id
stringlengths
33
34
title
stringlengths
6
218
authors
stringlengths
5
6.98k
categories
stringlengths
5
188
summary
stringlengths
34
3.26k
link_pdf
stringlengths
33
34
updated
timestamp[s]
published
timestamp[s]
embeddings
dict
http://arxiv.org/abs/2310.10195v2
AdaLomo: Low-memory Optimization with Adaptive Learning Rate
Kai Lv, Hang Yan, Qipeng Guo, Haijun Lv, Xipeng Qiu
cs.LG, cs.CL
Large language models have achieved remarkable success, but their extensive parameter size necessitates substantial memory for training, thereby setting a high threshold. While the recently proposed low-memory optimization (LOMO) reduces memory footprint, its optimization technique, akin to stochastic gradient descent,...
http://arxiv.org/pdf/2310.10195v2
2023-10-22T13:37:33
2023-10-16T09:04:28
{ "title": [ -0.04437256, -0.004802704, -0.05117798, -0.04260254, 0.0074424744, 0.005218506, -0.04034424, -0.026885986, 0.0011348724, 0.037261963, 0.0052452087, 0.021621704, -0.014144897, -0.016677856, -0.033813477, 0.02456665, 0.0025978088, 0.02...
http://arxiv.org/abs/2310.10191v4
VIBE: Topic-Driven Temporal Adaptation for Twitter Classification
Yuji Zhang, Jing Li, Wenjie Li
cs.CL
Language features are evolving in real-world social media, resulting in the deteriorating performance of text classification in dynamics. To address this challenge, we study temporal adaptation, where models trained on past data are tested in the future. Most prior work focused on continued pretraining or knowledge upd...
http://arxiv.org/pdf/2310.10191v4
2023-11-15T12:41:57
2023-10-16T08:53:57
{ "title": [ 0.0055274963, -0.008056641, -0.03729248, -0.061584473, 0.01159668, -0.005203247, -0.008964539, 0.0007004738, 0.018371582, 0.044403076, -0.04055786, 0.03817749, -0.013824463, 0.019165039, -0.05508423, -0.0030136108, -0.01474762, 0.041...
http://arxiv.org/abs/2310.10190v1
Battle of the Large Language Models: Dolly vs LLaMA vs Vicuna vs Guanaco vs Bard vs ChatGPT -- A Text-to-SQL Parsing Comparison
Shuo Sun, Yuchen Zhang, Jiahuan Yan, Yuze Gao, Donovan Ong, Bin Chen, Jian Su
cs.CL, cs.AI
The success of ChatGPT has ignited an AI race, with researchers striving to develop new large language models (LLMs) that can match or surpass the language understanding and generation abilities of commercial ones. In recent times, a number of models have emerged, claiming performance near that of GPT-3.5 or GPT-4 thro...
http://arxiv.org/pdf/2310.10190v1
2023-10-16T08:52:41
2023-10-16T08:52:41
{ "title": [ -0.012626648, -0.02835083, -0.030944824, -0.011383057, -0.03125, -0.006084442, -0.044189453, 0.010803223, -0.016586304, 0.031982422, -0.0357666, -0.021362305, 0.004333496, -0.040008545, -0.009796143, -0.05609131, -0.015930176, 0.0452...
http://arxiv.org/abs/2310.10184v1
Continual Generalized Intent Discovery: Marching Towards Dynamic and Open-world Intent Recognition
Xiaoshuai Song, Yutao Mou, Keqing He, Yueyan Qiu, Pei Wang, Weiran Xu
cs.CL, cs.AI, cs.LG
In a practical dialogue system, users may input out-of-domain (OOD) queries. The Generalized Intent Discovery (GID) task aims to discover OOD intents from OOD queries and extend them to the in-domain (IND) classifier. However, GID only considers one stage of OOD learning, and needs to utilize the data in all previous s...
http://arxiv.org/pdf/2310.10184v1
2023-10-16T08:48:07
2023-10-16T08:48:07
{ "title": [ -0.003894806, 0.012054443, -0.05999756, -0.057739258, 0.0357666, -0.023986816, -0.008575439, -0.009597778, 0.0050239563, 0.021362305, -0.024780273, 0.0018777847, -0.0011997223, -0.022384644, -0.029327393, 0.0021476746, -0.012817383, ...
http://arxiv.org/abs/2310.10180v2
TRIGO: Benchmarking Formal Mathematical Proof Reduction for Generative Language Models
Jing Xiong, Jianhao Shen, Ye Yuan, Haiming Wang, Yichun Yin, Zhengying Liu, Lin Li, Zhijiang Guo, Qingxing Cao, Yinya Huang, Chuanyang Zheng, Xiaodan Liang, Ming Zhang, Qun Liu
cs.CL
Automated theorem proving (ATP) has become an appealing domain for exploring the reasoning ability of the recent successful generative language models. However, current ATP benchmarks mainly focus on symbolic inference, but rarely involve the understanding of complex number combination reasoning. In this work, we propo...
http://arxiv.org/pdf/2310.10180v2
2023-10-24T15:17:56
2023-10-16T08:42:39
{ "title": [ -0.023712158, -0.015411377, -0.039611816, -0.08306885, 0.006324768, 0.005832672, -0.06903076, 0.014266968, -0.0018882751, 0.034454346, -0.0012302399, 0.019683838, 0.0037879944, -0.01524353, -0.04144287, 0.011917114, -0.01927185, 0.03...
http://arxiv.org/abs/2310.10176v1
Large Language Models Meet Open-World Intent Discovery and Recognition: An Evaluation of ChatGPT
Xiaoshuai Song, Keqing He, Pei Wang, Guanting Dong, Yutao Mou, Jingang Wang, Yunsen Xian, Xunliang Cai, Weiran Xu
cs.CL, cs.AI, cs.LG
The tasks of out-of-domain (OOD) intent discovery and generalized intent discovery (GID) aim to extend a closed intent classifier to open-world intent sets, which is crucial to task-oriented dialogue (TOD) systems. Previous methods address them by fine-tuning discriminative models. Recently, although some studies have ...
http://arxiv.org/pdf/2310.10176v1
2023-10-16T08:34:44
2023-10-16T08:34:44
{ "title": [ -0.008453369, 0.0022258759, -0.02381897, -0.0340271, 0.002779007, -0.0036525726, -0.027877808, -0.006450653, -0.00630188, 0.015075684, -0.0115737915, 0.0030345917, -0.0004644394, -0.018341064, -0.020812988, 0.0019102097, -0.027908325, ...
http://arxiv.org/abs/2310.10169v1
DemoNSF: A Multi-task Demonstration-based Generative Framework for Noisy Slot Filling Task
Guanting Dong, Tingfeng Hui, Zhuoma GongQue, Jinxu Zhao, Daichi Guo, Gang Zhao, Keqing He, Weiran Xu
cs.CL, cs.AI, cs.IR, cs.LG
Recently, prompt-based generative frameworks have shown impressive capabilities in sequence labeling tasks. However, in practical dialogue scenarios, relying solely on simplistic templates and traditional corpora presents a challenge for these methods in generalizing to unknown input perturbations. To address this gap,...
http://arxiv.org/pdf/2310.10169v1
2023-10-16T08:16:53
2023-10-16T08:16:53
{ "title": [ 0.0026474, -0.004573822, -0.03479004, -0.053588867, -0.015060425, -0.004261017, -0.06188965, -0.026275635, 0.028137207, 0.025543213, -0.042633057, 0.02027893, -0.027450562, -0.013504028, -0.044403076, -0.020080566, -0.013427734, 0.02...
http://arxiv.org/abs/2310.10159v1
Joint Music and Language Attention Models for Zero-shot Music Tagging
Xingjian Du, Zhesong Yu, Jiaju Lin, Bilei Zhu, Qiuqiang Kong
cs.SD, cs.CL, eess.AS
Music tagging is a task to predict the tags of music recordings. However, previous music tagging research primarily focuses on close-set music tagging tasks which can not be generalized to new tags. In this work, we propose a zero-shot music tagging system modeled by a joint music and language attention (JMLA) model to...
http://arxiv.org/pdf/2310.10159v1
2023-10-16T08:00:16
2023-10-16T08:00:16
{ "title": [ 0.019622803, -0.0020961761, -0.045715332, -0.068115234, 0.0051002502, -0.027145386, -0.026184082, 0.015838623, 0.0029411316, 0.048583984, -0.017654419, -0.0006546974, 0.01386261, -0.014808655, -0.044158936, 0.0037517548, -0.006385803, ...
http://arxiv.org/abs/2310.10158v1
Character-LLM: A Trainable Agent for Role-Playing
Yunfan Shao, Linyang Li, Junqi Dai, Xipeng Qiu
cs.CL, cs.AI
Large language models (LLMs) can be used to serve as agents to simulate human behaviors, given the powerful ability to understand human instructions and provide high-quality generated texts. Such ability stimulates us to wonder whether LLMs can simulate a person in a higher form than simple human behaviors. Therefore, ...
http://arxiv.org/pdf/2310.10158v1
2023-10-16T07:58:56
2023-10-16T07:58:56
{ "title": [ -0.012252808, -0.01171875, -0.033325195, -0.03543091, -0.003967285, -0.010978699, -0.012817383, -0.014221191, -0.018432617, 0.022857666, -0.009490967, 0.01939392, 0.0019350052, -0.03439331, -0.01411438, 0.0029392242, -0.02154541, 0.0...
http://arxiv.org/abs/2310.10701v2
Theory of Mind for Multi-Agent Collaboration via Large Language Models
Huao Li, Yu Quan Chong, Simon Stepputtis, Joseph Campbell, Dana Hughes, Michael Lewis, Katia Sycara
cs.CL, cs.AI
While Large Language Models (LLMs) have demonstrated impressive accomplishments in both reasoning and planning, their abilities in multi-agent collaborations remains largely unexplored. This study evaluates LLM-based agents in a multi-agent cooperative text game with Theory of Mind (ToM) inference tasks, comparing thei...
http://arxiv.org/pdf/2310.10701v2
2023-10-22T03:49:34
2023-10-16T07:51:19
{ "title": [ -0.04309082, -0.022506714, -0.030075073, -0.030197144, -0.008796692, 0.0017375946, -0.031555176, -0.0011901855, -0.0047950745, 0.02671814, 0.0024166107, -0.008834839, 0.0055618286, -0.022766113, -0.039886475, -0.00095129013, -0.026473999...
http://arxiv.org/abs/2310.10151v1
DNA: Denoised Neighborhood Aggregation for Fine-grained Category Discovery
Wenbin An, Feng Tian, Wenkai Shi, Yan Chen, Qinghua Zheng, QianYing Wang, Ping Chen
cs.LG, cs.CL, cs.IR
Discovering fine-grained categories from coarsely labeled data is a practical and challenging task, which can bridge the gap between the demand for fine-grained analysis and the high annotation cost. Previous works mainly focus on instance-level discrimination to learn low-level features, but ignore semantic similariti...
http://arxiv.org/pdf/2310.10151v1
2023-10-16T07:43:30
2023-10-16T07:43:30
{ "title": [ -0.0034484863, -0.0056877136, -0.0692749, -0.07727051, 0.041778564, -0.016418457, -0.01008606, 0.013755798, 0.006362915, 0.0184021, -0.034423828, 0.009437561, -0.016586304, -0.008895874, -0.04449463, -0.015464783, 0.0035247803, 0.027...
http://arxiv.org/abs/2310.10138v1
Node-based Knowledge Graph Contrastive Learning for Medical Relationship Prediction
Zhiguang Fan, Yuedong Yang, Mingyuan Xu, Hongming Chen
cs.DB, cs.CL, q-bio.QM
The embedding of Biomedical Knowledge Graphs (BKGs) generates robust representations, valuable for a variety of artificial intelligence applications, including predicting drug combinations and reasoning disease-drug relationships. Meanwhile, contrastive learning (CL) is widely employed to enhance the distinctiveness of...
http://arxiv.org/pdf/2310.10138v1
2023-10-16T07:27:43
2023-10-16T07:27:43
{ "title": [ -0.009918213, -0.0078048706, -0.034362793, -0.07183838, -0.011703491, 0.003704071, -0.052642822, -0.015625, 0.025421143, 0.018981934, -0.01361084, 0.0034942627, -0.00021255016, -0.03010559, -0.038360596, 0.0010557175, 0.00022387505, ...
http://arxiv.org/abs/2310.13008v1
LoBaSS: Gauging Learnability in Supervised Fine-tuning Data
Haotian Zhou, Tingkai Liu, Qianli Ma, Jianbo Yuan, Pengfei Liu, Yang You, Hongxia Yang
cs.LG, cs.AI, cs.CL
Supervised Fine-Tuning (SFT) serves as a crucial phase in aligning Large Language Models (LLMs) to specific task prerequisites. The selection of fine-tuning data profoundly influences the model's performance, whose principle is traditionally grounded in data quality and distribution. In this paper, we introduce a new d...
http://arxiv.org/pdf/2310.13008v1
2023-10-16T07:26:24
2023-10-16T07:26:24
{ "title": [ -0.020462036, 0.0075950623, -0.018218994, -0.043426514, 0.012252808, 0.010848999, -0.024108887, -0.03765869, -0.0030059814, -0.005317688, -0.042388916, 0.017440796, -0.058288574, -0.0025672913, -0.04071045, -0.030151367, -0.0023727417, ...
http://arxiv.org/abs/2310.10134v1
CLIN: A Continually Learning Language Agent for Rapid Task Adaptation and Generalization
Bodhisattwa Prasad Majumder, Bhavana Dalvi Mishra, Peter Jansen, Oyvind Tafjord, Niket Tandon, Li Zhang, Chris Callison-Burch, Peter Clark
cs.CL, cs.AI, cs.LG
Language agents have shown some ability to interact with an external environment, e.g., a virtual world such as ScienceWorld, to perform complex tasks, e.g., growing a plant, without the startup costs of reinforcement learning. However, despite their zero-shot capabilities, these agents to date do not continually impro...
http://arxiv.org/pdf/2310.10134v1
2023-10-16T07:17:27
2023-10-16T07:17:27
{ "title": [ -0.0413208, -0.012191772, -0.04714966, -0.05517578, 0.0062217712, -0.0032234192, -0.03540039, 0.0066223145, 0.005176544, 0.03768921, -0.0071868896, -0.0029392242, -0.007247925, -0.012901306, -0.051757812, 0.006298065, -0.003332138, 0...
http://arxiv.org/abs/2310.10118v2
Learning to Rank Context for Named Entity Recognition Using a Synthetic Dataset
Arthur Amalvy, Vincent Labatut, Richard Dufour
cs.CL
While recent pre-trained transformer-based models can perform named entity recognition (NER) with great accuracy, their limited range remains an issue when applied to long documents such as whole novels. To alleviate this issue, a solution is to retrieve relevant context at the document level. Unfortunately, the lack o...
http://arxiv.org/pdf/2310.10118v2
2023-11-06T10:09:22
2023-10-16T06:53:12
{ "title": [ -0.0028438568, 0.012611389, -0.05606079, -0.06756592, 0.03439331, -0.0053634644, -0.022094727, -0.0020484924, -0.0004901886, 0.027450562, -0.045135498, 0.021087646, -0.008460999, 0.006526947, -0.030014038, -0.0143585205, -0.017959595, ...
http://arxiv.org/abs/2310.10106v1
End-to-end Multichannel Speaker-Attributed ASR: Speaker Guided Decoder and Input Feature Analysis
Can Cui, Imran Ahamad Sheikh, Mostafa Sadeghi, Emmanuel Vincent
cs.CL, cs.SD, eess.AS
We present an end-to-end multichannel speaker-attributed automatic speech recognition (MC-SA-ASR) system that combines a Conformer-based encoder with multi-frame crosschannel attention and a speaker-attributed Transformer-based decoder. To the best of our knowledge, this is the first model that efficiently integrates A...
http://arxiv.org/pdf/2310.10106v1
2023-10-16T06:40:18
2023-10-16T06:40:18
{ "title": [ -0.015899658, 0.006500244, -0.05606079, -0.07299805, -0.0013065338, -0.0049858093, -0.055786133, 0.014259338, -0.022872925, 0.028564453, 0.021270752, -0.0065307617, 0.034484863, -0.056671143, -0.0058670044, 0.0152282715, -0.0009455681, ...
http://arxiv.org/abs/2310.10103v1
Navigation with Large Language Models: Semantic Guesswork as a Heuristic for Planning
Dhruv Shah, Michael Equi, Blazej Osinski, Fei Xia, Brian Ichter, Sergey Levine
cs.RO, cs.AI, cs.CL, cs.LG
Navigation in unfamiliar environments presents a major challenge for robots: while mapping and planning techniques can be used to build up a representation of the world, quickly discovering a path to a desired goal in unfamiliar settings with such methods often requires lengthy mapping and exploration. Humans can rapid...
http://arxiv.org/pdf/2310.10103v1
2023-10-16T06:21:06
2023-10-16T06:21:06
{ "title": [ -0.03652954, -0.029876709, -0.021453857, -0.033843994, -0.022857666, -0.009941101, -0.032470703, -0.013046265, -0.008354187, 0.019088745, -0.035064697, 0.011810303, -0.039642334, -0.009735107, -0.023040771, -0.035736084, 0.0035381317, ...
http://arxiv.org/abs/2310.10094v1
Decomposed Prompt Tuning via Low-Rank Reparameterization
Yao Xiao, Lu Xu, Jiaxi Li, Wei Lu, Xiaoli Li
cs.CL
While prompt tuning approaches have achieved competitive performance with high efficiency, we observe that they invariably employ the same initialization process, wherein the soft prompt is either randomly initialized or derived from an existing embedding vocabulary. In contrast to these conventional methods, this stud...
http://arxiv.org/pdf/2310.10094v1
2023-10-16T05:56:06
2023-10-16T05:56:06
{ "title": [ -0.048828125, -0.011161804, -0.019958496, -0.05279541, -0.03765869, 0.027557373, -0.08538818, -0.017440796, -0.030899048, 0.025421143, -0.0012207031, -0.0031051636, 0.00043272972, -0.03945923, 0.0051879883, 0.007843018, 0.01424408, 0...
http://arxiv.org/abs/2310.10083v1
JMedLoRA:Medical Domain Adaptation on Japanese Large Language Models using Instruction-tuning
Issey Sukeda, Masahiro Suzuki, Hiroki Sakaji, Satoshi Kodera
cs.CL
In the ongoing wave of impact driven by large language models (LLMs) like ChatGPT, the adaptation of LLMs to medical domain has emerged as a crucial research frontier. Since mainstream LLMs tend to be designed for general-purpose applications, constructing a medical LLM through domain adaptation is a huge challenge. Wh...
http://arxiv.org/pdf/2310.10083v1
2023-10-16T05:28:28
2023-10-16T05:28:28
{ "title": [ -0.048034668, -0.016342163, -0.04333496, -0.066101074, -0.0045204163, 0.013809204, -0.0703125, -0.0020542145, -0.0063095093, 0.0284729, 0.026748657, -0.014297485, 0.019607544, -0.039276123, -0.033447266, 0.0076828003, -0.011810303, 0...
http://arxiv.org/abs/2310.10080v1
Let's reward step by step: Step-Level reward model as the Navigators for Reasoning
Qianli Ma, Haotian Zhou, Tingkai Liu, Jianbo Yuan, Pengfei Liu, Yang You, Hongxia Yang
cs.CL
Recent years have seen considerable advancements in multi-step reasoning with Large Language Models (LLMs). The previous studies have elucidated the merits of integrating feedback or search mechanisms during model inference to improve the reasoning accuracy. The Process-Supervised Reward Model (PRM), typically furnishe...
http://arxiv.org/pdf/2310.10080v1
2023-10-16T05:21:50
2023-10-16T05:21:50
{ "title": [ 0.017303467, -0.010147095, -0.03656006, -0.035308838, -0.0018339157, -0.024307251, -0.0063972473, -0.023468018, -0.034454346, -0.006881714, 0.013031006, -0.015380859, -0.030578613, -0.061462402, 0.010940552, -0.012374878, 0.0047073364, ...
http://arxiv.org/abs/2310.10077v1
Prompt Packer: Deceiving LLMs through Compositional Instruction with Hidden Attacks
Shuyu Jiang, Xingshu Chen, Rui Tang
cs.CL
Recently, Large language models (LLMs) with powerful general capabilities have been increasingly integrated into various Web applications, while undergoing alignment training to ensure that the generated content aligns with user intent and ethics. Unfortunately, they remain the risk of generating harmful content like h...
http://arxiv.org/pdf/2310.10077v1
2023-10-16T05:19:25
2023-10-16T05:19:25
{ "title": [ -0.051574707, -0.006614685, -0.05517578, -0.019073486, -0.030700684, -0.00983429, -0.059783936, -0.014770508, -0.05090332, 0.018920898, -0.0017261505, 0.032226562, -0.015823364, -0.042938232, 0.0035648346, -0.027908325, -0.023376465, ...
http://arxiv.org/abs/2310.10076v1
Verbosity Bias in Preference Labeling by Large Language Models
Keita Saito, Akifumi Wachi, Koki Wataoka, Youhei Akimoto
cs.CL, cs.AI
In recent years, Large Language Models (LLMs) have witnessed a remarkable surge in prevalence, altering the landscape of natural language processing and machine learning. One key factor in improving the performance of LLMs is alignment with humans achieved with Reinforcement Learning from Human Feedback (RLHF), as for ...
http://arxiv.org/pdf/2310.10076v1
2023-10-16T05:19:02
2023-10-16T05:19:02
{ "title": [ -0.006122589, -0.0017871857, -0.048339844, -0.053619385, 0.0011730194, -0.01878357, -0.019439697, -0.0066566467, 0.008430481, 0.020462036, -0.025009155, -0.012588501, -0.017852783, -0.018234253, -0.040985107, -0.013420105, -0.02468872, ...
http://arxiv.org/abs/2310.10698v2
Bridging Code Semantic and LLMs: Semantic Chain-of-Thought Prompting for Code Generation
Yingwei Ma, Yue Yu, Shanshan Li, Yu Jiang, Yong Guo, Yuanliang Zhang, Yutao Xie, Xiangke Liao
cs.CL
Large language models (LLMs) have showcased remarkable prowess in code generation. However, automated code generation is still challenging since it requires a high-level semantic mapping between natural language requirements and codes. Most existing LLMs-based approaches for code generation rely on decoder-only causal ...
http://arxiv.org/pdf/2310.10698v2
2023-10-22T10:21:12
2023-10-16T05:09:58
{ "title": [ -0.041259766, -0.036224365, -0.046173096, -0.042419434, -0.04788208, -0.00041913986, -0.062164307, 0.023971558, 0.003572464, 0.00067424774, -0.053863525, 0.017562866, -0.008094788, -0.027877808, 0.012886047, -0.023391724, -0.00023937225,...
http://arxiv.org/abs/2310.10072v2
Fine-tuning ChatGPT for Automatic Scoring
Ehsan Latif, Xiaoming Zhai
cs.CL, cs.AI
This study highlights the potential of fine-tuned ChatGPT (GPT-3.5) for automatically scoring student written constructed responses using example assessment tasks in science education. Recent studies on OpenAI's generative model GPT-3.5 proved its superiority in predicting the natural language with high accuracy and hu...
http://arxiv.org/pdf/2310.10072v2
2023-11-13T02:38:51
2023-10-16T05:09:16
{ "title": [ 0.0015468597, -0.007873535, -0.008308411, -0.031311035, -0.0009050369, 0.0037250519, -0.00017011166, -0.020721436, -0.009513855, 0.016693115, -0.024154663, 0.01121521, -0.028442383, -0.028793335, -0.021881104, 0.0046653748, -0.011924744,...
http://arxiv.org/abs/2310.10062v1
A Comprehensive Evaluation of Tool-Assisted Generation Strategies
Alon Jacovi, Avi Caciularu, Jonathan Herzig, Roee Aharoni, Bernd Bohnet, Mor Geva
cs.CL, cs.AI
A growing area of research investigates augmenting language models with tools (e.g., search engines, calculators) to overcome their shortcomings (e.g., missing or incorrect knowledge, incorrect logical inferences). Various few-shot tool-usage strategies have been proposed. However, there is no systematic and fair compa...
http://arxiv.org/pdf/2310.10062v1
2023-10-16T04:53:22
2023-10-16T04:53:22
{ "title": [ 0.009490967, -0.0017347336, -0.034973145, -0.04864502, -0.02281189, -0.017745972, -0.035217285, -0.00390625, -0.0068855286, 0.046295166, 0.041870117, -0.015853882, 0.024398804, -0.053131104, -0.01864624, 0.039123535, 0.011146545, 0.0...
http://arxiv.org/abs/2310.10054v1
NASH: A Simple Unified Framework of Structured Pruning for Accelerating Encoder-Decoder Language Models
Jongwoo Ko, Seungjoon Park, Yujin Kim, Sumyeong Ahn, Du-Seong Chang, Euijai Ahn, Se-Young Yun
cs.CL, cs.AI, cs.LG
Structured pruning methods have proven effective in reducing the model size and accelerating inference speed in various network architectures such as Transformers. Despite the versatility of encoder-decoder models in numerous NLP tasks, the structured pruning methods on such models are relatively less explored compared...
http://arxiv.org/pdf/2310.10054v1
2023-10-16T04:27:36
2023-10-16T04:27:36
{ "title": [ -0.048065186, -0.018554688, -0.037750244, -0.068603516, -0.020462036, 0.014839172, -0.07635498, 0.027801514, 0.012786865, 0.03250122, -0.0014257431, 0.0029125214, 0.035217285, -0.020736694, -0.032470703, 0.02407837, -0.0014667511, 0....
http://arxiv.org/abs/2310.10050v1
EfficientOCR: An Extensible, Open-Source Package for Efficiently Digitizing World Knowledge
Tom Bryan, Jacob Carlson, Abhishek Arora, Melissa Dell
cs.CV, cs.CL, econ.GN, q-fin.EC
Billions of public domain documents remain trapped in hard copy or lack an accurate digitization. Modern natural language processing methods cannot be used to index, retrieve, and summarize their texts; conduct computational textual analyses; or extract information for statistical analyses, and these texts cannot be in...
http://arxiv.org/pdf/2310.10050v1
2023-10-16T04:20:16
2023-10-16T04:20:16
{ "title": [ -0.04852295, -0.009887695, -0.06982422, -0.06762695, 0.011909485, -0.004447937, -0.030517578, 0.0006480217, 0.008728027, 0.026535034, -0.028686523, 0.018127441, 0.01171875, -0.020187378, -0.029403687, 0.008293152, 0.02003479, 0.03689...
http://arxiv.org/abs/2310.10047v1
Improving Large Language Model Fine-tuning for Solving Math Problems
Yixin Liu, Avi Singh, C. Daniel Freeman, John D. Co-Reyes, Peter J. Liu
cs.CL
Despite their success in many natural language tasks, solving math problems remains a significant challenge for large language models (LLMs). A large gap exists between LLMs' pass-at-one and pass-at-N performance in solving math problems, suggesting LLMs might be close to finding correct solutions, motivating our explo...
http://arxiv.org/pdf/2310.10047v1
2023-10-16T04:11:19
2023-10-16T04:11:19
{ "title": [ -0.016708374, -0.009506226, -0.018447876, -0.05053711, -0.0020751953, -0.0090408325, -0.023223877, 0.008804321, -0.0026187897, 0.035491943, -0.015655518, 0.006038666, -0.0060310364, -0.0052337646, -0.042114258, 0.0028762817, -0.010620117...
http://arxiv.org/abs/2310.10035v1
Empirical Study of Zero-Shot NER with ChatGPT
Tingyu Xie, Qi Li, Jian Zhang, Yan Zhang, Zuozhu Liu, Hongwei Wang
cs.CL
Large language models (LLMs) exhibited powerful capability in various natural language processing tasks. This work focuses on exploring LLM performance on zero-shot information extraction, with a focus on the ChatGPT and named entity recognition (NER) task. Inspired by the remarkable reasoning capability of LLM on symb...
http://arxiv.org/pdf/2310.10035v1
2023-10-16T03:40:03
2023-10-16T03:40:03
{ "title": [ 0.019882202, -0.013282776, -0.0013723373, -0.010139465, -0.029525757, -0.0074272156, 0.005706787, -0.011489868, -0.0053100586, 0.017059326, -0.014701843, 0.0127334595, 0.011024475, -0.048431396, 0.012817383, 0.019515991, -0.019195557, ...
http://arxiv.org/abs/2310.09983v1
Farzi Data: Autoregressive Data Distillation
Noveen Sachdeva, Zexue He, Wang-Cheng Kang, Jianmo Ni, Derek Zhiyuan Cheng, Julian McAuley
cs.LG, cs.AI, cs.CL, cs.IR
We study data distillation for auto-regressive machine learning tasks, where the input and output have a strict left-to-right causal structure. More specifically, we propose Farzi, which summarizes an event sequence dataset into a small number of synthetic sequences -- Farzi Data -- which are optimized to maintain (if ...
http://arxiv.org/pdf/2310.09983v1
2023-10-15T23:23:27
2023-10-15T23:23:27
{ "title": [ -0.014656067, 0.00667572, -0.06707764, -0.039520264, -0.0018787384, -0.021911621, -0.05706787, -0.0121536255, 0.0026187897, 0.03918457, -0.055725098, -0.03488159, 0.024612427, -0.044281006, -0.034118652, -0.031707764, 0.028823853, 0....
http://arxiv.org/abs/2310.09949v2
Chameleon: a heterogeneous and disaggregated accelerator system for retrieval-augmented language models
Wenqi Jiang, Marco Zeller, Roger Waleffe, Torsten Hoefler, Gustavo Alonso
cs.LG, cs.AI, cs.AR, cs.CL
A Retrieval-Augmented Language Model (RALM) augments a generative language model by retrieving context-specific knowledge from an external database. This strategy facilitates impressive text generation quality even with smaller models, thus reducing orders of magnitude of computational demands. However, RALMs introduce...
http://arxiv.org/pdf/2310.09949v2
2023-11-09T18:23:59
2023-10-15T20:57:25
{ "title": [ -0.0116119385, -0.0015268326, -0.04623413, -0.054107666, -0.0023994446, -0.0020389557, -0.04763794, 0.00033926964, 0.015151978, 0.05038452, -0.007888794, 0.0013780594, 0.024490356, -0.011154175, -0.05606079, 0.018493652, -0.022384644, ...
http://arxiv.org/abs/2310.09946v1
UvA-MT's Participation in the WMT23 General Translation Shared Task
Di Wu, Shaomu Tan, David Stap, Ali Araabi, Christof Monz
cs.CL, cs.LG
This paper describes the UvA-MT's submission to the WMT 2023 shared task on general machine translation. We participate in the constrained track in two directions: English <-> Hebrew. In this competition, we show that by using one model to handle bidirectional tasks, as a minimal setting of Multilingual Machine Transla...
http://arxiv.org/pdf/2310.09946v1
2023-10-15T20:49:31
2023-10-15T20:49:31
{ "title": [ -0.004512787, -0.0065689087, -0.033294678, -0.048309326, -0.01991272, -0.0009822845, -0.04043579, 0.014060974, -0.012939453, 0.066711426, 0.019943237, -0.013954163, 0.0690918, -0.04901123, 0.0036945343, 0.029586792, 0.012023926, 0.04...
http://arxiv.org/abs/2310.09930v1
FiLM: Fill-in Language Models for Any-Order Generation
Tianxiao Shen, Hao Peng, Ruoqi Shen, Yao Fu, Zaid Harchaoui, Yejin Choi
cs.CL
Language models have become the backbone of today's AI systems. However, their predominant left-to-right generation limits the use of bidirectional context, which is essential for tasks that involve filling text in the middle. We propose the Fill-in Language Model (FiLM), a new language modeling approach that allows fo...
http://arxiv.org/pdf/2310.09930v1
2023-10-15T19:37:39
2023-10-15T19:37:39
{ "title": [ -0.03289795, -0.012718201, -0.042114258, -0.09320068, 0.011985779, 0.02684021, -0.08026123, 0.016311646, 0.0071411133, 0.03475952, -0.0135650635, 0.004398346, 0.001865387, -0.022079468, -0.036376953, 0.005115509, -0.018585205, 0.0243...
http://arxiv.org/abs/2310.09929v1
Prompting Scientific Names for Zero-Shot Species Recognition
Shubham Parashar, Zhiqiu Lin, Yanan Li, Shu Kong
cs.CV, cs.CL
Trained on web-scale image-text pairs, Vision-Language Models (VLMs) such as CLIP can recognize images of common objects in a zero-shot fashion. However, it is underexplored how to use CLIP for zero-shot recognition of highly specialized concepts, e.g., species of birds, plants, and animals, for which their scientific ...
http://arxiv.org/pdf/2310.09929v1
2023-10-15T19:36:43
2023-10-15T19:36:43
{ "title": [ 0.011505127, -0.016586304, -0.01637268, 0.008628845, -0.058654785, -0.039855957, 0.016235352, -0.019470215, 0.006462097, 0.019882202, -0.026901245, 0.009407043, -0.01574707, -0.040252686, -0.0037384033, -0.005870819, 0.025146484, 0.0...
http://arxiv.org/abs/2310.09925v1
Homophone Disambiguation Reveals Patterns of Context Mixing in Speech Transformers
Hosein Mohebbi, Grzegorz Chrupała, Willem Zuidema, Afra Alishahi
cs.CL, cs.AI, cs.LG
Transformers have become a key architecture in speech processing, but our understanding of how they build up representations of acoustic and linguistic structure is limited. In this study, we address this gap by investigating how measures of 'context-mixing' developed for text models can be adapted and applied to model...
http://arxiv.org/pdf/2310.09925v1
2023-10-15T19:24:13
2023-10-15T19:24:13
{ "title": [ 0.010513306, 0.021133423, -0.04928589, -0.0602417, 0.020065308, -0.015914917, -0.030532837, 0.009399414, -0.047027588, 0.07550049, -0.024124146, 0.030853271, 0.047729492, -0.009414673, -0.016342163, 0.006515503, -0.011802673, 0.04608...
http://arxiv.org/abs/2310.09917v1
Empirical study of pretrained multilingual language models for zero-shot cross-lingual generation
Nadezhda Chirkova, Sheng Liang, Vassilina Nikoulina
cs.CL
Zero-shot cross-lingual generation assumes finetuning the multilingual pretrained language model (mPLM) on a generation task in one language and then using it to make predictions for this task in other languages. Previous works notice a frequent problem of generation in a wrong language and propose approaches to addres...
http://arxiv.org/pdf/2310.09917v1
2023-10-15T18:58:53
2023-10-15T18:58:53
{ "title": [ -0.017791748, -0.01689148, -0.0335083, -0.056762695, -0.014007568, 0.002149582, -0.06964111, 0.0057144165, 0.004600525, 0.050323486, 0.011734009, 0.006614685, 0.037384033, -0.02998352, -0.02746582, 0.030029297, -0.018508911, 0.060455...
http://arxiv.org/abs/2310.09909v2
Can GPT-4V(ision) Serve Medical Applications? Case Studies on GPT-4V for Multimodal Medical Diagnosis
Chaoyi Wu, Jiayu Lei, Qiaoyu Zheng, Weike Zhao, Weixiong Lin, Xiaoman Zhang, Xiao Zhou, Ziheng Zhao, Ya Zhang, Yanfeng Wang, Weidi Xie
cs.CV, cs.CL
Driven by the large foundation models, the development of artificial intelligence has witnessed tremendous progress lately, leading to a surge of general interest from the public. In this study, we aim to assess the performance of OpenAI's newest model, GPT-4V(ision), specifically in the realm of multimodal medical dia...
http://arxiv.org/pdf/2310.09909v2
2023-10-17T03:41:09
2023-10-15T18:32:27
{ "title": [ 0.0030899048, 0.020690918, -0.025543213, -0.023666382, 0.0054473877, -0.0036678314, -0.026168823, -0.020645142, -0.0138168335, 0.035736084, 0.0020713806, -0.022384644, 0.023956299, -0.044952393, -0.019592285, -0.008155823, -0.0060424805,...
http://arxiv.org/abs/2310.09897v1
Reformulating NLP tasks to Capture Longitudinal Manifestation of Language Disorders in People with Dementia
Dimitris Gkoumas, Matthew Purver, Maria Liakata
cs.CL
Dementia is associated with language disorders which impede communication. Here, we automatically learn linguistic disorder patterns by making use of a moderately-sized pre-trained language model and forcing it to focus on reformulated natural language processing (NLP) tasks and associated linguistic patterns. Our expe...
http://arxiv.org/pdf/2310.09897v1
2023-10-15T17:58:47
2023-10-15T17:58:47
{ "title": [ -0.014694214, -0.0014381409, -0.059753418, -0.045074463, -0.0049209595, -0.022460938, -0.031311035, 0.0082092285, 0.0068588257, 0.011947632, -0.043701172, 0.016906738, -0.007259369, -0.02331543, -0.023757935, -0.0053367615, -0.027786255,...
http://arxiv.org/abs/2310.09886v2
Lifelong Sequence Generation with Dynamic Module Expansion and Adaptation
Chengwei Qin, Chen Chen, Shafiq Joty
cs.CL, cs.AI
Lifelong sequence generation (LSG), a problem in continual learning, aims to continually train a model on a sequence of generation tasks to learn constantly emerging new generation patterns while avoiding the forgetting of previous knowledge. Existing LSG methods mainly focus on maintaining old knowledge while paying l...
http://arxiv.org/pdf/2310.09886v2
2023-10-20T07:37:48
2023-10-15T16:51:11
{ "title": [ -0.027877808, 0.00065374374, -0.082214355, -0.09820557, 0.018829346, -0.0044784546, -0.066223145, 0.02923584, 0.0034427643, 0.042388916, -0.023117065, 0.032958984, -0.023269653, -0.009231567, -0.04321289, 0.017913818, 0.0074386597, 0...
http://arxiv.org/abs/2310.09881v2
In-Context Learning with Iterative Demonstration Selection
Chengwei Qin, Aston Zhang, Anirudh Dagar, Wenming Ye
cs.CL, cs.AI
Spurred by advancements in scale, large language models (LLMs) have demonstrated strong few-shot learning ability via in-context learning (ICL). However, the performance of ICL has been shown to be highly sensitive to the selection of few-shot demonstrations. Selecting the most suitable examples as context remains an o...
http://arxiv.org/pdf/2310.09881v2
2023-10-22T13:40:02
2023-10-15T16:40:19
{ "title": [ 0.01083374, 0.023971558, -0.050750732, -0.06378174, 0.030838013, -0.02973938, -0.019638062, -0.020187378, 0.0051841736, 0.019866943, -0.053527832, 0.024215698, -0.030838013, -0.0021152496, -0.021835327, -0.016815186, -0.0027828217, 0...
http://arxiv.org/abs/2310.09876v1
Bounding and Filling: A Fast and Flexible Framework for Image Captioning
Zheng Ma, Changxin Wang, Bo Huang, Zixuan Zhu, Jianbing Zhang
cs.CV, cs.CL
Most image captioning models following an autoregressive manner suffer from significant inference latency. Several models adopted a non-autoregressive manner to speed up the process. However, the vanilla non-autoregressive manner results in subpar performance, since it generates all words simultaneously, which fails to...
http://arxiv.org/pdf/2310.09876v1
2023-10-15T16:17:20
2023-10-15T16:17:20
{ "title": [ 0.0071868896, 0.00032138824, -0.009109497, -0.058044434, 0.016296387, -0.0013122559, -0.016677856, -0.031921387, 0.019638062, 0.0524292, -0.043121338, 0.033172607, 0.0012435913, 0.013786316, -0.051971436, 0.003932953, -0.0033359528, ...
http://arxiv.org/abs/2310.09848v1
Enhancing Stance Classification with Quantified Moral Foundations
Hong Zhang, Prasanta Bhattacharya, Wei Gao, Liang Ze Wong, Brandon Siyuan Loh, Joseph J. P. Simons, Jisun An
cs.CL
This study enhances stance detection on social media by incorporating deeper psychological attributes, specifically individuals' moral foundations. These theoretically-derived dimensions aim to provide a comprehensive profile of an individual's moral concerns which, in recent work, has been linked to behaviour in a ran...
http://arxiv.org/pdf/2310.09848v1
2023-10-15T14:40:57
2023-10-15T14:40:57
{ "title": [ 0.005882263, 0.017974854, -0.04248047, -0.055541992, 0.014503479, -0.02519226, -0.022155762, -0.022369385, 0.018188477, 0.026779175, -0.040008545, 0.003982544, -0.033233643, -0.003452301, -0.052246094, -0.021499634, -0.0054397583, 0....
http://arxiv.org/abs/2310.09832v2
Merging Experts into One: Improving Computational Efficiency of Mixture of Experts
Shwai He, Run-Ze Fan, Liang Ding, Li Shen, Tianyi Zhou, Dacheng Tao
cs.CL
Scaling the size of language models usually leads to remarkable advancements in NLP tasks. But it often comes with a price of growing computational cost. Although a sparse Mixture of Experts (MoE) can reduce the cost by activating a small subset of parameters (e.g., one expert) for each input, its computation escalates...
http://arxiv.org/pdf/2310.09832v2
2023-10-22T19:39:58
2023-10-15T13:28:42
{ "title": [ -0.027328491, 0.004047394, -0.05630493, -0.087524414, 0.022094727, -0.00065279007, -0.026992798, 0.0061798096, 0.016830444, 0.023635864, -0.01184845, 0.013221741, -0.0028705597, -0.013870239, -0.029632568, 0.018310547, 0.0062446594, ...
http://arxiv.org/abs/2310.10690v1
Large Language Models for In-Context Student Modeling: Synthesizing Student's Behavior in Visual Programming from One-Shot Observation
Manh Hung Nguyen, Sebastian Tschiatschek, Adish Singla
cs.CL, cs.AI
Student modeling is central to many educational technologies as it enables the prediction of future learning outcomes and targeted instructional strategies. However, open-ended learning environments pose challenges for accurately modeling students due to the diverse behaviors exhibited by students and the absence of a ...
http://arxiv.org/pdf/2310.10690v1
2023-10-15T12:56:13
2023-10-15T12:56:13
{ "title": [ -0.010063171, -0.024642944, -0.00868988, -0.029693604, -0.030044556, -0.011772156, -0.023406982, -0.007106781, 0.00793457, 0.029144287, -0.042510986, -0.01626587, 0.0023593903, -0.012458801, -0.044525146, -0.02960205, -0.012313843, 0...
http://arxiv.org/abs/2310.09820v1
Assessing the Reliability of Large Language Model Knowledge
Weixuan Wang, Barry Haddow, Alexandra Birch, Wei Peng
cs.CL
Large language models (LLMs) have been treated as knowledge bases due to their strong performance in knowledge probing tasks. LLMs are typically evaluated using accuracy, yet this metric does not capture the vulnerability of LLMs to hallucination-inducing factors like prompt and context variability. How do we evaluate ...
http://arxiv.org/pdf/2310.09820v1
2023-10-15T12:40:30
2023-10-15T12:40:30
{ "title": [ -0.032440186, -0.018432617, -0.049316406, -0.082214355, -0.0031261444, 0.0009737015, -0.07513428, 0.008621216, 0.011528015, 0.022918701, -0.01499939, -0.011993408, 0.00006687641, -0.025390625, -0.052825928, 0.006038666, -0.0113220215, ...
http://arxiv.org/abs/2310.09773v1
RSVP: Customer Intent Detection via Agent Response Contrastive and Generative Pre-Training
Yu-Chien Tang, Wei-Yao Wang, An-Zi Yen, Wen-Chih Peng
cs.CL
The dialogue systems in customer services have been developed with neural models to provide users with precise answers and round-the-clock support in task-oriented conversations by detecting customer intents based on their utterances. Existing intent detection approaches have highly relied on adaptively pre-training la...
http://arxiv.org/pdf/2310.09773v1
2023-10-15T08:21:38
2023-10-15T08:21:38
{ "title": [ 0.03189087, 0.004627228, -0.01876831, -0.036010742, 0.009117126, -0.028411865, -0.0047798157, -0.049316406, 0.025665283, 0.016342163, -0.039276123, 0.014602661, 0.010368347, -0.014297485, -0.030578613, -0.010887146, -0.016342163, 0.0...
http://arxiv.org/abs/2310.09772v1
Revisiting Graph Meaning Representations through Decoupling Contextual Representation Learning and Structural Information Propagation
Li Zhou, Wenyu Chen, Dingyi Zeng, Hong Qu, Daniel Hershcovich
cs.CL
In the field of natural language understanding, the intersection of neural models and graph meaning representations (GMRs) remains a compelling area of research. Despite the growing interest, a critical gap persists in understanding the exact influence of GMRs, particularly concerning relation extraction tasks. Address...
http://arxiv.org/pdf/2310.09772v1
2023-10-15T08:07:56
2023-10-15T08:07:56
{ "title": [ -0.012802124, -0.011619568, -0.030303955, -0.031463623, -0.019989014, -0.0009860992, -0.043884277, -0.009017944, 0.0010118484, 0.032043457, 0.0026741028, 0.0025081635, 0.003358841, -0.035949707, -0.026916504, 0.0118637085, -0.0015687943,...
http://arxiv.org/abs/2310.09767v1
VLIS: Unimodal Language Models Guide Multimodal Language Generation
Jiwan Chung, Youngjae Yu
cs.CL, cs.AI
Multimodal language generation, which leverages the synergy of language and vision, is a rapidly expanding field. However, existing vision-language models face challenges in tasks that require complex linguistic understanding. To address this issue, we introduce Visual-Language models as Importance Sampling weights (VL...
http://arxiv.org/pdf/2310.09767v1
2023-10-15T07:58:52
2023-10-15T07:58:52
{ "title": [ -0.047027588, -0.011940002, -0.05227661, -0.04827881, -0.02079773, 0.009407043, -0.076538086, 0.0023899078, -0.017425537, 0.04071045, 0.014015198, -0.00020742416, 0.026794434, -0.02810669, -0.03125, 0.016464233, -0.009933472, 0.04931...
http://arxiv.org/abs/2310.09765v1
Improving Access to Justice for the Indian Population: A Benchmark for Evaluating Translation of Legal Text to Indian Languages
Sayan Mahapatra, Debtanu Datta, Shubham Soni, Adrijit Goswami, Saptarshi Ghosh
cs.CL, cs.AI
Most legal text in the Indian judiciary is written in complex English due to historical reasons. However, only about 10% of the Indian population is comfortable in reading English. Hence legal text needs to be made available in various Indian languages, possibly by translating the available legal text from English. Tho...
http://arxiv.org/pdf/2310.09765v1
2023-10-15T07:49:56
2023-10-15T07:49:56
{ "title": [ -0.05581665, -0.0026245117, -0.030883789, -0.009933472, -0.030303955, 0.0118255615, -0.053771973, 0.0027179718, 0.008163452, 0.034179688, -0.01663208, 0.028411865, 0.03213501, -0.017944336, -0.030273438, 0.0024757385, -0.0053215027, ...
http://arxiv.org/abs/2310.09762v1
Diversifying the Mixture-of-Experts Representation for Language Models with Orthogonal Optimizer
Boan Liu, Liang Ding, Li Shen, Keqin Peng, Yu Cao, Dazhao Cheng, Dacheng Tao
cs.CL, cs.AI
The Mixture of Experts (MoE) has emerged as a highly successful technique in deep learning, based on the principle of divide-and-conquer to maximize model capacity without significant additional computational cost. Even in the era of large-scale language models (LLMs), MoE continues to play a crucial role, as some rese...
http://arxiv.org/pdf/2310.09762v1
2023-10-15T07:20:28
2023-10-15T07:20:28
{ "title": [ -0.043792725, 0.005672455, -0.04675293, -0.055877686, -0.0037975311, 0.0044822693, -0.07373047, 0.006542206, -0.013160706, 0.05319214, 0.010284424, 0.0034427643, 0.033477783, -0.020706177, -0.04348755, 0.015205383, -0.008468628, 0.03...
http://arxiv.org/abs/2310.09753v1
When can transformers reason with abstract symbols?
Enric Boix-Adsera, Omid Saremi, Emmanuel Abbe, Samy Bengio, Etai Littwin, Joshua Susskind
cs.CL, cs.AI, cs.LG
We investigate the capabilities of transformer large language models (LLMs) on relational reasoning tasks involving abstract symbols. Such tasks have long been studied in the neuroscience literature as fundamental building blocks for more complex abilities in programming, mathematics, and verbal reasoning. For (i) regr...
http://arxiv.org/pdf/2310.09753v1
2023-10-15T06:45:38
2023-10-15T06:45:38
{ "title": [ 0.028244019, 0.007423401, 0.0040740967, -0.0030117035, -0.00907135, 0.009269714, 0.0049057007, -0.039794922, -0.015571594, 0.032043457, -0.019195557, -0.0289917, -0.012016296, -0.032348633, -0.0008058548, -0.023284912, 0.028182983, 0...
http://arxiv.org/abs/2310.09748v1
Large Language Model-Aware In-Context Learning for Code Generation
Jia Li, Ge Li, Chongyang Tao, Jia Li, Huangzhao Zhang, Fang Liu, Zhi Jin
cs.SE, cs.CL
Large language models (LLMs) have shown impressive in-context learning (ICL) ability in code generation. LLMs take a prompt consisting of requirement-code examples and a new requirement as input, and output new programs. Existing studies have found that ICL is highly dominated by the examples and thus arises research o...
http://arxiv.org/pdf/2310.09748v1
2023-10-15T06:12:58
2023-10-15T06:12:58
{ "title": [ -0.015365601, -0.028045654, -0.040985107, -0.06726074, -0.0041542053, 0.0025901794, -0.037322998, 0.036743164, 0.011230469, 0.026565552, -0.03930664, 0.009124756, 0.0051841736, -0.010475159, -0.01939392, -0.0028018951, -0.008880615, ...
http://arxiv.org/abs/2310.09736v1
Domain-Specific Language Model Post-Training for Indonesian Financial NLP
Ni Putu Intan Maharani, Yoga Yustiawan, Fauzy Caesar Rochim, Ayu Purwarianti
cs.CL, cs.AI
BERT and IndoBERT have achieved impressive performance in several NLP tasks. There has been several investigation on its adaption in specialized domains especially for English language. We focus on financial domain and Indonesian language, where we perform post-training on pre-trained IndoBERT for financial domain usin...
http://arxiv.org/pdf/2310.09736v1
2023-10-15T05:07:08
2023-10-15T05:07:08
{ "title": [ -0.003932953, -0.008285522, -0.041900635, -0.050628662, -0.021102905, -0.012458801, -0.057006836, 0.012397766, -0.027908325, 0.032806396, 0.002374649, -0.00983429, 0.015197754, -0.031982422, -0.024993896, -0.006717682, -0.028121948, ...
http://arxiv.org/abs/2310.12172v2
Overview of ImageArg-2023: The First Shared Task in Multimodal Argument Mining
Zhexiong Liu, Mohamed Elaraby, Yang Zhong, Diane Litman
cs.CL
This paper presents an overview of the ImageArg shared task, the first multimodal Argument Mining shared task co-located with the 10th Workshop on Argument Mining at EMNLP 2023. The shared task comprises two classification subtasks - (1) Subtask-A: Argument Stance Classification; (2) Subtask-B: Image Persuasiveness Cla...
http://arxiv.org/pdf/2310.12172v2
2023-10-24T14:58:13
2023-10-15T04:52:04
{ "title": [ 0.033050537, -0.010391235, -0.014228821, -0.066711426, -0.0015869141, -0.0075912476, -0.03616333, -0.013824463, -0.01979065, 0.021896362, 0.0060272217, 0.03048706, -0.03845215, -0.029922485, -0.01448822, 0.003129959, -0.013656616, 0....
http://arxiv.org/abs/2310.09725v2
KGQuiz: Evaluating the Generalization of Encoded Knowledge in Large Language Models
Yuyang Bai, Shangbin Feng, Vidhisha Balachandran, Zhaoxuan Tan, Shiqi Lou, Tianxing He, Yulia Tsvetkov
cs.CL
Large language models (LLMs) demonstrate remarkable performance on knowledge-intensive tasks, suggesting that real-world knowledge is encoded in their model parameters. However, besides explorations on a few probing tasks in limited knowledge domains, it is not well understood how to evaluate LLMs' knowledge systematic...
http://arxiv.org/pdf/2310.09725v2
2023-10-24T19:14:23
2023-10-15T04:00:36
{ "title": [ -0.05267334, -0.027236938, -0.045135498, -0.053741455, -0.013435364, 0.007659912, -0.056915283, 0.008621216, 0.012306213, 0.04534912, -0.015808105, 0.015098572, 0.012573242, 0.004611969, -0.063964844, 0.011878967, -0.012802124, 0.051...
http://arxiv.org/abs/2310.09720v1
HiCL: Hierarchical Contrastive Learning of Unsupervised Sentence Embeddings
Zhuofeng Wu, Chaowei Xiao, VG Vinod Vydiswaran
cs.CL, cs.LG
In this paper, we propose a hierarchical contrastive learning framework, HiCL, which considers local segment-level and global sequence-level relationships to improve training efficiency and effectiveness. Traditional methods typically encode a sequence in its entirety for contrast with others, often neglecting local re...
http://arxiv.org/pdf/2310.09720v1
2023-10-15T03:14:33
2023-10-15T03:14:33
{ "title": [ -0.0026664734, 0.0054740906, -0.032836914, -0.060546875, 0.006690979, -0.0009188652, -0.034423828, -0.023834229, 0.026016235, 0.025253296, -0.022537231, 0.03378296, 0.005760193, -0.0093307495, -0.036987305, 0.010887146, -0.025558472, ...
http://arxiv.org/abs/2310.09716v2
Enhancing Conversational Search: Large Language Model-Aided Informative Query Rewriting
Fanghua Ye, Meng Fang, Shenghui Li, Emine Yilmaz
cs.HC, cs.AI, cs.CL, cs.IR
Query rewriting plays a vital role in enhancing conversational search by transforming context-dependent user queries into standalone forms. Existing approaches primarily leverage human-rewritten queries as labels to train query rewriting models. However, human rewrites may lack sufficient information for optimal retrie...
http://arxiv.org/pdf/2310.09716v2
2023-10-18T13:48:03
2023-10-15T03:04:17
{ "title": [ -0.004611969, -0.015670776, -0.017349243, -0.08831787, -0.0061950684, 0.025238037, -0.08215332, -0.0072288513, 0.032043457, 0.042907715, -0.020187378, -0.005016327, 0.023788452, -0.0056495667, -0.053222656, 0.011474609, -0.008094788, ...
http://arxiv.org/abs/2310.09680v3
Improved Contextual Recognition In Automatic Speech Recognition Systems By Semantic Lattice Rescoring
Ankitha Sudarshan, Vinay Samuel, Parth Patwa, Ibtihel Amara, Aman Chadha
cs.CL, cs.AI, cs.SD, eess.AS
Automatic Speech Recognition (ASR) has witnessed a profound research interest. Recent breakthroughs have given ASR systems different prospects such as faithfully transcribing spoken language, which is a pivotal advancement in building conversational agents. However, there is still an imminent challenge of accurately di...
http://arxiv.org/pdf/2310.09680v3
2023-10-27T18:56:51
2023-10-14T23:16:05
{ "title": [ -0.0053634644, 0.008369446, -0.0541687, -0.07965088, 0.009109497, -0.0037345886, -0.051849365, 0.017700195, -0.0072631836, 0.074035645, -0.024536133, 0.010253906, 0.009635925, -0.0019760132, -0.06744385, 0.004924774, -0.004383087, 0....
http://arxiv.org/abs/2310.09675v1
Efficient Model-Agnostic Multi-Group Equivariant Networks
Razan Baltaji, Sourya Basu, Lav R. Varshney
cs.LG, cs.AI, cs.CL
Constructing model-agnostic group equivariant networks, such as equitune (Basu et al., 2023b) and its generalizations (Kim et al., 2023), can be computationally expensive for large product groups. We address this by providing efficient model-agnostic equivariant designs for two related problems: one where the network h...
http://arxiv.org/pdf/2310.09675v1
2023-10-14T22:24:26
2023-10-14T22:24:26
{ "title": [ -0.03463745, -0.007572174, -0.07659912, -0.061187744, 0.016998291, -0.01953125, -0.046417236, 0.020431519, 0.0010099411, 0.029418945, 0.006263733, 0.023620605, -0.0016679764, -0.018096924, -0.015686035, 0.02519226, -0.0019035339, 0.0...
http://arxiv.org/abs/2310.09668v1
Beyond Testers' Biases: Guiding Model Testing with Knowledge Bases using LLMs
Chenyang Yang, Rishabh Rustogi, Rachel Brower-Sinning, Grace A. Lewis, Christian Kästner, Tongshuang Wu
cs.CL, cs.SE
Current model testing work has mostly focused on creating test cases. Identifying what to test is a step that is largely ignored and poorly supported. We propose Weaver, an interactive tool that supports requirements elicitation for guiding model testing. Weaver uses large language models to generate knowledge bases an...
http://arxiv.org/pdf/2310.09668v1
2023-10-14T21:24:03
2023-10-14T21:24:03
{ "title": [ -0.05731201, -0.020019531, -0.041137695, -0.06149292, -0.0262146, 0.020004272, -0.08758545, -0.0025978088, 0.012161255, 0.0030651093, -0.062927246, 0.021255493, -0.024505615, -0.013900757, -0.024795532, -0.018295288, 0.0032672882, -0...
http://arxiv.org/abs/2310.09661v1
Legend at ArAIEval Shared Task: Persuasion Technique Detection using a Language-Agnostic Text Representation Model
Olumide E. Ojo, Olaronke O. Adebanji, Hiram Calvo, Damian O. Dieke, Olumuyiwa E. Ojo, Seye E. Akinsanya, Tolulope O. Abiola, Anna Feldman
cs.CL
In this paper, we share our best performing submission to the Arabic AI Tasks Evaluation Challenge (ArAIEval) at ArabicNLP 2023. Our focus was on Task 1, which involves identifying persuasion techniques in excerpts from tweets and news articles. The persuasion technique in Arabic texts was detected using a training loo...
http://arxiv.org/pdf/2310.09661v1
2023-10-14T20:27:04
2023-10-14T20:27:04
{ "title": [ 0.027648926, 0.021759033, -0.012840271, 0.021026611, -0.0022239685, -0.023498535, -0.0056381226, -0.0541687, -0.059661865, 0.046905518, -0.0060653687, 0.015029907, 0.007297516, -0.0413208, 0.0043296814, -0.02659607, -0.039123535, 0.1...
http://arxiv.org/abs/2310.09651v1
Lexical Entrainment for Conversational Systems
Zhengxiang Shi, Procheta Sen, Aldo Lipani
cs.CL, cs.AI, cs.LG
Conversational agents have become ubiquitous in assisting with daily tasks, and are expected to possess human-like features. One such feature is lexical entrainment (LE), a phenomenon in which speakers in human-human conversations tend to naturally and subconsciously align their lexical choices with those of their inte...
http://arxiv.org/pdf/2310.09651v1
2023-10-14T19:47:37
2023-10-14T19:47:37
{ "title": [ -0.015960693, 0.003364563, -0.036712646, -0.053649902, 0.004547119, 0.00061130524, -0.046173096, 0.009559631, 0.029541016, 0.037322998, -0.044555664, 0.013214111, 0.009941101, -0.0071029663, -0.04425049, 0.019958496, -0.007335663, 0....
http://arxiv.org/abs/2310.09634v1
An End-to-End System for Reproducibility Assessment of Source Code Repositories via Their Readmes
Eyüp Kaan Akdeniz, Selma Tekir, Malik Nizar Asad Al Hinnawi
cs.CL, cs.SE
Increased reproducibility of machine learning research has been a driving force for dramatic improvements in learning performances. The scientific community further fosters this effort by including reproducibility ratings in reviewer forms and considering them as a crucial factor for the overall evaluation of papers. A...
http://arxiv.org/pdf/2310.09634v1
2023-10-14T18:01:11
2023-10-14T18:01:11
{ "title": [ -0.021072388, -0.029907227, -0.024215698, -0.049621582, -0.025268555, -0.014053345, -0.03086853, 0.020767212, 0.030410767, 0.035614014, -0.027862549, 0.010070801, 0.021759033, -0.02659607, -0.025634766, 0.022994995, 0.000032246113, 0...
http://arxiv.org/abs/2310.09624v2
ASSERT: Automated Safety Scenario Red Teaming for Evaluating the Robustness of Large Language Models
Alex Mei, Sharon Levy, William Yang Wang
cs.CL, cs.AI, cs.LG
As large language models are integrated into society, robustness toward a suite of prompts is increasingly important to maintain reliability in a high-variance environment.Robustness evaluations must comprehensively encapsulate the various settings in which a user may invoke an intelligent system. This paper proposes A...
http://arxiv.org/pdf/2310.09624v2
2023-11-11T05:30:34
2023-10-14T17:10:28
{ "title": [ -0.013801575, -0.003929138, -0.048309326, -0.09979248, 0.008926392, -0.0008149147, -0.061553955, 0.013694763, -0.030090332, 0.010719299, -0.032806396, 0.032592773, -0.008880615, -0.032226562, 0.0002849102, -0.02268982, -0.021652222, ...
http://arxiv.org/abs/2310.09623v1
A Digital Language Coherence Marker for Monitoring Dementia
Dimitris Gkoumas, Adam Tsakalidis, Maria Liakata
cs.CL
The use of spontaneous language to derive appropriate digital markers has become an emergent, promising and non-intrusive method to diagnose and monitor dementia. Here we propose methods to capture language coherence as a cost-effective, human-interpretable digital marker for monitoring cognitive changes in people with...
http://arxiv.org/pdf/2310.09623v1
2023-10-14T17:10:19
2023-10-14T17:10:19
{ "title": [ -0.010627747, 0.001666069, -0.038970947, -0.045440674, -0.0018873215, -0.021194458, -0.038970947, -0.017807007, -0.017044067, 0.050750732, 0.0058937073, -0.012741089, 0.03363037, -0.03967285, -0.032196045, 0.007873535, -0.012161255, ...
http://arxiv.org/abs/2310.10688v1
A decoder-only foundation model for time-series forecasting
Abhimanyu Das, Weihao Kong, Rajat Sen, Yichen Zhou
cs.CL, cs.AI, cs.LG
Motivated by recent advances in large language models for Natural Language Processing (NLP), we design a time-series foundation model for forecasting whose out-of-the-box zero-shot performance on a variety of public datasets comes close to the accuracy of state-of-the-art supervised forecasting models for each individu...
http://arxiv.org/pdf/2310.10688v1
2023-10-14T17:01:37
2023-10-14T17:01:37
{ "title": [ -0.025146484, -0.005138397, -0.05026245, -0.073913574, 0.00037407875, -0.0030822754, -0.056732178, -0.006462097, 0.010192871, 0.044281006, -0.025970459, -0.0025482178, -0.012786865, -0.009010315, -0.06903076, 0.0034103394, 0.014175415, ...
http://arxiv.org/abs/2310.09619v3
An Expression Tree Decoding Strategy for Mathematical Equation Generation
Wenqi Zhang, Yongliang Shen, Qingpeng Nong, Zeqi Tan, Yanna Ma, Weiming Lu
cs.CL
Generating mathematical equations from natural language requires an accurate understanding of the relations among math expressions. Existing approaches can be broadly categorized into token-level and expression-level generation. The former treats equations as a mathematical language, sequentially generating math tokens...
http://arxiv.org/pdf/2310.09619v3
2023-10-18T09:28:54
2023-10-14T17:00:28
{ "title": [ -0.046966553, -0.003522873, -0.02470398, -0.040649414, -0.00869751, 0.013885498, -0.036376953, -0.0008034706, 0.02041626, 0.038208008, 0.020263672, -0.018951416, 0.012184143, -0.035949707, -0.048034668, 0.020645142, -0.0025081635, 0....
http://arxiv.org/abs/2310.09618v1
Moral consensus and divergence in partisan language use
Nakwon Rim, Marc G. Berman, Yuan Chang Leong
cs.CL
Polarization has increased substantially in political discourse, contributing to a widening partisan divide. In this paper, we analyzed large-scale, real-world language use in Reddit communities (294,476,146 comments) and in news outlets (6,749,781 articles) to uncover psychological dimensions along which partisan lang...
http://arxiv.org/pdf/2310.09618v1
2023-10-14T16:50:26
2023-10-14T16:50:26
{ "title": [ -0.0025405884, 0.010299683, -0.0473938, -0.016906738, -0.008361816, -0.040618896, 0.020721436, -0.004348755, -0.06616211, 0.052459717, -0.028625488, 0.015838623, -0.0026569366, -0.006290436, -0.025436401, -0.01473999, -0.00945282, 0....
http://arxiv.org/abs/2310.09596v1
RethinkingTMSC: An Empirical Study for Target-Oriented Multimodal Sentiment Classification
Junjie Ye, Jie Zhou, Junfeng Tian, Rui Wang, Qi Zhang, Tao Gui, Xuanjing Huang
cs.CL
Recently, Target-oriented Multimodal Sentiment Classification (TMSC) has gained significant attention among scholars. However, current multimodal models have reached a performance bottleneck. To investigate the causes of this problem, we perform extensive empirical evaluation and in-depth analysis of the datasets to an...
http://arxiv.org/pdf/2310.09596v1
2023-10-14T14:52:37
2023-10-14T14:52:37
{ "title": [ -0.003610611, 0.0002092123, -0.037078857, -0.04510498, 0.010444641, -0.016143799, -0.01928711, -0.00762558, 0.01234436, 0.03744507, -0.0124435425, 0.003074646, 0.0036258698, -0.010826111, -0.06048584, -0.0011882782, -0.018371582, 0.0...
http://arxiv.org/abs/2310.09590v1
Solving Math Word Problems with Reexamination
Yi Bin, Wenhao Shi, Yujuan Ding, Yang Yang, See-Kiong Ng
cs.CL, cs.AI
Math word problem (MWP) solving aims to understand the descriptive math problem and calculate the result, for which previous efforts are mostly devoted to upgrade different technical modules. This paper brings a different perspective of \textit{reexamination process} during training by introducing a pseudo-dual task to...
http://arxiv.org/pdf/2310.09590v1
2023-10-14T14:23:44
2023-10-14T14:23:44
{ "title": [ 0.019546509, -0.005470276, 0.006160736, -0.0020656586, -0.051696777, -0.014213562, -0.014137268, -0.030029297, -0.011367798, 0.03753662, -0.05126953, 0.0023040771, 0.0063095093, -0.013259888, 0.013954163, -0.00014770031, 0.02180481, ...
http://arxiv.org/abs/2310.10686v1
Autonomous Tree-search Ability of Large Language Models
Zheyu Zhang, Zhuorui Ye, Yikang Shen, Chuang Gan
cs.CL, cs.AI
Large Language Models have excelled in remarkable reasoning capabilities with advanced prompting techniques, but they fall short on tasks that require exploration, strategic foresight, and sequential decision-making. Recent works propose to utilize external programs to define search logic, such that LLMs can perform pa...
http://arxiv.org/pdf/2310.10686v1
2023-10-14T14:14:38
2023-10-14T14:14:38
{ "title": [ -0.034606934, -0.0067214966, -0.022109985, -0.059783936, -0.013809204, 0.008888245, -0.055145264, 0.01499939, 0.018875122, 0.039245605, -0.0284729, -0.001411438, 0.0014772415, -0.0070991516, -0.06536865, 0.009857178, -0.015625, 0.023...
http://arxiv.org/abs/2310.09573v1
Self-Detoxifying Language Models via Toxification Reversal
Chak Tou Leong, Yi Cheng, Jiashuo Wang, Jian Wang, Wenjie Li
cs.CL
Language model detoxification aims to minimize the risk of generating offensive or harmful content in pretrained language models (PLMs) for safer deployment. Existing methods can be roughly categorized as finetuning-based and decoding-based. However, the former is often resource-intensive, while the latter relies on ad...
http://arxiv.org/pdf/2310.09573v1
2023-10-14T12:51:38
2023-10-14T12:51:38
{ "title": [ 0.0032100677, 0.013618469, -0.038085938, -0.05380249, 0.009925842, -0.009010315, -0.023757935, -0.010467529, 0.008354187, 0.05038452, 0.019927979, -0.017181396, 0.015090942, -0.037506104, -0.042816162, 0.025482178, 0.0005121231, 0.04...
http://arxiv.org/abs/2311.03365v1
Leveraging Generative AI: Improving Software Metadata Classification with Generated Code-Comment Pairs
Samah Syed, Angel Deborah S
cs.SE, cs.AI, cs.CL, cs.LG
In software development, code comments play a crucial role in enhancing code comprehension and collaboration. This research paper addresses the challenge of objectively classifying code comments as "Useful" or "Not Useful." We propose a novel solution that harnesses contextualized embeddings, particularly BERT, to auto...
http://arxiv.org/pdf/2311.03365v1
2023-10-14T12:09:43
2023-10-14T12:09:43
{ "title": [ -0.020004272, -0.022949219, -0.032928467, -0.046875, -0.015357971, -0.02168274, -0.035949707, -0.0042381287, 0.018661499, 0.039245605, -0.027359009, 0.016860962, -0.010665894, -0.011909485, -0.052612305, -0.0062446594, -0.0209198, 0....
http://arxiv.org/abs/2310.13714v1
A study of the impact of generative AI-based data augmentation on software metadata classification
Tripti Kumari, Chakali Sai Charan, Ayan Das
cs.SE, cs.AI, cs.CL, cs.LG
This paper presents the system submitted by the team from IIT(ISM) Dhanbad in FIRE IRSE 2023 shared task 1 on the automatic usefulness prediction of code-comment pairs as well as the impact of Large Language Model(LLM) generated data on original base data towards an associated source code. We have developed a framework...
http://arxiv.org/pdf/2310.13714v1
2023-10-14T10:47:10
2023-10-14T10:47:10
{ "title": [ -0.010063171, -0.006252289, -0.05432129, -0.050231934, 0.00033783913, -0.03729248, -0.013023376, 0.0037498474, 0.0128479, 0.034484863, -0.0340271, 0.021484375, 0.007156372, -0.018829346, -0.037353516, 0.004459381, -0.016723633, 0.047...
http://arxiv.org/abs/2310.09550v1
Can Large Language Model Comprehend Ancient Chinese? A Preliminary Test on ACLUE
Yixuan Zhang, Haonan Li
cs.CL
Large language models (LLMs) have showcased remarkable capabilities in understanding and generating language. However, their ability in comprehending ancient languages, particularly ancient Chinese, remains largely unexplored. To bridge this gap, we present ACLUE, an evaluation benchmark designed to assess the capabili...
http://arxiv.org/pdf/2310.09550v1
2023-10-14T10:06:39
2023-10-14T10:06:39
{ "title": [ -0.027175903, 0.010543823, -0.0340271, -0.017608643, -0.008743286, -0.017852783, -0.03378296, -0.016220093, 0.019378662, 0.028701782, -0.012763977, -0.012451172, 0.028656006, -0.033355713, -0.040405273, 0.026748657, -0.013580322, 0.0...
http://arxiv.org/abs/2310.09536v1
CarExpert: Leveraging Large Language Models for In-Car Conversational Question Answering
Md Rashad Al Hasan Rony, Christian Suess, Sinchana Ramakanth Bhat, Viju Sudhi, Julia Schneider, Maximilian Vogel, Roman Teucher, Ken E. Friedl, Soumya Sahoo
cs.CL, cs.IR, cs.LG
Large language models (LLMs) have demonstrated remarkable performance by following natural language instructions without fine-tuning them on domain-specific tasks and data. However, leveraging LLMs for domain-specific question answering suffers from severe limitations. The generated answer tends to hallucinate due to t...
http://arxiv.org/pdf/2310.09536v1
2023-10-14T08:46:24
2023-10-14T08:46:24
{ "title": [ -0.02758789, -0.03265381, -0.038085938, -0.04385376, -0.010116577, -0.009857178, -0.045532227, 0.015312195, -0.0005931854, 0.027389526, -0.035858154, 0.012252808, 0.0011758804, -0.00674057, -0.04043579, -0.003583908, -0.008590698, 0....
http://arxiv.org/abs/2310.09520v3
Reward-Augmented Decoding: Efficient Controlled Text Generation With a Unidirectional Reward Model
Haikang Deng, Colin Raffel
cs.CL
While large language models have proven effective in a huge range of downstream applications, they often generate text that is problematic or lacks a desired attribute. In this paper, we introduce Reward-Augmented Decoding (RAD), a text generation procedure that uses a small unidirectional reward model to encourage a l...
http://arxiv.org/pdf/2310.09520v3
2023-10-26T20:04:47
2023-10-14T07:19:47
{ "title": [ -0.036895752, -0.012077332, -0.02357483, -0.034606934, -0.015281677, 0.013374329, -0.041503906, -0.026473999, -0.0052719116, 0.048461914, -0.0012159348, 0.018234253, 0.017562866, -0.010810852, -0.045654297, 0.012329102, -0.024536133, ...
http://arxiv.org/abs/2310.09518v1
Instruction Tuning with Human Curriculum
Bruce W. Lee, Hyunsoo Cho, Kang Min Yoo
cs.CL, cs.AI, cs.LG
The dominant paradigm for instruction tuning is the random-shuffled training of maximally diverse instruction-response pairs. This paper explores the potential benefits of applying a structured cognitive learning approach to instruction tuning in contemporary large language models like ChatGPT and GPT-4. Unlike the pre...
http://arxiv.org/pdf/2310.09518v1
2023-10-14T07:16:08
2023-10-14T07:16:08
{ "title": [ -0.03466797, -0.0059928894, -0.02003479, 0.0041122437, -0.055389404, -0.031051636, -0.035369873, -0.01184845, -0.04937744, 0.043121338, -0.033203125, 0.0069999695, -0.024002075, -0.013839722, 0.0016222, -0.024627686, 0.0043411255, 0....
http://arxiv.org/abs/2310.09512v1
Attentive Multi-Layer Perceptron for Non-autoregressive Generation
Shuyang Jiang, Jun Zhang, Jiangtao Feng, Lin Zheng, Lingpeng Kong
cs.CL
Autoregressive~(AR) generation almost dominates sequence generation for its efficacy. Recently, non-autoregressive~(NAR) generation gains increasing popularity for its efficiency and growing efficacy. However, its efficiency is still bottlenecked by quadratic complexity in sequence lengths, which is prohibitive for sca...
http://arxiv.org/pdf/2310.09512v1
2023-10-14T06:44:24
2023-10-14T06:44:24
{ "title": [ -0.013366699, -0.007545471, -0.03778076, -0.040527344, 0.00680542, -0.011756897, -0.022109985, -0.0050811768, 0.012062073, 0.026443481, -0.009460449, 0.0154953, -0.010635376, -0.022598267, -0.043884277, 0.0007839203, 0.008079529, 0.0...
http://arxiv.org/abs/2310.09501v1
DepNeCTI: Dependency-based Nested Compound Type Identification for Sanskrit
Jivnesh Sandhan, Yaswanth Narsupalli, Sreevatsa Muppirala, Sriram Krishnan, Pavankumar Satuluri, Amba Kulkarni, Pawan Goyal
cs.CL
Multi-component compounding is a prevalent phenomenon in Sanskrit, and understanding the implicit structure of a compound's components is crucial for deciphering its meaning. Earlier approaches in Sanskrit have focused on binary compounds and neglected the multi-component compound setting. This work introduces the nove...
http://arxiv.org/pdf/2310.09501v1
2023-10-14T06:11:53
2023-10-14T06:11:53
{ "title": [ -0.018920898, 0.009132385, -0.03942871, -0.115112305, 0.014152527, 0.032226562, -0.09692383, 0.01876831, -0.018692017, 0.00945282, -0.028945923, 0.017822266, -0.019470215, -0.021499634, -0.02861023, -0.018035889, -0.020324707, 0.0506...
http://arxiv.org/abs/2310.09499v1
One-Shot Sensitivity-Aware Mixed Sparsity Pruning for Large Language Models
Hang Shao, Bei Liu, Yanmin Qian
cs.CL, cs.AI
Various Large Language Models(LLMs) from the Generative Pretrained Transformer~(GPT) family have achieved outstanding performances in a wide range of text generation tasks. However, the enormous model sizes have hindered their practical use in real-world applications due to high inference latency. Therefore, improving ...
http://arxiv.org/pdf/2310.09499v1
2023-10-14T05:43:09
2023-10-14T05:43:09
{ "title": [ -0.027450562, -0.022659302, -0.0236969, -0.068847656, -0.020507812, 0.00756073, -0.07537842, 0.024963379, 0.042510986, 0.020935059, -0.01184845, 0.0040245056, 0.011627197, -0.03137207, -0.04272461, 0.006450653, 0.00027012825, 0.02461...
http://arxiv.org/abs/2310.09494v1
Computational analyses of linguistic features with schizophrenic and autistic traits along with formal thought disorders
Takeshi Saga, Hiroki Tanaka, Satoshi Nakamura
cs.CL, J.4; J.3; I.2.1; I.2.7
[See full abstract in the pdf] Formal Thought Disorder (FTD), which is a group of symptoms in cognition that affects language and thought, can be observed through language. FTD is seen across such developmental or psychiatric disorders as Autism Spectrum Disorder (ASD) or Schizophrenia, and its related Schizotypal Pers...
http://arxiv.org/pdf/2310.09494v1
2023-10-14T05:05:11
2023-10-14T05:05:11
{ "title": [ -0.012931824, 0.006099701, -0.021881104, -0.020614624, -0.004096985, -0.026245117, -0.014160156, -0.009132385, 0.0060653687, 0.03375244, 0.023635864, -0.02796936, 0.038635254, -0.05831909, -0.013450623, 0.016235352, -0.026473999, 0.0...
http://arxiv.org/abs/2310.10683v1
Large Language Model Unlearning
Yuanshun Yao, Xiaojun Xu, Yang Liu
cs.CL, cs.AI, cs.LG
We study how to perform unlearning, i.e. forgetting undesirable (mis)behaviors, on large language models (LLMs). We show at least three scenarios of aligning LLMs with human preferences can benefit from unlearning: (1) removing harmful responses, (2) erasing copyright-protected content as requested, and (3) eliminating...
http://arxiv.org/pdf/2310.10683v1
2023-10-14T00:32:55
2023-10-14T00:32:55
{ "title": [ -0.01121521, -0.000849247, -0.04537964, -0.03353882, -0.03817749, -0.016952515, -0.041015625, 0.004787445, 0.021453857, 0.05670166, -0.03060913, -0.0050239563, 0.036743164, -0.03540039, -0.006767273, 0.024780273, 0.008644104, 0.04647...
http://arxiv.org/abs/2310.09436v1
Sub-network Discovery and Soft-masking for Continual Learning of Mixed Tasks
Zixuan Ke, Bing Liu, Wenhan Xiong, Asli Celikyilmaz, Haoran Li
cs.CL, cs.AI, cs.LG, cs.NE
Continual learning (CL) has two main objectives: preventing catastrophic forgetting (CF) and encouraging knowledge transfer (KT). The existing literature mainly focused on overcoming CF. Some work has also been done on KT when the tasks are similar. To our knowledge, only one method has been proposed to learn a sequenc...
http://arxiv.org/pdf/2310.09436v1
2023-10-13T23:00:39
2023-10-13T23:00:39
{ "title": [ -0.06201172, 0.009178162, -0.07611084, -0.007411957, -0.0053482056, -0.03616333, -0.011734009, 0.012756348, -0.04296875, 0.028533936, -0.020401001, 0.018508911, -0.010871887, -0.033355713, 0.0031986237, -0.018936157, 0.005191803, 0.0...
http://arxiv.org/abs/2310.09432v1
Enhancing BERT-Based Visual Question Answering through Keyword-Driven Sentence Selection
Davide Napolitano, Lorenzo Vaiani, Luca Cagliero
cs.CL, cs.AI, cs.LG
The Document-based Visual Question Answering competition addresses the automatic detection of parent-child relationships between elements in multi-page documents. The goal is to identify the document elements that answer a specific question posed in natural language. This paper describes the PoliTo's approach to addres...
http://arxiv.org/pdf/2310.09432v1
2023-10-13T22:43:55
2023-10-13T22:43:55
{ "title": [ -0.0132751465, -0.013572693, -0.0064430237, -0.026306152, -0.03463745, 0.0063819885, -0.013198853, -0.0029277802, 0.03515625, 0.035461426, -0.026611328, 0.024658203, 0.0013151169, 0.0078086853, -0.078552246, 0.01008606, -0.0002758503, ...
http://arxiv.org/abs/2310.09430v3
A Systematic Evaluation of Large Language Models on Out-of-Distribution Logical Reasoning Tasks
Qiming Bao, Gael Gendron, Alex Yuxuan Peng, Wanjun Zhong, Neset Tan, Yang Chen, Michael Witbrock, Jiamou Liu
cs.CL, cs.AI
Large language models (LLMs), such as GPT-3.5 and GPT-4, have greatly advanced the performance of artificial systems on various natural language processing tasks to human-like levels. However, their generalisation and robustness to perform logical reasoning remain under-evaluated. To probe this ability, we propose thre...
http://arxiv.org/pdf/2310.09430v3
2023-10-18T22:46:12
2023-10-13T22:29:15
{ "title": [ -0.019119263, -0.014678955, -0.046051025, -0.06188965, 0.0013217926, -0.004711151, -0.054382324, 0.008270264, 0.01675415, 0.019958496, -0.024368286, -0.010910034, 0.0043792725, -0.02079773, -0.040405273, -0.0029125214, -0.026535034, ...
http://arxiv.org/abs/2310.09424v1
SALM: Speech-augmented Language Model with In-context Learning for Speech Recognition and Translation
Zhehuai Chen, He Huang, Andrei Andrusenko, Oleksii Hrinchuk, Krishna C. Puvvada, Jason Li, Subhankar Ghosh, Jagadeesh Balam, Boris Ginsburg
cs.CL, cs.HC, cs.SD, eess.AS, 68T10, I.2.7
We present a novel Speech Augmented Language Model (SALM) with {\em multitask} and {\em in-context} learning capabilities. SALM comprises a frozen text LLM, a audio encoder, a modality adapter module, and LoRA layers to accommodate speech input and associated task instructions. The unified SALM not only achieves perfor...
http://arxiv.org/pdf/2310.09424v1
2023-10-13T22:07:33
2023-10-13T22:07:33
{ "title": [ -0.023742676, -0.0044288635, -0.0491333, -0.052947998, 0.0050354004, 0.0026130676, -0.052490234, 0.009887695, -0.02508545, 0.04727173, -0.011428833, 0.005130768, 0.025817871, -0.021133423, -0.036224365, -0.0008969307, -0.006729126, 0...
http://arxiv.org/abs/2310.09411v1
Surveying the Landscape of Text Summarization with Deep Learning: A Comprehensive Review
Guanghua Wang, Weili Wu
cs.CL, cs.AI, cs.LG
In recent years, deep learning has revolutionized natural language processing (NLP) by enabling the development of models that can learn complex representations of language data, leading to significant improvements in performance across a wide range of NLP tasks. Deep learning models for NLP typically use large amounts...
http://arxiv.org/pdf/2310.09411v1
2023-10-13T21:24:37
2023-10-13T21:24:37
{ "title": [ -0.0074386597, 0.012168884, -0.01802063, -0.06512451, 0.0011358261, 0.0066490173, -0.040527344, -0.014694214, 0.0019044876, 0.049560547, -0.004245758, 0.032287598, 0.0070228577, 0.0002105236, -0.045898438, 0.019927979, -0.0048713684, ...
http://arxiv.org/abs/2310.09357v1
A Computational Approach to Style in American Poetry
David M. Kaplan, David M. Blei
cs.CL, J.5; I.2.7
We develop a quantitative method to assess the style of American poems and to visualize a collection of poems in relation to one another. Qualitative poetry criticism helped guide our development of metrics that analyze various orthographic, syntactic, and phonemic features. These features are used to discover comprehe...
http://arxiv.org/pdf/2310.09357v1
2023-10-13T18:49:14
2023-10-13T18:49:14
{ "title": [ -0.027038574, -0.0072402954, -0.021362305, -0.016189575, -0.007980347, -0.019058228, 0.005504608, 0.0076179504, 0.016403198, 0.038604736, -0.024627686, 0.00022482872, 0.013511658, -0.025436401, -0.033691406, 0.0061912537, -0.0019817352, ...
http://arxiv.org/abs/2310.09350v1
Unsupervised Domain Adaption for Neural Information Retrieval
Carlos Dominguez, Jon Ander Campos, Eneko Agirre, Gorka Azkune
cs.CL, cs.AI
Neural information retrieval requires costly annotated data for each target domain to be competitive. Synthetic annotation by query generation using Large Language Models or rule-based string manipulation has been proposed as an alternative, but their relative merits have not been analysed. In this paper, we compare bo...
http://arxiv.org/pdf/2310.09350v1
2023-10-13T18:27:33
2023-10-13T18:27:33
{ "title": [ -0.005554199, 0.01991272, -0.042266846, -0.018173218, 0.017944336, -0.0309906, 0.0074501038, -0.026306152, 0.010368347, 0.035461426, -0.0073394775, -0.015609741, -0.0039138794, -0.012435913, -0.058563232, 0.004962921, -0.0074043274, ...
http://arxiv.org/abs/2310.09343v2
Dialogue Chain-of-Thought Distillation for Commonsense-aware Conversational Agents
Hyungjoo Chae, Yongho Song, Kai Tzu-iunn Ong, Taeyoon Kwon, Minjin Kim, Youngjae Yu, Dongha Lee, Dongyeop Kang, Jinyoung Yeo
cs.CL, cs.AI
Human-like chatbots necessitate the use of commonsense reasoning in order to effectively comprehend and respond to implicit information present within conversations. Achieving such coherence and informativeness in responses, however, is a non-trivial task. Even for large language models (LLMs), the task of identifying ...
http://arxiv.org/pdf/2310.09343v2
2023-10-22T09:16:52
2023-10-13T18:17:23
{ "title": [ 0.022369385, -0.015823364, -0.019119263, -0.061828613, -0.004425049, -0.011108398, -0.024642944, 0.0041885376, 0.049926758, 0.026123047, -0.03756714, -0.018493652, 0.006378174, -0.020477295, -0.05557251, -0.0042800903, -0.030303955, ...
http://arxiv.org/abs/2310.09342v2
Ranking LLM-Generated Loop Invariants for Program Verification
Saikat Chakraborty, Shuvendu K. Lahiri, Sarah Fakhoury, Madanlal Musuvathi, Akash Lal, Aseem Rastogi, Aditya Senthilnathan, Rahul Sharma, Nikhil Swamy
cs.PL, cs.AI, cs.CL, cs.SE
Synthesizing inductive loop invariants is fundamental to automating program verification. In this work, we observe that Large Language Models (such as gpt-3.5 or gpt-4) are capable of synthesizing loop invariants for a class of programs in a 0-shot setting, yet require several samples to generate the correct invariants...
http://arxiv.org/pdf/2310.09342v2
2023-10-18T18:18:43
2023-10-13T18:13:52
{ "title": [ -0.030761719, 0.0035419464, -0.05117798, -0.039489746, -0.004837036, -0.00894928, -0.03768921, 0.0026187897, -0.009407043, 0.019515991, -0.009651184, 0.027832031, -0.022521973, -0.021453857, -0.019378662, 0.009887695, -0.020904541, -...
http://arxiv.org/abs/2310.09266v1
User Inference Attacks on Large Language Models
Nikhil Kandpal, Krishna Pillutla, Alina Oprea, Peter Kairouz, Christopher A. Choquette-Choo, Zheng Xu
cs.CR, cs.CL, cs.LG
Fine-tuning is a common and effective method for tailoring large language models (LLMs) to specialized tasks and applications. In this paper, we study the privacy implications of fine-tuning LLMs on user data. To this end, we define a realistic threat model, called user inference, wherein an attacker infers whether or ...
http://arxiv.org/pdf/2310.09266v1
2023-10-13T17:24:52
2023-10-13T17:24:52
{ "title": [ -0.018066406, -0.014221191, -0.046173096, -0.061035156, 0.004787445, -0.011512756, -0.037017822, 0.016189575, -0.024047852, 0.019058228, -0.009613037, -0.0027885437, 0.0104522705, -0.019042969, -0.024017334, -0.0024108887, -0.020309448, ...
http://arxiv.org/abs/2310.09265v1
PromptRE: Weakly-Supervised Document-Level Relation Extraction via Prompting-Based Data Programming
Chufan Gao, Xulin Fan, Jimeng Sun, Xuan Wang
cs.CL, cs.LG
Relation extraction aims to classify the relationships between two entities into pre-defined categories. While previous research has mainly focused on sentence-level relation extraction, recent studies have expanded the scope to document-level relation extraction. Traditional relation extraction methods heavily rely on...
http://arxiv.org/pdf/2310.09265v1
2023-10-13T17:23:17
2023-10-13T17:23:17
{ "title": [ -0.0236969, -0.025344849, -0.046844482, -0.09039307, -0.01727295, 0.009307861, -0.07574463, 0.011207581, 0.028823853, -0.014465332, -0.047027588, 0.010566711, -0.010353088, -0.010391235, -0.023895264, -0.023529053, -0.004966736, 0.02...
http://arxiv.org/abs/2310.09263v1
Table-GPT: Table-tuned GPT for Diverse Table Tasks
Peng Li, Yeye He, Dror Yashar, Weiwei Cui, Song Ge, Haidong Zhang, Danielle Rifinski Fainman, Dongmei Zhang, Surajit Chaudhuri
cs.CL, cs.AI, cs.DB
Language models, such as GPT-3.5 and ChatGPT, demonstrate remarkable abilities to follow diverse human instructions and perform a wide range of tasks. However, when probing language models using a range of basic table-understanding tasks, we observe that today's language models are still sub-optimal in many table-relat...
http://arxiv.org/pdf/2310.09263v1
2023-10-13T17:20:56
2023-10-13T17:20:56
{ "title": [ -0.019485474, 0.0014677048, -0.05517578, -0.056549072, -0.0028934479, -0.010444641, -0.0625, -0.021652222, 0.0049705505, 0.03829956, -0.02633667, 0.012786865, 0.013397217, -0.018585205, -0.0418396, 0.012458801, 0.006790161, 0.0105514...
http://arxiv.org/abs/2310.09256v1
Political claim identification and categorization in a multilingual setting: First experiments
Urs Zaberer, Sebastian Padó, Gabriella Lapesa
cs.CL
The identification and classification of political claims is an important step in the analysis of political newspaper reports; however, resources for this task are few and far between. This paper explores different strategies for the cross-lingual projection of political claims analysis. We conduct experiments on a Ger...
http://arxiv.org/pdf/2310.09256v1
2023-10-13T17:13:00
2023-10-13T17:13:00
{ "title": [ 0.020706177, 0.0022563934, -0.01361084, -0.049438477, 0.008544922, -0.027175903, -0.0079193115, -0.030258179, -0.0423584, 0.02468872, 0.011016846, 0.017730713, -0.015930176, -0.032440186, -0.015975952, -0.0055618286, -0.03466797, 0.0...
http://arxiv.org/abs/2310.09247v1
Hypernymy Understanding Evaluation of Text-to-Image Models via WordNet Hierarchy
Anton Baryshnikov, Max Ryabinin
cs.CV, cs.CL, cs.LG
Text-to-image synthesis has recently attracted widespread attention due to rapidly improving quality and numerous practical applications. However, the language understanding capabilities of text-to-image models are still poorly understood, which makes it difficult to reason about prompt formulations that a given model ...
http://arxiv.org/pdf/2310.09247v1
2023-10-13T16:53:25
2023-10-13T16:53:25
{ "title": [ 0.0047950745, 0.009170532, -0.011955261, -0.034240723, -0.0020370483, -0.008239746, -0.008834839, -0.01977539, 0.022659302, 0.047210693, -0.05078125, 0.011894226, 0.056793213, -0.014099121, -0.023529053, 0.008758545, -0.0016393661, 0...
http://arxiv.org/abs/2310.09241v1
Precedent-Enhanced Legal Judgment Prediction with LLM and Domain-Model Collaboration
Yiquan Wu, Siying Zhou, Yifei Liu, Weiming Lu, Xiaozhong Liu, Yating Zhang, Changlong Sun, Fei Wu, Kun Kuang
cs.CL
Legal Judgment Prediction (LJP) has become an increasingly crucial task in Legal AI, i.e., predicting the judgment of the case in terms of case fact description. Precedents are the previous legal cases with similar facts, which are the basis for the judgment of the subsequent case in national legal systems. Thus, it is...
http://arxiv.org/pdf/2310.09241v1
2023-10-13T16:47:20
2023-10-13T16:47:20
{ "title": [ -0.036071777, -0.0068588257, -0.054107666, -0.07946777, -0.0118637085, 0.0032634735, -0.06335449, 0.028060913, 0.01876831, 0.02368164, -0.033691406, 0.0090789795, 0.0079956055, -0.009674072, -0.03692627, -0.00068330765, -0.029663086, ...
http://arxiv.org/abs/2310.09238v2
BanglaNLP at BLP-2023 Task 2: Benchmarking different Transformer Models for Sentiment Analysis of Bangla Social Media Posts
Saumajit Saha, Albert Nanda
cs.CL
Bangla is the 7th most widely spoken language globally, with a staggering 234 million native speakers primarily hailing from India and Bangladesh. This morphologically rich language boasts a rich literary tradition, encompassing diverse dialects and language-specific challenges. Despite its linguistic richness and hist...
http://arxiv.org/pdf/2310.09238v2
2023-10-18T03:51:38
2023-10-13T16:46:38
{ "title": [ 0.0064735413, 0.003326416, -0.01033783, -0.003967285, -0.01134491, 0.0011672974, -0.0149383545, -0.03225708, -0.002653122, 0.045837402, -0.020355225, 0.022445679, -0.011703491, -0.018157959, -0.042663574, -0.015182495, -0.009117126, ...
http://arxiv.org/abs/2310.09233v1
AgentCF: Collaborative Learning with Autonomous Language Agents for Recommender Systems
Junjie Zhang, Yupeng Hou, Ruobing Xie, Wenqi Sun, Julian McAuley, Wayne Xin Zhao, Leyu Lin, Ji-Rong Wen
cs.IR, cs.CL
Recently, there has been an emergence of employing LLM-powered agents as believable human proxies, based on their remarkable decision-making capability. However, existing studies mainly focus on simulating human dialogue. Human non-verbal behaviors, such as item clicking in recommender systems, although implicitly exhi...
http://arxiv.org/pdf/2310.09233v1
2023-10-13T16:37:14
2023-10-13T16:37:14
{ "title": [ -0.019561768, -0.016799927, -0.03338623, -0.036071777, 0.0027256012, -0.009651184, -0.00038599968, 0.0037078857, -0.00907135, 0.029449463, -0.021621704, 0.027694702, -0.02670288, -0.0054016113, -0.033325195, -0.008598328, -0.029037476, ...