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/2309.04269v1
From Sparse to Dense: GPT-4 Summarization with Chain of Density Prompting
Griffin Adams, Alexander Fabbri, Faisal Ladhak, Eric Lehman, Noémie Elhadad
cs.CL
Selecting the ``right'' amount of information to include in a summary is a difficult task. A good summary should be detailed and entity-centric without being overly dense and hard to follow. To better understand this tradeoff, we solicit increasingly dense GPT-4 summaries with what we refer to as a ``Chain of Density''...
http://arxiv.org/pdf/2309.04269v1
2023-09-08T11:31:08
2023-09-08T11:31:08
{ "title": [ -0.0039901733, 0.0044784546, -0.016403198, -0.07348633, 0.0032196045, 0.009429932, -0.04638672, -0.015281677, 0.0071792603, 0.026412964, -0.007858276, 0.020111084, -0.01574707, -0.020370483, -0.044830322, 0.02330017, 0.009567261, 0.0...
http://arxiv.org/abs/2309.06384v1
Towards Reliable and Fluent Large Language Models: Incorporating Feedback Learning Loops in QA Systems
Dongyub Lee, Taesun Whang, Chanhee Lee, Heuiseok Lim
cs.CL, cs.AI
Large language models (LLMs) have emerged as versatile tools in various daily applications. However, they are fraught with issues that undermine their utility and trustworthiness. These include the incorporation of erroneous references (citation), the generation of hallucinated information (correctness), and the inclus...
http://arxiv.org/pdf/2309.06384v1
2023-09-08T09:39:53
2023-09-08T09:39:53
{ "title": [ -0.02609253, 0.00068092346, -0.028244019, -0.05883789, -0.0031986237, 0.0038700104, -0.054382324, -0.0016460419, 0.010688782, 0.0368042, -0.03527832, -0.0025424957, -0.00047254562, -0.0027217865, -0.06222534, -0.012290955, -0.01651001, ...
http://arxiv.org/abs/2309.04213v2
UQ at #SMM4H 2023: ALEX for Public Health Analysis with Social Media
Yan Jiang, Ruihong Qiu, Yi Zhang, Zi Huang
cs.CL, cs.AI
As social media becomes increasingly popular, more and more activities related to public health emerge. Current techniques for public health analysis involve popular models such as BERT and large language models (LLMs). However, the costs of training in-domain LLMs for public health are especially expensive. Furthermor...
http://arxiv.org/pdf/2309.04213v2
2023-09-12T07:19:22
2023-09-08T08:54:55
{ "title": [ 0.028793335, 0.014884949, -0.018508911, -0.040985107, 0.0060195923, -0.0012617111, -0.01108551, -0.00592041, -0.04925537, 0.03591919, 0.05682373, 0.017547607, 0.02305603, -0.07531738, 0.056427002, -0.016464233, 0.008163452, -0.005352...
http://arxiv.org/abs/2309.04198v2
The CALLA Dataset: Probing LLMs' Interactive Knowledge Acquisition from Chinese Medical Literature
Yanrui Du, Sendong Zhao, Muzhen Cai, Jianyu Chen, Haochun Wang, Yuhan Chen, Haoqiang Guo, Bing Qin
cs.CL
The application of Large Language Models (LLMs) to the medical domain has stimulated the interest of researchers. Recent studies have focused on constructing Instruction Fine-Tuning (IFT) data through medical knowledge graphs to enrich the interactive medical knowledge of LLMs. However, the medical literature serving a...
http://arxiv.org/pdf/2309.04198v2
2023-09-12T13:51:14
2023-09-08T08:20:46
{ "title": [ -0.0423584, -0.00793457, -0.05230713, -0.01789856, -0.020568848, -0.01737976, -0.026763916, -0.010284424, -0.014549255, 0.015434265, -0.03591919, 0.039489746, -0.027374268, -0.019378662, -0.019073486, -0.01361084, -0.0008587837, 0.01...
http://arxiv.org/abs/2309.04175v1
Knowledge-tuning Large Language Models with Structured Medical Knowledge Bases for Reliable Response Generation in Chinese
Haochun Wang, Sendong Zhao, Zewen Qiang, Zijian Li, Nuwa Xi, Yanrui Du, MuZhen Cai, Haoqiang Guo, Yuhan Chen, Haoming Xu, Bing Qin, Ting Liu
cs.CL, cs.AI
Large Language Models (LLMs) have demonstrated remarkable success in diverse natural language processing (NLP) tasks in general domains. However, LLMs sometimes generate responses with the hallucination about medical facts due to limited domain knowledge. Such shortcomings pose potential risks in the utilization of LLM...
http://arxiv.org/pdf/2309.04175v1
2023-09-08T07:42:57
2023-09-08T07:42:57
{ "title": [ -0.03756714, -0.020706177, -0.013809204, -0.061584473, -0.037628174, 0.020477295, -0.07867432, 0.00014436245, -0.011947632, 0.03439331, -0.0019550323, 0.002878189, 0.00983429, -0.03656006, -0.037384033, 0.0022029877, 0.018951416, 0.0...
http://arxiv.org/abs/2309.04174v1
Manifold-based Verbalizer Space Re-embedding for Tuning-free Prompt-based Classification
Haochun Wang, Sendong Zhao, Chi Liu, Nuwa Xi, Muzhen Cai, Bing Qin, Ting Liu
cs.CL, cs.AI
Prompt-based classification adapts tasks to a cloze question format utilizing the [MASK] token and the filled tokens are then mapped to labels through pre-defined verbalizers. Recent studies have explored the use of verbalizer embeddings to reduce labor in this process. However, all existing studies require a tuning pr...
http://arxiv.org/pdf/2309.04174v1
2023-09-08T07:42:29
2023-09-08T07:42:29
{ "title": [ 0.017929077, -0.0018959045, -0.01449585, -0.074645996, -0.01727295, 0.0021953583, -0.049743652, -0.0131073, 0.00843811, 0.019607544, -0.049346924, 0.015449524, -0.0004620552, -0.014022827, -0.030426025, -0.021499634, -0.00067567825, ...
http://arxiv.org/abs/2309.04162v1
GLS-CSC: A Simple but Effective Strategy to Mitigate Chinese STM Models' Over-Reliance on Superficial Clue
Yanrui Du, Sendong Zhao, Yuhan Chen, Rai Bai, Jing Liu, Hua Wu, Haifeng Wang, Bing Qin
cs.CL
Pre-trained models have achieved success in Chinese Short Text Matching (STM) tasks, but they often rely on superficial clues, leading to a lack of robust predictions. To address this issue, it is crucial to analyze and mitigate the influence of superficial clues on STM models. Our study aims to investigate their over-...
http://arxiv.org/pdf/2309.04162v1
2023-09-08T07:10:57
2023-09-08T07:10:57
{ "title": [ -0.0023059845, -0.012817383, 0.0019683838, -0.005760193, -0.055541992, -0.0143585205, -0.045410156, -0.04916382, -0.0077323914, 0.009208679, 0.012527466, -0.008415222, 0.00037622452, -0.062194824, 0.0073051453, -0.020187378, -0.003040313...
http://arxiv.org/abs/2309.04156v1
Cross-Utterance Conditioned VAE for Speech Generation
Yang Li, Cheng Yu, Guangzhi Sun, Weiqin Zu, Zheng Tian, Ying Wen, Wei Pan, Chao Zhang, Jun Wang, Yang Yang, Fanglei Sun
cs.SD, cs.CL, eess.AS
Speech synthesis systems powered by neural networks hold promise for multimedia production, but frequently face issues with producing expressive speech and seamless editing. In response, we present the Cross-Utterance Conditioned Variational Autoencoder speech synthesis (CUC-VAE S2) framework to enhance prosody and ens...
http://arxiv.org/pdf/2309.04156v1
2023-09-08T06:48:41
2023-09-08T06:48:41
{ "title": [ -0.0028305054, -0.017196655, -0.05718994, -0.07885742, 0.0062789917, -0.017425537, -0.031158447, 0.03540039, 0.00020098686, 0.050842285, -0.0026798248, 0.004459381, 0.041778564, -0.034210205, -0.019683838, 0.025253296, -0.024093628, ...
http://arxiv.org/abs/2309.04146v1
NESTLE: a No-Code Tool for Statistical Analysis of Legal Corpus
Kyoungyeon Cho, Seungkum Han, Wonseok Hwang
cs.CL, cs.AI
The statistical analysis of large scale legal corpus can provide valuable legal insights. For such analysis one needs to (1) select a subset of the corpus using document retrieval tools, (2) structuralize text using information extraction (IE) systems, and (3) visualize the data for the statistical analysis. Each proce...
http://arxiv.org/pdf/2309.04146v1
2023-09-08T06:23:25
2023-09-08T06:23:25
{ "title": [ -0.037139893, -0.004180908, -0.062194824, -0.03918457, 0.002855301, -0.0060806274, -0.043945312, -0.009536743, 0.011924744, 0.018432617, -0.05432129, 0.008453369, 0.015991211, -0.0132369995, -0.023223877, 0.0063476562, -0.022888184, ...
http://arxiv.org/abs/2309.04141v1
RST-style Discourse Parsing Guided by Document-level Content Structures
Ming Li, Ruihong Huang
cs.CL
Rhetorical Structure Theory based Discourse Parsing (RST-DP) explores how clauses, sentences, and large text spans compose a whole discourse and presents the rhetorical structure as a hierarchical tree. Existing RST parsing pipelines construct rhetorical structures without the knowledge of document-level content struct...
http://arxiv.org/pdf/2309.04141v1
2023-09-08T05:50:27
2023-09-08T05:50:27
{ "title": [ 0.0009036064, -0.00504303, -0.006690979, -0.097351074, -0.006641388, 0.017364502, -0.07305908, 0.0036258698, -0.0071907043, 0.01309967, -0.009010315, 0.0048980713, -0.023239136, -0.013580322, -0.04071045, 0.0003914833, -0.012580872, ...
http://arxiv.org/abs/2309.06415v2
Down the Toxicity Rabbit Hole: Investigating PaLM 2 Guardrails
Adel Khorramrouz, Sujan Dutta, Arka Dutta, Ashiqur R. KhudaBukhsh
cs.CL, cs.CY
This paper conducts a robustness audit of the safety feedback of PaLM 2 through a novel toxicity rabbit hole framework introduced here. Starting with a stereotype, the framework instructs PaLM 2 to generate more toxic content than the stereotype. Every subsequent iteration it continues instructing PaLM 2 to generate mo...
http://arxiv.org/pdf/2309.06415v2
2023-09-18T16:56:40
2023-09-08T03:59:02
{ "title": [ -0.005508423, -0.026626587, -0.06793213, -0.047851562, -0.029693604, -0.034576416, -0.03656006, 0.023101807, -0.10290527, 0.006855011, -0.008369446, 0.056762695, -0.059570312, -0.027801514, 0.045898438, -0.05340576, -0.014419556, 0.0...
http://arxiv.org/abs/2309.04106v2
Meta predictive learning model of languages in neural circuits
Chan Li, Junbin Qiu, Haiping Huang
cs.CL, q-bio.NC
Large language models based on self-attention mechanisms have achieved astonishing performances not only in natural language itself, but also in a variety of tasks of different nature. However, regarding processing language, our human brain may not operate using the same principle. Then, a debate is established on the ...
http://arxiv.org/pdf/2309.04106v2
2023-10-09T08:09:17
2023-09-08T03:58:05
{ "title": [ -0.009719849, 0.001786232, -0.04055786, -0.009750366, -0.031585693, -0.026062012, -0.027023315, -0.01928711, -0.019210815, 0.050628662, 0.0017404556, -0.024414062, 0.0071258545, -0.04937744, -0.03387451, 0.0035533905, 0.005466461, 0....
http://arxiv.org/abs/2309.04087v1
Unsupervised Multi-document Summarization with Holistic Inference
Haopeng Zhang, Sangwoo Cho, Kaiqiang Song, Xiaoyang Wang, Hongwei Wang, Jiawei Zhang, Dong Yu
cs.CL
Multi-document summarization aims to obtain core information from a collection of documents written on the same topic. This paper proposes a new holistic framework for unsupervised multi-document extractive summarization. Our method incorporates the holistic beam search inference method associated with the holistic mea...
http://arxiv.org/pdf/2309.04087v1
2023-09-08T02:56:30
2023-09-08T02:56:30
{ "title": [ -0.01222229, 0.0030918121, -0.050354004, -0.08508301, -0.005748749, 0.0075416565, -0.07507324, -0.0047073364, 0.021362305, 0.027038574, -0.029754639, 0.021865845, 0.005580902, 0.00054740906, -0.023132324, 0.00894928, -0.0024871826, 0...
http://arxiv.org/abs/2309.11506v1
Matching Table Metadata with Business Glossaries Using Large Language Models
Elita Lobo, Oktie Hassanzadeh, Nhan Pham, Nandana Mihindukulasooriya, Dharmashankar Subramanian, Horst Samulowitz
cs.IR, cs.AI, cs.CL
Enterprises often own large collections of structured data in the form of large databases or an enterprise data lake. Such data collections come with limited metadata and strict access policies that could limit access to the data contents and, therefore, limit the application of classic retrieval and analysis solutions...
http://arxiv.org/pdf/2309.11506v1
2023-09-08T02:23:59
2023-09-08T02:23:59
{ "title": [ -0.030532837, -0.001701355, -0.07312012, -0.07373047, 0.012756348, -0.007183075, -0.057159424, 0.016189575, 0.004043579, 0.056671143, -0.05114746, -0.007160187, 0.045013428, 0.010307312, -0.04550171, -0.010520935, 0.0016727448, 0.028...
http://arxiv.org/abs/2309.04041v1
Evaluation and Mitigation of Agnosia in Multimodal Large Language Models
Jiaying Lu, Jinmeng Rao, Kezhen Chen, Xiaoyuan Guo, Yawen Zhang, Baochen Sun, Carl Yang, Jie Yang
cs.CV, cs.CL
While Multimodal Large Language Models (MLLMs) are widely used for a variety of vision-language tasks, one observation is that they sometimes misinterpret visual inputs or fail to follow textual instructions even in straightforward cases, leading to irrelevant responses, mistakes, and ungrounded claims. This observatio...
http://arxiv.org/pdf/2309.04041v1
2023-09-07T22:59:56
2023-09-07T22:59:56
{ "title": [ -0.019561768, 0.011955261, -0.053344727, -0.03338623, 0.0034637451, -0.022659302, -0.030838013, -0.008369446, 0.014480591, 0.04537964, -0.0041923523, -0.013702393, 0.028503418, -0.021820068, -0.062927246, 0.008796692, -0.014625549, 0...
http://arxiv.org/abs/2309.04031v1
Multiple Representation Transfer from Large Language Models to End-to-End ASR Systems
Takuma Udagawa, Masayuki Suzuki, Gakuto Kurata, Masayasu Muraoka, George Saon
cs.CL, cs.SD, eess.AS
Transferring the knowledge of large language models (LLMs) is a promising technique to incorporate linguistic knowledge into end-to-end automatic speech recognition (ASR) systems. However, existing works only transfer a single representation of LLM (e.g. the last layer of pretrained BERT), while the representation of a...
http://arxiv.org/pdf/2309.04031v1
2023-09-07T21:57:39
2023-09-07T21:57:39
{ "title": [ -0.009971619, -0.010093689, -0.046691895, -0.09954834, 0.00554657, 0.0057373047, -0.05569458, 0.04006958, -0.005329132, 0.03475952, -0.0051002502, -0.0014019012, 0.013633728, -0.021499634, -0.026031494, 0.007385254, -0.0025405884, 0....
http://arxiv.org/abs/2309.04027v1
TIDE: Textual Identity Detection for Evaluating and Augmenting Classification and Language Models
Emmanuel Klu, Sameer Sethi
cs.CL, cs.LG
Machine learning models can perpetuate unintended biases from unfair and imbalanced datasets. Evaluating and debiasing these datasets and models is especially hard in text datasets where sensitive attributes such as race, gender, and sexual orientation may not be available. When these models are deployed into society, ...
http://arxiv.org/pdf/2309.04027v1
2023-09-07T21:44:42
2023-09-07T21:44:42
{ "title": [ -0.010284424, 0.0032291412, -0.050231934, -0.05267334, 0.02897644, -0.004180908, -0.029281616, -0.01209259, -0.003129959, 0.022750854, -0.008277893, 0.020767212, 0.016159058, -0.015991211, -0.035736084, 0.0017223358, -0.028137207, 0....
http://arxiv.org/abs/2309.04019v1
Evaluation of large language models for discovery of gene set function
Mengzhou Hu, Sahar Alkhairy, Ingoo Lee, Rudolf T. Pillich, Robin Bachelder, Trey Ideker, Dexter Pratt
q-bio.GN, cs.AI, cs.CL, q-bio.MN
Gene set analysis is a mainstay of functional genomics, but it relies on manually curated databases of gene functions that are incomplete and unaware of biological context. Here we evaluate the ability of OpenAI's GPT-4, a Large Language Model (LLM), to develop hypotheses about common gene functions from its embedded b...
http://arxiv.org/pdf/2309.04019v1
2023-09-07T21:10:48
2023-09-07T21:10:48
{ "title": [ -0.059173584, -0.0069274902, -0.070739746, -0.045532227, -0.00070381165, -0.0018444061, -0.07006836, 0.006839752, 0.0066986084, 0.020629883, -0.014305115, -0.015838623, -0.007820129, -0.026443481, -0.044158936, 0.006603241, -0.011199951,...
http://arxiv.org/abs/2309.03992v2
ConDA: Contrastive Domain Adaptation for AI-generated Text Detection
Amrita Bhattacharjee, Tharindu Kumarage, Raha Moraffah, Huan Liu
cs.CL, cs.AI, cs.LG
Large language models (LLMs) are increasingly being used for generating text in a variety of use cases, including journalistic news articles. Given the potential malicious nature in which these LLMs can be used to generate disinformation at scale, it is important to build effective detectors for such AI-generated text....
http://arxiv.org/pdf/2309.03992v2
2023-09-20T22:17:30
2023-09-07T19:51:30
{ "title": [ -0.03125, -0.012634277, -0.02659607, -0.008392334, -0.009269714, -0.017715454, 0.010185242, -0.01574707, 0.02041626, 0.033233643, -0.023925781, 0.010292053, 0.0023517609, -0.0033721924, -0.049072266, -0.0021457672, -0.03152466, 0.064...
http://arxiv.org/abs/2309.03978v1
LanSER: Language-Model Supported Speech Emotion Recognition
Taesik Gong, Josh Belanich, Krishna Somandepalli, Arsha Nagrani, Brian Eoff, Brendan Jou
cs.CL, cs.LG, cs.SD, eess.AS
Speech emotion recognition (SER) models typically rely on costly human-labeled data for training, making scaling methods to large speech datasets and nuanced emotion taxonomies difficult. We present LanSER, a method that enables the use of unlabeled data by inferring weak emotion labels via pre-trained large language m...
http://arxiv.org/pdf/2309.03978v1
2023-09-07T19:21:08
2023-09-07T19:21:08
{ "title": [ -0.0256958, -0.008094788, -0.037628174, -0.058502197, -0.0029640198, -0.0015506744, -0.05621338, 0.006008148, 0.028930664, 0.031921387, -0.0209198, -0.0149383545, 0.031585693, -0.022338867, -0.05441284, 0.014945984, -0.014320374, 0.0...
http://arxiv.org/abs/2309.03905v2
ImageBind-LLM: Multi-modality Instruction Tuning
Jiaming Han, Renrui Zhang, Wenqi Shao, Peng Gao, Peng Xu, Han Xiao, Kaipeng Zhang, Chris Liu, Song Wen, Ziyu Guo, Xudong Lu, Shuai Ren, Yafei Wen, Xiaoxin Chen, Xiangyu Yue, Hongsheng Li, Yu Qiao
cs.MM, cs.CL, cs.CV, cs.LG, cs.SD, eess.AS
We present ImageBind-LLM, a multi-modality instruction tuning method of large language models (LLMs) via ImageBind. Existing works mainly focus on language and image instruction tuning, different from which, our ImageBind-LLM can respond to multi-modality conditions, including audio, 3D point clouds, video, and their e...
http://arxiv.org/pdf/2309.03905v2
2023-09-11T20:25:16
2023-09-07T17:59:45
{ "title": [ -0.0602417, -0.016021729, -0.033172607, -0.026535034, -0.035827637, -0.0006303787, -0.051239014, -0.0012407303, -0.008125305, 0.03793335, -0.030731201, 0.03591919, 0.004386902, -0.022018433, -0.012565613, 0.008056641, 0.01209259, 0.0...
http://arxiv.org/abs/2309.03886v1
A Function Interpretation Benchmark for Evaluating Interpretability Methods
Sarah Schwettmann, Tamar Rott Shaham, Joanna Materzynska, Neil Chowdhury, Shuang Li, Jacob Andreas, David Bau, Antonio Torralba
cs.CL, cs.AI, cs.LG
Labeling neural network submodules with human-legible descriptions is useful for many downstream tasks: such descriptions can surface failures, guide interventions, and perhaps even explain important model behaviors. To date, most mechanistic descriptions of trained networks have involved small models, narrowly delimit...
http://arxiv.org/pdf/2309.03886v1
2023-09-07T17:47:26
2023-09-07T17:47:26
{ "title": [ -0.027053833, 0.023452759, -0.027740479, -0.0769043, 0.024398804, 0.013435364, -0.074157715, -0.038269043, 0.0024337769, -0.001159668, -0.008605957, -0.014907837, -0.04736328, -0.031799316, -0.040649414, -0.015388489, -0.0032424927, ...
http://arxiv.org/abs/2309.03884v1
Zero-Shot Audio Captioning via Audibility Guidance
Tal Shaharabany, Ariel Shaulov, Lior Wolf
cs.SD, cs.CL, eess.AS
The task of audio captioning is similar in essence to tasks such as image and video captioning. However, it has received much less attention. We propose three desiderata for captioning audio -- (i) fluency of the generated text, (ii) faithfulness of the generated text to the input audio, and the somewhat related (iii) ...
http://arxiv.org/pdf/2309.03884v1
2023-09-07T17:45:58
2023-09-07T17:45:58
{ "title": [ 0.004047394, -0.016403198, -0.024032593, -0.05328369, -0.03149414, -0.016204834, -0.016342163, 0.011177063, 0.0036087036, 0.05847168, -0.015083313, 0.00983429, 0.029876709, -0.013000488, -0.0340271, 0.030715942, -0.0040893555, 0.0262...
http://arxiv.org/abs/2309.03883v1
DoLa: Decoding by Contrasting Layers Improves Factuality in Large Language Models
Yung-Sung Chuang, Yujia Xie, Hongyin Luo, Yoon Kim, James Glass, Pengcheng He
cs.CL, cs.AI, cs.LG
Despite their impressive capabilities, large language models (LLMs) are prone to hallucinations, i.e., generating content that deviates from facts seen during pretraining. We propose a simple decoding strategy for reducing hallucinations with pretrained LLMs that does not require conditioning on retrieved external know...
http://arxiv.org/pdf/2309.03883v1
2023-09-07T17:45:31
2023-09-07T17:45:31
{ "title": [ -0.026168823, 0.01020813, -0.064453125, -0.06933594, -0.007835388, -0.02130127, -0.07702637, -0.0129470825, -0.019226074, 0.04840088, -0.0019836426, 0.019805908, 0.027328491, -0.019714355, -0.022644043, 0.022399902, -0.01739502, 0.04...
http://arxiv.org/abs/2309.03882v3
Large Language Models Are Not Robust Multiple Choice Selectors
Chujie Zheng, Hao Zhou, Fandong Meng, Jie Zhou, Minlie Huang
cs.CL
Multiple choice questions (MCQs) serve as a common yet important task format in the research of large language models (LLMs). This work shows that LLMs are vulnerable to option position changes in MCQs due to their inherent "selection bias", namely, they prefer to select specific option IDs as answers (like "Option A")...
http://arxiv.org/pdf/2309.03882v3
2023-10-06T08:27:26
2023-09-07T17:44:56
{ "title": [ -0.0049057007, 0.005153656, -0.013290405, -0.00019598007, -0.02684021, -0.029571533, 0.0077705383, -0.02684021, -0.013687134, 0.028839111, -0.050689697, -0.013511658, -0.0067710876, -0.029159546, -0.011703491, -0.028656006, -0.012290955,...
http://arxiv.org/abs/2309.03877v1
Introducing "Forecast Utterance" for Conversational Data Science
Md Mahadi Hassan, Alex Knipper, Shubhra Kanti Karmaker
cs.CL, cs.HC
Envision an intelligent agent capable of assisting users in conducting forecasting tasks through intuitive, natural conversations, without requiring in-depth knowledge of the underlying machine learning (ML) processes. A significant challenge for the agent in this endeavor is to accurately comprehend the user's predict...
http://arxiv.org/pdf/2309.03877v1
2023-09-07T17:41:41
2023-09-07T17:41:41
{ "title": [ -0.008384705, -0.017349243, -0.061676025, -0.08642578, -0.0056533813, -0.013633728, -0.056518555, 0.029815674, 0.01902771, 0.042144775, -0.03604126, -0.008834839, 0.025924683, -0.012466431, -0.055725098, -0.012168884, 0.008956909, 0....
http://arxiv.org/abs/2309.03876v1
OpinionGPT: Modelling Explicit Biases in Instruction-Tuned LLMs
Patrick Haller, Ansar Aynetdinov, Alan Akbik
cs.CL, cs.AI, cs.CY, cs.LG
Instruction-tuned Large Language Models (LLMs) have recently showcased remarkable ability to generate fitting responses to natural language instructions. However, an open research question concerns the inherent biases of trained models and their responses. For instance, if the data used to tune an LLM is dominantly wri...
http://arxiv.org/pdf/2309.03876v1
2023-09-07T17:41:01
2023-09-07T17:41:01
{ "title": [ -0.0597229, -0.02305603, -0.030548096, -0.025939941, -0.035705566, 0.002336502, -0.05532837, -0.003944397, 0.001540184, 0.01576233, -0.06048584, 0.015655518, -0.019134521, -0.016708374, -0.02268982, -0.0141067505, -0.027908325, 0.033...
http://arxiv.org/abs/2309.03852v2
FLM-101B: An Open LLM and How to Train It with $100K Budget
Xiang Li, Yiqun Yao, Xin Jiang, Xuezhi Fang, Xuying Meng, Siqi Fan, Peng Han, Jing Li, Li Du, Bowen Qin, Zheng Zhang, Aixin Sun, Yequan Wang
cs.CL, cs.AI
Large language models (LLMs) have achieved remarkable success in NLP and multimodal tasks, among others. Despite these successes, two main challenges remain in developing LLMs: (i) high computational cost, and (ii) fair and objective evaluations. In this paper, we report a solution to significantly reduce LLM training ...
http://arxiv.org/pdf/2309.03852v2
2023-09-17T07:38:10
2023-09-07T17:07:36
{ "title": [ -0.038330078, -0.010520935, -0.053741455, -0.023529053, -0.036468506, -0.010223389, -0.058441162, 0.011329651, -0.06964111, 0.015808105, -0.032043457, 0.030715942, -0.01727295, -0.053863525, 0.048706055, -0.051849365, -0.026367188, 0...
http://arxiv.org/abs/2309.03831v1
Uncovering Drift in Textual Data: An Unsupervised Method for Detecting and Mitigating Drift in Machine Learning Models
Saeed Khaki, Akhouri Abhinav Aditya, Zohar Karnin, Lan Ma, Olivia Pan, Samarth Marudheri Chandrashekar
cs.CL, cs.AI, cs.LG
Drift in machine learning refers to the phenomenon where the statistical properties of data or context, in which the model operates, change over time leading to a decrease in its performance. Therefore, maintaining a constant monitoring process for machine learning model performance is crucial in order to proactively p...
http://arxiv.org/pdf/2309.03831v1
2023-09-07T16:45:42
2023-09-07T16:45:42
{ "title": [ -0.02128601, 0.013885498, -0.039489746, -0.035339355, 0.020187378, -0.02017212, -0.018005371, -0.029342651, 0.015945435, 0.013298035, -0.030181885, -0.0024528503, -0.010986328, -0.017837524, -0.049560547, -0.009140015, -0.010696411, ...
http://arxiv.org/abs/2309.03787v2
USA: Universal Sentiment Analysis Model & Construction of Japanese Sentiment Text Classification and Part of Speech Dataset
Chengguang Gan, Qinghao Zhang, Tatsunori Mori
cs.CL
Sentiment analysis is a pivotal task in the domain of natural language processing. It encompasses both text-level sentiment polarity classification and word-level Part of Speech(POS) sentiment polarity determination. Such analysis challenges models to understand text holistically while also extracting nuanced informati...
http://arxiv.org/pdf/2309.03787v2
2023-09-14T05:53:45
2023-09-07T15:35:00
{ "title": [ -0.005455017, -0.000500679, -0.04067993, -0.051086426, 0.0025138855, -0.0053367615, -0.05670166, -0.010215759, 0.008857727, 0.056732178, -0.034576416, -0.003047943, 0.03137207, -0.015106201, -0.058410645, -0.007873535, -0.016082764, ...
http://arxiv.org/abs/2309.03748v1
Enhancing Pipeline-Based Conversational Agents with Large Language Models
Mina Foosherian, Hendrik Purwins, Purna Rathnayake, Touhidul Alam, Rui Teimao, Klaus-Dieter Thoben
cs.CL, cs.AI, cs.LG, stat.ML, 68T50, I.2.7
The latest advancements in AI and deep learning have led to a breakthrough in large language model (LLM)-based agents such as GPT-4. However, many commercial conversational agent development tools are pipeline-based and have limitations in holding a human-like conversation. This paper investigates the capabilities of L...
http://arxiv.org/pdf/2309.03748v1
2023-09-07T14:43:17
2023-09-07T14:43:17
{ "title": [ -0.013175964, -0.012817383, -0.036102295, -0.066467285, 0.0060310364, 0.0049552917, -0.05340576, 0.011314392, 0.0068740845, 0.029190063, -0.02217102, -0.0011711121, 0.009002686, -0.011352539, -0.041931152, 0.002073288, -0.021469116, ...
http://arxiv.org/abs/2309.03747v1
The Daunting Dilemma with Sentence Encoders: Success on Standard Benchmarks, Failure in Capturing Basic Semantic Properties
Yash Mahajan, Naman Bansal, Shubhra Kanti Karmaker
cs.CL
In this paper, we adopted a retrospective approach to examine and compare five existing popular sentence encoders, i.e., Sentence-BERT, Universal Sentence Encoder (USE), LASER, InferSent, and Doc2vec, in terms of their performance on downstream tasks versus their capability to capture basic semantic properties. Initial...
http://arxiv.org/pdf/2309.03747v1
2023-09-07T14:42:35
2023-09-07T14:42:35
{ "title": [ -0.022949219, -0.00033545494, 0.0015163422, 0.011405945, -0.029220581, -0.01399231, -0.0059318542, -0.041229248, 0.025421143, 0.04800415, -0.04699707, 0.040985107, 0.010917664, 0.0058174133, -0.04498291, 0.017456055, 0.00056886673, 0...
http://arxiv.org/abs/2309.03713v1
Word segmentation granularity in Korean
Jungyeul Park, Mija Kim
cs.CL
This paper describes word {segmentation} granularity in Korean language processing. From a word separated by blank space, which is termed an eojeol, to a sequence of morphemes in Korean, there are multiple possible levels of word segmentation granularity in Korean. For specific language processing and corpus annotation...
http://arxiv.org/pdf/2309.03713v1
2023-09-07T13:42:05
2023-09-07T13:42:05
{ "title": [ -0.028320312, -0.007736206, -0.02998352, -0.02897644, -0.06021118, -0.003616333, -0.08892822, -0.009529114, -0.03086853, 0.062164307, -0.03869629, 0.021896362, 0.045043945, -0.012283325, -0.0055999756, 0.0047721863, -0.00086021423, 0...
http://arxiv.org/abs/2309.03667v2
Exploring an LM to generate Prolog Predicates from Mathematics Questions
Xiaocheng Yang, Yik-Cheung Tam
cs.CL
Recently, there has been a surge in interest in NLP driven by ChatGPT. ChatGPT, a transformer-based generative language model of substantial scale, exhibits versatility in performing various tasks based on natural language. Nevertheless, large language models often exhibit poor performance in solving mathematics questi...
http://arxiv.org/pdf/2309.03667v2
2023-09-08T03:46:33
2023-09-07T12:10:47
{ "title": [ -0.010765076, -0.0003297329, -0.030975342, -0.051452637, 0.0017385483, 0.0042381287, -0.031173706, -0.00522995, -0.0016155243, 0.00075531006, -0.018188477, 0.0076179504, -0.014083862, -0.031204224, 0.0069503784, -0.0032215118, -0.0019016...
http://arxiv.org/abs/2309.03658v1
BNS-Net: A Dual-channel Sarcasm Detection Method Considering Behavior-level and Sentence-level Conflicts
Liming Zhou, Xiaowei Xu, Xiaodong Wang
cs.CL
Sarcasm detection is a binary classification task that aims to determine whether a given utterance is sarcastic. Over the past decade, sarcasm detection has evolved from classical pattern recognition to deep learning approaches, where features such as user profile, punctuation and sentiment words have been commonly emp...
http://arxiv.org/pdf/2309.03658v1
2023-09-07T11:55:11
2023-09-07T11:55:11
{ "title": [ 0.0074005127, -0.0007586479, -0.029953003, -0.0335083, -0.0031318665, -0.014007568, -0.023910522, -0.022766113, 0.0050239563, 0.011695862, -0.02810669, -0.039093018, 0.027450562, -0.032592773, -0.053375244, -0.030639648, -0.03173828, ...
http://arxiv.org/abs/2309.03613v1
Evaluating ChatGPT as a Recommender System: A Rigorous Approach
Dario Di Palma, Giovanni Maria Biancofiore, Vito Walter Anelli, Fedelucio Narducci, Tommaso Di Noia, Eugenio Di Sciascio
cs.IR, cs.AI, cs.CL
Recent popularity surrounds large AI language models due to their impressive natural language capabilities. They contribute significantly to language-related tasks, including prompt-based learning, making them valuable for various specific tasks. This approach unlocks their full potential, enhancing precision and gener...
http://arxiv.org/pdf/2309.03613v1
2023-09-07T10:13:09
2023-09-07T10:13:09
{ "title": [ 0.010154724, -0.0062332153, -0.016021729, -0.01890564, 0.009643555, -0.0006661415, 0.002254486, -0.02330017, -0.016815186, 0.020980835, -0.012809753, 0.024734497, -0.03274536, -0.007106781, -0.037872314, 0.0070228577, -0.037719727, 0...
http://arxiv.org/abs/2309.03595v1
Loquacity and Visible Emotion: ChatGPT as a Policy Advisor
Claudia Biancotti, Carolina Camassa
cs.CL, cs.HC, J.4; K.4.1
ChatGPT, a software seeking to simulate human conversational abilities, is attracting increasing attention. It is sometimes portrayed as a groundbreaking productivity aid, including for creative work. In this paper, we run an experiment to assess its potential in complex writing tasks. We ask the software to compose a ...
http://arxiv.org/pdf/2309.03595v1
2023-09-07T09:40:12
2023-09-07T09:40:12
{ "title": [ 0.000436306, -0.024780273, 0.0026130676, -0.019363403, -0.029571533, -0.005847931, 0.0042533875, -0.021072388, -0.011459351, 0.013160706, -0.043182373, 0.0079956055, 0.0015573502, -0.012229919, -0.024459839, -0.023025513, -0.014289856, ...
http://arxiv.org/abs/2309.03564v2
Supervised Learning and Large Language Model Benchmarks on Mental Health Datasets: Cognitive Distortions and Suicidal Risks in Chinese Social Media
Hongzhi Qi, Qing Zhao, Changwei Song, Wei Zhai, Dan Luo, Shuo Liu, Yi Jing Yu, Fan Wang, Huijing Zou, Bing Xiang Yang, Jianqiang Li, Guanghui Fu
cs.CL, cs.LG
In the realm of social media, users frequently convey personal sentiments, with some potentially indicating cognitive distortions or suicidal tendencies. Timely recognition of such signs is pivotal for effective interventions. In response, we introduce two novel annotated datasets from Chinese social media, focused on ...
http://arxiv.org/pdf/2309.03564v2
2023-11-01T10:15:34
2023-09-07T08:50:46
{ "title": [ 0.02407837, 0.020355225, 0.0029029846, -0.033050537, 0.004547119, -0.022109985, -0.03616333, -0.027648926, -0.022033691, 0.03427124, -0.031585693, 0.032104492, 0.024261475, -0.027862549, 0.00522995, -0.023651123, -0.0063934326, 0.027...
http://arxiv.org/abs/2309.03563v2
All Labels Together: Low-shot Intent Detection with an Efficient Label Semantic Encoding Paradigm
Jiangshu Du, Congying Xia, Wenpeng Yin, Tingting Liang, Philip S. Yu
cs.CL
In intent detection tasks, leveraging meaningful semantic information from intent labels can be particularly beneficial for few-shot scenarios. However, existing few-shot intent detection methods either ignore the intent labels, (e.g. treating intents as indices) or do not fully utilize this information (e.g. only usin...
http://arxiv.org/pdf/2309.03563v2
2023-09-08T01:00:52
2023-09-07T08:50:45
{ "title": [ 0.00032567978, -0.0009775162, -0.044281006, -0.042236328, 0.018371582, -0.030136108, 0.0026741028, -0.016906738, 0.003435135, 0.017150879, -0.0037117004, 0.014427185, 0.000065386295, -0.027404785, -0.019729614, 0.011978149, -0.0028743744...
http://arxiv.org/abs/2309.03559v1
An Anchor Learning Approach for Citation Field Learning
Zilin Yuan, Borun Chen, Yimeng Dai, Yinghui Li, Hai-Tao Zheng, Rui Zhang
cs.CL
Citation field learning is to segment a citation string into fields of interest such as author, title, and venue. Extracting such fields from citations is crucial for citation indexing, researcher profile analysis, etc. User-generated resources like academic homepages and Curriculum Vitae, provide rich citation field i...
http://arxiv.org/pdf/2309.03559v1
2023-09-07T08:42:40
2023-09-07T08:42:40
{ "title": [ -0.018356323, 0.013053894, -0.060943604, -0.02986145, 0.025375366, -0.024429321, 0.0015296936, -0.0021457672, -0.023834229, 0.023513794, -0.028152466, -0.008338928, -0.028671265, -0.015296936, -0.04269409, -0.04031372, -0.016647339, ...
http://arxiv.org/abs/2309.06578v2
Can Large Language Models Discern Evidence for Scientific Hypotheses? Case Studies in the Social Sciences
Sai Koneru, Jian Wu, Sarah Rajtmajer
cs.CL, cs.AI
Hypothesis formulation and testing are central to empirical research. A strong hypothesis is a best guess based on existing evidence and informed by a comprehensive view of relevant literature. However, with exponential increase in the number of scientific articles published annually, manual aggregation and synthesis o...
http://arxiv.org/pdf/2309.06578v2
2023-10-25T04:57:41
2023-09-07T04:15:17
{ "title": [ -0.03491211, 0.0055732727, -0.05038452, -0.025497437, -0.0115356445, -0.025039673, -0.028305054, -0.021850586, -0.023040771, 0.021194458, -0.015449524, -0.028198242, 0.012771606, -0.041870117, -0.02947998, -0.014450073, -0.0035095215, ...
http://arxiv.org/abs/2309.03470v1
Machine Learning for Tangible Effects: Natural Language Processing for Uncovering the Illicit Massage Industry & Computer Vision for Tactile Sensing
Rui Ouyang
cs.CL, cs.CY, cs.SI
I explore two questions in this thesis: how can computer science be used to fight human trafficking? And how can computer vision create a sense of touch? I use natural language processing (NLP) to monitor the United States illicit massage industry (IMI), a multi-billion dollar industry that offers not just therapeuti...
http://arxiv.org/pdf/2309.03470v1
2023-09-07T04:04:01
2023-09-07T04:04:01
{ "title": [ -0.0036849976, -0.00024092197, -0.023452759, 0.021728516, -0.030303955, -0.036987305, 0.017044067, -0.018981934, -0.025512695, 0.019134521, -0.037963867, 0.00831604, 0.022277832, -0.033996582, 0.015548706, -0.029571533, -0.0051612854, ...
http://arxiv.org/abs/2309.03450v1
XGen-7B Technical Report
Erik Nijkamp, Tian Xie, Hiroaki Hayashi, Bo Pang, Congying Xia, Chen Xing, Jesse Vig, Semih Yavuz, Philippe Laban, Ben Krause, Senthil Purushwalkam, Tong Niu, Wojciech Kryściński, Lidiya Murakhovs'ka, Prafulla Kumar Choubey, Alex Fabbri, Ye Liu, Rui Meng, Lifu Tu, Meghana Bhat, Chien-Sheng Wu, Silvio Savarese, Yingbo Z...
cs.CL, cs.AI, cs.LG
Large Language Models (LLMs) have become ubiquitous across various domains, transforming the way we interact with information and conduct research. However, most high-performing LLMs remain confined behind proprietary walls, hindering scientific progress. Most open-source LLMs, on the other hand, are limited in their a...
http://arxiv.org/pdf/2309.03450v1
2023-09-07T02:20:03
2023-09-07T02:20:03
{ "title": [ -0.0068969727, 0.014511108, -0.028625488, -0.010299683, 0.00069761276, -0.006111145, -0.040740967, -0.02633667, -0.04510498, 0.047851562, -0.033569336, -0.010543823, 0.049713135, -0.022766113, 0.051574707, -0.020935059, 0.004875183, ...
http://arxiv.org/abs/2309.03433v1
Improving Open Information Extraction with Large Language Models: A Study on Demonstration Uncertainty
Chen Ling, Xujiang Zhao, Xuchao Zhang, Yanchi Liu, Wei Cheng, Haoyu Wang, Zhengzhang Chen, Takao Osaki, Katsushi Matsuda, Haifeng Chen, Liang Zhao
cs.CL
Open Information Extraction (OIE) task aims at extracting structured facts from unstructured text, typically in the form of (subject, relation, object) triples. Despite the potential of large language models (LLMs) like ChatGPT as a general task solver, they lag behind state-of-the-art (supervised) methods in OIE tasks...
http://arxiv.org/pdf/2309.03433v1
2023-09-07T01:35:24
2023-09-07T01:35:24
{ "title": [ -0.043518066, 0.017105103, -0.055877686, -0.0579834, 0.0064582825, 0.0035133362, -0.06915283, -0.017166138, -0.00052928925, 0.015068054, -0.034118652, 0.00573349, -0.014846802, -0.008934021, -0.03439331, -0.016952515, -0.018661499, 0...
http://arxiv.org/abs/2309.03412v2
From Base to Conversational: Japanese Instruction Dataset and Tuning Large Language Models
Masahiro Suzuki, Masanori Hirano, Hiroki Sakaji
cs.CL
Instruction tuning is essential for large language models (LLMs) to become interactive. While many instruction tuning datasets exist in English, there is a noticeable lack in other languages. Also, their effectiveness has not been well verified in non-English languages. We construct a Japanese instruction dataset by ex...
http://arxiv.org/pdf/2309.03412v2
2023-11-05T06:32:30
2023-09-07T00:14:37
{ "title": [ -0.048187256, 0.00071907043, -0.039398193, -0.049316406, -0.011474609, 0.0030136108, -0.08685303, -0.0062408447, 0.0036811829, 0.020050049, -0.027801514, 0.0011711121, 0.00724411, -0.012184143, -0.03817749, 0.0002603531, -0.011772156, ...
http://arxiv.org/abs/2309.03409v1
Large Language Models as Optimizers
Chengrun Yang, Xuezhi Wang, Yifeng Lu, Hanxiao Liu, Quoc V. Le, Denny Zhou, Xinyun Chen
cs.LG, cs.AI, cs.CL
Optimization is ubiquitous. While derivative-based algorithms have been powerful tools for various problems, the absence of gradient imposes challenges on many real-world applications. In this work, we propose Optimization by PROmpting (OPRO), a simple and effective approach to leverage large language models (LLMs) as ...
http://arxiv.org/pdf/2309.03409v1
2023-09-07T00:07:15
2023-09-07T00:07:15
{ "title": [ -0.040222168, -0.0127334595, -0.028961182, -0.062927246, -0.004459381, 0.021636963, -0.07684326, 0.0051078796, -0.00434494, 0.034820557, -0.0064468384, -0.011489868, -0.0006875992, -0.020233154, -0.035614014, 0.0109939575, -0.0068893433,...
http://arxiv.org/abs/2309.03378v2
RoDia: A New Dataset for Romanian Dialect Identification from Speech
Codrut Rotaru, Nicolae-Catalin Ristea, Radu Tudor Ionescu
cs.CL, cs.SD, eess.AS
Dialect identification is a critical task in speech processing and language technology, enhancing various applications such as speech recognition, speaker verification, and many others. While most research studies have been dedicated to dialect identification in widely spoken languages, limited attention has been given...
http://arxiv.org/pdf/2309.03378v2
2023-09-12T14:07:54
2023-09-06T21:56:24
{ "title": [ -0.007381439, 0.002948761, -0.05734253, -0.031677246, 0.023529053, -0.016662598, -0.02168274, -0.008117676, -0.026611328, 0.048828125, -0.0011672974, -0.01158905, 0.03414917, -0.029388428, -0.044067383, 0.0022392273, -0.0076904297, 0...
http://arxiv.org/abs/2309.03340v1
Parameter Efficient Audio Captioning With Faithful Guidance Using Audio-text Shared Latent Representation
Arvind Krishna Sridhar, Yinyi Guo, Erik Visser, Rehana Mahfuz
cs.CL, cs.MM, cs.SD
There has been significant research on developing pretrained transformer architectures for multimodal-to-text generation tasks. Albeit performance improvements, such models are frequently overparameterized, hence suffer from hallucination and large memory footprint making them challenging to deploy on edge devices. In ...
http://arxiv.org/pdf/2309.03340v1
2023-09-06T19:42:52
2023-09-06T19:42:52
{ "title": [ 0.011962891, 0.00013554096, -0.02532959, -0.055541992, -0.004371643, -0.01675415, -0.014205933, -0.0052719116, -0.010314941, 0.06762695, -0.022979736, 0.034454346, 0.029434204, -0.005302429, -0.0236969, 0.02758789, -0.0063819885, 0.0...
http://arxiv.org/abs/2309.03175v1
Gender-specific Machine Translation with Large Language Models
Eduardo Sánchez, Pierre Andrews, Pontus Stenetorp, Mikel Artetxe, Marta R. Costa-jussà
cs.CL
Decoder-only Large Language Models (LLMs) have demonstrated potential in machine translation (MT), albeit with performance slightly lagging behind traditional encoder-decoder Neural Machine Translation (NMT) systems. However, LLMs offer a unique advantage: the ability to control the properties of the output through pro...
http://arxiv.org/pdf/2309.03175v1
2023-09-06T17:24:06
2023-09-06T17:24:06
{ "title": [ -0.06237793, -0.009101868, -0.03894043, -0.047088623, -0.011001587, 0.010147095, -0.04815674, 0.017852783, 0.008934021, 0.056793213, -0.008522034, -0.011955261, 0.049835205, -0.009895325, -0.051483154, 0.020996094, -0.021591187, 0.03...
http://arxiv.org/abs/2309.03079v1
GPT-InvestAR: Enhancing Stock Investment Strategies through Annual Report Analysis with Large Language Models
Udit Gupta
q-fin.ST, cs.CL, cs.LG
Annual Reports of publicly listed companies contain vital information about their financial health which can help assess the potential impact on Stock price of the firm. These reports are comprehensive in nature, going up to, and sometimes exceeding, 100 pages. Analysing these reports is cumbersome even for a single fi...
http://arxiv.org/pdf/2309.03079v1
2023-09-06T17:18:55
2023-09-06T17:18:55
{ "title": [ -0.022537231, -0.014350891, -0.06707764, -0.046142578, -0.006286621, -0.020812988, -0.038085938, 0.00012290478, 0.002620697, 0.027023315, -0.045532227, -0.0044670105, 0.010887146, -0.019195557, -0.026992798, -0.01864624, -0.006587982, ...
http://arxiv.org/abs/2309.03164v1
J-Guard: Journalism Guided Adversarially Robust Detection of AI-generated News
Tharindu Kumarage, Amrita Bhattacharjee, Djordje Padejski, Kristy Roschke, Dan Gillmor, Scott Ruston, Huan Liu, Joshua Garland
cs.CL, cs.AI
The rapid proliferation of AI-generated text online is profoundly reshaping the information landscape. Among various types of AI-generated text, AI-generated news presents a significant threat as it can be a prominent source of misinformation online. While several recent efforts have focused on detecting AI-generated t...
http://arxiv.org/pdf/2309.03164v1
2023-09-06T17:06:31
2023-09-06T17:06:31
{ "title": [ -0.017150879, -0.0040016174, -0.032165527, -0.025802612, 0.017440796, -0.00060749054, 0.00573349, -0.02848816, -0.04348755, 0.013122559, -0.025375366, 0.038879395, -0.019470215, -0.0068969727, -0.022735596, -0.021881104, -0.036621094, ...
http://arxiv.org/abs/2309.06600v2
Narrative as a Dynamical System
Isidoros Doxas, James Meiss, Steven Bottone, Tom Strelich, Andrew Plummer, Adrienne Breland, Simon Dennis, Kathy Garvin-Doxas, Michael Klymkowsky
cs.CL, J.5; I.2.7
There is increasing evidence that human activity in general, and narrative in particular, can be treated as a dynamical system in the physics sense; a system whose evolution is described by an action integral, such that the average of all possible paths from point A to point B is given by the extremum of the action. We...
http://arxiv.org/pdf/2309.06600v2
2023-09-29T15:41:37
2023-09-06T16:56:33
{ "title": [ 0.0017328262, 0.0019950867, -0.018066406, -0.03640747, -0.034210205, -0.015602112, -0.027297974, -0.0138168335, -0.007972717, 0.027175903, -0.044525146, -0.0128479, -0.01473999, -0.018951416, -0.023925781, -0.019485474, 0.016571045, ...
http://arxiv.org/abs/2309.03126v2
Everyone Deserves A Reward: Learning Customized Human Preferences
Pengyu Cheng, Jiawen Xie, Ke Bai, Yong Dai, Nan Du
cs.CL
Reward models (RMs) are essential for aligning large language models (LLMs) with human preferences to improve interaction quality. However, the real world is pluralistic, which leads to diversified human preferences with respect to different religions, politics, cultures, etc. Moreover, each individual can have their u...
http://arxiv.org/pdf/2309.03126v2
2023-09-15T09:24:30
2023-09-06T16:03:59
{ "title": [ 0.009284973, 0.012207031, -0.020187378, -0.045959473, 0.027420044, -0.021606445, 0.0119018555, -0.009010315, -0.039001465, 0.0027751923, -0.012748718, 0.005619049, -0.030715942, -0.027786255, -0.010528564, -0.017974854, -0.01876831, ...
http://arxiv.org/abs/2309.03118v1
Knowledge Solver: Teaching LLMs to Search for Domain Knowledge from Knowledge Graphs
Chao Feng, Xinyu Zhang, Zichu Fei
cs.CL
Large language models (LLMs), such as ChatGPT and GPT-4, are versatile and can solve different tasks due to their emergent ability and generalizability. However, LLMs sometimes lack domain-specific knowledge to perform tasks, which would also cause hallucination during inference. In some previous works, additional modu...
http://arxiv.org/pdf/2309.03118v1
2023-09-06T15:55:01
2023-09-06T15:55:01
{ "title": [ -0.057403564, -0.02218628, -0.059387207, -0.029571533, -0.029296875, -0.007095337, -0.051818848, 0.00060224533, 0.0009560585, -0.003692627, -0.045928955, 0.017227173, -0.014282227, -0.023040771, -0.01260376, -0.023529053, -0.0129776, ...
http://arxiv.org/abs/2309.03103v1
ContrastWSD: Enhancing Metaphor Detection with Word Sense Disambiguation Following the Metaphor Identification Procedure
Mohamad Elzohbi, Richard Zhao
cs.CL, cs.LG
This paper presents ContrastWSD, a RoBERTa-based metaphor detection model that integrates the Metaphor Identification Procedure (MIP) and Word Sense Disambiguation (WSD) to extract and contrast the contextual meaning with the basic meaning of a word to determine whether it is used metaphorically in a sentence. By utili...
http://arxiv.org/pdf/2309.03103v1
2023-09-06T15:41:38
2023-09-06T15:41:38
{ "title": [ -0.007789612, -0.008522034, -0.0057754517, -0.00409317, 0.0071258545, -0.0056419373, 0.0032978058, -0.042175293, 0.015159607, 0.03866577, -0.021652222, 0.010238647, 0.043548584, -0.008575439, -0.04272461, 0.014518738, -0.016479492, 0...
http://arxiv.org/abs/2309.03064v1
A Multimodal Analysis of Influencer Content on Twitter
Danae Sánchez Villegas, Catalina Goanta, Nikolaos Aletras
cs.CL, cs.AI, cs.CV
Influencer marketing involves a wide range of strategies in which brands collaborate with popular content creators (i.e., influencers) to leverage their reach, trust, and impact on their audience to promote and endorse products or services. Because followers of influencers are more likely to buy a product after receivi...
http://arxiv.org/pdf/2309.03064v1
2023-09-06T15:07:23
2023-09-06T15:07:23
{ "title": [ 0.0014638901, 0.01411438, -0.034851074, -0.04953003, 0.015853882, -0.0063476562, -0.018981934, -0.00025510788, -0.017562866, 0.04272461, -0.023361206, 0.039154053, -0.007633209, -0.00045919418, -0.036346436, 0.00008779764, -0.017730713, ...
http://arxiv.org/abs/2309.06364v2
Framework-Based Qualitative Analysis of Free Responses of Large Language Models: Algorithmic Fidelity
Aliya Amirova, Theodora Fteropoulli, Nafiso Ahmed, Martin R. Cowie, Joel Z. Leibo
cs.CL, cs.AI
Today, using Large-scale generative Language Models (LLMs) it is possible to simulate free responses to interview questions like those traditionally analyzed using qualitative research methods. Qualitative methodology encompasses a broad family of techniques involving manual analysis of open-ended interviews or convers...
http://arxiv.org/pdf/2309.06364v2
2023-10-27T12:10:54
2023-09-06T15:00:44
{ "title": [ -0.018600464, -0.004497528, -0.04989624, -0.058929443, -0.007698059, -0.019989014, -0.061584473, -0.010253906, -0.02204895, 0.048095703, -0.0034561157, 0.005508423, 0.003501892, -0.01939392, -0.060943604, 0.007843018, -0.01637268, 0....
http://arxiv.org/abs/2309.06550v1
Synthetic Text Generation using Hypergraph Representations
Natraj Raman, Sameena Shah
cs.CL, cs.AI
Generating synthetic variants of a document is often posed as text-to-text transformation. We propose an alternate LLM based method that first decomposes a document into semantic frames and then generates text using this interim sparse format. The frames are modeled using a hypergraph, which allows perturbing the frame...
http://arxiv.org/pdf/2309.06550v1
2023-09-06T14:14:37
2023-09-06T14:14:37
{ "title": [ -0.04345703, -0.0069732666, -0.040863037, -0.056365967, -0.006893158, 0.023544312, -0.06939697, -0.004043579, -0.0024356842, 0.042236328, 0.0026340485, 0.013893127, -0.01701355, -0.00868988, -0.057006836, -0.0029735565, -0.024368286, ...
http://arxiv.org/abs/2309.02915v1
Persona-aware Generative Model for Code-mixed Language
Ayan Sengupta, Md Shad Akhtar, Tanmoy Chakraborty
cs.CL, cs.LG
Code-mixing and script-mixing are prevalent across online social networks and multilingual societies. However, a user's preference toward code-mixing depends on the socioeconomic status, demographics of the user, and the local context, which existing generative models mostly ignore while generating code-mixed texts. In...
http://arxiv.org/pdf/2309.02915v1
2023-09-06T11:20:41
2023-09-06T11:20:41
{ "title": [ -0.027557373, -0.023101807, -0.033721924, -0.051879883, -0.022033691, 0.0028514862, -0.057159424, 0.029327393, 0.0077819824, 0.039367676, 0.0038414001, 0.007411957, 0.023757935, -0.032348633, -0.0357666, 0.012794495, -0.043762207, 0....
http://arxiv.org/abs/2309.02914v1
Leave no Place Behind: Improved Geolocation in Humanitarian Documents
Enrico M. Belliardo, Kyriaki Kalimeri, Yelena Mejova
cs.CL
Geographical location is a crucial element of humanitarian response, outlining vulnerable populations, ongoing events, and available resources. Latest developments in Natural Language Processing may help in extracting vital information from the deluge of reports and documents produced by the humanitarian sector. Howeve...
http://arxiv.org/pdf/2309.02914v1
2023-09-06T11:20:02
2023-09-06T11:20:02
{ "title": [ 0.006340027, 0.0073013306, -0.01979065, -0.046417236, 0.017730713, -0.017349243, -0.012634277, -0.03262329, -0.021194458, 0.032104492, -0.01259613, -0.0016527176, 0.013824463, -0.02519226, -0.022460938, -0.0110321045, 0.0047416687, 0...
http://arxiv.org/abs/2309.02912v1
On the Challenges of Building Datasets for Hate Speech Detection
Vitthal Bhandari
cs.CL, cs.AI
Detection of hate speech has been formulated as a standalone application of NLP and different approaches have been adopted for identifying the target groups, obtaining raw data, defining the labeling process, choosing the detection algorithm, and evaluating the performance in the desired setting. However, unlike other ...
http://arxiv.org/pdf/2309.02912v1
2023-09-06T11:15:47
2023-09-06T11:15:47
{ "title": [ 0.017150879, 0.006500244, -0.038024902, -0.045043945, 0.026794434, -0.028335571, 0.0071525574, -0.007205963, -0.009124756, 0.028884888, -0.03161621, 0.013244629, -0.0006656647, -0.0073127747, -0.043640137, -0.008041382, -0.008232117, ...
http://arxiv.org/abs/2309.02902v1
ViCGCN: Graph Convolutional Network with Contextualized Language Models for Social Media Mining in Vietnamese
Chau-Thang Phan, Quoc-Nam Nguyen, Chi-Thanh Dang, Trong-Hop Do, Kiet Van Nguyen
cs.CL
Social media processing is a fundamental task in natural language processing with numerous applications. As Vietnamese social media and information science have grown rapidly, the necessity of information-based mining on Vietnamese social media has become crucial. However, state-of-the-art research faces several signif...
http://arxiv.org/pdf/2309.02902v1
2023-09-06T10:51:34
2023-09-06T10:51:34
{ "title": [ -0.016830444, -0.01574707, -0.027359009, -0.035491943, -0.0151901245, 0.0060539246, -0.038085938, 0.0016260147, -0.015640259, 0.043914795, -0.04711914, 0.025634766, -0.008979797, -0.0037631989, -0.03564453, -0.013900757, -0.018814087, ...
http://arxiv.org/abs/2309.06572v1
Addressing the Blind Spots in Spoken Language Processing
Amit Moryossef
eess.AS, cs.CL, cs.SD
This paper explores the critical but often overlooked role of non-verbal cues, including co-speech gestures and facial expressions, in human communication and their implications for Natural Language Processing (NLP). We argue that understanding human communication requires a more holistic approach that goes beyond text...
http://arxiv.org/pdf/2309.06572v1
2023-09-06T10:29:25
2023-09-06T10:29:25
{ "title": [ 0.0046424866, -0.0057258606, -0.023345947, -0.052246094, -0.0002617836, -0.0076828003, -0.01979065, 0.0034313202, -0.0012197495, 0.026168823, -0.033477783, -0.01360321, -0.010925293, -0.0135650635, -0.04901123, -0.0013408661, -0.00956726...
http://arxiv.org/abs/2309.02887v1
A deep Natural Language Inference predictor without language-specific training data
Lorenzo Corradi, Alessandro Manenti, Francesca Del Bonifro, Francesco Setti, Dario Del Sorbo
cs.CL, cs.AI
In this paper we present a technique of NLP to tackle the problem of inference relation (NLI) between pairs of sentences in a target language of choice without a language-specific training dataset. We exploit a generic translation dataset, manually translated, along with two instances of the same pre-trained model - th...
http://arxiv.org/pdf/2309.02887v1
2023-09-06T10:20:59
2023-09-06T10:20:59
{ "title": [ -0.01007843, -0.008728027, -0.041931152, -0.07501221, -0.00080394745, 0.0027370453, -0.061706543, 0.0054130554, 0.0084991455, 0.013374329, -0.016708374, 0.0053367615, 0.004108429, -0.013435364, -0.021377563, 0.010192871, -0.008140564, ...
http://arxiv.org/abs/2309.02884v2
Aligning Large Language Models for Clinical Tasks
Supun Manathunga, Isuru Hettigoda
cs.CL, I.2, I.7, J.3
Large Language Models (LLMs) have demonstrated remarkable adaptability, showcasing their capacity to excel in tasks for which they were not explicitly trained. However, despite their impressive natural language processing (NLP) capabilities, effective alignment of LLMs remains a crucial challenge when deploying them fo...
http://arxiv.org/pdf/2309.02884v2
2023-09-07T01:52:33
2023-09-06T10:20:06
{ "title": [ -0.04119873, -0.0103302, -0.058013916, -0.078125, 0.005115509, 0.012008667, -0.09277344, 0.0014181137, 0.007091522, 0.04095459, 0.006828308, -0.019882202, 0.012275696, -0.025512695, -0.04345703, 0.015640259, 0.002248764, 0.024597168,...
http://arxiv.org/abs/2309.02823v1
Promoting Open-domain Dialogue Generation through Learning Pattern Information between Contexts and Responses
Mengjuan Liu, Chenyang Liu, Yunfan Yang, Jiang Liu, Mohan Jing
cs.CL, cs.AI
Recently, utilizing deep neural networks to build the opendomain dialogue models has become a hot topic. However, the responses generated by these models suffer from many problems such as responses not being contextualized and tend to generate generic responses that lack information content, damaging the user's experie...
http://arxiv.org/pdf/2309.02823v1
2023-09-06T08:11:39
2023-09-06T08:11:39
{ "title": [ 0.012176514, 0.0045661926, -0.055358887, -0.07672119, 0.023239136, -0.011390686, -0.041259766, -0.00020992756, -0.013679504, 0.01890564, 0.014144897, 0.017959595, -0.026519775, -0.02897644, -0.026901245, 0.0042648315, -0.009750366, 0...
http://arxiv.org/abs/2309.02812v1
Agent-based simulation of pedestrians' earthquake evacuation; application to Beirut, Lebanon
Rouba Iskandar, Kamel Allaw, Julie Dugdale, Elise Beck, Jocelyne Adjizian-Gérard, Cécile Cornou, Jacques Harb, Pascal Lacroix, Nada Badaro-Saliba, Stéphane Cartier, Rita Zaarour
cs.CL
Most seismic risk assessment methods focus on estimating the damages to the built environment and the consequent socioeconomic losses without fully taking into account the social aspect of risk. Yet, human behaviour is a key element in predicting the human impact of an earthquake, therefore, it is important to include ...
http://arxiv.org/pdf/2309.02812v1
2023-09-06T08:00:17
2023-09-06T08:00:17
{ "title": [ 0.013130188, 0.021377563, -0.029754639, -0.034118652, 0.013221741, -0.026062012, 0.029281616, -0.023895264, -0.031311035, 0.041015625, -0.023025513, 0.023101807, -0.029785156, -0.018966675, -0.038482666, -0.019332886, 0.008430481, -0...
http://arxiv.org/abs/2309.02784v1
Norm Tweaking: High-performance Low-bit Quantization of Large Language Models
Liang Li, Qingyuan Li, Bo Zhang, Xiangxiang Chu
cs.LG, cs.AI, cs.CL
As the size of large language models (LLMs) continues to grow, model compression without sacrificing accuracy has become a crucial challenge for deployment. While some quantization methods, such as GPTQ, have made progress in achieving acceptable 4-bit weight-only quantization, attempts at lower bit quantization often ...
http://arxiv.org/pdf/2309.02784v1
2023-09-06T06:51:15
2023-09-06T06:51:15
{ "title": [ -0.052703857, -0.014945984, -0.051330566, -0.049316406, -0.0012788773, -0.0075263977, -0.061676025, 0.013801575, -0.0020179749, 0.07043457, 0.0010194778, -0.0059165955, 0.025131226, -0.007648468, -0.06958008, 0.025802612, 0.0010547638, ...
http://arxiv.org/abs/2309.02780v2
GRASS: Unified Generation Model for Speech-to-Semantic Tasks
Aobo Xia, Shuyu Lei, Yushu Yang, Xiang Guo, Hua Chai
cs.CL, cs.SD, eess.AS
This paper explores the instruction fine-tuning technique for speech-to-semantic tasks by introducing a unified end-to-end (E2E) framework that generates target text conditioned on a task-related prompt for audio data. We pre-train the model using large and diverse data, where instruction-speech pairs are constructed v...
http://arxiv.org/pdf/2309.02780v2
2023-09-11T09:35:14
2023-09-06T06:44:26
{ "title": [ -0.028045654, -0.014465332, -0.049804688, -0.087890625, 0.011497498, 0.009414673, -0.07330322, 0.03475952, 0.006095886, 0.04949951, -0.0053863525, 0.0029659271, 0.02255249, -0.01713562, -0.04937744, 0.026275635, -0.013252258, 0.03961...
http://arxiv.org/abs/2309.02772v2
Improving Code Generation by Dynamic Temperature Sampling
Yuqi Zhu, Jia Li, Ge Li, YunFei Zhao, Jia Li, Zhi Jin, Hong Mei
cs.SE, cs.CL
Recently, Large Language Models (LLMs) have shown impressive results in code generation. However, existing decoding strategies are designed for Natural Language (NL) generation, overlooking the differences between NL and programming languages (PL). Due to this oversight, a better decoding strategy for code generation r...
http://arxiv.org/pdf/2309.02772v2
2023-10-14T09:27:51
2023-09-06T06:27:33
{ "title": [ -0.029541016, -0.008026123, -0.046722412, -0.07525635, 0.0034294128, 0.007331848, -0.032196045, 0.010986328, -0.012611389, 0.035949707, -0.0013103485, 0.02722168, -0.019836426, -0.02357483, -0.019210815, 0.013908386, 0.018859863, -0....
http://arxiv.org/abs/2309.03241v2
GPT Can Solve Mathematical Problems Without a Calculator
Zhen Yang, Ming Ding, Qingsong Lv, Zhihuan Jiang, Zehai He, Yuyi Guo, Jinfeng Bai, Jie Tang
cs.LG, cs.AI, cs.CL
Previous studies have typically assumed that large language models are unable to accurately perform arithmetic operations, particularly multiplication of >8 digits, and operations involving decimals and fractions, without the use of calculator tools. This paper aims to challenge this misconception. With sufficient trai...
http://arxiv.org/pdf/2309.03241v2
2023-09-12T11:01:25
2023-09-06T06:18:16
{ "title": [ 0.044647217, 0.009994507, -0.011955261, -0.0077209473, -0.02859497, -0.034423828, -0.017974854, -0.04345703, 0.0041236877, 0.037017822, -0.0072402954, -0.03387451, 0.0035953522, -0.030776978, -0.005207062, 0.0020065308, 0.026565552, ...
http://arxiv.org/abs/2309.02740v1
Rubric-Specific Approach to Automated Essay Scoring with Augmentation Training
Brian Cho, Youngbin Jang, Jaewoong Yoon
cs.CL, cs.AI, I.2.7
Neural based approaches to automatic evaluation of subjective responses have shown superior performance and efficiency compared to traditional rule-based and feature engineering oriented solutions. However, it remains unclear whether the suggested neural solutions are sufficient replacements of human raters as we find ...
http://arxiv.org/pdf/2309.02740v1
2023-09-06T05:51:19
2023-09-06T05:51:19
{ "title": [ -0.0070724487, 0.022079468, -0.026184082, -0.008399963, -0.009017944, -0.013946533, 0.018417358, -0.032958984, -0.022888184, 0.029815674, -0.03274536, 0.02268982, 0.00031876564, -0.01612854, -0.007713318, -0.011817932, -0.008323669, ...
http://arxiv.org/abs/2309.02731v1
HC3 Plus: A Semantic-Invariant Human ChatGPT Comparison Corpus
Zhenpeng Su, Xing Wu, Wei Zhou, Guangyuan Ma, Songlin Hu
cs.CL, cs.AI
ChatGPT has gained significant interest due to its impressive performance, but people are increasingly concerned about its potential risks, particularly around the detection of AI-generated content (AIGC), which is often difficult for untrained humans to identify. Current datasets utilized for detecting ChatGPT-generat...
http://arxiv.org/pdf/2309.02731v1
2023-09-06T05:33:57
2023-09-06T05:33:57
{ "title": [ -0.017196655, -0.010444641, -0.024658203, -0.03842163, 0.0004029274, 0.0012273788, -0.043762207, -0.00020349026, -0.003068924, 0.037506104, -0.029388428, 0.027755737, 0.0025749207, 0.0013542175, -0.044006348, 0.019714355, -0.022445679, ...
http://arxiv.org/abs/2309.02726v1
Large Language Models for Automated Open-domain Scientific Hypotheses Discovery
Zonglin Yang, Xinya Du, Junxian Li, Jie Zheng, Soujanya Poria, Erik Cambria
cs.CL, cs.AI
Hypothetical induction is recognized as the main reasoning type when scientists make observations about the world and try to propose hypotheses to explain those observations. Past research on hypothetical induction has a limited setting that (1) the observation annotations of the dataset are not raw web corpus but are ...
http://arxiv.org/pdf/2309.02726v1
2023-09-06T05:19:41
2023-09-06T05:19:41
{ "title": [ -0.046539307, -0.006965637, -0.052368164, -0.05105591, -0.009544373, 0.006462097, -0.04650879, -0.0059394836, 0.009651184, 0.030136108, 0.009841919, -0.021636963, 0.013755798, -0.03338623, -0.031433105, 0.027130127, -0.0057678223, 0....
http://arxiv.org/abs/2309.02724v1
Offensive Hebrew Corpus and Detection using BERT
Nagham Hamad, Mustafa Jarrar, Mohammad Khalilia, Nadim Nashif
cs.CL, cs.AI, cs.LG, I.2.1; I.2.6; I.2.7; I.5.1
Offensive language detection has been well studied in many languages, but it is lagging behind in low-resource languages, such as Hebrew. In this paper, we present a new offensive language corpus in Hebrew. A total of 15,881 tweets were retrieved from Twitter. Each was labeled with one or more of five classes (abusive,...
http://arxiv.org/pdf/2309.02724v1
2023-09-06T05:18:43
2023-09-06T05:18:43
{ "title": [ 0.021820068, -0.0025100708, -0.025802612, -0.06549072, 0.0057678223, -0.010559082, -0.017715454, -0.00004673004, -0.011146545, 0.045837402, -0.0152282715, -0.008049011, 0.03933716, -0.028915405, -0.033050537, 0.008773804, -0.02482605, ...
http://arxiv.org/abs/2309.02706v4
HAE-RAE Bench: Evaluation of Korean Knowledge in Language Models
Guijin Son, Hanwool Lee, Suwan Kim, Huiseo Kim, Jaecheol Lee, Je Won Yeom, Jihyu Jung, Jung Woo Kim, Songseong Kim
cs.CL
Large Language Models (LLMs) trained on massive corpora demonstrate impressive capabilities in a wide range of tasks. While there are ongoing efforts to adapt these models to languages beyond English, the attention given to their evaluation methodologies remains limited. Current multilingual benchmarks often rely on ba...
http://arxiv.org/pdf/2309.02706v4
2023-09-23T07:44:06
2023-09-06T04:38:16
{ "title": [ -0.06274414, -0.0154953, -0.04647827, -0.03652954, -0.027160645, 0.001783371, -0.0769043, -0.008201599, 0.0056533813, 0.03543091, -0.013618469, -0.0016794205, 0.0124435425, -0.030822754, -0.05026245, 0.0051574707, -0.0069465637, 0.02...
http://arxiv.org/abs/2309.02705v1
Certifying LLM Safety against Adversarial Prompting
Aounon Kumar, Chirag Agarwal, Suraj Srinivas, Soheil Feizi, Hima Lakkaraju
cs.CL, cs.AI, cs.CR, cs.LG
Large language models (LLMs) released for public use incorporate guardrails to ensure their output is safe, often referred to as "model alignment." An aligned language model should decline a user's request to produce harmful content. However, such safety measures are vulnerable to adversarial prompts, which contain mal...
http://arxiv.org/pdf/2309.02705v1
2023-09-06T04:37:20
2023-09-06T04:37:20
{ "title": [ -0.003753662, 0.0140686035, -0.023880005, -0.027832031, -0.034820557, 0.0071144104, -0.04837036, -0.02822876, -0.072265625, 0.01852417, 0.00070238113, 0.03491211, -0.020751953, -0.04916382, 0.036743164, -0.01663208, -0.00699234, 0.00...
http://arxiv.org/abs/2309.02691v1
A Joint Study of Phrase Grounding and Task Performance in Vision and Language Models
Noriyuki Kojima, Hadar Averbuch-Elor, Yoav Artzi
cs.CL, cs.CV
Key to tasks that require reasoning about natural language in visual contexts is grounding words and phrases to image regions. However, observing this grounding in contemporary models is complex, even if it is generally expected to take place if the task is addressed in a way that is conductive to generalization. We pr...
http://arxiv.org/pdf/2309.02691v1
2023-09-06T03:54:57
2023-09-06T03:54:57
{ "title": [ -0.016204834, 0.004634857, -0.036315918, -0.021316528, -0.011245728, -0.044158936, -0.024978638, -0.0093307495, 0.02784729, 0.04309082, -0.03265381, 0.000995636, 0.029830933, -0.0067481995, -0.061676025, 0.0027580261, 0.0001975298, 0...
http://arxiv.org/abs/2309.03238v1
Implicit Design Choices and Their Impact on Emotion Recognition Model Development and Evaluation
Mimansa Jaiswal
cs.LG, cs.CL, cs.SD, eess.AS
Emotion recognition is a complex task due to the inherent subjectivity in both the perception and production of emotions. The subjectivity of emotions poses significant challenges in developing accurate and robust computational models. This thesis examines critical facets of emotion recognition, beginning with the coll...
http://arxiv.org/pdf/2309.03238v1
2023-09-06T02:45:42
2023-09-06T02:45:42
{ "title": [ -0.021331787, -0.0054969788, -0.02558899, -0.0105896, -0.0060691833, -0.0236969, -0.010719299, -0.030166626, 0.02709961, 0.009170532, -0.0368042, -0.020812988, -0.015220642, -0.02861023, -0.052124023, -0.015975952, -0.005970001, 0.02...
http://arxiv.org/abs/2309.02654v3
Zero-Resource Hallucination Prevention for Large Language Models
Junyu Luo, Cao Xiao, Fenglong Ma
cs.CL
The prevalent use of large language models (LLMs) in various domains has drawn attention to the issue of "hallucination," which refers to instances where LLMs generate factually inaccurate or ungrounded information. Existing techniques for hallucination detection in language assistants rely on intricate fuzzy, specific...
http://arxiv.org/pdf/2309.02654v3
2023-10-08T02:08:41
2023-09-06T01:57:36
{ "title": [ 0.005519867, 0.0029907227, -0.0345459, -0.07879639, 0.0045547485, 0.0046157837, -0.049957275, 0.010894775, 0.0059280396, 0.039733887, 0.0035972595, -0.004638672, 0.022003174, -0.017532349, -0.04019165, 0.027496338, -0.012535095, 0.04...
http://arxiv.org/abs/2309.02640v1
Epi-Curriculum: Episodic Curriculum Learning for Low-Resource Domain Adaptation in Neural Machine Translation
Keyu Chen, Di Zhuang, Mingchen Li, J. Morris Chang
cs.LG, cs.CL
Neural Machine Translation (NMT) models have become successful, but their performance remains poor when translating on new domains with a limited number of data. In this paper, we present a novel approach Epi-Curriculum to address low-resource domain adaptation (DA), which contains a new episodic training framework alo...
http://arxiv.org/pdf/2309.02640v1
2023-09-06T00:59:27
2023-09-06T00:59:27
{ "title": [ -0.014190674, -0.021896362, -0.0093688965, -0.040649414, -0.013343811, -0.01108551, -0.007408142, 0.010322571, -0.013755798, 0.022125244, 0.014801025, 0.002653122, 0.0033130646, -0.036865234, 0.00020730495, 0.0047950745, -0.016067505, ...
http://arxiv.org/abs/2309.02591v1
Scaling Autoregressive Multi-Modal Models: Pretraining and Instruction Tuning
Lili Yu, Bowen Shi, Ramakanth Pasunuru, Benjamin Muller, Olga Golovneva, Tianlu Wang, Arun Babu, Binh Tang, Brian Karrer, Shelly Sheynin, Candace Ross, Adam Polyak, Russell Howes, Vasu Sharma, Puxin Xu, Hovhannes Tamoyan, Oron Ashual, Uriel Singer, Shang-Wen Li, Susan Zhang, Richard James, Gargi Ghosh, Yaniv Taigman, M...
cs.LG, cs.CL, cs.CV
We present CM3Leon (pronounced "Chameleon"), a retrieval-augmented, token-based, decoder-only multi-modal language model capable of generating and infilling both text and images. CM3Leon uses the CM3 multi-modal architecture but additionally shows the extreme benefits of scaling up and tuning on more diverse instructio...
http://arxiv.org/pdf/2309.02591v1
2023-09-05T21:27:27
2023-09-05T21:27:27
{ "title": [ -0.03567505, 0.0016145706, -0.03768921, -0.039245605, -0.0012607574, -0.0021076202, -0.070373535, -0.031585693, -0.0155181885, 0.048736572, -0.020828247, 0.006866455, -0.005264282, -0.015686035, -0.04763794, -0.009216309, 0.010047913, ...
http://arxiv.org/abs/2309.02553v3
Automating Behavioral Testing in Machine Translation
Javier Ferrando, Matthias Sperber, Hendra Setiawan, Dominic Telaar, Saša Hasan
cs.CL, cs.AI
Behavioral testing in NLP allows fine-grained evaluation of systems by examining their linguistic capabilities through the analysis of input-output behavior. Unfortunately, existing work on behavioral testing in Machine Translation (MT) is currently restricted to largely handcrafted tests covering a limited range of ca...
http://arxiv.org/pdf/2309.02553v3
2023-11-03T00:13:25
2023-09-05T19:40:45
{ "title": [ -0.024017334, -0.018478394, 0.0036182404, -0.06573486, -0.032043457, 0.0236969, -0.054534912, 0.008087158, -0.017654419, 0.050964355, -0.02420044, 0.021255493, 0.01927185, -0.011154175, -0.040893555, 0.013381958, 0.0011749268, 0.0279...
http://arxiv.org/abs/2309.02427v2
Cognitive Architectures for Language Agents
Theodore R. Sumers, Shunyu Yao, Karthik Narasimhan, Thomas L. Griffiths
cs.AI, cs.CL, cs.LG, cs.SC
Recent efforts have augmented large language models (LLMs) with external resources (e.g., the Internet) or internal control flows (e.g., prompt chaining) for tasks requiring grounding or reasoning, leading to a new class of language agents. While these agents have achieved substantial empirical success, we lack a syste...
http://arxiv.org/pdf/2309.02427v2
2023-09-27T15:27:25
2023-09-05T17:56:20
{ "title": [ -0.030838013, -0.011932373, -0.006958008, -0.036987305, -0.0070228577, 0.00944519, -0.024780273, -0.0049095154, 0.0064201355, 0.04272461, -0.03375244, -0.005607605, 0.011062622, -0.0021266937, -0.058563232, 0.0031013489, -0.013450623, ...
http://arxiv.org/abs/2309.02403v2
Substitution-based Semantic Change Detection using Contextual Embeddings
Dallas Card
cs.CL
Measuring semantic change has thus far remained a task where methods using contextual embeddings have struggled to improve upon simpler techniques relying only on static word vectors. Moreover, many of the previously proposed approaches suffer from downsides related to scalability and ease of interpretation. We present...
http://arxiv.org/pdf/2309.02403v2
2023-09-06T04:02:55
2023-09-05T17:33:59
{ "title": [ -0.0047416687, 0.010269165, -0.046447754, -0.06378174, 0.02494812, -0.005870819, -0.044403076, -0.025650024, 0.015975952, 0.01777649, -0.03164673, 0.010139465, 0.03366089, -0.017456055, -0.007358551, 0.0036125183, -0.00039172173, 0.0...
http://arxiv.org/abs/2309.02373v2
nanoT5: A PyTorch Framework for Pre-training and Fine-tuning T5-style Models with Limited Resources
Piotr Nawrot
cs.CL
State-of-the-art language models like T5 have revolutionized the NLP landscape, but their computational demands hinder a large portion of the research community. To address this challenge, we present nanoT5, a specially-optimized PyTorch framework for efficient pre-training and fine-tuning of T5 models. Drawing on insi...
http://arxiv.org/pdf/2309.02373v2
2023-10-24T14:53:50
2023-09-05T16:35:41
{ "title": [ -0.010444641, -0.00182724, -0.026519775, -0.06185913, -0.014183044, 0.0016412735, -0.07800293, -0.03692627, -0.0043411255, 0.019104004, -0.01537323, 0.02468872, 0.0027236938, -0.023712158, -0.002193451, 0.019500732, 0.008102417, 0.00...
http://arxiv.org/abs/2309.02311v1
Weigh Your Own Words: Improving Hate Speech Counter Narrative Generation via Attention Regularization
Helena Bonaldi, Giuseppe Attanasio, Debora Nozza, Marco Guerini
cs.CL
Recent computational approaches for combating online hate speech involve the automatic generation of counter narratives by adapting Pretrained Transformer-based Language Models (PLMs) with human-curated data. This process, however, can produce in-domain overfitting, resulting in models generating acceptable narratives ...
http://arxiv.org/pdf/2309.02311v1
2023-09-05T15:27:22
2023-09-05T15:27:22
{ "title": [ 0.031799316, 0.0077323914, -0.0115737915, -0.048339844, 0.0030136108, -0.016189575, -0.004234314, -0.022079468, -0.018066406, 0.035339355, -0.006629944, 0.036468506, -0.008552551, -0.008262634, -0.048858643, -0.0020160675, -0.017929077, ...
http://arxiv.org/abs/2310.06846v1
Exploiting Language Models as a Source of Knowledge for Cognitive Agents
James R. Kirk, Robert E. Wray, John E. Laird
cs.AI, cs.CL, I.2.7; I.2.11
Large language models (LLMs) provide capabilities far beyond sentence completion, including question answering, summarization, and natural-language inference. While many of these capabilities have potential application to cognitive systems, our research is exploiting language models as a source of task knowledge for co...
http://arxiv.org/pdf/2310.06846v1
2023-09-05T15:18:04
2023-09-05T15:18:04
{ "title": [ -0.033325195, -0.014724731, -0.018096924, -0.028167725, -0.0062789917, -0.0052108765, -0.035064697, -0.019927979, 0.017059326, 0.031280518, -0.025268555, -0.015792847, 0.0050735474, -0.007507324, -0.061187744, 0.0021476746, -0.007297516,...
http://arxiv.org/abs/2309.02285v2
PromptTTS 2: Describing and Generating Voices with Text Prompt
Yichong Leng, Zhifang Guo, Kai Shen, Xu Tan, Zeqian Ju, Yanqing Liu, Yufei Liu, Dongchao Yang, Leying Zhang, Kaitao Song, Lei He, Xiang-Yang Li, Sheng Zhao, Tao Qin, Jiang Bian
eess.AS, cs.CL, cs.LG, cs.SD
Speech conveys more information than text, as the same word can be uttered in various voices to convey diverse information. Compared to traditional text-to-speech (TTS) methods relying on speech prompts (reference speech) for voice variability, using text prompts (descriptions) is more user-friendly since speech prompt...
http://arxiv.org/pdf/2309.02285v2
2023-10-12T03:05:36
2023-09-05T14:45:27
{ "title": [ -0.024337769, -0.01134491, -0.06604004, -0.057281494, -0.0032539368, -0.0044441223, -0.057800293, -0.0029182434, -0.0003876686, 0.050598145, -0.028717041, -0.0011968613, 0.04156494, -0.024017334, -0.031402588, 0.012313843, -0.008605957, ...
http://arxiv.org/abs/2309.02240v1
Dialog Action-Aware Transformer for Dialog Policy Learning
Huimin Wang, Wai-Chung Kwan, Kam-Fai Wong
cs.CL
Recent works usually address Dialog policy learning DPL by training a reinforcement learning (RL) agent to determine the best dialog action. However, existing works on deep RL require a large volume of agent-user interactions to achieve acceptable performance. In this paper, we propose to make full use of the plain tex...
http://arxiv.org/pdf/2309.02240v1
2023-09-05T13:47:25
2023-09-05T13:47:25
{ "title": [ 0.02394104, 0.007194519, -0.042510986, -0.074035645, 0.011238098, -0.019714355, -0.024749756, 0.005947113, -0.012161255, 0.03427124, -0.01676941, 0.021484375, -0.004501343, -0.024093628, -0.01398468, -0.005584717, -0.015388489, 0.032...
http://arxiv.org/abs/2309.06495v1
AGIBench: A Multi-granularity, Multimodal, Human-referenced, Auto-scoring Benchmark for Large Language Models
Fei Tang, Wanling Gao, Luzhou Peng, Jianfeng Zhan
cs.CL, cs.AI, cs.PF
Large language models (LLMs) like ChatGPT have revealed amazing intelligence. How to evaluate the question-solving abilities of LLMs and their degrees of intelligence is a hot-spot but challenging issue. First, the question-solving abilities are interlaced with different ability branches like understanding and massive ...
http://arxiv.org/pdf/2309.06495v1
2023-09-05T13:43:37
2023-09-05T13:43:37
{ "title": [ -0.035308838, 0.001405716, -0.03086853, -0.06781006, -0.00566864, 0.019866943, -0.09564209, -0.012672424, 0.014152527, 0.032928467, -0.015457153, 0.0025577545, 0.02267456, -0.016815186, -0.050872803, 0.009864807, -0.0070762634, 0.013...
http://arxiv.org/abs/2309.02233v1
Augmenting Black-box LLMs with Medical Textbooks for Clinical Question Answering
Yubo Wang, Xueguang Ma, Wenhu Chen
cs.CL, cs.AI
Large-scale language models (LLMs), such as ChatGPT, are capable of generating human-like responses for various downstream tasks, such as task-oriented dialogues and question answering. However, applying LLMs to medical domains remains challenging due to their inability to leverage domain-specific knowledge. In this st...
http://arxiv.org/pdf/2309.02233v1
2023-09-05T13:39:38
2023-09-05T13:39:38
{ "title": [ -0.027389526, -0.011566162, -0.038238525, -0.059417725, -0.026016235, 0.012954712, -0.08081055, -0.008995056, -0.0029582977, 0.009628296, -0.041900635, 0.024963379, -0.016952515, -0.0053482056, -0.020965576, -0.024810791, 0.004081726, ...
http://arxiv.org/abs/2309.02189v1
Leveraging BERT Language Models for Multi-Lingual ESG Issue Identification
Elvys Linhares Pontes, Mohamed Benjannet, Lam Kim Ming
cs.CL, cs.AI, cs.LG
Environmental, Social, and Governance (ESG) has been used as a metric to measure the negative impacts and enhance positive outcomes of companies in areas such as the environment, society, and governance. Recently, investors have increasingly recognized the significance of ESG criteria in their investment choices, leadi...
http://arxiv.org/pdf/2309.02189v1
2023-09-05T12:48:21
2023-09-05T12:48:21
{ "title": [ -0.009292603, -0.020736694, -0.040802002, -0.070007324, 0.00047540665, 0.010177612, -0.06530762, 0.023422241, -0.019515991, 0.029006958, -0.009559631, -0.0060424805, 0.0077438354, -0.015914917, -0.03668213, -0.008148193, -0.023742676, ...
http://arxiv.org/abs/2309.02188v1
Incorporating Dictionaries into a Neural Network Architecture to Extract COVID-19 Medical Concepts From Social Media
Abul Hasan, Mark Levene, David Weston
cs.CL, cs.SI
We investigate the potential benefit of incorporating dictionary information into a neural network architecture for natural language processing. In particular, we make use of this architecture to extract several concepts related to COVID-19 from an on-line medical forum. We use a sample from the forum to manually curat...
http://arxiv.org/pdf/2309.02188v1
2023-09-05T12:47:44
2023-09-05T12:47:44
{ "title": [ -0.0067481995, 0.0021953583, -0.0435791, -0.053222656, 0.008460999, -0.011192322, -0.01449585, -0.008728027, -0.013114929, 0.03933716, -0.03717041, 0.010009766, 0.013404846, -0.0039520264, -0.030426025, -0.015914917, 0.0033721924, 0....
http://arxiv.org/abs/2309.02162v1
Advancing Text-to-GLOSS Neural Translation Using a Novel Hyper-parameter Optimization Technique
Younes Ouargani, Noussaima El Khattabi
cs.CL, I.2.7
In this paper, we investigate the use of transformers for Neural Machine Translation of text-to-GLOSS for Deaf and Hard-of-Hearing communication. Due to the scarcity of available data and limited resources for text-to-GLOSS translation, we treat the problem as a low-resource language task. We use our novel hyper-parame...
http://arxiv.org/pdf/2309.02162v1
2023-09-05T11:59:31
2023-09-05T11:59:31
{ "title": [ -0.025878906, -0.00089740753, -0.015403748, -0.039916992, -0.0045814514, 0.009361267, -0.04055786, -0.005947113, -0.017852783, 0.062316895, 0.00774765, 0.012054443, 0.04043579, -0.017242432, -0.027175903, 0.024871826, 0.00623703, 0.0...
http://arxiv.org/abs/2309.02145v1
Bring the Noise: Introducing Noise Robustness to Pretrained Automatic Speech Recognition
Patrick Eickhoff, Matthias Möller, Theresa Pekarek Rosin, Johannes Twiefel, Stefan Wermter
cs.CL, cs.SD, eess.AS
In recent research, in the domain of speech processing, large End-to-End (E2E) systems for Automatic Speech Recognition (ASR) have reported state-of-the-art performance on various benchmarks. These systems intrinsically learn how to handle and remove noise conditions from speech. Previous research has shown, that it is...
http://arxiv.org/pdf/2309.02145v1
2023-09-05T11:34:21
2023-09-05T11:34:21
{ "title": [ 0.0058517456, 0.00982666, -0.040618896, -0.080566406, 0.0317688, -0.005554199, -0.03189087, 0.00010365248, -0.008720398, 0.034057617, -0.032470703, 0.022415161, -0.001824379, -0.008308411, -0.053222656, -0.0079422, 0.0065460205, 0.04...
http://arxiv.org/abs/2309.02144v1
Making Large Language Models Better Reasoners with Alignment
Peiyi Wang, Lei Li, Liang Chen, Feifan Song, Binghuai Lin, Yunbo Cao, Tianyu Liu, Zhifang Sui
cs.CL, cs.AI, cs.LG
Reasoning is a cognitive process of using evidence to reach a sound conclusion. The reasoning capability is essential for large language models (LLMs) to serve as the brain of the artificial general intelligence agent. Recent studies reveal that fine-tuning LLMs on data with the chain of thought (COT) reasoning process...
http://arxiv.org/pdf/2309.02144v1
2023-09-05T11:32:48
2023-09-05T11:32:48
{ "title": [ -0.02116394, -0.008392334, -0.038909912, -0.08251953, 0.0044822693, 0.012786865, -0.07196045, 0.014701843, 0.011009216, 0.032196045, -0.020736694, -0.005531311, 0.00868988, -0.006515503, -0.042785645, 0.010932922, -0.01473999, 0.0325...
http://arxiv.org/abs/2309.02133v1
Evaluating Methods for Ground-Truth-Free Foreign Accent Conversion
Wen-Chin Huang, Tomoki Toda
cs.SD, cs.CL, eess.AS
Foreign accent conversion (FAC) is a special application of voice conversion (VC) which aims to convert the accented speech of a non-native speaker to a native-sounding speech with the same speaker identity. FAC is difficult since the native speech from the desired non-native speaker to be used as the training target i...
http://arxiv.org/pdf/2309.02133v1
2023-09-05T11:22:08
2023-09-05T11:22:08
{ "title": [ 0.003917694, 0.003238678, -0.008583069, -0.022232056, -0.0059814453, -0.008270264, -0.009933472, -0.0262146, -0.01260376, 0.05984497, 0.010635376, -0.015289307, 0.032409668, -0.03491211, -0.047973633, 0.025619507, 0.0046081543, 0.022...
http://arxiv.org/abs/2309.02110v3
Wordle: A Microcosm of Life. Luck, Skill, Cheating, Loyalty, and Influence!
James P. Dilger
math.HO, cs.CL
Wordle is a popular, online word game offered by the New York Times (nytimes.com). Currently there are some 2 million players of the English version worldwide. Players have 6 attempts to guess the daily word (target word) and after each attempt, the player receives color-coded information about the correctness and posi...
http://arxiv.org/pdf/2309.02110v3
2023-10-07T19:31:14
2023-09-05T10:38:53
{ "title": [ -0.007709503, -0.033081055, -0.026153564, -0.029037476, -0.024978638, -0.03378296, -0.014450073, 0.01676941, 0.01197052, 0.013702393, -0.03375244, 0.03048706, 0.010955811, -0.025634766, 0.028442383, -0.0135269165, -0.02418518, 0.0337...
http://arxiv.org/abs/2309.02106v1
Leveraging Label Information for Multimodal Emotion Recognition
Peiying Wang, Sunlu Zeng, Junqing Chen, Lu Fan, Meng Chen, Youzheng Wu, Xiaodong He
cs.CL, cs.AI, cs.LG, eess.AS
Multimodal emotion recognition (MER) aims to detect the emotional status of a given expression by combining the speech and text information. Intuitively, label information should be capable of helping the model locate the salient tokens/frames relevant to the specific emotion, which finally facilitates the MER task. In...
http://arxiv.org/pdf/2309.02106v1
2023-09-05T10:26:32
2023-09-05T10:26:32
{ "title": [ -0.021438599, -0.00009286404, -0.0473938, -0.044433594, 0.0019836426, -0.021850586, -0.022750854, -0.0014190674, 0.015609741, 0.02154541, -0.016235352, -0.00869751, 0.0044403076, -0.026062012, -0.04647827, 0.0024280548, -0.00080919266, ...