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/2306.17575v1
Augmenting Holistic Review in University Admission using Natural Language Processing for Essays and Recommendation Letters
Jinsook Lee, Bradon Thymes, Joyce Zhou, Thorsten Joachims, Rene F. Kizilcec
cs.CL, cs.AI, cs.LG
University admission at many highly selective institutions uses a holistic review process, where all aspects of the application, including protected attributes (e.g., race, gender), grades, essays, and recommendation letters are considered, to compose an excellent and diverse class. In this study, we empirically evalua...
http://arxiv.org/pdf/2306.17575v1
2023-06-30T11:51:08
2023-06-30T11:51:08
{ "title": [ 0.0074691772, -0.018875122, -0.0041275024, 0.0063934326, -0.031433105, -0.0070991516, 0.01763916, -0.030197144, 0.008605957, 0.01499939, -0.070739746, 0.019088745, 0.014549255, 0.007911682, -0.02357483, -0.028884888, -0.029571533, 0....
http://arxiv.org/abs/2306.17564v1
A Cost-aware Study of Depression Language on Social Media using Topic and Affect Contextualization
Andrea Laguna, Oscar Araque
cs.CL, 68T50
Depression is a growing issue in society's mental health that affects all areas of life and can even lead to suicide. Fortunately, prevention programs can be effective in its treatment. In this context, this work proposes an automatic system for detecting depression on social media based on machine learning and natural...
http://arxiv.org/pdf/2306.17564v1
2023-06-30T11:34:48
2023-06-30T11:34:48
{ "title": [ 0.018035889, -0.002670288, -0.010215759, -0.06439209, 0.007045746, 0.0019798279, -0.040252686, -0.010650635, -0.029541016, 0.045196533, -0.018295288, 0.041229248, 0.033691406, -0.019241333, -0.009361267, 0.0013818741, -0.0052833557, ...
http://arxiv.org/abs/2306.17563v1
Large Language Models are Effective Text Rankers with Pairwise Ranking Prompting
Zhen Qin, Rolf Jagerman, Kai Hui, Honglei Zhuang, Junru Wu, Jiaming Shen, Tianqi Liu, Jialu Liu, Donald Metzler, Xuanhui Wang, Michael Bendersky
cs.IR, cs.CL, cs.LG
Ranking documents using Large Language Models (LLMs) by directly feeding the query and candidate documents into the prompt is an interesting and practical problem. However, there has been limited success so far, as researchers have found it difficult to outperform fine-tuned baseline rankers on benchmark datasets. We a...
http://arxiv.org/pdf/2306.17563v1
2023-06-30T11:32:25
2023-06-30T11:32:25
{ "title": [ 0.008918762, 0.0035648346, -0.0074386597, -0.029953003, -0.029129028, -0.021743774, -0.03390503, -0.03387451, 0.0124435425, 0.035064697, -0.023468018, -0.0031833649, 0.023254395, -0.026794434, -0.026626587, -0.018478394, -0.0140686035, ...
http://arxiv.org/abs/2306.17558v2
Towards the extraction of robust sign embeddings for low resource sign language recognition
Mathieu De Coster, Ellen Rushe, Ruth Holmes, Anthony Ventresque, Joni Dambre
cs.CV, cs.CL
Isolated Sign Language Recognition (SLR) has mostly been applied on datasets containing signs executed slowly and clearly by a limited group of signers. In real-world scenarios, however, we are met with challenging visual conditions, coarticulated signing, small datasets, and the need for signer independent models. To ...
http://arxiv.org/pdf/2306.17558v2
2023-08-16T08:46:52
2023-06-30T11:21:40
{ "title": [ -0.026473999, 0.011482239, -0.03942871, -0.037719727, 0.00082302094, 0.0012350082, -0.0435791, -0.018707275, 0.023666382, 0.014518738, -0.055389404, 0.0055007935, -0.024856567, -0.0070495605, -0.054382324, -0.02671814, -0.002998352, ...
http://arxiv.org/abs/2306.17519v2
GPT-FinRE: In-context Learning for Financial Relation Extraction using Large Language Models
Pawan Kumar Rajpoot, Ankur Parikh
cs.CL
Relation extraction (RE) is a crucial task in natural language processing (NLP) that aims to identify and classify relationships between entities mentioned in text. In the financial domain, relation extraction plays a vital role in extracting valuable information from financial documents, such as news articles, earning...
http://arxiv.org/pdf/2306.17519v2
2023-07-21T06:57:49
2023-06-30T10:12:30
{ "title": [ -0.010276794, -0.0010814667, -0.05404663, -0.060180664, 0.0067329407, 0.0009045601, -0.044036865, -0.0044937134, 0.022628784, 0.029388428, -0.039978027, -0.008300781, 0.019332886, -0.004180908, -0.049957275, -0.008277893, -0.023208618, ...
http://arxiv.org/abs/2306.17492v1
Preference Ranking Optimization for Human Alignment
Feifan Song, Bowen Yu, Minghao Li, Haiyang Yu, Fei Huang, Yongbin Li, Houfeng Wang
cs.CL, cs.AI
Large language models (LLMs) often contain misleading content, emphasizing the need to align them with human values to ensure secur AI systems. Reinforcement learning from human feedback (RLHF) has been employed to achieve this alignment by combining a reward model, typically based on Bradley-Terry paired comparison, w...
http://arxiv.org/pdf/2306.17492v1
2023-06-30T09:07:37
2023-06-30T09:07:37
{ "title": [ -0.015571594, 0.015640259, -0.03289795, -0.029937744, 0.00843811, -0.011100769, -0.020141602, -0.030532837, -0.0051651, 0.054595947, -0.005870819, -0.0070495605, 0.026062012, -0.017333984, -0.04925537, 0.0042877197, 0.0012617111, -0....
http://arxiv.org/abs/2307.01211v1
An automated method for the ontological representation of security directives
Giampaolo Bella, Gianpietro Castiglione, Daniele Francesco Santamaria
cs.AI, cs.CL
Large documents written in juridical language are difficult to interpret, with long sentences leading to intricate and intertwined relations between the nouns. The present paper frames this problem in the context of recent European security directives. The complexity of their language is here thwarted by automating the...
http://arxiv.org/pdf/2307.01211v1
2023-06-30T09:04:47
2023-06-30T09:04:47
{ "title": [ -0.006336212, 0.012893677, -0.051818848, -0.07501221, 0.020385742, -0.009811401, -0.038024902, -0.014259338, -0.018203735, 0.035308838, -0.010414124, -0.009925842, 0.023239136, -0.031555176, -0.019134521, 0.009162903, -0.010757446, 0...
http://arxiv.org/abs/2306.17472v1
Knowledge Base Completion for Long-Tail Entities
Lihu Chen, Simon Razniewski, Gerhard Weikum
cs.CL
Despite their impressive scale, knowledge bases (KBs), such as Wikidata, still contain significant gaps. Language models (LMs) have been proposed as a source for filling these gaps. However, prior works have focused on prominent entities with rich coverage by LMs, neglecting the crucial case of long-tail entities. In t...
http://arxiv.org/pdf/2306.17472v1
2023-06-30T08:37:55
2023-06-30T08:37:55
{ "title": [ 0.00003552437, 0.010513306, -0.049743652, -0.10266113, 0.031463623, 0.0131073, -0.04046631, 0.015556335, 0.014816284, 0.023086548, -0.036590576, 0.005207062, 0.006111145, -0.010345459, -0.031951904, -0.0004916191, -0.0021457672, 0.02...
http://arxiv.org/abs/2306.17459v1
Harnessing LLMs in Curricular Design: Using GPT-4 to Support Authoring of Learning Objectives
Pragnya Sridhar, Aidan Doyle, Arav Agarwal, Christopher Bogart, Jaromir Savelka, Majd Sakr
cs.AI, cs.CL
We evaluated the capability of a generative pre-trained transformer (GPT-4) to automatically generate high-quality learning objectives (LOs) in the context of a practically oriented university course on Artificial Intelligence. Discussions of opportunities (e.g., content generation, explanation) and risks (e.g., cheati...
http://arxiv.org/pdf/2306.17459v1
2023-06-30T08:15:18
2023-06-30T08:15:18
{ "title": [ -0.039886475, -0.0012750626, -0.05038452, -0.0018863678, -0.026901245, -0.023727417, -0.030960083, -0.023254395, -0.026931763, -0.0023860931, -0.020309448, 0.0063323975, -0.039276123, -0.041748047, -0.0068206787, -0.031341553, -0.0382385...
http://arxiv.org/abs/2306.17447v2
Progressive Multi-task Learning Framework for Chinese Text Error Correction
Shirong Ma, Yinghui Li, Haojing Huang, Shulin Huang, Yangning Li, Hai-Tao Zheng, Ying Shen
cs.CL
Chinese Text Error Correction (CTEC) aims to detect and correct errors in the input text, which benefits human's daily life and various downstream tasks. Recent approaches mainly employ Pre-trained Language Models (PLMs) to resolve CTEC task and achieve tremendous success. However, previous approaches suffer from issue...
http://arxiv.org/pdf/2306.17447v2
2023-07-03T05:29:59
2023-06-30T07:44:49
{ "title": [ -0.008468628, 0.017120361, -0.045043945, -0.05630493, 0.008071899, -0.012016296, -0.03753662, 0.0043258667, 0.008575439, 0.05831909, -0.011268616, -0.0050315857, 0.021438599, -0.009674072, -0.06286621, -0.0009794235, -0.019699097, 0....
http://arxiv.org/abs/2306.17439v2
Provable Robust Watermarking for AI-Generated Text
Xuandong Zhao, Prabhanjan Ananth, Lei Li, Yu-Xiang Wang
cs.CL, cs.LG
We study the problem of watermarking large language models (LLMs) generated text -- one of the most promising approaches for addressing the safety challenges of LLM usage. In this paper, we propose a rigorous theoretical framework to quantify the effectiveness and robustness of LLM watermarks. We propose a robust and h...
http://arxiv.org/pdf/2306.17439v2
2023-10-13T04:50:04
2023-06-30T07:24:32
{ "title": [ -0.0284729, 0.017456055, -0.020935059, -0.038909912, 0.020965576, 0.0029067993, -0.012550354, -0.022277832, -0.010856628, 0.03640747, 0.0024337769, 0.02142334, 0.015098572, -0.025314331, -0.022827148, 0.013137817, -0.017730713, 0.021...
http://arxiv.org/abs/2306.17408v2
LMBot: Distilling Graph Knowledge into Language Model for Graph-less Deployment in Twitter Bot Detection
Zijian Cai, Zhaoxuan Tan, Zhenyu Lei, Zifeng Zhu, Hongrui Wang, Qinghua Zheng, Minnan Luo
cs.AI, cs.CL, cs.SI
As malicious actors employ increasingly advanced and widespread bots to disseminate misinformation and manipulate public opinion, the detection of Twitter bots has become a crucial task. Though graph-based Twitter bot detection methods achieve state-of-the-art performance, we find that their inference depends on the ne...
http://arxiv.org/pdf/2306.17408v2
2023-07-03T11:22:09
2023-06-30T05:50:26
{ "title": [ 0.009613037, -0.018966675, -0.019577026, -0.042541504, 0.00082206726, -0.0054893494, 0.00013101101, -0.00074863434, -0.018112183, 0.03451538, -0.049194336, 0.019882202, 0.006198883, -0.0053367615, -0.032318115, -0.018859863, -0.005886078...
http://arxiv.org/abs/2306.17399v1
Japanese Lexical Complexity for Non-Native Readers: A New Dataset
Yusuke Ide, Masato Mita, Adam Nohejl, Hiroki Ouchi, Taro Watanabe
cs.CL
Lexical complexity prediction (LCP) is the task of predicting the complexity of words in a text on a continuous scale. It plays a vital role in simplifying or annotating complex words to assist readers. To study lexical complexity in Japanese, we construct the first Japanese LCP dataset. Our dataset provides separate c...
http://arxiv.org/pdf/2306.17399v1
2023-06-30T04:37:43
2023-06-30T04:37:43
{ "title": [ -0.03353882, 0.005252838, -0.01689148, -0.029632568, -0.007774353, -0.0055122375, -0.045715332, -0.013305664, -0.0017690659, 0.035980225, -0.02923584, 0.016830444, -0.005214691, 0.0051460266, -0.07836914, -0.014541626, -0.010528564, ...
http://arxiv.org/abs/2306.17384v1
SummQA at MEDIQA-Chat 2023:In-Context Learning with GPT-4 for Medical Summarization
Yash Mathur, Sanketh Rangreji, Raghav Kapoor, Medha Palavalli, Amanda Bertsch, Matthew R. Gormley
cs.CL
Medical dialogue summarization is challenging due to the unstructured nature of medical conversations, the use of medical terminology in gold summaries, and the need to identify key information across multiple symptom sets. We present a novel system for the Dialogue2Note Medical Summarization tasks in the MEDIQA 2023 S...
http://arxiv.org/pdf/2306.17384v1
2023-06-30T03:14:04
2023-06-30T03:14:04
{ "title": [ 0.010848999, 0.013191223, -0.026321411, -0.055999756, 0.0022773743, -0.00006484985, -0.025756836, -0.020751953, -0.023849487, 0.057495117, 0.0009560585, 0.0064582825, 0.020584106, -0.034851074, -0.030227661, 0.009666443, 0.007068634, ...
http://arxiv.org/abs/2307.01209v1
Multi-Dialectal Representation Learning of Sinitic Phonology
Zhibai Jia
cs.CL, cs.LG, cs.SI
Machine learning techniques have shown their competence for representing and reasoning in symbolic systems such as language and phonology. In Sinitic Historical Phonology, notable tasks that could benefit from machine learning include the comparison of dialects and reconstruction of proto-languages systems. Motivated b...
http://arxiv.org/pdf/2307.01209v1
2023-06-30T02:37:25
2023-06-30T02:37:25
{ "title": [ 0.013542175, 0.024749756, -0.031341553, -0.031341553, -0.011932373, -0.02166748, -0.024765015, 0.0043792725, -0.014526367, 0.054473877, -0.021347046, 0.016983032, 0.011238098, -0.031951904, -0.035705566, -0.00045204163, -0.0047035217, ...
http://arxiv.org/abs/2306.17322v1
Citations as Queries: Source Attribution Using Language Models as Rerankers
Ryan Muther, David Smith
cs.CL
This paper explores new methods for locating the sources used to write a text, by fine-tuning a variety of language models to rerank candidate sources. After retrieving candidates sources using a baseline BM25 retrieval model, a variety of reranking methods are tested to see how effective they are at the task of source...
http://arxiv.org/pdf/2306.17322v1
2023-06-29T22:13:38
2023-06-29T22:13:38
{ "title": [ -0.011482239, -0.0054473877, -0.020996094, -0.043151855, 0.0061035156, 0.005844116, -0.039093018, -0.04284668, -0.009094238, 0.030319214, -0.020111084, 0.008522034, 0.006072998, -0.0061149597, -0.03515625, -0.004207611, -0.016067505, ...
http://arxiv.org/abs/2306.17290v1
Towards Open-Domain Topic Classification
Hantian Ding, Jinrui Yang, Yuqian Deng, Hongming Zhang, Dan Roth
cs.CL
We introduce an open-domain topic classification system that accepts user-defined taxonomy in real time. Users will be able to classify a text snippet with respect to any candidate labels they want, and get instant response from our web interface. To obtain such flexibility, we build the backend model in a zero-shot wa...
http://arxiv.org/pdf/2306.17290v1
2023-06-29T20:25:28
2023-06-29T20:25:28
{ "title": [ -0.006526947, 0.022781372, -0.041412354, -0.03451538, 0.023849487, -0.01398468, -0.014007568, -0.047180176, 0.009803772, 0.01574707, -0.010826111, 0.020141602, -0.031280518, -0.019088745, -0.028167725, 0.0017518997, -0.015068054, 0.0...
http://arxiv.org/abs/2306.17257v1
Prediction of COVID-19 Patients' Emergency Room Revisit using Multi-Source Transfer Learning
Yuelyu Ji, Yuhe Gao, Runxue Bao, Qi Li, Disheng Liu, Yiming Sun, Ye Ye
cs.LG, cs.AI, cs.CL
The coronavirus disease 2019 (COVID-19) has led to a global pandemic of significant severity. In addition to its high level of contagiousness, COVID-19 can have a heterogeneous clinical course, ranging from asymptomatic carriers to severe and potentially life-threatening health complications. Many patients have to revi...
http://arxiv.org/pdf/2306.17257v1
2023-06-29T18:51:42
2023-06-29T18:51:42
{ "title": [ 0.027694702, 0.019821167, -0.038970947, -0.046722412, 0.015129089, -0.026672363, 0.004940033, -0.022964478, -0.029052734, 0.042785645, -0.018341064, 0.028182983, -0.00045871735, -0.022872925, -0.012550354, -0.008613586, 0.016693115, ...
http://arxiv.org/abs/2306.17256v3
Towards Personalized Cold-Start Recommendation with Prompts
Xuansheng Wu, Huachi Zhou, Yucheng Shi, Wenlin Yao, Xiao Huang, Ninghao Liu
cs.IR, cs.AI, cs.CL, cs.SI
Recommender systems play a crucial role in helping users discover information that aligns with their interests based on their past behaviors. However, developing personalized recommendation systems becomes challenging when historical records of user-item interactions are unavailable, leading to what is known as the sys...
http://arxiv.org/pdf/2306.17256v3
2023-10-30T20:03:22
2023-06-29T18:50:12
{ "title": [ 0.0037498474, 0.011581421, -0.031280518, -0.03942871, 0.019104004, -0.0008916855, -0.015289307, -0.02053833, -0.03060913, 0.03338623, -0.027297974, 0.040100098, -0.022567749, -0.0098724365, -0.030075073, 0.0028495789, -0.0028190613, ...
http://arxiv.org/abs/2306.17156v3
Generative AI for Programming Education: Benchmarking ChatGPT, GPT-4, and Human Tutors
Tung Phung, Victor-Alexandru Pădurean, José Cambronero, Sumit Gulwani, Tobias Kohn, Rupak Majumdar, Adish Singla, Gustavo Soares
cs.CY, cs.AI, cs.CL
Generative AI and large language models hold great promise in enhancing computing education by powering next-generation educational technologies for introductory programming. Recent works have studied these models for different scenarios relevant to programming education; however, these works are limited for several re...
http://arxiv.org/pdf/2306.17156v3
2023-08-01T00:03:25
2023-06-29T17:57:40
{ "title": [ -0.02583313, -0.022872925, -0.0029773712, -0.008872986, -0.027313232, -0.011741638, 0.0025844574, -0.019805908, 0.01133728, 0.019683838, -0.028244019, 0.029647827, -0.019699097, -0.016662598, -0.03152466, -0.003068924, -0.020996094, ...
http://arxiv.org/abs/2306.17107v1
LLaVAR: Enhanced Visual Instruction Tuning for Text-Rich Image Understanding
Yanzhe Zhang, Ruiyi Zhang, Jiuxiang Gu, Yufan Zhou, Nedim Lipka, Diyi Yang, Tong Sun
cs.CV, cs.CL
Instruction tuning unlocks the superior capability of Large Language Models (LLM) to interact with humans. Furthermore, recent instruction-following datasets include images as visual inputs, collecting responses for image-based instructions. However, visual instruction-tuned models cannot comprehend textual details wit...
http://arxiv.org/pdf/2306.17107v1
2023-06-29T17:08:16
2023-06-29T17:08:16
{ "title": [ -0.029907227, -0.0036067963, -0.00049972534, -0.032104492, -0.014274597, -0.0007414818, -0.020645142, -0.03286743, 0.010093689, 0.0070114136, -0.04812622, 0.027511597, -0.025131226, -0.0019550323, -0.02406311, -0.015686035, 0.003353119, ...
http://arxiv.org/abs/2306.17103v2
LyricWhiz: Robust Multilingual Zero-shot Lyrics Transcription by Whispering to ChatGPT
Le Zhuo, Ruibin Yuan, Jiahao Pan, Yinghao Ma, Yizhi LI, Ge Zhang, Si Liu, Roger Dannenberg, Jie Fu, Chenghua Lin, Emmanouil Benetos, Wenhu Chen, Wei Xue, Yike Guo
cs.CL, cs.SD, eess.AS
We introduce LyricWhiz, a robust, multilingual, and zero-shot automatic lyrics transcription method achieving state-of-the-art performance on various lyrics transcription datasets, even in challenging genres such as rock and metal. Our novel, training-free approach utilizes Whisper, a weakly supervised robust speech re...
http://arxiv.org/pdf/2306.17103v2
2023-07-07T16:32:26
2023-06-29T17:01:51
{ "title": [ -0.0340271, -0.047821045, -0.033050537, -0.020889282, -0.061431885, -0.007019043, -0.044128418, 0.034606934, 0.0023422241, 0.046691895, -0.03253174, 0.023101807, 0.009864807, -0.0020942688, -0.028671265, 0.008956909, -0.010635376, 0....
http://arxiv.org/abs/2306.17089v2
Concept-Oriented Deep Learning with Large Language Models
Daniel T. Chang
cs.LG, cs.CL
Large Language Models (LLMs) have been successfully used in many natural-language tasks and applications including text generation and AI chatbots. They also are a promising new technology for concept-oriented deep learning (CODL). However, the prerequisite is that LLMs understand concepts and ensure conceptual consist...
http://arxiv.org/pdf/2306.17089v2
2023-09-19T21:15:52
2023-06-29T16:47:11
{ "title": [ -0.0066375732, -0.0131073, -0.03555298, -0.06738281, 0.008262634, -0.009094238, -0.038269043, 0.0045280457, 0.018615723, 0.028945923, -0.0062294006, -0.0027179718, 0.004787445, -0.016601562, -0.032958984, 0.017196655, -0.010437012, 0...
http://arxiv.org/abs/2306.17059v2
The mapKurator System: A Complete Pipeline for Extracting and Linking Text from Historical Maps
Jina Kim, Zekun Li, Yijun Lin, Min Namgung, Leeje Jang, Yao-Yi Chiang
cs.AI, cs.CL, cs.CV
Scanned historical maps in libraries and archives are valuable repositories of geographic data that often do not exist elsewhere. Despite the potential of machine learning tools like the Google Vision APIs for automatically transcribing text from these maps into machine-readable formats, they do not work well with larg...
http://arxiv.org/pdf/2306.17059v2
2023-07-03T19:38:37
2023-06-29T16:05:40
{ "title": [ -0.015960693, 0.012550354, -0.023025513, -0.010398865, 0.009422302, -0.014907837, -0.00018894672, -0.05709839, 0.038848877, 0.026733398, -0.021209717, 0.0076065063, -0.00655365, -0.014572144, -0.06286621, 0.009002686, 0.017623901, 0....
http://arxiv.org/abs/2306.17042v1
Towards Grammatical Tagging for the Legal Language of Cybersecurity
Gianpietro Castiglione, Giampaolo Bella, Daniele Francesco Santamaria
cs.CL, cs.CY
Legal language can be understood as the language typically used by those engaged in the legal profession and, as such, it may come both in spoken or written form. Recent legislation on cybersecurity obviously uses legal language in writing, thus inheriting all its interpretative complications due to the typical abundan...
http://arxiv.org/pdf/2306.17042v1
2023-06-29T15:39:20
2023-06-29T15:39:20
{ "title": [ -0.017974854, -0.0066375732, -0.0541687, -0.049743652, 0.005844116, -0.009048462, -0.02748108, 0.011726379, -0.046081543, 0.029586792, -0.025634766, 0.02406311, 0.0046844482, -0.01739502, -0.009803772, -0.010383606, -0.020324707, 0.0...
http://arxiv.org/abs/2306.17034v1
Exploring & Exploiting High-Order Graph Structure for Sparse Knowledge Graph Completion
Tao He, Ming Liu, Yixin Cao, Zekun Wang, Zihao Zheng, Zheng Chu, Bing Qin
cs.AI, cs.CL
Sparse knowledge graph (KG) scenarios pose a challenge for previous Knowledge Graph Completion (KGC) methods, that is, the completion performance decreases rapidly with the increase of graph sparsity. This problem is also exacerbated because of the widespread existence of sparse KGs in practical applications. To allevi...
http://arxiv.org/pdf/2306.17034v1
2023-06-29T15:35:34
2023-06-29T15:35:34
{ "title": [ -0.04525757, -0.020507812, -0.08728027, -0.09887695, 0.009010315, 0.010551453, -0.07110596, 0.043823242, 0.022079468, 0.015991211, -0.021835327, 0.0128479, 0.0013751984, -0.019134521, -0.022644043, 0.010643005, 0.01184082, 0.01176452...
http://arxiv.org/abs/2306.17020v2
Classifying Crime Types using Judgment Documents from Social Media
Haoxuan Xu, Zeyu He, Mengfan Shen, Songning Lai, Ziqiang Han, Yifan Peng
cs.CL, cs.AI
The task of determining crime types based on criminal behavior facts has become a very important and meaningful task in social science. But the problem facing the field now is that the data samples themselves are unevenly distributed, due to the nature of the crime itself. At the same time, data sets in the judicial fi...
http://arxiv.org/pdf/2306.17020v2
2023-10-21T14:44:41
2023-06-29T15:12:24
{ "title": [ 0.0056114197, 0.0018148422, -0.029434204, -0.028564453, 0.0135269165, -0.025665283, 0.016952515, -0.005016327, -0.03463745, 0.017700195, -0.035064697, 0.027954102, -0.019760132, -0.018249512, 0.0016851425, -0.025177002, -0.015411377, ...
http://arxiv.org/abs/2306.17005v1
High-Quality Automatic Voice Over with Accurate Alignment: Supervision through Self-Supervised Discrete Speech Units
Junchen Lu, Berrak Sisman, Mingyang Zhang, Haizhou Li
eess.AS, cs.CL, cs.SD
The goal of Automatic Voice Over (AVO) is to generate speech in sync with a silent video given its text script. Recent AVO frameworks built upon text-to-speech synthesis (TTS) have shown impressive results. However, the current AVO learning objective of acoustic feature reconstruction brings in indirect supervision for...
http://arxiv.org/pdf/2306.17005v1
2023-06-29T15:02:22
2023-06-29T15:02:22
{ "title": [ 0.007820129, 0.009407043, -0.07745361, -0.080322266, 0.026672363, -0.046051025, -0.019317627, 0.019882202, -0.037322998, 0.06750488, -0.015174866, 0.015449524, 0.037384033, -0.020019531, -0.032592773, 0.008079529, -0.011726379, 0.033...
http://arxiv.org/abs/2306.16956v1
MEMD-ABSA: A Multi-Element Multi-Domain Dataset for Aspect-Based Sentiment Analysis
Hongjie Cai, Nan Song, Zengzhi Wang, Qiming Xie, Qiankun Zhao, Ke Li, Siwei Wu, Shijie Liu, Jianfei Yu, Rui Xia
cs.CL, cs.AI
Aspect-based sentiment analysis is a long-standing research interest in the field of opinion mining, and in recent years, researchers have gradually shifted their focus from simple ABSA subtasks to end-to-end multi-element ABSA tasks. However, the datasets currently used in the research are limited to individual elemen...
http://arxiv.org/pdf/2306.16956v1
2023-06-29T14:03:49
2023-06-29T14:03:49
{ "title": [ 0.0050735474, -0.014213562, -0.04083252, -0.06427002, 0.0138549805, -0.009155273, -0.024490356, 0.00894165, 0.026687622, 0.053009033, 0.0027980804, 0.0000016093254, -0.007045746, -0.008598328, -0.08270264, 0.02041626, -0.019470215, 0...
http://arxiv.org/abs/2306.16955v1
Predicting Music Hierarchies with a Graph-Based Neural Decoder
Francesco Foscarin, Daniel Harasim, Gerhard Widmer
cs.SD, cs.CL, eess.AS
This paper describes a data-driven framework to parse musical sequences into dependency trees, which are hierarchical structures used in music cognition research and music analysis. The parsing involves two steps. First, the input sequence is passed through a transformer encoder to enrich it with contextual information...
http://arxiv.org/pdf/2306.16955v1
2023-06-29T13:59:18
2023-06-29T13:59:18
{ "title": [ -0.006252289, -0.012123108, -0.042877197, -0.06958008, 0.0035114288, -0.010444641, -0.055633545, 0.002105713, -0.0085372925, 0.06341553, 0.016525269, -0.009666443, 0.043426514, -0.028823853, -0.054534912, 0.01878357, -0.0048217773, 0...
http://arxiv.org/abs/2306.16931v1
UMASS_BioNLP at MEDIQA-Chat 2023: Can LLMs generate high-quality synthetic note-oriented doctor-patient conversations?
Junda Wang, Zonghai Yao, Avijit Mitra, Samuel Osebe, Zhichao Yang, Hong Yu
cs.CL, cs.AI
This paper presents UMASS_BioNLP team participation in the MEDIQA-Chat 2023 shared task for Task-A and Task-C. We focus especially on Task-C and propose a novel LLMs cooperation system named a doctor-patient loop to generate high-quality conversation data sets. The experiment results demonstrate that our approaches yie...
http://arxiv.org/pdf/2306.16931v1
2023-06-29T13:30:41
2023-06-29T13:30:41
{ "title": [ -0.040740967, -0.029190063, 0.0032024384, -0.015602112, -0.05392456, 0.02947998, -0.054748535, 0.017974854, -0.023712158, 0.009666443, -0.0077667236, 0.01499939, -0.004878998, -0.045166016, 0.020065308, -0.037475586, 0.017059326, 0.0...
http://arxiv.org/abs/2306.16903v1
Leveraging Cross-Utterance Context For ASR Decoding
Robert Flynn, Anton Ragni
cs.CL, cs.SD, eess.AS
While external language models (LMs) are often incorporated into the decoding stage of automated speech recognition systems, these models usually operate with limited context. Cross utterance information has been shown to be beneficial during second pass re-scoring, however this limits the hypothesis space based on the...
http://arxiv.org/pdf/2306.16903v1
2023-06-29T12:48:25
2023-06-29T12:48:25
{ "title": [ -0.016921997, -0.0056495667, -0.05593872, -0.082458496, 0.009857178, -0.005722046, -0.047576904, 0.022460938, -0.00049495697, 0.0569458, -0.028335571, 0.01499176, 0.039154053, -0.0029392242, -0.033813477, 0.027679443, -0.002916336, 0...
http://arxiv.org/abs/2306.16900v2
Surveying (Dis)Parities and Concerns of Compute Hungry NLP Research
Ji-Ung Lee, Haritz Puerto, Betty van Aken, Yuki Arase, Jessica Zosa Forde, Leon Derczynski, Andreas Rücklé, Iryna Gurevych, Roy Schwartz, Emma Strubell, Jesse Dodge
cs.CL
Many recent improvements in NLP stem from the development and use of large pre-trained language models (PLMs) with billions of parameters. Large model sizes makes computational cost one of the main limiting factors for training and evaluating such models; and has raised severe concerns about the sustainability, reprodu...
http://arxiv.org/pdf/2306.16900v2
2023-11-09T13:52:45
2023-06-29T12:44:53
{ "title": [ -0.011764526, -0.029159546, -0.00919342, -0.036743164, -0.024810791, -0.01020813, -0.001991272, 0.001326561, -0.012893677, 0.011932373, -0.073913574, 0.0140686035, -0.0035972595, -0.006526947, -0.024353027, -0.021102905, -0.016601562, ...
http://arxiv.org/abs/2306.16842v1
Tokenization and the Noiseless Channel
Vilém Zouhar, Clara Meister, Juan Luis Gastaldi, Li Du, Mrinmaya Sachan, Ryan Cotterell
cs.CL, cs.IT, math.IT
Subword tokenization is a key part of many NLP pipelines. However, little is known about why some tokenizer and hyperparameter combinations lead to better downstream model performance than others. We propose that good tokenizers lead to \emph{efficient} channel usage, where the channel is the means by which some input ...
http://arxiv.org/pdf/2306.16842v1
2023-06-29T10:32:09
2023-06-29T10:32:09
{ "title": [ -0.037353516, 0.00374794, -0.042144775, -0.03363037, -0.0017929077, 0.002298355, -0.013710022, -0.01108551, -0.000052809715, 0.051849365, -0.024978638, -0.012306213, 0.03074646, -0.033294678, -0.012588501, 0.022750854, -0.005252838, ...
http://arxiv.org/abs/2306.16837v1
A Formal Perspective on Byte-Pair Encoding
Vilém Zouhar, Clara Meister, Juan Luis Gastaldi, Li Du, Tim Vieira, Mrinmaya Sachan, Ryan Cotterell
cs.CL, math.OC
Byte-Pair Encoding (BPE) is a popular algorithm used for tokenizing data in NLP, despite being devised initially as a compression method. BPE appears to be a greedy algorithm at face value, but the underlying optimization problem that BPE seeks to solve has not yet been laid down. We formalize BPE as a combinatorial op...
http://arxiv.org/pdf/2306.16837v1
2023-06-29T10:29:23
2023-06-29T10:29:23
{ "title": [ -0.044952393, -0.0023498535, -0.040771484, -0.075927734, 0.010139465, 0.0137786865, -0.050598145, 0.020202637, 0.0059280396, 0.043548584, 0.011695862, -0.016662598, 0.005710602, -0.02557373, -0.044830322, 0.022567749, -0.0057907104, ...
http://arxiv.org/abs/2306.16805v2
CLIPAG: Towards Generator-Free Text-to-Image Generation
Roy Ganz, Michael Elad
cs.CV, cs.CL, cs.LG
Perceptually Aligned Gradients (PAG) refer to an intriguing property observed in robust image classification models, wherein their input gradients align with human perception and pose semantic meanings. While this phenomenon has gained significant research attention, it was solely studied in the context of unimodal vis...
http://arxiv.org/pdf/2306.16805v2
2023-09-01T12:53:44
2023-06-29T09:35:53
{ "title": [ -0.00894928, -0.0028476715, -0.02571106, -0.0519104, 0.0044822693, 0.0049972534, -0.025939941, -0.027893066, 0.018417358, 0.050567627, -0.017028809, 0.021469116, -0.001821518, -0.001458168, -0.052734375, 0.0103302, 0.0021896362, 0.01...
http://arxiv.org/abs/2306.16793v1
Benchmarking Large Language Model Capabilities for Conditional Generation
Joshua Maynez, Priyanka Agrawal, Sebastian Gehrmann
cs.CL
Pre-trained large language models (PLMs) underlie most new developments in natural language processing. They have shifted the field from application-specific model pipelines to a single model that is adapted to a wide range of tasks. Autoregressive PLMs like GPT-3 or PaLM, alongside techniques like few-shot learning, h...
http://arxiv.org/pdf/2306.16793v1
2023-06-29T08:59:40
2023-06-29T08:59:40
{ "title": [ -0.04586792, -0.011764526, -0.06097412, -0.11651611, 0.009048462, 0.02758789, -0.11798096, 0.030166626, 0.01348114, 0.025756836, -0.02784729, 0.00096940994, 0.0036334991, -0.019119263, -0.04638672, 0.0128479, -0.0032444, 0.031234741,...
http://arxiv.org/abs/2306.16774v1
Stop Pre-Training: Adapt Visual-Language Models to Unseen Languages
Yasmine Karoui, Rémi Lebret, Negar Foroutan, Karl Aberer
cs.CL
Vision-Language Pre-training (VLP) has advanced the performance of many vision-language tasks, such as image-text retrieval, visual entailment, and visual reasoning. The pre-training mostly utilizes lexical databases and image queries in English. Previous work has demonstrated that the pre-training in English does not ...
http://arxiv.org/pdf/2306.16774v1
2023-06-29T08:20:57
2023-06-29T08:20:57
{ "title": [ -0.0059509277, -0.008804321, -0.013465881, -0.03616333, -0.037902832, -0.006023407, -0.050750732, -0.035491943, 0.00919342, 0.009796143, -0.032196045, -0.013168335, -0.009857178, -0.029876709, -0.020950317, -0.025177002, 0.023742676, ...
http://arxiv.org/abs/2306.16770v1
DialoGPS: Dialogue Path Sampling in Continuous Semantic Space for Data Augmentation in Multi-Turn Conversations
Ang Lv, Jinpeng Li, Yuhan Chen, Xing Gao, Ji Zhang, Rui Yan
cs.CL, cs.AI
In open-domain dialogue generation tasks, contexts and responses in most datasets are one-to-one mapped, violating an important many-to-many characteristic: a context leads to various responses, and a response answers multiple contexts. Without such patterns, models poorly generalize and prefer responding safely. Many ...
http://arxiv.org/pdf/2306.16770v1
2023-06-29T08:12:47
2023-06-29T08:12:47
{ "title": [ 0.004043579, -0.011161804, -0.06335449, -0.07678223, 0.0076789856, -0.024124146, -0.03060913, 0.027450562, 0.022750854, 0.019073486, -0.045776367, 0.032043457, -0.012084961, -0.0030918121, -0.047943115, -0.012039185, -0.017425537, 0....
http://arxiv.org/abs/2306.17199v1
Learning Multilingual Expressive Speech Representation for Prosody Prediction without Parallel Data
Jarod Duret, Titouan Parcollet, Yannick Estève
eess.AS, cs.CL, cs.SD
We propose a method for speech-to-speech emotionpreserving translation that operates at the level of discrete speech units. Our approach relies on the use of multilingual emotion embedding that can capture affective information in a language-independent manner. We show that this embedding can be used to predict the pit...
http://arxiv.org/pdf/2306.17199v1
2023-06-29T08:06:54
2023-06-29T08:06:54
{ "title": [ -0.00730896, 0.004425049, -0.040161133, -0.09136963, -0.009536743, 0.011474609, -0.079833984, 0.0362854, -0.0052452087, 0.048950195, 0.02319336, -0.019165039, 0.015525818, -0.033111572, -0.046844482, 0.006931305, -0.0044441223, 0.051...
http://arxiv.org/abs/2306.16762v1
Unified Language Representation for Question Answering over Text, Tables, and Images
Bowen Yu, Cheng Fu, Haiyang Yu, Fei Huang, Yongbin Li
cs.CL
When trying to answer complex questions, people often rely on multiple sources of information, such as visual, textual, and tabular data. Previous approaches to this problem have focused on designing input features or model structure in the multi-modal space, which is inflexible for cross-modal reasoning or data-effici...
http://arxiv.org/pdf/2306.16762v1
2023-06-29T08:02:23
2023-06-29T08:02:23
{ "title": [ -0.0045700073, -0.008338928, -0.019882202, -0.06188965, 0.007419586, 0.0030994415, -0.036254883, -0.006000519, 0.018295288, 0.044403076, -0.02432251, 0.012588501, 0.016159058, 0.0033569336, -0.048614502, 0.013320923, 0.0066223145, 0....
http://arxiv.org/abs/2306.16722v1
Evaluating Paraphrastic Robustness in Textual Entailment Models
Dhruv Verma, Yash Kumar Lal, Shreyashee Sinha, Benjamin Van Durme, Adam Poliak
cs.CL, cs.AI
We present PaRTE, a collection of 1,126 pairs of Recognizing Textual Entailment (RTE) examples to evaluate whether models are robust to paraphrasing. We posit that if RTE models understand language, their predictions should be consistent across inputs that share the same meaning. We use the evaluation set to determine ...
http://arxiv.org/pdf/2306.16722v1
2023-06-29T06:48:22
2023-06-29T06:48:22
{ "title": [ -0.0042495728, -0.00034093857, -0.022918701, -0.07434082, 0.0072021484, 0.012191772, -0.050201416, -0.012954712, 0.006881714, 0.045562744, -0.023773193, 0.012321472, 0.025939941, 0.00086688995, -0.05331421, 0.026275635, -0.031677246, ...
http://arxiv.org/abs/2306.16710v1
Automatic Speech Recognition of Non-Native Child Speech for Language Learning Applications
Simone Wills, Yu Bai, Cristian Tejedor-Garcia, Catia Cucchiarini, Helmer Strik
cs.CL, cs.SD, eess.AS
Voicebots have provided a new avenue for supporting the development of language skills, particularly within the context of second language learning. Voicebots, though, have largely been geared towards native adult speakers. We sought to assess the performance of two state-of-the-art ASR systems, Wav2Vec2.0 and Whisper ...
http://arxiv.org/pdf/2306.16710v1
2023-06-29T06:14:26
2023-06-29T06:14:26
{ "title": [ -0.026565552, -0.0022945404, -0.032440186, -0.045135498, -0.006729126, -0.00028848648, -0.03237915, 0.011940002, -0.029006958, 0.052581787, -0.037109375, 0.0030937195, 0.0034179688, -0.009895325, -0.048095703, -0.020446777, -0.011932373,...
http://arxiv.org/abs/2306.16650v1
Multi-source Semantic Graph-based Multimodal Sarcasm Explanation Generation
Liqiang Jing, Xuemeng Song, Kun Ouyang, Mengzhao Jia, Liqiang Nie
cs.CL, cs.AI
Multimodal Sarcasm Explanation (MuSE) is a new yet challenging task, which aims to generate a natural language sentence for a multimodal social post (an image as well as its caption) to explain why it contains sarcasm. Although the existing pioneer study has achieved great success with the BART backbone, it overlooks t...
http://arxiv.org/pdf/2306.16650v1
2023-06-29T03:26:10
2023-06-29T03:26:10
{ "title": [ -0.023544312, -0.012039185, -0.036956787, -0.066467285, -0.03479004, 0.0064888, -0.06542969, 0.003501892, -0.0015459061, 0.017791748, -0.0003798008, -0.017791748, 0.01751709, -0.03567505, -0.027420044, -0.0068206787, -0.026290894, 0....
http://arxiv.org/abs/2306.16649v1
ZeroGen: Zero-shot Multimodal Controllable Text Generation with Multiple Oracles
Haoqin Tu, Bowen Yang, Xianfeng Zhao
cs.CL
Automatically generating textual content with desired attributes is an ambitious task that people have pursued long. Existing works have made a series of progress in incorporating unimodal controls into language models (LMs), whereas how to generate controllable sentences with multimodal signals and high efficiency rem...
http://arxiv.org/pdf/2306.16649v1
2023-06-29T03:22:43
2023-06-29T03:22:43
{ "title": [ -0.03062439, -0.01486969, -0.053466797, -0.06274414, -0.016784668, 0.0077438354, -0.05441284, 0.012832642, 0.013008118, 0.047454834, -0.0065841675, 0.014175415, 0.02230835, -0.008621216, -0.047668457, 0.02180481, -0.021240234, 0.0338...
http://arxiv.org/abs/2306.16644v2
Probabilistic Linguistic Knowledge and Token-level Text Augmentation
Zhengxiang Wang
cs.CL
This paper investigates the effectiveness of token-level text augmentation and the role of probabilistic linguistic knowledge within a linguistically-motivated evaluation context. Two text augmentation programs, REDA and REDA$_{NG}$, were developed, both implementing five token-level text editing operations: Synonym Re...
http://arxiv.org/pdf/2306.16644v2
2023-07-03T05:01:12
2023-06-29T03:02:04
{ "title": [ -0.0052261353, -0.014472961, -0.033813477, -0.061065674, -0.0022621155, 0.0011959076, -0.03250122, 0.00856781, 0.004447937, 0.013694763, -0.024932861, 0.02420044, 0.0019454956, -0.007896423, -0.02734375, 0.009117126, -0.008857727, 0....
http://arxiv.org/abs/2306.16638v1
A negation detection assessment of GPTs: analysis with the xNot360 dataset
Ha Thanh Nguyen, Randy Goebel, Francesca Toni, Kostas Stathis, Ken Satoh
cs.CL
Negation is a fundamental aspect of natural language, playing a critical role in communication and comprehension. Our study assesses the negation detection performance of Generative Pre-trained Transformer (GPT) models, specifically GPT-2, GPT-3, GPT-3.5, and GPT-4. We focus on the identification of negation in natural...
http://arxiv.org/pdf/2306.16638v1
2023-06-29T02:27:48
2023-06-29T02:27:48
{ "title": [ 0.04849243, 0.014801025, 0.0041618347, -0.053527832, 0.015586853, -0.015510559, -0.028961182, -0.047546387, -0.0069084167, 0.02154541, -0.054382324, -0.018859863, 0.023345947, -0.024353027, -0.019989014, -0.021087646, -0.022872925, 0...
http://arxiv.org/abs/2306.16636v1
CMATH: Can Your Language Model Pass Chinese Elementary School Math Test?
Tianwen Wei, Jian Luan, Wei Liu, Shuang Dong, Bin Wang
cs.CL, cs.AI, cs.LG
We present the Chinese Elementary School Math Word Problems (CMATH) dataset, comprising 1.7k elementary school-level math word problems with detailed annotations, source from actual Chinese workbooks and exams. This dataset aims to provide a benchmark tool for assessing the following question: to what grade level of el...
http://arxiv.org/pdf/2306.16636v1
2023-06-29T02:19:50
2023-06-29T02:19:50
{ "title": [ -0.016799927, -0.012580872, -0.01209259, 0.024230957, -0.03552246, -0.011436462, 0.0029830933, -0.016983032, -0.03289795, 0.03503418, -0.018005371, 0.0006656647, 0.017410278, -0.037628174, 0.01737976, -0.040252686, -0.011871338, 0.01...
http://arxiv.org/abs/2307.00028v1
Seeing in Words: Learning to Classify through Language Bottlenecks
Khalid Saifullah, Yuxin Wen, Jonas Geiping, Micah Goldblum, Tom Goldstein
cs.CV, cs.AI, cs.CL, cs.LG
Neural networks for computer vision extract uninterpretable features despite achieving high accuracy on benchmarks. In contrast, humans can explain their predictions using succinct and intuitive descriptions. To incorporate explainability into neural networks, we train a vision model whose feature representations are t...
http://arxiv.org/pdf/2307.00028v1
2023-06-29T00:24:42
2023-06-29T00:24:42
{ "title": [ -0.024002075, -0.0041503906, -0.030715942, -0.033996582, 0.0096588135, -0.009849548, -0.016784668, -0.012840271, 0.009010315, 0.011856079, -0.043945312, 0.012832642, -0.020629883, -0.004711151, -0.0491333, -0.011367798, -0.005760193, ...
http://arxiv.org/abs/2306.16601v1
An Efficient Sparse Inference Software Accelerator for Transformer-based Language Models on CPUs
Haihao Shen, Hengyu Meng, Bo Dong, Zhe Wang, Ofir Zafrir, Yi Ding, Yu Luo, Hanwen Chang, Qun Gao, Ziheng Wang, Guy Boudoukh, Moshe Wasserblat
cs.LG, cs.AI, cs.CL
In recent years, Transformer-based language models have become the standard approach for natural language processing tasks. However, stringent throughput and latency requirements in industrial applications are limiting their adoption. To mitigate the gap, model compression techniques such as structured pruning are bein...
http://arxiv.org/pdf/2306.16601v1
2023-06-28T23:55:51
2023-06-28T23:55:51
{ "title": [ -0.05593872, -0.031463623, -0.030166626, -0.06665039, -0.021697998, 0.019439697, -0.0793457, 0.027633667, 0.0060043335, 0.031585693, -0.024963379, -0.0021972656, 0.04067993, -0.025527954, -0.026824951, 0.018341064, 0.0090408325, 0.01...
http://arxiv.org/abs/2306.16564v3
Automatic Calibration and Error Correction for Generative Large Language Models via Pareto Optimal Self-Supervision
Theodore Zhao, Mu Wei, J. Samuel Preston, Hoifung Poon
cs.CL, stat.ML
Generative Large language models (LLMs) have demonstrated remarkable capabilities for a wide range of applications, but reducing ungrounded or erroneous responses remains a major growth area. Unlike task-specific models, there lack an effective method to calibrate the confidence level of LLM responses to indicate poten...
http://arxiv.org/pdf/2306.16564v3
2023-10-26T18:27:00
2023-06-28T21:11:15
{ "title": [ -0.0340271, -0.007297516, -0.018127441, -0.05819702, -0.022735596, 0.0135650635, -0.0881958, -0.010444641, 0.01864624, 0.035064697, -0.028152466, 0.009155273, 0.010231018, -0.015602112, -0.052520752, 0.0065689087, -0.0074806213, 0.03...
http://arxiv.org/abs/2306.16533v1
ICSVR: Investigating Compositional and Semantic Understanding in Video Retrieval Models
Avinash Madasu, Vasudev Lal
cs.CV, cs.AI, cs.CL
Video retrieval (VR) involves retrieving the ground truth video from the video database given a text caption or vice-versa. The two important components of compositionality: objects \& attributes and actions are joined using correct semantics to form a proper text query. These components (objects \& attributes, actions...
http://arxiv.org/pdf/2306.16533v1
2023-06-28T20:06:36
2023-06-28T20:06:36
{ "title": [ 0.014778137, 0.014099121, -0.02609253, -0.05480957, 0.02029419, -0.010917664, -0.014442444, -0.022247314, 0.027175903, 0.038330078, -0.037872314, -0.009933472, -0.00001603365, -0.0013818741, -0.07727051, -0.0135269165, -0.010398865, ...
http://arxiv.org/abs/2306.16478v1
Pre-Training Multi-Modal Dense Retrievers for Outside-Knowledge Visual Question Answering
Alireza Salemi, Mahta Rafiee, Hamed Zamani
cs.IR, cs.CL, cs.CV
This paper studies a category of visual question answering tasks, in which accessing external knowledge is necessary for answering the questions. This category is called outside-knowledge visual question answering (OK-VQA). A major step in developing OK-VQA systems is to retrieve relevant documents for the given multi-...
http://arxiv.org/pdf/2306.16478v1
2023-06-28T18:06:40
2023-06-28T18:06:40
{ "title": [ 0.014144897, -0.0007648468, -0.0054244995, -0.066833496, -0.0036354065, 0.0064048767, -0.040802002, -0.0051574707, 0.026046753, 0.027069092, -0.03555298, 0.014762878, 0.007144928, -0.0054092407, -0.042907715, -0.0042533875, 0.027786255, ...
http://arxiv.org/abs/2306.16413v1
MultiZoo & MultiBench: A Standardized Toolkit for Multimodal Deep Learning
Paul Pu Liang, Yiwei Lyu, Xiang Fan, Arav Agarwal, Yun Cheng, Louis-Philippe Morency, Ruslan Salakhutdinov
cs.LG, cs.AI, cs.CL, cs.CV, cs.MM
Learning multimodal representations involves integrating information from multiple heterogeneous sources of data. In order to accelerate progress towards understudied modalities and tasks while ensuring real-world robustness, we release MultiZoo, a public toolkit consisting of standardized implementations of > 20 core ...
http://arxiv.org/pdf/2306.16413v1
2023-06-28T17:59:10
2023-06-28T17:59:10
{ "title": [ -0.024261475, 0.0060768127, -0.052093506, -0.06323242, 0.016983032, -0.006614685, -0.048431396, -0.0076026917, 0.008079529, 0.02255249, -0.011131287, 0.012969971, -0.013633728, -0.024841309, -0.026443481, 0.007194519, -0.0044059753, ...
http://arxiv.org/abs/2306.16410v1
Towards Language Models That Can See: Computer Vision Through the LENS of Natural Language
William Berrios, Gautam Mittal, Tristan Thrush, Douwe Kiela, Amanpreet Singh
cs.CL, cs.CV
We propose LENS, a modular approach for tackling computer vision problems by leveraging the power of large language models (LLMs). Our system uses a language model to reason over outputs from a set of independent and highly descriptive vision modules that provide exhaustive information about an image. We evaluate the a...
http://arxiv.org/pdf/2306.16410v1
2023-06-28T17:57:10
2023-06-28T17:57:10
{ "title": [ -0.042999268, -0.013450623, -0.023651123, -0.039215088, -0.009246826, -0.003944397, -0.035003662, -0.008605957, 0.022781372, 0.01914978, -0.051605225, -0.0093307495, 0.009109497, -0.0056037903, -0.058166504, -0.0063819885, -0.004261017, ...
http://arxiv.org/abs/2306.17194v2
On the Exploitability of Instruction Tuning
Manli Shu, Jiongxiao Wang, Chen Zhu, Jonas Geiping, Chaowei Xiao, Tom Goldstein
cs.CR, cs.CL, cs.LG
Instruction tuning is an effective technique to align large language models (LLMs) with human intents. In this work, we investigate how an adversary can exploit instruction tuning by injecting specific instruction-following examples into the training data that intentionally changes the model's behavior. For example, an...
http://arxiv.org/pdf/2306.17194v2
2023-10-28T18:04:36
2023-06-28T17:54:04
{ "title": [ -0.037719727, -0.0020580292, -0.0020713806, -0.047424316, -0.00598526, 0.013572693, -0.0149002075, 0.0030117035, -0.017364502, 0.029815674, -0.029785156, 0.009239197, -0.041625977, -0.0038013458, -0.04434204, -0.0073547363, 0.0037593842,...
http://arxiv.org/abs/2306.16388v1
Towards Measuring the Representation of Subjective Global Opinions in Language Models
Esin Durmus, Karina Nyugen, Thomas I. Liao, Nicholas Schiefer, Amanda Askell, Anton Bakhtin, Carol Chen, Zac Hatfield-Dodds, Danny Hernandez, Nicholas Joseph, Liane Lovitt, Sam McCandlish, Orowa Sikder, Alex Tamkin, Janel Thamkul, Jared Kaplan, Jack Clark, Deep Ganguli
cs.CL, cs.AI
Large language models (LLMs) may not equitably represent diverse global perspectives on societal issues. In this paper, we develop a quantitative framework to evaluate whose opinions model-generated responses are more similar to. We first build a dataset, GlobalOpinionQA, comprised of questions and answers from cross-n...
http://arxiv.org/pdf/2306.16388v1
2023-06-28T17:31:53
2023-06-28T17:31:53
{ "title": [ -0.024047852, -0.006450653, -0.035949707, -0.03857422, -0.007419586, -0.0029907227, -0.03665161, -0.003850937, 0.038757324, 0.030975342, -0.04171753, -0.014274597, 0.017730713, -0.010185242, -0.061676025, 0.010650635, -0.01424408, 0....
http://arxiv.org/abs/2306.16367v1
Multi-Site Clinical Federated Learning using Recursive and Attentive Models and NVFlare
Won Joon Yun, Samuel Kim, Joongheon Kim
cs.LG, cs.CL, cs.DC
The prodigious growth of digital health data has precipitated a mounting interest in harnessing machine learning methodologies, such as natural language processing (NLP), to scrutinize medical records, clinical notes, and other text-based health information. Although NLP techniques have exhibited substantial potential ...
http://arxiv.org/pdf/2306.16367v1
2023-06-28T17:00:32
2023-06-28T17:00:32
{ "title": [ 0.0061454773, 0.026535034, -0.053985596, -0.0869751, 0.03845215, -0.0074005127, -0.051483154, -0.0049858093, -0.037872314, 0.031433105, -0.026168823, 0.019821167, -0.016860962, -0.015197754, -0.006839752, -0.0178833, -0.009613037, 0....
http://arxiv.org/abs/2306.16326v1
Representation Learning via Variational Bayesian Networks
Oren Barkan, Avi Caciularu, Idan Rejwan, Ori Katz, Jonathan Weill, Itzik Malkiel, Noam Koenigstein
cs.LG, cs.AI, cs.CL
We present Variational Bayesian Network (VBN) - a novel Bayesian entity representation learning model that utilizes hierarchical and relational side information and is particularly useful for modeling entities in the ``long-tail'', where the data is scarce. VBN provides better modeling for long-tail entities via two co...
http://arxiv.org/pdf/2306.16326v1
2023-06-28T16:00:45
2023-06-28T16:00:45
{ "title": [ -0.0015821457, 0.009384155, -0.050201416, -0.07525635, 0.017532349, -0.008804321, -0.043121338, -0.008094788, 0.024475098, 0.016906738, -0.013656616, -0.0047454834, -0.028198242, -0.02758789, -0.05606079, -0.00047135353, 0.0078086853, ...
http://arxiv.org/abs/2306.16322v1
Taqyim: Evaluating Arabic NLP Tasks Using ChatGPT Models
Zaid Alyafeai, Maged S. Alshaibani, Badr AlKhamissi, Hamzah Luqman, Ebrahim Alareqi, Ali Fadel
cs.CL
Large language models (LLMs) have demonstrated impressive performance on various downstream tasks without requiring fine-tuning, including ChatGPT, a chat-based model built on top of LLMs such as GPT-3.5 and GPT-4. Despite having a lower training proportion compared to English, these models also exhibit remarkable capa...
http://arxiv.org/pdf/2306.16322v1
2023-06-28T15:54:29
2023-06-28T15:54:29
{ "title": [ -0.030761719, -0.028884888, -0.011238098, -0.017227173, -0.01890564, 0.008369446, -0.019058228, 0.002954483, -0.00957489, 0.010406494, -0.0077438354, 0.009712219, -0.033294678, -0.033599854, -0.04171753, -0.006351471, -0.018447876, 0...
http://arxiv.org/abs/2306.16313v1
An Adversarial Multi-Task Learning Method for Chinese Text Correction with Semantic Detection
Fanyu Wang, Zhenping Xie
cs.CL, cs.AI
Text correction, especially the semantic correction of more widely used scenes, is strongly required to improve, for the fluency and writing efficiency of the text. An adversarial multi-task learning method is proposed to enhance the modeling and detection ability of character polysemy in Chinese sentence context. Wher...
http://arxiv.org/pdf/2306.16313v1
2023-06-28T15:46:00
2023-06-28T15:46:00
{ "title": [ -0.028213501, 0.018051147, -0.024627686, -0.0059661865, 0.005176544, -0.0062217712, -0.02003479, -0.028945923, -0.009819031, 0.058654785, -0.006259918, 0.01259613, 0.013183594, -0.0074691772, -0.05239868, -0.0098724365, -0.025985718, ...
http://arxiv.org/abs/2306.16275v1
Leveraging GPT-4 for Food Effect Summarization to Enhance Product-Specific Guidance Development via Iterative Prompting
Yiwen Shi, Ping Ren, Jing Wang, Biao Han, Taha ValizadehAslani, Felix Agbavor, Yi Zhang, Meng Hu, Liang Zhao, Hualou Liang
cs.CL, cs.AI, cs.LG
Food effect summarization from New Drug Application (NDA) is an essential component of product-specific guidance (PSG) development and assessment. However, manual summarization of food effect from extensive drug application review documents is time-consuming, which arouses a need to develop automated methods. Recent ad...
http://arxiv.org/pdf/2306.16275v1
2023-06-28T14:55:13
2023-06-28T14:55:13
{ "title": [ 0.013160706, -0.0024261475, -0.02696228, -0.06524658, -0.012260437, -0.01777649, -0.048797607, -0.001958847, -0.007369995, 0.008369446, 0.0032901764, 0.025970459, -0.023895264, -0.03704834, -0.014083862, 0.008979797, 0.0015192032, 0....
http://arxiv.org/abs/2306.16268v1
Emotion Analysis of Tweets Banning Education in Afghanistan
Mohammad Ali Hussiny, Lilja Øvrelid
cs.CL, cs.AI
This paper introduces the first emotion annotated dataset for the Dari variant of Persian spoken in Afghanistan. The LetHerLearn dataset contains 7,600 tweets posted in reaction to the Taliban ban of women rights to education in 2022 and has been manually annotated according to Ekman emotion categories. We here detail ...
http://arxiv.org/pdf/2306.16268v1
2023-06-28T14:50:49
2023-06-28T14:50:49
{ "title": [ 0.03765869, 0.025390625, -0.017318726, 0.028060913, -0.009506226, -0.031799316, 0.03579712, -0.037109375, -0.008995056, 0.040374756, -0.047943115, 0.014587402, 0.024246216, -0.028915405, 0.0006337166, -0.013900757, 0.0071144104, 0.02...
http://arxiv.org/abs/2306.16244v1
CBBQ: A Chinese Bias Benchmark Dataset Curated with Human-AI Collaboration for Large Language Models
Yufei Huang, Deyi Xiong
cs.CL, cs.AI
Holistically measuring societal biases of large language models is crucial for detecting and reducing ethical risks in highly capable AI models. In this work, we present a Chinese Bias Benchmark dataset that consists of over 100K questions jointly constructed by human experts and generative language models, covering st...
http://arxiv.org/pdf/2306.16244v1
2023-06-28T14:14:44
2023-06-28T14:14:44
{ "title": [ -0.035125732, -0.0060310364, -0.028808594, -0.029129028, -0.023147583, -0.010314941, -0.04272461, -0.0026855469, 0.021469116, 0.032318115, -0.019927979, -0.0016794205, 0.029541016, -0.024536133, -0.045776367, 0.008735657, -0.027511597, ...
http://arxiv.org/abs/2306.16207v1
Inferring the Goals of Communicating Agents from Actions and Instructions
Lance Ying, Tan Zhi-Xuan, Vikash Mansinghka, Joshua B. Tenenbaum
cs.AI, cs.CL, cs.RO
When humans cooperate, they frequently coordinate their activity through both verbal communication and non-verbal actions, using this information to infer a shared goal and plan. How can we model this inferential ability? In this paper, we introduce a model of a cooperative team where one agent, the principal, may comm...
http://arxiv.org/pdf/2306.16207v1
2023-06-28T13:43:46
2023-06-28T13:43:46
{ "title": [ -0.031799316, 0.00071525574, -0.032165527, -0.042541504, 0.00957489, -0.008728027, -0.032989502, -0.019226074, -0.003238678, 0.014442444, -0.0072364807, 0.0034656525, -0.0018014908, -0.030853271, -0.0078086853, -0.00010091066, -0.0001288...
http://arxiv.org/abs/2306.16195v1
Enhancing Dialogue Generation via Dynamic Graph Knowledge Aggregation
Chen Tang, Hongbo Zhang, Tyler Loakman, Chenghua Lin, Frank Guerin
cs.CL, cs.AI
Incorporating external graph knowledge into neural chatbot models has been proven effective for enhancing dialogue generation. However, in conventional graph neural networks (GNNs), message passing on a graph is independent from text, resulting in the graph representation hidden space differing from that of the text. T...
http://arxiv.org/pdf/2306.16195v1
2023-06-28T13:21:00
2023-06-28T13:21:00
{ "title": [ -0.016952515, -0.019042969, -0.078552246, -0.09320068, -0.003314972, -0.002916336, -0.06829834, 0.037872314, 0.0033245087, 0.021362305, 0.0033950806, 0.0022964478, -0.01725769, -0.02746582, -0.047912598, 0.00012719631, -0.00066280365, ...
http://arxiv.org/abs/2306.16176v1
SkillNet-X: A Multilingual Multitask Model with Sparsely Activated Skills
Zhangyin Feng, Yong Dai, Fan Zhang, Duyu Tang, Xiaocheng Feng, Shuangzhi Wu, Bing Qin, Yunbo Cao, Shuming Shi
cs.CL
Traditional multitask learning methods basically can only exploit common knowledge in task- or language-wise, which lose either cross-language or cross-task knowledge. This paper proposes a general multilingual multitask model, named SkillNet-X, which enables a single model to tackle many different tasks from different...
http://arxiv.org/pdf/2306.16176v1
2023-06-28T12:53:30
2023-06-28T12:53:30
{ "title": [ -0.014831543, -0.0132751465, -0.044525146, -0.04144287, -0.0049591064, -0.012008667, -0.045013428, -0.0041007996, 0.004840851, 0.03302002, 0.010864258, -0.009284973, 0.0054244995, -0.031051636, -0.025482178, 0.0132751465, -0.019104004, ...
http://arxiv.org/abs/2306.16143v3
Generative User-Experience Research for Developing Domain-specific Natural Language Processing Applications
Anastasia Zhukova, Lukas von Sperl, Christian E. Matt, Bela Gipp
cs.CL, cs.HC
User experience (UX) is a part of human-computer interaction (HCI) research and focuses on increasing intuitiveness, transparency, simplicity, and trust for system users. Most of the UX research for machine learning (ML) or natural language processing (NLP) focuses on a data-driven methodology, i.e., it fails to focus ...
http://arxiv.org/pdf/2306.16143v3
2023-09-14T11:41:40
2023-06-28T12:17:45
{ "title": [ -0.0032520294, -0.017532349, -0.029815674, -0.037719727, -0.00749588, -0.017440796, -0.018341064, 0.0011920929, 0.002916336, 0.039611816, -0.022155762, 0.007297516, 0.015777588, -0.010826111, -0.050231934, 0.0077667236, -0.038146973, ...
http://arxiv.org/abs/2306.16125v2
A Framework for Identifying Depression on Social Media: MentalRiskES@IberLEF 2023
Simon Sanchez Viloria, Daniel Peix del Río, Rubén Bermúdez Cabo, Guillermo Arturo Arrojo Fuentes, Isabel Segura-Bedmar
cs.CL, cs.AI
This paper describes our participation in the MentalRiskES task at IberLEF 2023. The task involved predicting the likelihood of an individual experiencing depression based on their social media activity. The dataset consisted of conversations from 175 Telegram users, each labeled according to their evidence of sufferin...
http://arxiv.org/pdf/2306.16125v2
2023-06-29T07:02:59
2023-06-28T11:53:07
{ "title": [ 0.030059814, -0.00014412403, -0.01335907, -0.03491211, -0.0052871704, -0.0047683716, -0.026000977, -0.01133728, -0.027893066, 0.040283203, -0.007873535, 0.0030155182, 0.023086548, -0.05053711, -0.010536194, -0.022201538, 0.02255249, ...
http://arxiv.org/abs/2306.16108v2
Is ChatGPT a Biomedical Expert? -- Exploring the Zero-Shot Performance of Current GPT Models in Biomedical Tasks
Samy Ateia, Udo Kruschwitz
cs.CL
We assessed the performance of commercial Large Language Models (LLMs) GPT-3.5-Turbo and GPT-4 on tasks from the 2023 BioASQ challenge. In Task 11b Phase B, which is focused on answer generation, both models demonstrated competitive abilities with leading systems. Remarkably, they achieved this with simple zero-shot le...
http://arxiv.org/pdf/2306.16108v2
2023-07-24T08:14:44
2023-06-28T11:24:48
{ "title": [ -0.0044670105, -0.012863159, -0.0016040802, -0.013450623, -0.029510498, -0.0071105957, -0.02494812, -0.026901245, 0.0039634705, 0.024551392, 0.0039787292, 0.01109314, 0.010383606, -0.045654297, -0.029785156, 0.026138306, -0.009437561, ...
http://arxiv.org/abs/2306.16092v1
ChatLaw: Open-Source Legal Large Language Model with Integrated External Knowledge Bases
Jiaxi Cui, Zongjian Li, Yang Yan, Bohua Chen, Li Yuan
cs.CL
Large Language Models (LLMs) have shown the potential to revolutionize natural language processing tasks in various domains, sparking great interest in vertical-specific large models. However, unlike proprietary models such as BloombergGPT and FinGPT, which have leveraged their unique data accumulations to make strides...
http://arxiv.org/pdf/2306.16092v1
2023-06-28T10:48:34
2023-06-28T10:48:34
{ "title": [ -0.057373047, -0.022903442, -0.0501709, -0.07751465, -0.009803772, 0.03277588, -0.08380127, 0.014923096, 0.0069503784, 0.040771484, -0.02619934, 0.012481689, 0.022338867, -0.0074539185, -0.034423828, 0.008392334, -0.017745972, 0.0435...
http://arxiv.org/abs/2306.16071v1
Long-term Conversation Analysis: Exploring Utility and Privacy
Francesco Nespoli, Jule Pohlhausen, Patrick A. Naylor, Joerg Bitzer
eess.AS, cs.CL, cs.SD
The analysis of conversations recorded in everyday life requires privacy protection. In this contribution, we explore a privacy-preserving feature extraction method based on input feature dimension reduction, spectral smoothing and the low-cost speaker anonymization technique based on McAdams coefficient. We assess the...
http://arxiv.org/pdf/2306.16071v1
2023-06-28T10:10:57
2023-06-28T10:10:57
{ "title": [ -0.006175995, 0.012329102, -0.044799805, -0.064697266, 0.03152466, -0.010948181, -0.024276733, 0.001156807, -0.0017175674, 0.016708374, -0.009643555, 0.017028809, -0.020690918, -0.030212402, -0.024215698, 0.0043678284, -0.01121521, 0...
http://arxiv.org/abs/2306.16049v1
What Sentiment and Fun Facts We Learnt Before FIFA World Cup Qatar 2022 Using Twitter and AI
James She, Kamilla Swart-Arries, Mohammad Belal, Simon Wong
cs.CL, cs.SI
Twitter is a social media platform bridging most countries and allows real-time news discovery. Since the tweets on Twitter are usually short and express public feelings, thus provide a source for opinion mining and sentiment analysis for global events. This paper proposed an effective solution, in providing a sentimen...
http://arxiv.org/pdf/2306.16049v1
2023-06-28T09:29:23
2023-06-28T09:29:23
{ "title": [ -0.053649902, -0.015975952, -0.018157959, -0.0038337708, -0.032165527, 0.008598328, 0.025039673, 0.0236969, 0.009109497, 0.020233154, -0.05303955, 0.04446411, -0.013298035, -0.0013055801, -0.013801575, -0.027938843, 0.011779785, 0.03...
http://arxiv.org/abs/2306.16031v1
Exploring Spatial-Temporal Variations of Public Discourse on Social Media: A Case Study on the First Wave of the Coronavirus Pandemic in Italy
Anslow Michael, Galletti Martina
cs.CL
This paper proposes a methodology for exploring how linguistic behaviour on social media can be used to explore societal reactions to important events such as those that transpired during the SARS CoV2 pandemic. In particular, where spatial and temporal aspects of events are important features. Our methodology consists...
http://arxiv.org/pdf/2306.16031v1
2023-06-28T08:59:50
2023-06-28T08:59:50
{ "title": [ 0.035705566, -0.006587982, -0.043945312, -0.02760315, 0.0061531067, -0.03933716, 0.005771637, -0.004497528, -0.028366089, 0.052886963, -0.022705078, 0.020385742, 0.022399902, -0.008003235, -0.038269043, -0.022735596, 0.009750366, 0.0...
http://arxiv.org/abs/2306.16009v1
Accelerating Transducers through Adjacent Token Merging
Yuang Li, Yu Wu, Jinyu Li, Shujie Liu
cs.CL, eess.AS, eess.SP
Recent end-to-end automatic speech recognition (ASR) systems often utilize a Transformer-based acoustic encoder that generates embedding at a high frame rate. However, this design is inefficient, particularly for long speech signals due to the quadratic computation of self-attention. To address this, we propose a new m...
http://arxiv.org/pdf/2306.16009v1
2023-06-28T08:33:13
2023-06-28T08:33:13
{ "title": [ 0.004337311, 0.0063819885, -0.052337646, -0.090270996, 0.009468079, -0.0043029785, -0.05331421, 0.0034656525, -0.012802124, 0.041778564, -0.03515625, -0.009223938, 0.036071777, -0.022476196, 0.007160187, 0.004470825, -0.000497818, 0....
http://arxiv.org/abs/2306.16007v1
Prompting Large Language Models for Zero-Shot Domain Adaptation in Speech Recognition
Yuang Li, Yu Wu, Jinyu Li, Shujie Liu
cs.CL, eess.AS, eess.SP
The integration of Language Models (LMs) has proven to be an effective way to address domain shifts in speech recognition. However, these approaches usually require a significant amount of target domain text data for the training of LMs. Different from these methods, in this work, with only a domain-specific text promp...
http://arxiv.org/pdf/2306.16007v1
2023-06-28T08:29:00
2023-06-28T08:29:00
{ "title": [ -0.008865356, -0.011985779, -0.0395813, -0.077697754, -0.004421234, -0.0008687973, -0.05645752, 0.034820557, 0.000869751, 0.020568848, -0.014312744, -0.002035141, 0.0014543533, -0.021377563, -0.034576416, 0.0036144257, 0.0045280457, ...
http://arxiv.org/abs/2306.16001v1
Streamlining Social Media Information Retrieval for Public Health Research with Deep Learning
Yining Hua, Shixu Lin, Minghui Li, Yujie Zhang, Peilin Zhou, Ying-Chih Lo, Li Zhou, Jie Yang
cs.CL, cs.AI, cs.IR
The utilization of social media in epidemic surveillance has been well established. Nonetheless, bias is often introduced when pre-defined lexicons are used to retrieve relevant corpus. This study introduces a framework aimed at curating extensive dictionaries of medical colloquialisms and Unified Medical Language Syst...
http://arxiv.org/pdf/2306.16001v1
2023-06-28T08:20:35
2023-06-28T08:20:35
{ "title": [ 0.025802612, 0.015213013, -0.030090332, -0.0635376, 0.03338623, -0.017227173, -0.021362305, -0.014350891, -0.020004272, 0.029953003, -0.037902832, 0.039154053, -0.032196045, 0.004432678, -0.03277588, -0.010658264, -0.0034828186, 0.01...
http://arxiv.org/abs/2306.15978v1
Sentence-to-Label Generation Framework for Multi-task Learning of Japanese Sentence Classification and Named Entity Recognition
Chengguang Gan, Qinghao Zhang, Tatsunori Mori
cs.CL
Information extraction(IE) is a crucial subfield within natural language processing. In this study, we introduce a Sentence Classification and Named Entity Recognition Multi-task (SCNM) approach that combines Sentence Classification (SC) and Named Entity Recognition (NER). We develop a Sentence-to-Label Generation (SLG...
http://arxiv.org/pdf/2306.15978v1
2023-06-28T07:29:44
2023-06-28T07:29:44
{ "title": [ -0.03112793, 0.016296387, -0.038848877, -0.038635254, -0.0048446655, -0.0007915497, -0.05657959, -0.009727478, 0.0033550262, 0.017196655, -0.0146865845, 0.018249512, -0.00598526, -0.0124053955, -0.047698975, -0.0060272217, -0.017333984, ...
http://arxiv.org/abs/2306.15933v1
You Can Generate It Again: Data-to-text Generation with Verification and Correction Prompting
Xuan Ren, Lingqiao Liu
cs.CL, cs.AI, cs.LG
Despite significant advancements in existing models, generating text descriptions from structured data input, known as data-to-text generation, remains a challenging task. In this paper, we propose a novel approach that goes beyond traditional one-shot generation methods by introducing a multi-step process consisting o...
http://arxiv.org/pdf/2306.15933v1
2023-06-28T05:34:25
2023-06-28T05:34:25
{ "title": [ -0.013832092, -0.018371582, -0.035186768, -0.05505371, -0.019744873, -0.005531311, -0.052490234, 0.008918762, 0.0021572113, 0.048736572, -0.0473938, 0.006290436, 0.011627197, -0.0057373047, -0.041809082, -0.017242432, -0.001001358, 0...
http://arxiv.org/abs/2306.15926v1
Most Language Models can be Poets too: An AI Writing Assistant and Constrained Text Generation Studio
Allen Roush, Sanjay Basu, Akshay Moorthy, Dmitry Dubovoy
cs.CL, cs.AI, cs.LG
Despite rapid advancement in the field of Constrained Natural Language Generation, little time has been spent on exploring the potential of language models which have had their vocabularies lexically, semantically, and/or phonetically constrained. We find that most language models generate compelling text even under si...
http://arxiv.org/pdf/2306.15926v1
2023-06-28T05:10:51
2023-06-28T05:10:51
{ "title": [ -0.048675537, -0.038604736, -0.030807495, 0.012939453, -0.059143066, 0.0010433197, -0.0132751465, 0.0078125, 0.015090942, 0.028182983, -0.02633667, -0.004917145, 0.015838623, -0.03173828, -0.017822266, -0.033294678, -0.00969696, 0.04...
http://arxiv.org/abs/2306.15917v1
Confidence-Calibrated Ensemble Dense Phrase Retrieval
William Yang, Noah Bergam, Arnav Jain, Nima Sheikhoslami
cs.CL, cs.IR
In this paper, we consider the extent to which the transformer-based Dense Passage Retrieval (DPR) algorithm, developed by (Karpukhin et. al. 2020), can be optimized without further pre-training. Our method involves two particular insights: we apply the DPR context encoder at various phrase lengths (e.g. one-sentence v...
http://arxiv.org/pdf/2306.15917v1
2023-06-28T04:40:21
2023-06-28T04:40:21
{ "title": [ -0.0037117004, 0.013557434, -0.05166626, -0.05380249, 0.028152466, -0.013252258, -0.022155762, -0.016937256, 0.033233643, 0.05203247, -0.04257202, 0.020431519, 0.015037537, 0.010398865, -0.066589355, 0.023117065, -0.003074646, -0.000...
http://arxiv.org/abs/2306.15895v2
Large Language Model as Attributed Training Data Generator: A Tale of Diversity and Bias
Yue Yu, Yuchen Zhuang, Jieyu Zhang, Yu Meng, Alexander Ratner, Ranjay Krishna, Jiaming Shen, Chao Zhang
cs.CL, cs.AI, cs.LG
Large language models (LLMs) have been recently leveraged as training data generators for various natural language processing (NLP) tasks. While previous research has explored different approaches to training models using generated data, they generally rely on simple class-conditional prompts, which may limit the diver...
http://arxiv.org/pdf/2306.15895v2
2023-10-18T02:07:12
2023-06-28T03:31:31
{ "title": [ -0.026916504, -0.0109939575, -0.060638428, -0.048461914, -0.007865906, -0.0063972473, -0.045837402, 0.004749298, -0.020339966, 0.022644043, -0.005783081, -0.0105896, -0.003622055, -0.030349731, -0.03967285, -0.0042381287, -0.024230957, ...
http://arxiv.org/abs/2306.15837v1
Symbol emergence as interpersonal cross-situational learning: the emergence of lexical knowledge with combinatoriality
Yoshinobu Hagiwara, Kazuma Furukawa, Takafumi Horie, Akira Taniguchi, Tadahiro Taniguchi
cs.CL, cs.AI
We present a computational model for a symbol emergence system that enables the emergence of lexical knowledge with combinatoriality among agents through a Metropolis-Hastings naming game and cross-situational learning. Many computational models have been proposed to investigate combinatoriality in emergent communicati...
http://arxiv.org/pdf/2306.15837v1
2023-06-27T23:55:33
2023-06-27T23:55:33
{ "title": [ -0.0025405884, 0.009376526, -0.026168823, -0.011581421, 0.0021743774, -0.02861023, 0.0073623657, -0.018508911, -0.000726223, 0.031982422, -0.022781372, 0.020004272, -0.0046577454, -0.011230469, -0.03390503, -0.0069122314, -0.007965088, ...
http://arxiv.org/abs/2306.15826v1
MAT: Mixed-Strategy Game of Adversarial Training in Fine-tuning
Zhehua Zhong, Tianyi Chen, Zhen Wang
cs.CL, cs.AI
Fine-tuning large-scale pre-trained language models has been demonstrated effective for various natural language processing (NLP) tasks. Previous studies have established that incorporating adversarial training during the fine-tuning stage can significantly enhance model generalization and robustness. However, from the...
http://arxiv.org/pdf/2306.15826v1
2023-06-27T23:19:53
2023-06-27T23:19:53
{ "title": [ -0.010627747, -0.003704071, -0.022842407, -0.020568848, 0.0066223145, -0.023376465, 0.030288696, -0.024993896, -0.07684326, 0.039367676, 0.0064468384, 0.03604126, -0.030853271, -0.01663208, -0.016174316, -0.02923584, -0.0004930496, 0...
http://arxiv.org/abs/2306.15824v1
Confidence-based Ensembles of End-to-End Speech Recognition Models
Igor Gitman, Vitaly Lavrukhin, Aleksandr Laptev, Boris Ginsburg
eess.AS, cs.CL, cs.LG
The number of end-to-end speech recognition models grows every year. These models are often adapted to new domains or languages resulting in a proliferation of expert systems that achieve great results on target data, while generally showing inferior performance outside of their domain of expertise. We explore combinat...
http://arxiv.org/pdf/2306.15824v1
2023-06-27T23:13:43
2023-06-27T23:13:43
{ "title": [ -0.022659302, 0.02583313, -0.06378174, -0.07293701, 0.035614014, -0.004497528, -0.044311523, 0.013572693, 0.0053138733, 0.041656494, -0.02746582, 0.010971069, -0.014259338, -0.014709473, -0.05621338, 0.006980896, -0.012825012, 0.0316...
http://arxiv.org/abs/2306.15799v1
FLuRKA: Fast fused Low-Rank & Kernel Attention
Ahan Gupta, Yueming Yuan, Yanqi Zhou, Charith Mendis
cs.LG, cs.CL, cs.PF
Many efficient approximate self-attention techniques have become prevalent since the inception of the transformer architecture. Two popular classes of these techniques are low-rank and kernel methods. Each of these methods has its own strengths. We observe these strengths synergistically complement each other and explo...
http://arxiv.org/pdf/2306.15799v1
2023-06-27T20:58:41
2023-06-27T20:58:41
{ "title": [ -0.015068054, 0.0026340485, -0.043060303, -0.06185913, 0.032684326, -0.007217407, -0.015853882, -0.020858765, 0.024841309, 0.028274536, -0.025482178, 0.0015201569, 0.02116394, -0.011039734, -0.031433105, 0.017074585, -0.00042891502, ...
http://arxiv.org/abs/2306.15788v2
Evaluating GPT-3.5 and GPT-4 on Grammatical Error Correction for Brazilian Portuguese
Maria Carolina Penteado, Fábio Perez
cs.CL, cs.AI
We investigate the effectiveness of GPT-3.5 and GPT-4, two large language models, as Grammatical Error Correction (GEC) tools for Brazilian Portuguese and compare their performance against Microsoft Word and Google Docs. We introduce a GEC dataset for Brazilian Portuguese with four categories: Grammar, Spelling, Intern...
http://arxiv.org/pdf/2306.15788v2
2023-07-18T13:31:56
2023-06-27T20:37:54
{ "title": [ -0.0038852692, 0.012840271, -0.01701355, -0.014762878, -0.024490356, 0.007827759, -0.04586792, -0.021575928, -0.03451538, 0.034332275, -0.026367188, -0.004005432, 0.01977539, -0.027038574, -0.012008667, -0.013214111, -0.04724121, 0.0...
http://arxiv.org/abs/2306.15774v1
Next Steps for Human-Centered Generative AI: A Technical Perspective
Xiang 'Anthony' Chen, Jeff Burke, Ruofei Du, Matthew K. Hong, Jennifer Jacobs, Philippe Laban, Dingzeyu Li, Nanyun Peng, Karl D. D. Willis, Chien-Sheng Wu, Bolei Zhou
cs.HC, cs.CL, cs.CV, cs.LG
Through iterative, cross-disciplinary discussions, we define and propose next-steps for Human-centered Generative AI (HGAI) from a technical perspective. We contribute a roadmap that lays out future directions of Generative AI spanning three levels: Aligning with human values; Accommodating humans' expression of intent...
http://arxiv.org/pdf/2306.15774v1
2023-06-27T19:54:30
2023-06-27T19:54:30
{ "title": [ -0.011047363, -0.0046844482, -0.033691406, -0.036071777, 0.0044898987, -0.02406311, -0.0032043457, -0.021194458, -0.0016098022, 0.027572632, -0.015808105, 0.0131073, -0.0061035156, -0.02204895, -0.044769287, 0.009628296, 0.0019054413, ...
http://arxiv.org/abs/2306.15766v1
Large Language Models as Annotators: Enhancing Generalization of NLP Models at Minimal Cost
Parikshit Bansal, Amit Sharma
cs.CL, cs.LG
State-of-the-art supervised NLP models achieve high accuracy but are also susceptible to failures on inputs from low-data regimes, such as domains that are not represented in training data. As an approximation to collecting ground-truth labels for the specific domain, we study the use of large language models (LLMs) fo...
http://arxiv.org/pdf/2306.15766v1
2023-06-27T19:29:55
2023-06-27T19:29:55
{ "title": [ -0.0178833, -0.005344391, -0.022567749, -0.0770874, 0.0012254715, 0.01802063, -0.061828613, -0.0053596497, 0.015403748, 0.028366089, -0.019577026, 0.0031776428, 0.004386902, -0.0038032532, -0.049560547, 0.014259338, -0.014427185, 0.0...
http://arxiv.org/abs/2306.15745v3
Identity Construction in a Misogynist Incels Forum
Michael Miller Yoder, Chloe Perry, David West Brown, Kathleen M. Carley, Meredith L. Pruden
cs.CL
Online communities of involuntary celibates (incels) are a prominent source of misogynist hate speech. In this paper, we use quantitative text and network analysis approaches to examine how identity groups are discussed on incels-dot-is, the largest black-pilled incels forum. We find that this community produces a wide...
http://arxiv.org/pdf/2306.15745v3
2023-07-09T21:15:36
2023-06-27T18:56:28
{ "title": [ 0.062561035, -0.0042037964, -0.041168213, -0.049804688, -0.0035171509, -0.044067383, 0.031982422, 0.019165039, -0.056884766, 0.05532837, 0.018920898, 0.023086548, 0.012763977, -0.050476074, 0.012001038, -0.014266968, 0.006252289, 0.0...
http://arxiv.org/abs/2306.15736v2
DMNER: Biomedical Entity Recognition by Detection and Matching
Junyi Bian, Rongze Jiang, Weiqi Zhai, Tianyang Huang, Hong Zhou, Shanfeng Zhu
cs.CL
Biomedical named entity recognition (BNER) serves as the foundation for numerous biomedical text mining tasks. Unlike general NER, BNER require a comprehensive grasp of the domain, and incorporating external knowledge beyond training data poses a significant challenge. In this study, we propose a novel BNER framework c...
http://arxiv.org/pdf/2306.15736v2
2023-07-05T12:26:07
2023-06-27T18:32:07
{ "title": [ -0.026229858, 0.004535675, -0.056884766, -0.07727051, 0.02999878, 0.013137817, -0.048309326, -0.004787445, 0.0024814606, 0.019607544, -0.0004696846, -0.0046424866, 0.001159668, -0.022384644, -0.0368042, 0.004142761, -0.0054092407, 0....
http://arxiv.org/abs/2306.15732v1
A Weakly Supervised Classifier and Dataset of White Supremacist Language
Michael Miller Yoder, Ahmad Diab, David West Brown, Kathleen M. Carley
cs.CL
We present a dataset and classifier for detecting the language of white supremacist extremism, a growing issue in online hate speech. Our weakly supervised classifier is trained on large datasets of text from explicitly white supremacist domains paired with neutral and anti-racist data from similar domains. We demonstr...
http://arxiv.org/pdf/2306.15732v1
2023-06-27T18:19:32
2023-06-27T18:19:32
{ "title": [ 0.006412506, 0.018615723, -0.046417236, -0.05697632, 0.02104187, -0.028442383, -0.005874634, -0.0028858185, -0.010398865, 0.035705566, -0.008300781, 0.0022392273, 0.017608643, -0.029327393, -0.021484375, -0.0010585785, 0.0063171387, ...
http://arxiv.org/abs/2306.15724v4
REFLECT: Summarizing Robot Experiences for Failure Explanation and Correction
Zeyi Liu, Arpit Bahety, Shuran Song
cs.RO, cs.AI, cs.CL, cs.CV
The ability to detect and analyze failed executions automatically is crucial for an explainable and robust robotic system. Recently, Large Language Models (LLMs) have demonstrated strong reasoning abilities on textual inputs. To leverage the power of LLMs for robot failure explanation, we introduce REFLECT, a framework...
http://arxiv.org/pdf/2306.15724v4
2023-10-16T21:04:57
2023-06-27T18:03:15
{ "title": [ -0.0015077591, -0.025924683, 0.0027122498, -0.065979004, -0.02104187, 0.011146545, -0.03970337, -0.005657196, -0.0075187683, 0.027175903, -0.036315918, 0.006275177, 0.0074882507, -0.007534027, -0.02444458, -0.02180481, 0.000664711, 0...
http://arxiv.org/abs/2306.15656v3
SparseOptimizer: Sparsify Language Models through Moreau-Yosida Regularization and Accelerate via Compiler Co-design
Fu-Ming Guo
cs.LG, cs.AI, cs.CC, cs.CL, cs.MS
This paper introduces SparseOptimizer, a novel deep learning optimizer that exploits Moreau-Yosida regularization to naturally induce sparsity in large language models such as BERT, ALBERT and GPT. Key to the design of SparseOptimizer is an embedded shrinkage operator, which imparts sparsity directly within the optimiz...
http://arxiv.org/pdf/2306.15656v3
2023-07-18T17:52:28
2023-06-27T17:50:26
{ "title": [ -0.04498291, -0.016021729, -0.05090332, -0.04849243, -0.02784729, -0.017837524, -0.06719971, 0.01991272, -0.006198883, 0.031097412, -0.011550903, 0.0132369995, 0.003227234, -0.04058838, -0.020339966, 0.0038471222, 0.0019760132, 0.006...
http://arxiv.org/abs/2306.15644v1
Style-transfer based Speech and Audio-visual Scene Understanding for Robot Action Sequence Acquisition from Videos
Chiori Hori, Puyuan Peng, David Harwath, Xinyu Liu, Kei Ota, Siddarth Jain, Radu Corcodel, Devesh Jha, Diego Romeres, Jonathan Le Roux
cs.CL
To realize human-robot collaboration, robots need to execute actions for new tasks according to human instructions given finite prior knowledge. Human experts can share their knowledge of how to perform a task with a robot through multi-modal instructions in their demonstrations, showing a sequence of short-horizon ste...
http://arxiv.org/pdf/2306.15644v1
2023-06-27T17:37:53
2023-06-27T17:37:53
{ "title": [ -0.017868042, -0.007019043, -0.052642822, -0.030670166, -0.0069236755, -0.026977539, -0.011909485, 0.019866943, -0.0287323, 0.020614624, -0.027832031, 0.021011353, -0.0021743774, -0.025024414, -0.009590149, -0.027389526, 0.008483887, ...
http://arxiv.org/abs/2306.15634v2
Automatic Annotation of Direct Speech in Written French Narratives
Noé Durandard, Viet-Anh Tran, Gaspard Michel, Elena V. Epure
cs.CL
The automatic annotation of direct speech (AADS) in written text has been often used in computational narrative understanding. Methods based on either rules or deep neural networks have been explored, in particular for English or German languages. Yet, for French, our target language, not many works exist. Our goal is ...
http://arxiv.org/pdf/2306.15634v2
2023-06-28T07:44:53
2023-06-27T17:21:00
{ "title": [ -0.044036865, 0.0062675476, -0.032592773, -0.050872803, -0.014549255, 0.008125305, -0.04196167, 0.009857178, 0.00003618002, 0.043151855, 0.0039901733, -0.00048422813, 0.004928589, -0.028045654, -0.037902832, 0.018722534, -0.01449585, ...
http://arxiv.org/abs/2306.15604v1
Constructing Multilingual Code Search Dataset Using Neural Machine Translation
Ryo Sekizawa, Nan Duan, Shuai Lu, Hitomi Yanaka
cs.CL, cs.SE
Code search is a task to find programming codes that semantically match the given natural language queries. Even though some of the existing datasets for this task are multilingual on the programming language side, their query data are only in English. In this research, we create a multilingual code search dataset in f...
http://arxiv.org/pdf/2306.15604v1
2023-06-27T16:42:36
2023-06-27T16:42:36
{ "title": [ -0.022735596, -0.020874023, -0.0070610046, -0.036621094, -0.018096924, 0.005645752, -0.035308838, -0.0018510818, -0.00087976456, 0.059173584, -0.0053863525, -0.010566711, 0.015594482, -0.012931824, -0.049072266, 0.00869751, -0.018203735,...
http://arxiv.org/abs/2306.15595v2
Extending Context Window of Large Language Models via Positional Interpolation
Shouyuan Chen, Sherman Wong, Liangjian Chen, Yuandong Tian
cs.CL, cs.AI, cs.LG
We present Position Interpolation (PI) that extends the context window sizes of RoPE-based pretrained LLMs such as LLaMA models to up to 32768 with minimal fine-tuning (within 1000 steps), while demonstrating strong empirical results on various tasks that require long context, including passkey retrieval, language mode...
http://arxiv.org/pdf/2306.15595v2
2023-06-28T04:26:05
2023-06-27T16:26:26
{ "title": [ -0.009994507, 0.00030183792, -0.041015625, -0.07104492, 0.013977051, -0.0004799366, -0.052886963, -0.006931305, 0.02571106, 0.050811768, -0.034851074, -0.009780884, 0.023620605, -0.000030398369, -0.06604004, -0.0031776428, -0.0063095093,...
http://arxiv.org/abs/2306.15551v2
MyCrunchGPT: A chatGPT assisted framework for scientific machine learning
Varun Kumar, Leonard Gleyzer, Adar Kahana, Khemraj Shukla, George Em Karniadakis
cs.LG, cs.CL, physics.soc-ph
Scientific Machine Learning (SciML) has advanced recently across many different areas in computational science and engineering. The objective is to integrate data and physics seamlessly without the need of employing elaborate and computationally taxing data assimilation schemes. However, preprocessing, problem formulat...
http://arxiv.org/pdf/2306.15551v2
2023-07-31T21:39:01
2023-06-27T15:23:42
{ "title": [ -0.027954102, -0.020736694, -0.035125732, -0.03967285, -0.00039863586, 0.006290436, -0.021194458, -0.010658264, 0.0000564456, 0.0181427, -0.001750946, 0.018569946, -0.0113220215, -0.025634766, -0.027770996, 0.014167786, -0.014389038, ...
http://arxiv.org/abs/2306.15550v2
CamemBERT-bio: a Tasty French Language Model Better for your Health
Rian Touchent, Laurent Romary, Eric de la Clergerie
cs.CL, cs.AI
Clinical data in hospitals are increasingly accessible for research through clinical data warehouses, however these documents are unstructured. It is therefore necessary to extract information from medical reports to conduct clinical studies. Transfer learning with BERT-like models such as CamemBERT has allowed major a...
http://arxiv.org/pdf/2306.15550v2
2023-08-02T17:53:45
2023-06-27T15:23:14
{ "title": [ -0.007633209, -0.02810669, -0.045532227, -0.061340332, -0.024612427, -0.0014886856, -0.054901123, 0.010475159, -0.029800415, 0.03845215, 0.0016584396, 0.005290985, 0.033081055, -0.046569824, 0.010643005, -0.008201599, 0.008842468, 0....