, 0.0007829666, -0.004585266, -0.016647339, 0.045166016, 0.03060913, -0.04852295, -0.016433716, -0.0007958412, 0.004764557, -0.07366943, -0.006965637, -0.016540527, 0...
http://arxiv.org/abs/2311.04250v1
Unifying Structure and Language Semantic for Efficient Contrastive Knowledge Graph Completion with Structured Entity Anchors
Sang-Hyun Je, Wontae Choi, Kwangjin Oh
cs.AI, cs.CL, cs.LG
The goal of knowledge graph completion (KGC) is to predict missing links in a KG using trained facts that are already known. In recent, pre-trained language model (PLM) based methods that utilize both textual and structural information are emerging, but their performances lag behind state-of-the-art (SOTA) structure-ba...
http://arxiv.org/pdf/2311.04250v1
2023-11-07T11:17:55
2023-11-07T11:17:55
{ "title": [ -0.01159668, -0.012374878, -0.046905518, -0.0914917, 0.0022239685, 0.0011730194, -0.053863525, 0.0072669983, 0.020874023, 0.027328491, -0.025909424, 0.016494751, 0.019470215, -0.006378174, -0.032165527, -0.0016441345, -0.0024871826, ...
http://arxiv.org/abs/2311.03881v1
Sparse Contrastive Learning of Sentence Embeddings
Ruize An, Chen Zhang, Dawei Song
cs.CL
Recently, SimCSE has shown the feasibility of contrastive learning in training sentence embeddings and illustrates its expressiveness in spanning an aligned and uniform embedding space. However, prior studies have shown that dense models could contain harmful parameters that affect the model performance, and it is no w...
http://arxiv.org/pdf/2311.03881v1
2023-11-07T10:54:45
2023-11-07T10:54:45
{ "title": [ -0.0055007935, -0.005870819, -0.018356323, -0.046173096, 0.0002799034, -0.0031585693, -0.027450562, -0.004501343, 0.040008545, 0.02027893, -0.01966858, 0.012138367, -0.005821228, -0.014152527, -0.05392456, 0.008964539, -0.005836487, ...
http://arxiv.org/abs/2311.07585v1
Input Reconstruction Attack against Vertical Federated Large Language Models
Fei Zheng
cs.CL, cs.AI, cs.CR
Recently, large language models (LLMs) have drawn extensive attention from academia and the public, due to the advent of the ChatGPT. While LLMs show their astonishing ability in text generation for various tasks, privacy concerns limit their usage in real-life businesses. More specifically, either the user's inputs (t...
http://arxiv.org/pdf/2311.07585v1
2023-11-07T09:39:22
2023-11-07T09:39:22
{ "title": [ -0.010696411, -0.00258255, -0.045532227, -0.07879639, 0.009353638, -0.0026378632, -0.04232788, 0.022750854, -0.053894043, 0.040405273, 0.009025574, 0.0231781, 0.0051879883, -0.0124053955, -0.032104492, -0.0005054474, -0.022583008, 0....
http://arxiv.org/abs/2311.03839v2
Aspects of human memory and Large Language Models
Romuald A. Janik
cs.CL, cs.AI, cs.LG, q-bio.NC
Large Language Models (LLMs) are huge artificial neural networks which primarily serve to generate text, but also provide a very sophisticated probabilistic model of language use. Since generating a semantically consistent text requires a form of effective memory, we investigate the memory properties of LLMs and find s...
http://arxiv.org/pdf/2311.03839v2
2023-11-09T18:16:24
2023-11-07T09:39:12
{ "title": [ -0.00522995, -0.00844574, -0.039276123, -0.061431885, 0.0016107559, -0.020950317, -0.023132324, 0.012924194, 0.0020256042, 0.039245605, -0.01676941, -0.01373291, 0.008056641, -0.02255249, -0.046295166, 0.018630981, 0.0017700195, 0.03...
http://arxiv.org/abs/2311.03837v1
OLaLa: Ontology Matching with Large Language Models
Sven Hertling, Heiko Paulheim
cs.IR, cs.CL
Ontology (and more generally: Knowledge Graph) Matching is a challenging task where information in natural language is one of the most important signals to process. With the rise of Large Language Models, it is possible to incorporate this knowledge in a better way into the matching pipeline. A number of decisions stil...
http://arxiv.org/pdf/2311.03837v1
2023-11-07T09:34:20
2023-11-07T09:34:20
{ "title": [ -0.03250122, -0.00945282, -0.049804688, -0.08660889, 0.016067505, 0.015960693, -0.05630493, -0.00052404404, 0.016021729, 0.013534546, -0.013336182, -0.005306244, 0.01197052, -0.016983032, -0.021957397, 0.011558533, -0.015525818, 0.04...
http://arxiv.org/abs/2311.03812v1
Conversations in Galician: a Large Language Model for an Underrepresented Language
Eliseo Bao, Anxo Pérez, Javier Parapar
cs.CL
The recent proliferation of Large Conversation Language Models has highlighted the economic significance of widespread access to this type of AI technologies in the current information age. Nevertheless, prevailing models have primarily been trained on corpora consisting of documents written in popular languages. The d...
http://arxiv.org/pdf/2311.03812v1
2023-11-07T08:52:28
2023-11-07T08:52:28
{ "title": [ -0.0020618439, 0.0047721863, -0.027389526, -0.029510498, -0.012336731, -0.004283905, -0.051513672, -0.016220093, -0.011016846, 0.03692627, -0.022277832, 0.0030956268, 0.023162842, -0.033813477, -0.03112793, 0.0009384155, -0.018325806, ...
http://arxiv.org/abs/2311.03810v1
Rethinking and Improving Multi-task Learning for End-to-end Speech Translation
Yuhao Zhang, Chen Xu, Bei Li, Hao Chen, Tong Xiao, Chunliang Zhang, Jingbo Zhu
cs.CL, cs.AI, cs.SD, eess.AS
Significant improvements in end-to-end speech translation (ST) have been achieved through the application of multi-task learning. However, the extent to which auxiliary tasks are highly consistent with the ST task, and how much this approach truly helps, have not been thoroughly studied. In this paper, we investigate t...
http://arxiv.org/pdf/2311.03810v1
2023-11-07T08:48:46
2023-11-07T08:48:46
{ "title": [ -0.016784668, 0.0033245087, -0.030548096, -0.058929443, 0.000048696995, -0.010810852, -0.040985107, 0.015792847, -0.0236969, 0.054595947, 0.0011749268, 0.024108887, 0.0045394897, -0.020690918, -0.03970337, 0.008201599, -0.0035552979, ...
http://arxiv.org/abs/2311.03798v1
Noisy Pair Corrector for Dense Retrieval
Hang Zhang, Yeyun Gong, Xingwei He, Dayiheng Liu, Daya Guo, Jiancheng Lv, Jian Guo
cs.CL
Most dense retrieval models contain an implicit assumption: the training query-document pairs are exactly matched. Since it is expensive to annotate the corpus manually, training pairs in real-world applications are usually collected automatically, which inevitably introduces mismatched-pair noise. In this paper, we ex...
http://arxiv.org/pdf/2311.03798v1
2023-11-07T08:27:14
2023-11-07T08:27:14
{ "title": [ -0.0066719055, 0.017425537, -0.013687134, -0.032714844, 0.011108398, 0.0029296875, -0.024139404, -0.0340271, 0.05645752, 0.04171753, -0.028396606, -0.015159607, 0.021133423, -0.006893158, -0.0524292, 0.020095825, 0.0016794205, 0.0071...
http://arxiv.org/abs/2311.03792v1
Character-Level Bangla Text-to-IPA Transcription Using Transformer Architecture with Sequence Alignment
Jakir Hasan, Shrestha Datta, Ameya Debnath
cs.CL
The International Phonetic Alphabet (IPA) is indispensable in language learning and understanding, aiding users in accurate pronunciation and comprehension. Additionally, it plays a pivotal role in speech therapy, linguistic research, accurate transliteration, and the development of text-to-speech systems, making it an...
http://arxiv.org/pdf/2311.03792v1
2023-11-07T08:20:06
2023-11-07T08:20:06
{ "title": [ -0.02708435, -0.011222839, -0.040008545, -0.045410156, -0.036987305, -0.0004901886, -0.06762695, 0.019302368, 0.00072193146, 0.060546875, -0.01979065, -0.013336182, 0.056762695, -0.035614014, -0.034942627, -0.0036678314, 0.004501343, ...
http://arxiv.org/abs/2311.03788v1
Language Representation Projection: Can We Transfer Factual Knowledge across Languages in Multilingual Language Models?
Shaoyang Xu, Junzhuo Li, Deyi Xiong
cs.CL
Multilingual pretrained language models serve as repositories of multilingual factual knowledge. Nevertheless, a substantial performance gap of factual knowledge probing exists between high-resource languages and low-resource languages, suggesting limited implicit factual knowledge transfer across languages in multilin...
http://arxiv.org/pdf/2311.03788v1
2023-11-07T08:16:16
2023-11-07T08:16:16
{ "title": [ -0.02456665, -0.011833191, -0.0236969, -0.04586792, -0.025177002, -0.008651733, -0.045654297, -0.008369446, -0.012763977, 0.04736328, -0.00248909, 0.0055732727, 0.032714844, -0.028305054, -0.0317688, 0.007820129, -0.004962921, 0.0418...