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/2310.06770v1
SWE-bench: Can Language Models Resolve Real-World GitHub Issues?
Carlos E. Jimenez, John Yang, Alexander Wettig, Shunyu Yao, Kexin Pei, Ofir Press, Karthik Narasimhan
cs.CL, cs.AI, cs.SE
Language models have outpaced our ability to evaluate them effectively, but for their future development it is essential to study the frontier of their capabilities. We consider real-world software engineering to be a rich, sustainable, and challenging testbed for evaluating the next generation of language models. We t...
http://arxiv.org/pdf/2310.06770v1
2023-10-10T16:47:29
2023-10-10T16:47:29
{ "title": [ -0.0317688, -0.017089844, -0.051940918, -0.03237915, -0.013008118, -0.01109314, -0.034057617, 0.015144348, 0.0018415451, 0.042541504, -0.004333496, -0.0016269684, 0.04034424, -0.027664185, -0.019241333, 0.0069732666, 0.009094238, 0.0...
http://arxiv.org/abs/2310.06764v1
OmniLingo: Listening- and speaking-based language learning
Francis M. Tyers, Nicholas Howell
cs.CL
In this demo paper we present OmniLingo, an architecture for distributing data for listening- and speaking-based language learning applications and a demonstration client built using the architecture. The architecture is based on the Interplanetary Filesystem (IPFS) and puts at the forefront user sovereignty over data.
http://arxiv.org/pdf/2310.06764v1
2023-10-10T16:40:00
2023-10-10T16:40:00
{ "title": [ -0.03237915, -0.02406311, -0.027786255, -0.034240723, -0.02128601, 0.006412506, -0.037994385, -0.004749298, -0.031433105, 0.02029419, 0.0009407997, 0.018920898, -0.019210815, -0.03277588, -0.005039215, -0.0046920776, -0.014770508, 0....
http://arxiv.org/abs/2310.06762v1
TRACE: A Comprehensive Benchmark for Continual Learning in Large Language Models
Xiao Wang, Yuansen Zhang, Tianze Chen, Songyang Gao, Senjie Jin, Xianjun Yang, Zhiheng Xi, Rui Zheng, Yicheng Zou, Tao Gui, Qi Zhang, Xuanjing Huang
cs.CL
Aligned large language models (LLMs) demonstrate exceptional capabilities in task-solving, following instructions, and ensuring safety. However, the continual learning aspect of these aligned LLMs has been largely overlooked. Existing continual learning benchmarks lack sufficient challenge for leading aligned LLMs, owi...
http://arxiv.org/pdf/2310.06762v1
2023-10-10T16:38:49
2023-10-10T16:38:49
{ "title": [ -0.023254395, -0.011528015, -0.0390625, -0.085754395, 0.00818634, 0.0020656586, -0.0592041, 0.013908386, 0.010917664, 0.026550293, -0.015045166, 0.020217896, -0.0018863678, -0.016311646, -0.04638672, 0.018859863, -0.0054740906, 0.038...
http://arxiv.org/abs/2310.06714v1
Exploring Memorization in Fine-tuned Language Models
Shenglai Zeng, Yaxin Li, Jie Ren, Yiding Liu, Han Xu, Pengfei He, Yue Xing, Shuaiqiang Wang, Jiliang Tang, Dawei Yin
cs.AI, cs.CL, cs.LG
LLMs have shown great capabilities in various tasks but also exhibited memorization of training data, thus raising tremendous privacy and copyright concerns. While prior work has studied memorization during pre-training, the exploration of memorization during fine-tuning is rather limited. Compared with pre-training, f...
http://arxiv.org/pdf/2310.06714v1
2023-10-10T15:41:26
2023-10-10T15:41:26
{ "title": [ -0.030899048, -0.016098022, -0.022155762, -0.062042236, -0.02355957, 0.004032135, -0.043273926, 0.02381897, -0.008911133, 0.05706787, -0.016433716, 0.0140686035, 0.0012512207, -0.0007543564, -0.06414795, 0.0075531006, -0.0126953125, ...
http://arxiv.org/abs/2310.06707v1
Quality Control at Your Fingertips: Quality-Aware Translation Models
Christian Tomani, David Vilar, Markus Freitag, Colin Cherry, Subhajit Naskar, Mara Finkelstein, Daniel Cremers
cs.CL, cs.AI
Maximum-a-posteriori (MAP) decoding is the most widely used decoding strategy for neural machine translation (NMT) models. The underlying assumption is that model probability correlates well with human judgment, with better translations being more likely. However, research has shown that this assumption does not always...
http://arxiv.org/pdf/2310.06707v1
2023-10-10T15:33:51
2023-10-10T15:33:51
{ "title": [ -0.052764893, -0.008125305, -0.01424408, -0.048187256, -0.015525818, 0.02142334, -0.06945801, -0.021987915, -0.0090408325, 0.0496521, 0.0067596436, 0.006072998, 0.021728516, -0.025863647, -0.057373047, 0.021759033, -0.0032615662, 0.0...
http://arxiv.org/abs/2310.06702v1
Temporally Aligning Long Audio Interviews with Questions: A Case Study in Multimodal Data Integration
Piyush Singh Pasi, Karthikeya Battepati, Preethi Jyothi, Ganesh Ramakrishnan, Tanmay Mahapatra, Manoj Singh
cs.CL, cs.LG, cs.SD, eess.AS
The problem of audio-to-text alignment has seen significant amount of research using complete supervision during training. However, this is typically not in the context of long audio recordings wherein the text being queried does not appear verbatim within the audio file. This work is a collaboration with a non-governm...
http://arxiv.org/pdf/2310.06702v1
2023-10-10T15:25:33
2023-10-10T15:25:33
{ "title": [ -0.01285553, 0.012466431, -0.07244873, -0.070739746, 0.016738892, -0.029815674, -0.042114258, 0.013381958, 0.0064888, 0.03262329, -0.02142334, -0.011260986, -0.021499634, -0.020401001, -0.054779053, -0.030700684, -0.019378662, 0.0291...
http://arxiv.org/abs/2310.06694v1
Sheared LLaMA: Accelerating Language Model Pre-training via Structured Pruning
Mengzhou Xia, Tianyu Gao, Zhiyuan Zeng, Danqi Chen
cs.CL, cs.AI, cs.LG
The popularity of LLaMA (Touvron et al., 2023a;b) and other recently emerged moderate-sized large language models (LLMs) highlights the potential of building smaller yet powerful LLMs. Regardless, the cost of training such models from scratch on trillions of tokens remains high. In this work, we study structured prunin...
http://arxiv.org/pdf/2310.06694v1
2023-10-10T15:13:30
2023-10-10T15:13:30
{ "title": [ -0.027664185, -0.016662598, -0.032073975, -0.08123779, -0.009063721, 0.01335907, -0.068481445, 0.02494812, -0.010360718, 0.023956299, -0.013008118, 0.018875122, 0.018829346, -0.01977539, -0.0178833, 0.011009216, -0.0023155212, 0.0428...
http://arxiv.org/abs/2310.06692v2
Meta-CoT: Generalizable Chain-of-Thought Prompting in Mixed-task Scenarios with Large Language Models
Anni Zou, Zhuosheng Zhang, Hai Zhao, Xiangru Tang
cs.CL, cs.AI
Large language models (LLMs) have unveiled remarkable reasoning capabilities by exploiting chain-of-thought (CoT) prompting, which generates intermediate reasoning chains to serve as the rationale for deriving the answer. However, current CoT methods either simply employ general prompts such as Let's think step by step...
http://arxiv.org/pdf/2310.06692v2
2023-10-11T10:05:29
2023-10-10T15:10:03
{ "title": [ -0.015823364, -0.033966064, -0.01828003, -0.033203125, -0.031585693, -0.0022201538, -0.04815674, -0.012771606, -0.000010848045, 0.023498535, -0.020324707, 0.0043182373, 0.014480591, -0.022125244, -0.027618408, -0.00011181831, 0.003107071...
http://arxiv.org/abs/2310.06684v1
Learning Multiplex Embeddings on Text-rich Networks with One Text Encoder
Bowen Jin, Wentao Zhang, Yu Zhang, Yu Meng, Han Zhao, Jiawei Han
cs.CL, cs.LG
In real-world scenarios, texts in a network are often linked by multiple semantic relations (e.g., papers in an academic network are referenced by other publications, written by the same author, or published in the same venue), where text documents and their relations form a multiplex text-rich network. Mainstream text...
http://arxiv.org/pdf/2310.06684v1
2023-10-10T14:59:22
2023-10-10T14:59:22
{ "title": [ -0.03161621, -0.0045547485, -0.03149414, -0.029922485, -0.0046310425, 0.0045547485, -0.031677246, -0.0062026978, 0.003534317, 0.01965332, -0.0287323, 0.035095215, -0.023101807, -0.0102005005, -0.0101623535, 0.0036029816, -0.013076782, ...
http://arxiv.org/abs/2310.06675v2
SEER : A Knapsack approach to Exemplar Selection for In-Context HybridQA
Jonathan Tonglet, Manon Reusens, Philipp Borchert, Bart Baesens
cs.CL
Question answering over hybrid contexts is a complex task, which requires the combination of information extracted from unstructured texts and structured tables in various ways. Recently, In-Context Learning demonstrated significant performance advances for reasoning tasks. In this paradigm, a large language model perf...
http://arxiv.org/pdf/2310.06675v2
2023-10-20T08:02:25
2023-10-10T14:50:20
{ "title": [ -0.0007176399, -0.0116119385, -0.020217896, -0.0061149597, 0.0008049011, -0.0079422, -0.0010290146, -0.0262146, 0.010284424, 0.026535034, -0.04421997, 0.008781433, 0.021453857, -0.040039062, -0.019180298, -0.024307251, -0.014549255, ...
http://arxiv.org/abs/2310.06671v1
Making Large Language Models Perform Better in Knowledge Graph Completion
Yichi Zhang, Zhuo Chen, Wen Zhang, Huajun Chen
cs.CL
Large language model (LLM) based knowledge graph completion (KGC) aims to predict the missing triples in the KGs with LLMs and enrich the KGs to become better web infrastructure, which can benefit a lot of web-based automatic services. However, research about LLM-based KGC is limited and lacks effective utilization of ...
http://arxiv.org/pdf/2310.06671v1
2023-10-10T14:47:09
2023-10-10T14:47:09
{ "title": [ -0.03527832, -0.0181427, -0.053222656, -0.09399414, -0.0063323975, 0.016708374, -0.08087158, 0.030517578, 0.0067329407, 0.03201294, -0.03845215, 0.018249512, 0.0137786865, 0.0045928955, -0.036865234, 0.007762909, -0.009689331, 0.0489...
http://arxiv.org/abs/2310.06666v1
Unlock the Potential of Counterfactually-Augmented Data in Out-Of-Distribution Generalization
Caoyun Fan, Wenqing Chen, Jidong Tian, Yitian Li, Hao He, Yaohui Jin
cs.CL, cs.AI, cs.LG
Counterfactually-Augmented Data (CAD) -- minimal editing of sentences to flip the corresponding labels -- has the potential to improve the Out-Of-Distribution (OOD) generalization capability of language models, as CAD induces language models to exploit domain-independent causal features and exclude spurious correlation...
http://arxiv.org/pdf/2310.06666v1
2023-10-10T14:41:38
2023-10-10T14:41:38
{ "title": [ 0.078308105, 0.00415802, -0.01033783, -0.032440186, -0.014884949, -0.025024414, 0.011260986, -0.028656006, 0.016433716, 0.017501831, -0.070007324, 0.024414062, -0.021575928, -0.03060913, -0.0046844482, -0.03363037, -0.00932312, 0.062...
http://arxiv.org/abs/2310.06645v1
Self-Supervised Representation Learning for Online Handwriting Text Classification
Pouya Mehralian, Bagher BabaAli, Ashena Gorgan Mohammadi
cs.LG, cs.CL
Self-supervised learning offers an efficient way of extracting rich representations from various types of unlabeled data while avoiding the cost of annotating large-scale datasets. This is achievable by designing a pretext task to form pseudo labels with respect to the modality and domain of the data. Given the evolvin...
http://arxiv.org/pdf/2310.06645v1
2023-10-10T14:07:49
2023-10-10T14:07:49
{ "title": [ -0.013542175, 0.0052108765, -0.015464783, -0.022872925, -0.014717102, -0.007911682, -0.021240234, -0.0209198, 0.00258255, 0.03439331, -0.023254395, 0.0052223206, 0.008735657, -0.0074920654, -0.038909912, -0.013336182, 0.01600647, 0.0...
http://arxiv.org/abs/2310.06627v1
What If the TV Was Off? Examining Counterfactual Reasoning Abilities of Multi-modal Language Models
Letian Zhang, Xiaotong Zhai, Zhongkai Zhao, Xin Wen, Yongshuo Zong, Bingchen Zhao
cs.CL, cs.CV, cs.LG
Counterfactual reasoning ability is one of the core abilities of human intelligence. This reasoning process involves the processing of alternatives to observed states or past events, and this process can improve our ability for planning and decision-making. In this work, we focus on benchmarking the counterfactual reas...
http://arxiv.org/pdf/2310.06627v1
2023-10-10T13:45:59
2023-10-10T13:45:59
{ "title": [ 0.0021591187, -0.0067443848, -0.019851685, -0.01965332, -0.015914917, -0.015335083, -0.036315918, -0.035583496, 0.023971558, 0.024795532, -0.0076293945, -0.0027561188, 0.01687622, -0.033325195, -0.04449463, -0.0033874512, -0.012161255, ...
http://arxiv.org/abs/2310.06626v1
Topic-DPR: Topic-based Prompts for Dense Passage Retrieval
Qingfa Xiao, Shuangyin Li, Lei Chen
cs.CL, cs.AI
Prompt-based learning's efficacy across numerous natural language processing tasks has led to its integration into dense passage retrieval. Prior research has mainly focused on enhancing the semantic understanding of pre-trained language models by optimizing a single vector as a continuous prompt. This approach, howeve...
http://arxiv.org/pdf/2310.06626v1
2023-10-10T13:45:24
2023-10-10T13:45:24
{ "title": [ -0.016586304, -0.008407593, -0.04852295, -0.04360962, -0.0067443848, -0.009651184, -0.034332275, -0.024291992, 0.0491333, 0.029266357, -0.04751587, 0.03552246, 0.0064582825, -0.0018749237, -0.03942871, 0.016479492, 0.0071792603, -0.0...
http://arxiv.org/abs/2310.06590v1
No Pitch Left Behind: Addressing Gender Unbalance in Automatic Speech Recognition through Pitch Manipulation
Dennis Fucci, Marco Gaido, Matteo Negri, Mauro Cettolo, Luisa Bentivogli
cs.CL
Automatic speech recognition (ASR) systems are known to be sensitive to the sociolinguistic variability of speech data, in which gender plays a crucial role. This can result in disparities in recognition accuracy between male and female speakers, primarily due to the under-representation of the latter group in the trai...
http://arxiv.org/pdf/2310.06590v1
2023-10-10T12:55:22
2023-10-10T12:55:22
{ "title": [ -0.0206604, -0.006374359, -0.057037354, -0.055389404, 0.003189087, -0.02217102, -0.028381348, 0.018356323, -0.0151901245, 0.0513916, -0.027618408, 0.008544922, 0.029556274, -0.0021820068, -0.048828125, -0.011260986, -0.034240723, 0.0...
http://arxiv.org/abs/2310.06588v1
FTFT: efficient and robust Fine-Tuning by transFerring Training dynamics
Yupei Du, Albert Gatt, Dong Nguyen
cs.CL, cs.LG
Despite the massive success of fine-tuning large Pre-trained Language Models (PLMs) on a wide range of Natural Language Processing (NLP) tasks, they remain susceptible to out-of-distribution (OOD) and adversarial inputs. Data map (DM) is a simple yet effective dual-model approach that enhances the robustness of fine-tu...
http://arxiv.org/pdf/2310.06588v1
2023-10-10T12:53:48
2023-10-10T12:53:48
{ "title": [ 0.0025901794, -0.0024604797, -0.030349731, -0.059936523, 0.020263672, -0.011238098, -0.0071411133, -0.0129776, -0.032684326, 0.028305054, -0.0001950264, 0.024353027, -0.00630188, -0.022750854, -0.023040771, 0.00793457, -0.0006246567, ...
http://arxiv.org/abs/2310.06555v1
On Temporal References in Emergent Communication
Olaf Lipinski, Adam J. Sobey, Federico Cerutti, Timothy J. Norman
cs.CL, cs.AI, cs.LG, cs.MA
As humans, we use linguistic elements referencing time, such as before or tomorrow, to easily share past experiences and future predictions. While temporal aspects of the language have been considered in computational linguistics, no such exploration has been done within the field of emergent communication. We research...
http://arxiv.org/pdf/2310.06555v1
2023-10-10T12:10:40
2023-10-10T12:10:40
{ "title": [ -0.007507324, 0.0052948, -0.028076172, -0.05456543, 0.022735596, -0.0061454773, -0.0030021667, -0.0037670135, 0.015464783, 0.049194336, 0.00096559525, -0.008598328, 0.013641357, -0.01524353, -0.06536865, 0.004219055, -0.008346558, 0....
http://arxiv.org/abs/2310.06552v3
Automated clinical coding using off-the-shelf large language models
Joseph S. Boyle, Antanas Kascenas, Pat Lok, Maria Liakata, Alison Q. O'Neil
cs.AI, cs.CL, I.2.7; I.2.8
The task of assigning diagnostic ICD codes to patient hospital admissions is typically performed by expert human coders. Efforts towards automated ICD coding are dominated by supervised deep learning models. However, difficulties in learning to predict the large number of rare codes remain a barrier to adoption in clin...
http://arxiv.org/pdf/2310.06552v3
2023-11-13T12:38:00
2023-10-10T11:56:48
{ "title": [ -0.049468994, -0.026321411, -0.047088623, -0.08013916, -0.013710022, 0.020996094, -0.08581543, 0.016326904, 0.002105713, 0.04296875, 0.00982666, -0.0052871704, 0.026947021, -0.025177002, -0.028366089, 0.020187378, 0.002664566, 0.0190...
http://arxiv.org/abs/2310.06547v1
Rationale-Enhanced Language Models are Better Continual Relation Learners
Weimin Xiong, Yifan Song, Peiyi Wang, Sujian Li
cs.CL, cs.AI
Continual relation extraction (CRE) aims to solve the problem of catastrophic forgetting when learning a sequence of newly emerging relations. Recent CRE studies have found that catastrophic forgetting arises from the model's lack of robustness against future analogous relations. To address the issue, we introduce rati...
http://arxiv.org/pdf/2310.06547v1
2023-10-10T11:50:27
2023-10-10T11:50:27
{ "title": [ 0.021697998, -0.0038490295, -0.030380249, -0.0904541, -0.0184021, -0.019546509, -0.0657959, 0.009979248, 0.012931824, 0.017166138, -0.0025119781, -0.010688782, -0.0068893433, -0.04058838, -0.028289795, 0.020339966, -0.02027893, 0.056...
http://arxiv.org/abs/2310.06546v1
AutoCycle-VC: Towards Bottleneck-Independent Zero-Shot Cross-Lingual Voice Conversion
Haeyun Choi, Jio Gim, Yuho Lee, Youngin Kim, Young-Joo Suh
cs.SD, cs.CL, eess.AS
This paper proposes a simple and robust zero-shot voice conversion system with a cycle structure and mel-spectrogram pre-processing. Previous works suffer from information loss and poor synthesis quality due to their reliance on a carefully designed bottleneck structure. Moreover, models relying solely on self-reconstr...
http://arxiv.org/pdf/2310.06546v1
2023-10-10T11:50:16
2023-10-10T11:50:16
{ "title": [ -0.010215759, -0.020721436, -0.03665161, -0.07873535, -0.0018043518, 0.00020074844, -0.06335449, 0.021957397, 0.0056381226, 0.060180664, -0.014457703, -0.005016327, 0.04473877, -0.018722534, -0.04220581, 0.01789856, -0.0067253113, 0....
http://arxiv.org/abs/2310.06540v1
A Novel Contrastive Learning Method for Clickbait Detection on RoCliCo: A Romanian Clickbait Corpus of News Articles
Daria-Mihaela Broscoteanu, Radu Tudor Ionescu
cs.CL, cs.AI, cs.LG
To increase revenue, news websites often resort to using deceptive news titles, luring users into clicking on the title and reading the full news. Clickbait detection is the task that aims to automatically detect this form of false advertisement and avoid wasting the precious time of online users. Despite the importanc...
http://arxiv.org/pdf/2310.06540v1
2023-10-10T11:38:16
2023-10-10T11:38:16
{ "title": [ 0.000936985, -0.001543045, 0.0013284683, -0.02003479, 0.0037956238, -0.0074501038, 0.008056641, -0.028259277, 0.004085541, 0.04537964, -0.051940918, 0.026062012, 0.003660202, -0.00082969666, -0.044281006, -0.01448822, -0.038330078, 0...
http://arxiv.org/abs/2310.06536v1
EmoTwiCS: A Corpus for Modelling Emotion Trajectories in Dutch Customer Service Dialogues on Twitter
Sofie Labat, Thomas Demeester, Véronique Hoste
cs.CL
Due to the rise of user-generated content, social media is increasingly adopted as a channel to deliver customer service. Given the public character of these online platforms, the automatic detection of emotions forms an important application in monitoring customer satisfaction and preventing negative word-of-mouth. Th...
http://arxiv.org/pdf/2310.06536v1
2023-10-10T11:31:11
2023-10-10T11:31:11
{ "title": [ 0.00087833405, -0.018829346, -0.011924744, -0.064208984, -0.0138168335, 0.0104904175, -0.032806396, 0.0055503845, 0.038024902, 0.042266846, -0.04522705, -0.012664795, 0.03781128, -0.0060577393, -0.074645996, -0.002588272, -0.0096206665, ...
http://arxiv.org/abs/2310.06517v1
Toward Semantic Publishing in Non-Invasive Brain Stimulation: A Comprehensive Analysis of rTMS Studies
Swathi Anil, Jennifer D'Souza
cs.DL, cs.CL, cs.IT, math.IT
Noninvasive brain stimulation (NIBS) encompasses transcranial stimulation techniques that can influence brain excitability. These techniques have the potential to treat conditions like depression, anxiety, and chronic pain, and to provide insights into brain function. However, a lack of standardized reporting practices...
http://arxiv.org/pdf/2310.06517v1
2023-10-10T11:00:23
2023-10-10T11:00:23
{ "title": [ -0.0038280487, 0.009307861, -0.012237549, -0.03152466, -0.0020217896, -0.002916336, -0.038726807, -0.045928955, -0.024108887, 0.028533936, 0.016326904, 0.009742737, 0.022781372, -0.050842285, -0.017471313, 0.01637268, 0.0071754456, 0...
http://arxiv.org/abs/2310.06505v1
Evaluation of ChatGPT Feedback on ELL Writers' Coherence and Cohesion
Su-Youn Yoon, Eva Miszoglad, Lisa R. Pierce
cs.CL, cs.AI, I.2.7
Since its launch in November 2022, ChatGPT has had a transformative effect on education where students are using it to help with homework assignments and teachers are actively employing it in their teaching practices. This includes using ChatGPT as a tool for writing teachers to grade and generate feedback on students'...
http://arxiv.org/pdf/2310.06505v1
2023-10-10T10:25:56
2023-10-10T10:25:56
{ "title": [ -0.029891968, -0.0063056946, -0.010597229, -0.003124237, -0.03286743, 0.021987915, -0.045074463, -0.014511108, -0.047546387, 0.05529785, 0.018600464, 0.029159546, 0.03366089, -0.019546509, -0.0039253235, 0.012237549, -0.04321289, 0.0...
http://arxiv.org/abs/2310.06504v1
Revisit Input Perturbation Problems for LLMs: A Unified Robustness Evaluation Framework for Noisy Slot Filling Task
Guanting Dong, Jinxu Zhao, Tingfeng Hui, Daichi Guo, Wenlong Wan, Boqi Feng, Yueyan Qiu, Zhuoma Gongque, Keqing He, Zechen Wang, Weiran Xu
cs.CL, cs.AI, cs.LG
With the increasing capabilities of large language models (LLMs), these high-performance models have achieved state-of-the-art results on a wide range of natural language processing (NLP) tasks. However, the models' performance on commonly-used benchmark datasets often fails to accurately reflect their reliability and ...
http://arxiv.org/pdf/2310.06504v1
2023-10-10T10:22:05
2023-10-10T10:22:05
{ "title": [ 0.013336182, 0.020523071, 0.021530151, -0.015235901, -0.0102005005, -0.0041275024, -0.023498535, -0.06329346, 0.005519867, 0.004306793, -0.055786133, 0.027114868, -0.01197052, -0.016860962, -0.015525818, -0.030136108, -0.012886047, 0...
http://arxiv.org/abs/2310.06502v1
The Limits of ChatGPT in Extracting Aspect-Category-Opinion-Sentiment Quadruples: A Comparative Analysis
Xiancai Xu, Jia-Dong Zhang, Rongchang Xiao, Lei Xiong
cs.CL
Recently, ChatGPT has attracted great attention from both industry and academia due to its surprising abilities in natural language understanding and generation. We are particularly curious about whether it can achieve promising performance on one of the most complex tasks in aspect-based sentiment analysis, i.e., extr...
http://arxiv.org/pdf/2310.06502v1
2023-10-10T10:19:58
2023-10-10T10:19:58
{ "title": [ 0.014251709, -0.02104187, -0.019302368, -0.034484863, -0.0069618225, -0.00944519, -0.0043258667, -0.005558014, 0.0038146973, 0.050109863, -0.011711121, 0.009338379, -0.01448822, -0.0055618286, -0.06854248, 0.009475708, -0.019882202, ...
http://arxiv.org/abs/2310.06498v2
A New Benchmark and Reverse Validation Method for Passage-level Hallucination Detection
Shiping Yang, Renliang Sun, Xiaojun Wan
cs.CL
Large Language Models (LLMs) have shown their ability to collaborate effectively with humans in real-world scenarios. However, LLMs are apt to generate hallucinations, i.e., makeup incorrect text and unverified information, which can cause significant damage when deployed for mission-critical tasks. In this paper, we p...
http://arxiv.org/pdf/2310.06498v2
2023-10-24T01:37:10
2023-10-10T10:14:59
{ "title": [ 0.0051078796, 0.014221191, -0.017089844, -0.02671814, 0.0044670105, -0.021087646, -0.01209259, -0.046661377, 0.007019043, 0.04083252, -0.01474762, 0.03488159, 0.0073776245, -0.01247406, -0.05706787, 0.020629883, -0.011276245, 0.04745...
http://arxiv.org/abs/2310.06488v2
SpikeCLIP: A Contrastive Language-Image Pretrained Spiking Neural Network
Tianlong Li, Wenhao Liu, Changze Lv, Jianhan Xu, Cenyuan Zhang, Muling Wu, Xiaoqing Zheng, Xuanjing Huang
cs.NE, cs.CL, cs.CV, cs.LG
Spiking neural networks (SNNs) have demonstrated the capability to achieve comparable performance to deep neural networks (DNNs) in both visual and linguistic domains while offering the advantages of improved energy efficiency and adherence to biological plausibility. However, the extension of such single-modality SNNs...
http://arxiv.org/pdf/2310.06488v2
2023-10-12T03:23:40
2023-10-10T09:57:17
{ "title": [ 0.018630981, -0.0034542084, -0.008110046, -0.029907227, 0.003484726, -0.014289856, -0.0030021667, -0.035827637, 0.016174316, 0.024505615, 0.005077362, -0.0043029785, -0.012123108, -0.04083252, -0.04107666, 0.020263672, 0.0016269684, ...
http://arxiv.org/abs/2310.06474v1
Multilingual Jailbreak Challenges in Large Language Models
Yue Deng, Wenxuan Zhang, Sinno Jialin Pan, Lidong Bing
cs.CL
While large language models (LLMs) exhibit remarkable capabilities across a wide range of tasks, they pose potential safety concerns, such as the ``jailbreak'' problem, wherein malicious instructions can manipulate LLMs to exhibit undesirable behavior. Although several preventive measures have been developed to mitigat...
http://arxiv.org/pdf/2310.06474v1
2023-10-10T09:44:06
2023-10-10T09:44:06
{ "title": [ -0.04055786, -0.0060768127, -0.034484863, -0.040252686, -0.008102417, -0.0019798279, -0.048614502, 0.014205933, -0.023910522, 0.04095459, 0.017837524, -0.010047913, 0.02722168, -0.03765869, -0.012802124, 0.021896362, -0.020141602, 0....
http://arxiv.org/abs/2310.06458v1
Cultural Compass: Predicting Transfer Learning Success in Offensive Language Detection with Cultural Features
Li Zhou, Antonia Karamolegkou, Wenyu Chen, Daniel Hershcovich
cs.CL
The increasing ubiquity of language technology necessitates a shift towards considering cultural diversity in the machine learning realm, particularly for subjective tasks that rely heavily on cultural nuances, such as Offensive Language Detection (OLD). Current understanding underscores that these tasks are substantia...
http://arxiv.org/pdf/2310.06458v1
2023-10-10T09:29:38
2023-10-10T09:29:38
{ "title": [ 0.009986877, 0.009750366, -0.027511597, -0.012916565, 0.017486572, -0.04055786, 0.019378662, -0.027938843, 0.007194519, 0.019348145, -0.029006958, 0.013626099, -0.0003244877, -0.014282227, -0.044799805, -0.0126953125, -0.0149383545, ...
http://arxiv.org/abs/2310.06452v1
Understanding the Effects of RLHF on LLM Generalisation and Diversity
Robert Kirk, Ishita Mediratta, Christoforos Nalmpantis, Jelena Luketina, Eric Hambro, Edward Grefenstette, Roberta Raileanu
cs.LG, cs.AI, cs.CL
Large language models (LLMs) fine-tuned with reinforcement learning from human feedback (RLHF) have been used in some of the most widely deployed AI models to date, such as OpenAI's ChatGPT, Anthropic's Claude, or Meta's LLaMA-2. While there has been significant work developing these methods, our understanding of the b...
http://arxiv.org/pdf/2310.06452v1
2023-10-10T09:25:44
2023-10-10T09:25:44
{ "title": [ -0.020065308, -0.006576538, -0.075805664, -0.07635498, -0.03250122, -0.020004272, -0.07458496, 0.032287598, -0.052337646, 0.034942627, 0.0060577393, 0.02217102, -0.0137786865, -0.039245605, 0.01625061, -0.0066070557, -0.020767212, 0....
http://arxiv.org/abs/2310.06450v2
Constructive Large Language Models Alignment with Diverse Feedback
Tianshu Yu, Ting-En Lin, Yuchuan Wu, Min Yang, Fei Huang, Yongbin Li
cs.CL, cs.AI
In recent research on large language models (LLMs), there has been a growing emphasis on aligning these models with human values to reduce the impact of harmful content. However, current alignment methods often rely solely on singular forms of human feedback, such as preferences, annotated labels, or natural language c...
http://arxiv.org/pdf/2310.06450v2
2023-10-11T07:04:04
2023-10-10T09:20:14
{ "title": [ -0.04876709, 0.0065460205, -0.06298828, -0.061798096, 0.0058670044, 0.0073623657, -0.09277344, 0.0003786087, 0.011703491, 0.043121338, -0.015487671, -0.004650116, 0.012199402, -0.0065689087, -0.050964355, 0.012916565, -0.0056266785, ...
http://arxiv.org/abs/2310.06436v1
MemSum-DQA: Adapting An Efficient Long Document Extractive Summarizer for Document Question Answering
Nianlong Gu, Yingqiang Gao, Richard H. R. Hahnloser
cs.CL
We introduce MemSum-DQA, an efficient system for document question answering (DQA) that leverages MemSum, a long document extractive summarizer. By prefixing each text block in the parsed document with the provided question and question type, MemSum-DQA selectively extracts text blocks as answers from documents. On ful...
http://arxiv.org/pdf/2310.06436v1
2023-10-10T09:06:08
2023-10-10T09:06:08
{ "title": [ 0.007896423, -0.0076026917, -0.00002515316, -0.06323242, -0.013313293, 0.012031555, -0.04159546, -0.016403198, 0.03579712, 0.035064697, -0.00945282, 0.018539429, 0.0032310486, -0.004924774, -0.070251465, 0.021072388, -0.0013046265, 0...
http://arxiv.org/abs/2310.06434v2
Whispering LLaMA: A Cross-Modal Generative Error Correction Framework for Speech Recognition
Srijith Radhakrishnan, Chao-Han Huck Yang, Sumeer Ahmad Khan, Rohit Kumar, Narsis A. Kiani, David Gomez-Cabrero, Jesper N. Tegner
cs.CL, cs.AI, cs.MM, cs.SD, eess.AS
We introduce a new cross-modal fusion technique designed for generative error correction in automatic speech recognition (ASR). Our methodology leverages both acoustic information and external linguistic representations to generate accurate speech transcription contexts. This marks a step towards a fresh paradigm in ge...
http://arxiv.org/pdf/2310.06434v2
2023-10-16T21:32:56
2023-10-10T09:04:33
{ "title": [ -0.027053833, -0.0011510849, -0.044647217, -0.055023193, 0.0011129379, 0.0023612976, -0.0715332, -0.00970459, 0.011177063, 0.051818848, -0.03805542, 0.013084412, 0.022323608, -0.011390686, -0.047027588, 0.0023517609, -0.020141602, 0....
http://arxiv.org/abs/2310.06433v1
Retromorphic Testing: A New Approach to the Test Oracle Problem
Boxi Yu, Qiuyang Mang, Qingshuo Guo, Pinjia He
cs.SE, cs.AI, cs.CL, cs.CV, D.3.0; I.2.7; I.4.0
A test oracle serves as a criterion or mechanism to assess the correspondence between software output and the anticipated behavior for a given input set. In automated testing, black-box techniques, known for their non-intrusive nature in test oracle construction, are widely used, including notable methodologies like di...
http://arxiv.org/pdf/2310.06433v1
2023-10-10T09:03:01
2023-10-10T09:03:01
{ "title": [ -0.030670166, 0.008277893, -0.028701782, -0.07751465, 0.012458801, 0.019088745, -0.052947998, 0.0011987686, 0.013259888, 0.026473999, -0.020217896, 0.0073509216, -0.021499634, -0.025619507, -0.051940918, -0.0027103424, -0.011245728, ...
http://arxiv.org/abs/2310.06422v1
Large Language Models for Propaganda Detection
Kilian Sprenkamp, Daniel Gordon Jones, Liudmila Zavolokina
cs.CL, cs.AI
The prevalence of propaganda in our digital society poses a challenge to societal harmony and the dissemination of truth. Detecting propaganda through NLP in text is challenging due to subtle manipulation techniques and contextual dependencies. To address this issue, we investigate the effectiveness of modern Large Lan...
http://arxiv.org/pdf/2310.06422v1
2023-10-10T08:46:10
2023-10-10T08:46:10
{ "title": [ 0.013671875, 0.013336182, -0.015159607, -0.047821045, 0.022659302, -0.0071525574, -0.022201538, -0.026245117, -0.016525269, 0.025909424, -0.017196655, -0.0031147003, 0.003730774, -0.009712219, -0.029144287, -0.009056091, -0.030960083, ...
http://arxiv.org/abs/2310.06408v2
Humans and language models diverge when predicting repeating text
Aditya R. Vaidya, Javier Turek, Alexander G. Huth
cs.CL
Language models that are trained on the next-word prediction task have been shown to accurately model human behavior in word prediction and reading speed. In contrast with these findings, we present a scenario in which the performance of humans and LMs diverges. We collected a dataset of human next-word predictions for...
http://arxiv.org/pdf/2310.06408v2
2023-10-23T03:15:46
2023-10-10T08:24:28
{ "title": [ 0.0027923584, 0.011161804, -0.011451721, -0.031463623, -0.043029785, -0.01374054, -0.028244019, -0.011276245, -0.03189087, 0.061523438, -0.05053711, -0.009757996, 0.032989502, -0.0068588257, -0.037628174, -0.013710022, -0.0009393692, ...
http://arxiv.org/abs/2310.06404v2
Hexa: Self-Improving for Knowledge-Grounded Dialogue System
Daejin Jo, Daniel Wontae Nam, Gunsoo Han, Kyoung-Woon On, Taehwan Kwon, Seungeun Rho, Sungwoong Kim
cs.CL, cs.AI, cs.LG
A common practice in knowledge-grounded dialogue generation is to explicitly utilize intermediate steps (e.g., web-search, memory retrieval) with modular approaches. However, data for such steps are often inaccessible compared to those of dialogue responses as they are unobservable in an ordinary dialogue. To fill in t...
http://arxiv.org/pdf/2310.06404v2
2023-10-22T23:58:53
2023-10-10T08:15:24
{ "title": [ -0.02217102, -0.0025901794, -0.07702637, -0.076660156, 0.025634766, -0.015312195, -0.044647217, 0.018310547, 0.005558014, 0.034729004, -0.0007548332, 0.0022411346, -0.0063552856, -0.024597168, -0.052520752, 0.0016870499, -0.0042915344, ...
http://arxiv.org/abs/2310.06391v1
Improved prompting and process for writing user personas with LLMs, using qualitative interviews: Capturing behaviour and personality traits of users
Stefano De Paoli
cs.HC, cs.CL, cs.CY
This draft paper presents a workflow for creating User Personas with Large Language Models, using the results of a Thematic Analysis of qualitative interviews. The proposed workflow uses improved prompting and a larger pool of Themes, compared to previous work conducted by the author for the same task. This is possible...
http://arxiv.org/pdf/2310.06391v1
2023-10-10T07:54:24
2023-10-10T07:54:24
{ "title": [ -0.005882263, 0.014808655, -0.0309906, 0.030548096, -0.01109314, -0.02418518, 0.00007599592, -0.03616333, -0.017333984, 0.00042295456, -0.029663086, 0.03555298, -0.0127334595, -0.040863037, 0.021072388, -0.018600464, -0.031707764, 0....
http://arxiv.org/abs/2310.06390v1
P5: Plug-and-Play Persona Prompting for Personalized Response Selection
Joosung Lee, Minsik Oh, Donghun Lee
cs.CL, cs.AI, cs.IR
The use of persona-grounded retrieval-based chatbots is crucial for personalized conversations, but there are several challenges that need to be addressed. 1) In general, collecting persona-grounded corpus is very expensive. 2) The chatbot system does not always respond in consideration of persona at real applications....
http://arxiv.org/pdf/2310.06390v1
2023-10-10T07:53:36
2023-10-10T07:53:36
{ "title": [ 0.005771637, 0.000248909, -0.022262573, -0.064697266, 0.0053138733, 0.0016746521, -0.033203125, -0.014732361, -0.020141602, 0.024093628, 0.0059280396, 0.041259766, -0.009857178, -0.029937744, -0.013420105, 0.01524353, -0.008415222, 0...
http://arxiv.org/abs/2310.06387v1
Jailbreak and Guard Aligned Language Models with Only Few In-Context Demonstrations
Zeming Wei, Yifei Wang, Yisen Wang
cs.LG, cs.AI, cs.CL, cs.CR
Large Language Models (LLMs) have shown remarkable success in various tasks, but concerns about their safety and the potential for generating malicious content have emerged. In this paper, we explore the power of In-Context Learning (ICL) in manipulating the alignment ability of LLMs. We find that by providing just few...
http://arxiv.org/pdf/2310.06387v1
2023-10-10T07:50:29
2023-10-10T07:50:29
{ "title": [ -0.024963379, -0.005996704, -0.06915283, -0.010116577, -0.025909424, -0.031341553, -0.021972656, 0.016937256, -0.030807495, 0.017654419, -0.024856567, -0.015327454, 0.03289795, -0.053894043, 0.040985107, -0.012214661, 0.0065689087, -...
http://arxiv.org/abs/2310.06374v2
Rethinking Model Selection and Decoding for Keyphrase Generation with Pre-trained Sequence-to-Sequence Models
Di Wu, Wasi Uddin Ahmad, Kai-Wei Chang
cs.CL
Keyphrase Generation (KPG) is a longstanding task in NLP with widespread applications. The advent of sequence-to-sequence (seq2seq) pre-trained language models (PLMs) has ushered in a transformative era for KPG, yielding promising performance improvements. However, many design decisions remain unexplored and are often ...
http://arxiv.org/pdf/2310.06374v2
2023-10-22T08:37:43
2023-10-10T07:34:45
{ "title": [ -0.041259766, -0.013000488, -0.023483276, -0.044647217, -0.01524353, 0.009376526, -0.06188965, -0.0041007996, 0.0037250519, 0.052856445, -0.0008716583, 0.017654419, 0.017089844, -0.015686035, -0.05142212, 0.020858765, -0.015991211, 0...
http://arxiv.org/abs/2310.06365v1
Multi-Modal Knowledge Graph Transformer Framework for Multi-Modal Entity Alignment
Qian Li, Cheng Ji, Shu Guo, Zhaoji Liang, Lihong Wang, Jianxin Li
cs.CL
Multi-Modal Entity Alignment (MMEA) is a critical task that aims to identify equivalent entity pairs across multi-modal knowledge graphs (MMKGs). However, this task faces challenges due to the presence of different types of information, including neighboring entities, multi-modal attributes, and entity types. Directly ...
http://arxiv.org/pdf/2310.06365v1
2023-10-10T07:06:06
2023-10-10T07:06:06
{ "title": [ -0.029769897, 0.0041618347, -0.09051514, -0.07434082, 0.013618469, -0.012397766, -0.070739746, 0.0076026917, 0.01927185, 0.050048828, 0.0041885376, 0.012512207, 0.06414795, -0.02810669, -0.018356323, 0.02658081, -0.0007419586, 0.0236...
http://arxiv.org/abs/2310.06362v1
InfoCL: Alleviating Catastrophic Forgetting in Continual Text Classification from An Information Theoretic Perspective
Yifan Song, Peiyi Wang, Weimin Xiong, Dawei Zhu, Tianyu Liu, Zhifang Sui, Sujian Li
cs.CL
Continual learning (CL) aims to constantly learn new knowledge over time while avoiding catastrophic forgetting on old tasks. We focus on continual text classification under the class-incremental setting. Recent CL studies have identified the severe performance decrease on analogous classes as a key factor for catastro...
http://arxiv.org/pdf/2310.06362v1
2023-10-10T07:00:13
2023-10-10T07:00:13
{ "title": [ -0.014472961, 0.012573242, -0.0597229, -0.06359863, 0.019592285, -0.007774353, -0.025970459, -0.002752304, 0.015380859, 0.037902832, -0.032226562, 0.016967773, 0.0006556511, -0.00021231174, -0.058502197, -0.006549835, -0.010284424, 0...
http://arxiv.org/abs/2310.06356v1
A Semantic Invariant Robust Watermark for Large Language Models
Aiwei Liu, Leyi Pan, Xuming Hu, Shiao Meng, Lijie Wen
cs.CR, cs.CL, 68T50, I.2.7
Watermark algorithms for large language models (LLMs) have achieved extremely high accuracy in detecting text generated by LLMs. Such algorithms typically involve adding extra watermark logits to the LLM's logits at each generation step. However, prior algorithms face a trade-off between attack robustness and security ...
http://arxiv.org/pdf/2310.06356v1
2023-10-10T06:49:43
2023-10-10T06:49:43
{ "title": [ -0.018066406, 0.021591187, -0.04711914, -0.04953003, 0.027069092, -0.010818481, -0.0519104, -0.024612427, 0.0062408447, 0.045959473, -0.0154418945, 0.0067329407, 0.012504578, -0.014823914, -0.043518066, 0.014328003, -0.013046265, 0.0...
http://arxiv.org/abs/2310.06302v1
Selective Demonstrations for Cross-domain Text-to-SQL
Shuaichen Chang, Eric Fosler-Lussier
cs.CL
Large language models (LLMs) with in-context learning have demonstrated impressive generalization capabilities in the cross-domain text-to-SQL task, without the use of in-domain annotations. However, incorporating in-domain demonstration examples has been found to greatly enhance LLMs' performance. In this paper, we de...
http://arxiv.org/pdf/2310.06302v1
2023-10-10T04:31:41
2023-10-10T04:31:41
{ "title": [ 0.022445679, 0.0070114136, -0.032836914, -0.08721924, 0.030685425, -0.01449585, -0.020355225, 0.008300781, 0.010223389, 0.03186035, -0.050933838, 0.012329102, 0.026016235, -0.013679504, -0.014511108, -0.0077781677, -0.009338379, 0.04...
http://arxiv.org/abs/2310.06272v1
Let Models Speak Ciphers: Multiagent Debate through Embeddings
Chau Pham, Boyi Liu, Yingxiang Yang, Zhengyu Chen, Tianyi Liu, Jianbo Yuan, Bryan A. Plummer, Zhaoran Wang, Hongxia Yang
cs.CL, cs.AI, cs.LG
Discussion and debate among Large Language Models (LLMs) have gained considerable attention due to their potential to enhance the reasoning ability of LLMs. Although natural language is an obvious choice for communication due to LLM's language understanding capability, the token sampling step needed when generating nat...
http://arxiv.org/pdf/2310.06272v1
2023-10-10T03:06:38
2023-10-10T03:06:38
{ "title": [ -0.018707275, 0.00077295303, -0.023880005, -0.017044067, -0.0049362183, -0.014266968, -0.006290436, -0.019897461, -0.020019531, 0.01576233, -0.0030460358, -0.004047394, -0.0072288513, -0.022338867, -0.022918701, -0.025115967, -0.01928711...
http://arxiv.org/abs/2310.06271v1
Towards Mitigating Hallucination in Large Language Models via Self-Reflection
Ziwei Ji, Tiezheng Yu, Yan Xu, Nayeon Lee, Etsuko Ishii, Pascale Fung
cs.CL, cs.AI
Large language models (LLMs) have shown promise for generative and knowledge-intensive tasks including question-answering (QA) tasks. However, the practical deployment still faces challenges, notably the issue of "hallucination", where models generate plausible-sounding but unfaithful or nonsensical information. This i...
http://arxiv.org/pdf/2310.06271v1
2023-10-10T03:05:44
2023-10-10T03:05:44
{ "title": [ -0.011672974, 0.0015459061, -0.030471802, -0.06222534, -0.0018587112, 0.002122879, -0.048980713, -0.00013363361, 0.0036888123, 0.037109375, 0.0034255981, -0.008918762, 0.01360321, -0.02015686, -0.060913086, 0.0129776, -0.0073890686, ...
http://arxiv.org/abs/2310.06260v1
An experiment on an automated literature survey of data-driven speech enhancement methods
Arthur dos Santos, Jayr Pereira, Rodrigo Nogueira, Bruno Masiero, Shiva Sander-Tavallaey, Elias Zea
cs.SD, cs.CL, eess.AS
The increasing number of scientific publications in acoustics, in general, presents difficulties in conducting traditional literature surveys. This work explores the use of a generative pre-trained transformer (GPT) model to automate a literature survey of 116 articles on data-driven speech enhancement methods. The mai...
http://arxiv.org/pdf/2310.06260v1
2023-10-10T02:07:24
2023-10-10T02:07:24
{ "title": [ -0.0018749237, 0.02558899, -0.06774902, -0.041778564, 0.017211914, -0.036834717, 0.00052928925, -0.00023329258, -0.036468506, 0.045928955, -0.018554688, 0.008811951, 0.007972717, -0.02961731, -0.035369873, -0.0031166077, -0.021636963, ...
http://arxiv.org/abs/2310.06254v1
Get the gist? Using large language models for few-shot decontextualization
Benjamin Kane, Lenhart Schubert
cs.CL, cs.AI
In many NLP applications that involve interpreting sentences within a rich context -- for instance, information retrieval systems or dialogue systems -- it is desirable to be able to preserve the sentence in a form that can be readily understood without context, for later reuse -- a process known as ``decontextualizati...
http://arxiv.org/pdf/2310.06254v1
2023-10-10T02:00:00
2023-10-10T02:00:00
{ "title": [ -0.013664246, -0.0055389404, -0.016555786, -0.036712646, -0.014915466, 0.0067634583, -0.057128906, -0.030807495, 0.006099701, 0.03024292, -0.016555786, 0.005432129, 0.004184723, -0.029144287, -0.044708252, 0.004173279, 0.0024051666, ...
http://arxiv.org/abs/2310.06245v1
We are what we repeatedly do: Inducing and deploying habitual schemas in persona-based responses
Benjamin Kane, Lenhart Schubert
cs.AI, cs.CL
Many practical applications of dialogue technology require the generation of responses according to a particular developer-specified persona. While a variety of personas can be elicited from recent large language models, the opaqueness and unpredictability of these models make it desirable to be able to specify persona...
http://arxiv.org/pdf/2310.06245v1
2023-10-10T01:44:47
2023-10-10T01:44:47
{ "title": [ 0.016723633, -0.0015926361, -0.02468872, -0.039398193, 0.013145447, -0.03024292, 0.0074501038, -0.00749588, 0.005004883, 0.025299072, 0.0029468536, 0.008049011, -0.0064926147, -0.040740967, -0.044799805, -0.0036792755, -0.015342712, ...
http://arxiv.org/abs/2310.06239v1
Model Tuning or Prompt Tuning? A Study of Large Language Models for Clinical Concept and Relation Extraction
Cheng Peng, Xi Yang, Kaleb E Smith, Zehao Yu, Aokun Chen, Jiang Bian, Yonghui Wu
cs.CL, cs.AI
Objective To develop soft prompt-based learning algorithms for large language models (LLMs), examine the shape of prompts, prompt-tuning using frozen/unfrozen LLMs, transfer learning, and few-shot learning abilities. Methods We developed a soft prompt-based LLM model and compared 4 training strategies including (1) fin...
http://arxiv.org/pdf/2310.06239v1
2023-10-10T01:27:08
2023-10-10T01:27:08
{ "title": [ -0.009635925, -0.010940552, -0.027511597, -0.093444824, -0.0069732666, 0.019683838, -0.07922363, -0.007003784, 0.016311646, 0.018569946, -0.010406494, -0.0068244934, 0.0073776245, -0.025741577, -0.039001465, 0.007980347, -0.015182495, ...
http://arxiv.org/abs/2310.06238v1
Tackling Data Bias in MUSIC-AVQA: Crafting a Balanced Dataset for Unbiased Question-Answering
Xiulong Liu, Zhikang Dong, Peng Zhang
cs.CV, cs.AI, cs.CL, cs.LG, cs.MM, cs.SD, eess.AS
In recent years, there has been a growing emphasis on the intersection of audio, vision, and text modalities, driving forward the advancements in multimodal research. However, strong bias that exists in any modality can lead to the model neglecting the others. Consequently, the model's ability to effectively reason acr...
http://arxiv.org/pdf/2310.06238v1
2023-10-10T01:22:41
2023-10-10T01:22:41
{ "title": [ 0.0039138794, -0.0063056946, -0.06896973, -0.0637207, 0.022064209, -0.041656494, -0.011245728, 0.013923645, -0.0027618408, 0.021881104, -0.033447266, -0.0070724487, 0.0077171326, -0.027618408, -0.037017822, -0.01977539, -0.007621765, ...
http://arxiv.org/abs/2310.06228v1
Evolution of Natural Language Processing Technology: Not Just Language Processing Towards General Purpose AI
Masahiro Yamamoto
cs.CL, cs.AI, I.2.7
Since the invention of computers, communication through natural language (actual human language) has been a dream technology. However, natural language is extremely difficult to mathematically formulate, making it difficult to realize as an algorithm without considering programming. While there have been numerous techn...
http://arxiv.org/pdf/2310.06228v1
2023-10-10T00:41:38
2023-10-10T00:41:38
{ "title": [ -0.032470703, -0.0040359497, -0.041625977, -0.043518066, 0.006702423, -0.018188477, -0.013931274, -0.0035800934, -0.009017944, 0.035705566, -0.036224365, -0.008460999, 0.012809753, -0.012062073, -0.027008057, -0.009094238, 0.0025501251, ...
http://arxiv.org/abs/2310.06213v1
GeoLLM: Extracting Geospatial Knowledge from Large Language Models
Rohin Manvi, Samar Khanna, Gengchen Mai, Marshall Burke, David Lobell, Stefano Ermon
cs.CL, cs.LG
The application of machine learning (ML) in a range of geospatial tasks is increasingly common but often relies on globally available covariates such as satellite imagery that can either be expensive or lack predictive power. Here we explore the question of whether the vast amounts of knowledge found in Internet langua...
http://arxiv.org/pdf/2310.06213v1
2023-10-10T00:03:23
2023-10-10T00:03:23
{ "title": [ -0.030044556, -0.030181885, -0.034301758, -0.055725098, -0.010253906, 0.0022335052, -0.05807495, -0.008857727, 0.01525116, 0.027542114, -0.020233154, -0.00969696, -0.0017766953, -0.015586853, -0.04208374, 0.00258255, -0.001209259, 0....
http://arxiv.org/abs/2310.06204v1
Estimating Numbers without Regression
Avijit Thawani, Jay Pujara, Ashwin Kalyan
cs.CL, cs.AI
Despite recent successes in language models, their ability to represent numbers is insufficient. Humans conceptualize numbers based on their magnitudes, effectively projecting them on a number line; whereas subword tokenization fails to explicitly capture magnitude by splitting numbers into arbitrary chunks. To allevia...
http://arxiv.org/pdf/2310.06204v1
2023-10-09T23:07:05
2023-10-09T23:07:05
{ "title": [ 0.012832642, 0.002532959, -0.01626587, -0.019882202, -0.0072784424, -0.032836914, -0.0012140274, -0.02456665, -0.0031986237, 0.00059604645, -0.053955078, -0.021087646, -0.08856201, -0.035827637, 0.0023555756, -0.031707764, 0.03967285, ...
http://arxiv.org/abs/2310.06202v1
GPT-who: An Information Density-based Machine-Generated Text Detector
Saranya Venkatraman, Adaku Uchendu, Dongwon Lee
cs.CL
The Uniform Information Density principle posits that humans prefer to spread information evenly during language production. In this work, we examine if the UID principle can help capture differences between Large Language Models (LLMs) and human-generated text. We propose GPT-who, the first psycholinguistically-aware ...
http://arxiv.org/pdf/2310.06202v1
2023-10-09T23:06:05
2023-10-09T23:06:05
{ "title": [ -0.016983032, 0.004470825, -0.024520874, -0.040100098, 0.0093688965, -0.0026435852, -0.027053833, -0.0335083, 0.0074157715, 0.037994385, -0.034240723, 0.010650635, 0.014099121, 0.0041275024, -0.057434082, 0.008308411, -0.023147583, 0...
http://arxiv.org/abs/2310.06201v1
Compressing Context to Enhance Inference Efficiency of Large Language Models
Yucheng Li, Bo Dong, Chenghua Lin, Frank Guerin
cs.CL
Large language models (LLMs) achieved remarkable performance across various tasks. However, they face challenges in managing long documents and extended conversations, due to significantly increased computational requirements, both in memory and inference time, and potential context truncation when the input exceeds th...
http://arxiv.org/pdf/2310.06201v1
2023-10-09T23:03:24
2023-10-09T23:03:24
{ "title": [ -0.0357666, -0.027648926, -0.05307007, -0.08166504, 0.0046806335, 0.012748718, -0.06048584, 0.03564453, 0.022445679, 0.029815674, -0.01852417, -0.007980347, 0.015274048, -0.01701355, -0.04333496, 0.018600464, 0.008491516, 0.024047852...
http://arxiv.org/abs/2310.06200v2
The Importance of Prompt Tuning for Automated Neuron Explanations
Justin Lee, Tuomas Oikarinen, Arjun Chatha, Keng-Chi Chang, Yilan Chen, Tsui-Wei Weng
cs.CL, cs.LG
Recent advances have greatly increased the capabilities of large language models (LLMs), but our understanding of the models and their safety has not progressed as fast. In this paper we aim to understand LLMs deeper by studying their individual neurons. We build upon previous work showing large language models such as...
http://arxiv.org/pdf/2310.06200v2
2023-10-11T05:00:10
2023-10-09T23:02:07
{ "title": [ 0.017150879, -0.011772156, -0.0045433044, -0.028076172, -0.053466797, -0.013298035, -0.029785156, -0.024414062, -0.023284912, 0.006164551, -0.0034332275, -0.029037476, -0.020980835, -0.05505371, -0.00043702126, -0.0050468445, 0.024993896...
http://arxiv.org/abs/2310.06165v2
CAW-coref: Conjunction-Aware Word-level Coreference Resolution
Karel D'Oosterlinck, Semere Kiros Bitew, Brandon Papineau, Christopher Potts, Thomas Demeester, Chris Develder
cs.CL, cs.AI
State-of-the-art coreference resolutions systems depend on multiple LLM calls per document and are thus prohibitively expensive for many use cases (e.g., information extraction with large corpora). The leading word-level coreference system (WL-coref) attains 96.6% of these SOTA systems' performance while being much mor...
http://arxiv.org/pdf/2310.06165v2
2023-10-19T17:31:14
2023-10-09T21:32:49
{ "title": [ 0.015014648, -0.015609741, -0.027709961, -0.103637695, 0.0030059814, 0.0017900467, -0.056854248, 0.02305603, 0.015617371, 0.04623413, -0.01235199, 0.02394104, 0.03765869, -0.0041542053, -0.036102295, 0.034820557, -0.025299072, 0.0498...
http://arxiv.org/abs/2310.06117v1
Take a Step Back: Evoking Reasoning via Abstraction in Large Language Models
Huaixiu Steven Zheng, Swaroop Mishra, Xinyun Chen, Heng-Tze Cheng, Ed H. Chi, Quoc V Le, Denny Zhou
cs.LG, cs.AI, cs.CL
We present Step-Back Prompting, a simple prompting technique that enables LLMs to do abstractions to derive high-level concepts and first principles from instances containing specific details. Using the concepts and principles to guide the reasoning steps, LLMs significantly improve their abilities in following a corre...
http://arxiv.org/pdf/2310.06117v1
2023-10-09T19:48:55
2023-10-09T19:48:55
{ "title": [ -0.021972656, -0.016708374, -0.049926758, -0.10467529, 0.011314392, 0.008117676, -0.05996704, 0.02947998, 0.014801025, 0.0184021, -0.02709961, -0.0059051514, -0.008903503, -0.01512146, -0.043426514, -0.0022277832, -0.014122009, 0.036...
http://arxiv.org/abs/2310.06111v1
BYOC: Personalized Few-Shot Classification with Co-Authored Class Descriptions
Arth Bohra, Govert Verkes, Artem Harutyunyan, Pascal Weinberger, Giovanni Campagna
cs.CL, cs.LG
Text classification is a well-studied and versatile building block for many NLP applications. Yet, existing approaches require either large annotated corpora to train a model with or, when using large language models as a base, require carefully crafting the prompt as well as using a long context that can fit many exam...
http://arxiv.org/pdf/2310.06111v1
2023-10-09T19:37:38
2023-10-09T19:37:38
{ "title": [ -0.0090408325, 0.005657196, -0.04083252, -0.045562744, 0.01386261, -0.013671875, -0.008010864, -0.012687683, -0.022720337, 0.002664566, -0.032073975, 0.034362793, -0.040618896, -0.015258789, -0.010345459, -0.034088135, -0.007396698, ...
http://arxiv.org/abs/2310.06103v1
Leveraging Multilingual Self-Supervised Pretrained Models for Sequence-to-Sequence End-to-End Spoken Language Understanding
Pavel Denisov, Ngoc Thang Vu
cs.CL, cs.SD, eess.AS
A number of methods have been proposed for End-to-End Spoken Language Understanding (E2E-SLU) using pretrained models, however their evaluation often lacks multilingual setup and tasks that require prediction of lexical fillers, such as slot filling. In this work, we propose a unified method that integrates multilingua...
http://arxiv.org/pdf/2310.06103v1
2023-10-09T19:22:51
2023-10-09T19:22:51
{ "title": [ 0.004070282, 0.005683899, -0.02822876, -0.10333252, -0.004421234, 0.008460999, -0.08880615, 0.025726318, -0.006351471, 0.016235352, -0.0012302399, -0.0018033981, -0.011795044, -0.02998352, -0.03668213, -0.0025405884, -0.0013580322, 0...
http://arxiv.org/abs/2310.06061v1
Auditing Gender Analyzers on Text Data
Siddharth D Jaiswal, Ankit Kumar Verma, Animesh Mukherjee
cs.CY, cs.CL
AI models have become extremely popular and accessible to the general public. However, they are continuously under the scanner due to their demonstrable biases toward various sections of the society like people of color and non-binary people. In this study, we audit three existing gender analyzers -- uClassify, Readabl...
http://arxiv.org/pdf/2310.06061v1
2023-10-09T18:13:07
2023-10-09T18:13:07
{ "title": [ -0.013450623, 0.0049324036, -0.04815674, -0.026657104, -0.01914978, -0.017669678, -0.016662598, 0.0048179626, -0.011116028, 0.058624268, -0.058044434, -0.010948181, 0.039123535, -0.002811432, -0.026473999, -0.0154953, -0.02607727, 0....
http://arxiv.org/abs/2310.06046v1
LLM for SoC Security: A Paradigm Shift
Dipayan Saha, Shams Tarek, Katayoon Yahyaei, Sujan Kumar Saha, Jingbo Zhou, Mark Tehranipoor, Farimah Farahmandi
cs.CR, cs.AI, cs.CL
As the ubiquity and complexity of system-on-chip (SoC) designs increase across electronic devices, the task of incorporating security into an SoC design flow poses significant challenges. Existing security solutions are inadequate to provide effective verification of modern SoC designs due to their limitations in scala...
http://arxiv.org/pdf/2310.06046v1
2023-10-09T18:02:38
2023-10-09T18:02:38
{ "title": [ -0.041015625, 0.0062713623, -0.050750732, -0.03729248, -0.02407837, -0.009597778, -0.03277588, 0.013618469, -0.0692749, -0.008163452, -0.021743774, 0.025741577, -0.042816162, -0.05117798, 0.02935791, -0.02243042, -0.004348755, 0.0029...
http://arxiv.org/abs/2310.05919v1
Few-Shot Spoken Language Understanding via Joint Speech-Text Models
Chung-Ming Chien, Mingjiamei Zhang, Ju-Chieh Chou, Karen Livescu
cs.CL, eess.AS
Recent work on speech representation models jointly pre-trained with text has demonstrated the potential of improving speech representations by encoding speech and text in a shared space. In this paper, we leverage such shared representations to address the persistent challenge of limited data availability in spoken la...
http://arxiv.org/pdf/2310.05919v1
2023-10-09T17:59:21
2023-10-09T17:59:21
{ "title": [ -0.010536194, -0.013252258, -0.01171875, -0.07080078, -0.01663208, 0.019897461, -0.06732178, 0.019989014, 0.0039405823, 0.020355225, -0.03201294, -0.006717682, -0.006504059, -0.00844574, -0.056671143, -0.013977051, -0.007858276, 0.01...
http://arxiv.org/abs/2310.05915v1
FireAct: Toward Language Agent Fine-tuning
Baian Chen, Chang Shu, Ehsan Shareghi, Nigel Collier, Karthik Narasimhan, Shunyu Yao
cs.CL, cs.AI, cs.LG
Recent efforts have augmented language models (LMs) with external tools or environments, leading to the development of language agents that can reason and act. However, most of these agents rely on few-shot prompting techniques with off-the-shelf LMs. In this paper, we investigate and argue for the overlooked direction...
http://arxiv.org/pdf/2310.05915v1
2023-10-09T17:58:38
2023-10-09T17:58:38
{ "title": [ -0.028152466, -0.018218994, -0.027191162, -0.07092285, 0.009628296, 0.011306763, -0.03579712, 0.015792847, -0.013374329, 0.028381348, -0.023239136, 0.021377563, -0.002149582, -0.0059776306, -0.03656006, 0.0050354004, -0.023269653, 0....
http://arxiv.org/abs/2310.05914v2
NEFTune: Noisy Embeddings Improve Instruction Finetuning
Neel Jain, Ping-yeh Chiang, Yuxin Wen, John Kirchenbauer, Hong-Min Chu, Gowthami Somepalli, Brian R. Bartoldson, Bhavya Kailkhura, Avi Schwarzschild, Aniruddha Saha, Micah Goldblum, Jonas Geiping, Tom Goldstein
cs.CL, cs.LG
We show that language model finetuning can be improved, sometimes dramatically, with a simple augmentation. NEFTune adds noise to the embedding vectors during training. Standard finetuning of LLaMA-2-7B using Alpaca achieves 29.79% on AlpacaEval, which rises to 64.69% using noisy embeddings. NEFTune also improves over ...
http://arxiv.org/pdf/2310.05914v2
2023-10-10T17:31:00
2023-10-09T17:58:34
{ "title": [ -0.015914917, -0.008529663, -0.01234436, -0.04055786, -0.0020656586, -0.006111145, -0.0064468384, -0.00315094, -0.00021457672, 0.008293152, -0.055633545, 0.042388916, -0.026138306, -0.003320694, -0.0022277832, -0.014282227, -0.012191772,...
http://arxiv.org/abs/2310.05910v1
SALMON: Self-Alignment with Principle-Following Reward Models
Zhiqing Sun, Yikang Shen, Hongxin Zhang, Qinhong Zhou, Zhenfang Chen, David Cox, Yiming Yang, Chuang Gan
cs.CL, cs.AI, cs.LG
Supervised Fine-Tuning (SFT) on response demonstrations combined with Reinforcement Learning from Human Feedback (RLHF) constitutes a powerful paradigm for aligning LLM-based AI agents. However, a significant limitation of such an approach is its dependency on high-quality human annotations, making its application to i...
http://arxiv.org/pdf/2310.05910v1
2023-10-09T17:56:53
2023-10-09T17:56:53
{ "title": [ 0.0012311935, 0.013244629, -0.022567749, -0.014572144, 0.0041389465, -0.018920898, 0.0027046204, -0.04940796, -0.014968872, 0.03213501, 0.0158844, -0.0053253174, -0.03277588, -0.033355713, -0.046325684, -0.0008215904, -0.0059280396, ...
http://arxiv.org/abs/2310.05884v1
A Meta-Learning Perspective on Transformers for Causal Language Modeling
Xinbo Wu, Lav R. Varshney
cs.LG, cs.AI, cs.CL
The Transformer architecture has become prominent in developing large causal language models. However, mechanisms to explain its capabilities are not well understood. Focused on the training process, here we establish a meta-learning view of the Transformer architecture when trained for the causal language modeling tas...
http://arxiv.org/pdf/2310.05884v1
2023-10-09T17:27:36
2023-10-09T17:27:36
{ "title": [ -0.01625061, -0.010269165, -0.026321411, -0.08074951, -0.019454956, 0.025436401, -0.0925293, 0.0022659302, -0.010421753, 0.02545166, -0.012916565, 0.013442993, 0.014160156, -0.027709961, -0.025756836, 0.004432678, -0.021987915, 0.081...
http://arxiv.org/abs/2310.05881v1
Controllable Chest X-Ray Report Generation from Longitudinal Representations
Francesco Dalla Serra, Chaoyang Wang, Fani Deligianni, Jeffrey Dalton, Alison Q O'Neil
cs.CV, cs.CL
Radiology reports are detailed text descriptions of the content of medical scans. Each report describes the presence/absence and location of relevant clinical findings, commonly including comparison with prior exams of the same patient to describe how they evolved. Radiology reporting is a time-consuming process, and s...
http://arxiv.org/pdf/2310.05881v1
2023-10-09T17:22:58
2023-10-09T17:22:58
{ "title": [ -0.0008678436, 0.01322937, -0.048950195, -0.050994873, 0.010826111, -0.0051345825, -0.019927979, -0.013168335, 0.0068511963, 0.031341553, -0.025405884, 0.017196655, -0.024978638, -0.026382446, -0.04852295, 0.004360199, -0.00027275085, ...
http://arxiv.org/abs/2310.13002v1
Are Large Language Models Geospatially Knowledgeable?
Prabin Bhandari, Antonios Anastasopoulos, Dieter Pfoser
cs.CL
Despite the impressive performance of Large Language Models (LLM) for various natural language processing tasks, little is known about their comprehension of geographic data and related ability to facilitate informed geospatial decision-making. This paper investigates the extent of geospatial knowledge, awareness, and ...
http://arxiv.org/pdf/2310.13002v1
2023-10-09T17:20:11
2023-10-09T17:20:11
{ "title": [ -0.0032253265, -0.0043678284, -0.017410278, -0.046081543, -0.03286743, -0.004219055, -0.047424316, -0.008705139, -0.002319336, 0.025131226, -0.025619507, -0.031280518, -0.0003335476, -0.021575928, -0.032684326, -0.021469116, 0.020614624,...
http://arxiv.org/abs/2310.05872v1
ViCor: Bridging Visual Understanding and Commonsense Reasoning with Large Language Models
Kaiwen Zhou, Kwonjoon Lee, Teruhisa Misu, Xin Eric Wang
cs.CV, cs.AI, cs.CL
In our work, we explore the synergistic capabilities of pre-trained vision-and-language models (VLMs) and large language models (LLMs) for visual commonsense reasoning (VCR). We categorize the problem of VCR into visual commonsense understanding (VCU) and visual commonsense inference (VCI). For VCU, which involves perc...
http://arxiv.org/pdf/2310.05872v1
2023-10-09T17:10:35
2023-10-09T17:10:35
{ "title": [ -0.012138367, -0.037322998, -0.017074585, -0.06390381, -0.023376465, 0.01109314, -0.043792725, -0.007888794, 0.0435791, 0.0101623535, -0.02709961, -0.012313843, -0.008346558, -0.022491455, -0.043670654, -0.0076408386, -0.0001950264, ...
http://arxiv.org/abs/2310.05861v1
Rephrase, Augment, Reason: Visual Grounding of Questions for Vision-Language Models
Archiki Prasad, Elias Stengel-Eskin, Mohit Bansal
cs.CL, cs.AI, cs.CV, cs.LG
An increasing number of vision-language tasks can be handled with little to no training, i.e., in a zero and few-shot manner, by marrying large language models (LLMs) to vision encoders, resulting in large vision-language models (LVLMs). While this has huge upsides, such as not requiring training data or custom archite...
http://arxiv.org/pdf/2310.05861v1
2023-10-09T16:57:57
2023-10-09T16:57:57
{ "title": [ -0.022827148, -0.015686035, -0.024902344, -0.04711914, -0.021087646, 0.000006854534, -0.036468506, -0.0067367554, 0.038085938, 0.015014648, -0.044830322, 0.010612488, -0.0023441315, -0.007095337, -0.053466797, 0.00013947487, 0.012451172,...
http://arxiv.org/abs/2310.05857v2
Improving Summarization with Human Edits
Zonghai Yao, Benjamin J Schloss, Sai P. Selvaraj
cs.CL, cs.AI, cs.LG
Recent work has shown the promise of learning with human feedback paradigms to produce human-determined high-quality text. Existing works use human feedback to train large language models (LLMs) in general domain abstractive summarization and have obtained summary quality exceeding traditional likelihood training. In t...
http://arxiv.org/pdf/2310.05857v2
2023-10-24T20:00:21
2023-10-09T16:52:07
{ "title": [ -0.005596161, 0.0048332214, -0.021865845, -0.06738281, -0.013198853, 0.0071105957, -0.055389404, -0.014961243, -0.006111145, 0.058013916, -0.010414124, 0.0016651154, -0.0060195923, 0.012687683, -0.074035645, 0.004512787, -0.009292603, ...
http://arxiv.org/abs/2310.05845v1
GraphLLM: Boosting Graph Reasoning Ability of Large Language Model
Ziwei Chai, Tianjie Zhang, Liang Wu, Kaiqiao Han, Xiaohai Hu, Xuanwen Huang, Yang Yang
cs.CL, cs.AI
The advancement of Large Language Models (LLMs) has remarkably pushed the boundaries towards artificial general intelligence (AGI), with their exceptional ability on understanding diverse types of information, including but not limited to images and audio. Despite this progress, a critical gap remains in empowering LLM...
http://arxiv.org/pdf/2310.05845v1
2023-10-09T16:42:00
2023-10-09T16:42:00
{ "title": [ -0.032287598, -0.030838013, -0.040924072, -0.0881958, -0.012504578, 0.023788452, -0.08404541, 0.030288696, 0.011199951, 0.0063285828, -0.006565094, 0.0056037903, -0.0019388199, -0.03286743, -0.008705139, 0.010772705, -0.01939392, 0.0...
http://arxiv.org/abs/2310.05824v1
Terminology-Aware Translation with Constrained Decoding and Large Language Model Prompting
Nikolay Bogoychev, Pinzhen Chen
cs.CL
Terminology correctness is important in the downstream application of machine translation, and a prevalent way to ensure this is to inject terminology constraints into a translation system. In our submission to the WMT 2023 terminology translation task, we adopt a translate-then-refine approach which can be domain-inde...
http://arxiv.org/pdf/2310.05824v1
2023-10-09T16:08:23
2023-10-09T16:08:23
{ "title": [ -0.04837036, -0.0063095093, -0.030776978, -0.06225586, -0.016418457, 0.016479492, -0.07513428, 0.012420654, 0.013420105, 0.044769287, 0.00068712234, 0.00554657, 0.044067383, -0.022216797, -0.041259766, 0.03656006, 0.011512756, 0.0106...
http://arxiv.org/abs/2310.05818v1
SC-Safety: A Multi-round Open-ended Question Adversarial Safety Benchmark for Large Language Models in Chinese
Liang Xu, Kangkang Zhao, Lei Zhu, Hang Xue
cs.CL
Large language models (LLMs), like ChatGPT and GPT-4, have demonstrated remarkable abilities in natural language understanding and generation. However, alongside their positive impact on our daily tasks, they can also produce harmful content that negatively affects societal perceptions. To systematically assess the saf...
http://arxiv.org/pdf/2310.05818v1
2023-10-09T16:03:22
2023-10-09T16:03:22
{ "title": [ -0.008255005, 0.0011529922, 0.005645752, -0.0680542, -0.0069007874, 0.020431519, -0.05596924, -0.0037136078, -0.026641846, 0.024459839, -0.002380371, 0.027908325, 0.007156372, -0.038269043, -0.007232666, -0.0023517609, -0.0058670044, ...
http://arxiv.org/abs/2310.05804v1
Learning Language-guided Adaptive Hyper-modality Representation for Multimodal Sentiment Analysis
Haoyu Zhang, Yu Wang, Guanghao Yin, Kejun Liu, Yuanyuan Liu, Tianshu Yu
cs.AI, cs.CL, cs.CV, cs.MM
Though Multimodal Sentiment Analysis (MSA) proves effective by utilizing rich information from multiple sources (e.g., language, video, and audio), the potential sentiment-irrelevant and conflicting information across modalities may hinder the performance from being further improved. To alleviate this, we present Adapt...
http://arxiv.org/pdf/2310.05804v1
2023-10-09T15:43:07
2023-10-09T15:43:07
{ "title": [ -0.00844574, 0.006336212, -0.044433594, -0.050689697, 0.0033607483, -0.009941101, -0.046875, -0.0076026917, 0.023147583, 0.05038452, -0.011650085, -0.005622864, 0.024673462, -0.0098724365, -0.056762695, 0.0107040405, 0.0005617142, 0....
http://arxiv.org/abs/2310.05797v2
Are Large Language Models Post Hoc Explainers?
Nicholas Kroeger, Dan Ley, Satyapriya Krishna, Chirag Agarwal, Himabindu Lakkaraju
cs.CL, cs.AI, cs.LG
Large Language Models (LLMs) are increasingly used as powerful tools for a plethora of natural language processing (NLP) applications. A recent innovation, in-context learning (ICL), enables LLMs to learn new tasks by supplying a few examples in the prompt during inference time, thereby eliminating the need for model f...
http://arxiv.org/pdf/2310.05797v2
2023-10-10T19:33:38
2023-10-09T15:31:03
{ "title": [ -0.016616821, -0.0050964355, 0.0019168854, -0.072387695, -0.040161133, 0.016967773, -0.066101074, 0.0013103485, -0.026687622, 0.0078086853, -0.011756897, -0.013496399, -0.01826477, -0.037841797, 0.0071411133, -0.019058228, 0.01058197, ...
http://arxiv.org/abs/2310.05793v2
DiffuSeq-v2: Bridging Discrete and Continuous Text Spaces for Accelerated Seq2Seq Diffusion Models
Shansan Gong, Mukai Li, Jiangtao Feng, Zhiyong Wu, Lingpeng Kong
cs.LG, cs.CL
Diffusion models have gained prominence in generating high-quality sequences of text. Nevertheless, current approaches predominantly represent discrete text within a continuous diffusion space, which incurs substantial computational overhead during training and results in slower sampling speeds. In this paper, we intro...
http://arxiv.org/pdf/2310.05793v2
2023-10-16T09:56:02
2023-10-09T15:29:10
{ "title": [ -0.039886475, -0.0037250519, -0.070251465, -0.030654907, 0.0070877075, -0.015960693, -0.03503418, -0.026153564, 0.008354187, 0.015731812, -0.032989502, 0.019012451, 0.003129959, -0.029327393, -0.018692017, 0.00086927414, -0.016738892, ...
http://arxiv.org/abs/2310.05791v1
Problem-Solving Guide: Predicting the Algorithm Tags and Difficulty for Competitive Programming Problems
Juntae Kim, Eunjung Cho, Dongwoo Kim, Dongbin Na
cs.CL
The recent program development industries have required problem-solving abilities for engineers, especially application developers. However, AI-based education systems to help solve computer algorithm problems have not yet attracted attention, while most big tech companies require the ability to solve algorithm problem...
http://arxiv.org/pdf/2310.05791v1
2023-10-09T15:26:07
2023-10-09T15:26:07
{ "title": [ -0.0011358261, -0.014717102, 0.0029640198, 0.014129639, -0.0289917, -0.031097412, 0.017486572, -0.014122009, -0.046447754, 0.030303955, -0.017196655, 0.0034446716, -0.021072388, -0.02897644, -0.00617218, -0.043121338, 0.014862061, 0....
http://arxiv.org/abs/2310.05782v1
Aligning Language Models with Human Preferences via a Bayesian Approach
Jiashuo Wang, Haozhao Wang, Shichao Sun, Wenjie Li
cs.CL
In the quest to advance human-centric natural language generation (NLG) systems, ensuring alignment between NLG models and human preferences is crucial. For this alignment, current popular methods leverage a reinforcement learning (RL) approach with a reward model trained on feedback from humans. However, inherent disa...
http://arxiv.org/pdf/2310.05782v1
2023-10-09T15:15:05
2023-10-09T15:15:05
{ "title": [ -0.019683838, 0.0044898987, -0.041015625, -0.03286743, 0.00077199936, -0.016342163, -0.039398193, -0.0019702911, -0.0010442734, 0.01737976, -0.0184021, -0.020080566, 0.0020446777, -0.018554688, -0.044189453, -0.0066947937, -0.014923096, ...
http://arxiv.org/abs/2310.05746v1
Put Your Money Where Your Mouth Is: Evaluating Strategic Planning and Execution of LLM Agents in an Auction Arena
Jiangjie Chen, Siyu Yuan, Rong Ye, Bodhisattwa Prasad Majumder, Kyle Richardson
cs.CL, cs.AI
Can Large Language Models (LLMs) simulate human behavior in complex environments? LLMs have recently been shown to exhibit advanced reasoning skills but much of NLP evaluation still relies on static benchmarks. Answering this requires evaluation environments that probe strategic reasoning in competitive, dynamic scenar...
http://arxiv.org/pdf/2310.05746v1
2023-10-09T14:22:09
2023-10-09T14:22:09
{ "title": [ -0.023529053, -0.007537842, -0.016830444, -0.005016327, -0.021606445, -0.015304565, -0.0154418945, -0.0093688965, -0.016113281, 0.015777588, -0.03704834, 0.026687622, -0.04296875, -0.013191223, -0.023162842, -0.03918457, -0.021347046, ...
http://arxiv.org/abs/2310.05736v1
LLMLingua: Compressing Prompts for Accelerated Inference of Large Language Models
Huiqiang Jiang, Qianhui Wu, Chin-Yew Lin, Yuqing Yang, Lili Qiu
cs.CL, cs.LG
Large language models (LLMs) have been applied in various applications due to their astonishing capabilities. With advancements in technologies such as chain-of-thought (CoT) prompting and in-context learning (ICL), the prompts fed to LLMs are becoming increasingly lengthy, even exceeding tens of thousands of tokens. T...
http://arxiv.org/pdf/2310.05736v1
2023-10-09T14:10:21
2023-10-09T14:10:21
{ "title": [ -0.044708252, -0.036956787, -0.05505371, -0.07293701, -0.0155181885, 0.012268066, -0.082214355, 0.03579712, 0.0059127808, 0.014823914, -0.011192322, -0.008529663, 0.0038776398, -0.02432251, -0.033996582, 0.016571045, 0.006378174, 0.0...
http://arxiv.org/abs/2310.05727v1
The Program Testing Ability of Large Language Models for Code
Weimin Xiong, Yiwen Guo, Hao Chen
cs.CL, cs.AI, cs.LG, cs.SE
Recent development of large language models (LLMs) for code like CodeX and CodeT5+ demonstrates tremendous promise in achieving code intelligence. Their ability of synthesizing code that completes a program for performing a pre-defined task has been intensively tested and verified on benchmark datasets including HumanE...
http://arxiv.org/pdf/2310.05727v1
2023-10-09T13:55:45
2023-10-09T13:55:45
{ "title": [ -0.0368042, -0.017410278, -0.030883789, -0.049468994, -0.017715454, -0.006996155, -0.043029785, 0.016555786, 0.0098724365, 0.030685425, -0.031982422, -0.011917114, -0.0012741089, -0.023590088, -0.03704834, -0.0041275024, -0.009635925, ...
http://arxiv.org/abs/2310.05707v1
Guiding Language Model Reasoning with Planning Tokens
Xinyi Wang, Lucas Caccia, Oleksiy Ostapenko, Xingdi Yuan, Alessandro Sordoni
cs.CL, cs.AI, cs.LG
Large language models (LLMs) have recently attracted considerable interest for their ability to perform complex reasoning tasks, such as chain-of-thought reasoning. However, most of the existing approaches to enhance this ability rely heavily on data-driven methods, while neglecting the structural aspects of the model'...
http://arxiv.org/pdf/2310.05707v1
2023-10-09T13:29:37
2023-10-09T13:29:37
{ "title": [ -0.034118652, -0.042114258, -0.03704834, -0.088134766, -0.034088135, 0.02810669, -0.08679199, 0.028045654, -0.017684937, 0.01108551, -0.03213501, -0.0013237, -0.014892578, -0.023834229, 0.0047187805, -0.017837524, -0.0046463013, 0.01...
http://arxiv.org/abs/2310.10673v1
Towards Emotion-Based Synthetic Consciousness: Using LLMs to Estimate Emotion Probability Vectors
David Sinclair, Willem Pye
cs.CL
This paper shows how LLMs (Large Language Models) may be used to estimate a summary of the emotional state associated with piece of text. The summary of emotional state is a dictionary of words used to describe emotion together with the probability of the word appearing after a prompt comprising the original text and a...
http://arxiv.org/pdf/2310.10673v1
2023-10-09T13:29:36
2023-10-09T13:29:36
{ "title": [ -0.03765869, 0.0010089874, -0.033233643, -0.008979797, -0.011131287, -0.01348114, -0.0029640198, -0.021575928, 0.018875122, 0.004890442, -0.049468994, -0.0032463074, 0.0032920837, -0.034729004, -0.017150879, -0.018630981, -0.0020141602, ...
http://arxiv.org/abs/2310.05703v2
An Attribution Method for Siamese Encoders
Lucas Möller, Dmitry Nikolaev, Sebastian Padó
cs.CL, cs.AI, cs.LG
Despite the success of Siamese encoder models such as sentence transformers (ST), little is known about the aspects of inputs they pay attention to. A barrier is that their predictions cannot be attributed to individual features, as they compare two inputs rather than processing a single one. This paper derives a local...
http://arxiv.org/pdf/2310.05703v2
2023-10-23T15:26:40
2023-10-09T13:24:44
{ "title": [ -0.010215759, -0.00071287155, -0.02456665, -0.020996094, -0.001578331, -0.008178711, -0.01638794, -0.031234741, 0.00012010336, 0.025878906, -0.005962372, -0.002456665, -0.01071167, -0.031433105, -0.017532349, 0.011436462, -0.009483337, ...
http://arxiv.org/abs/2310.05694v1
A Survey of Large Language Models for Healthcare: from Data, Technology, and Applications to Accountability and Ethics
Kai He, Rui Mao, Qika Lin, Yucheng Ruan, Xiang Lan, Mengling Feng, Erik Cambria
cs.CL
The utilization of large language models (LLMs) in the Healthcare domain has generated both excitement and concern due to their ability to effectively respond to freetext queries with certain professional knowledge. This survey outlines the capabilities of the currently developed LLMs for Healthcare and explicates thei...
http://arxiv.org/pdf/2310.05694v1
2023-10-09T13:15:23
2023-10-09T13:15:23
{ "title": [ -0.025665283, -0.011932373, -0.059814453, -0.07550049, 0.014694214, -0.0129470825, -0.06341553, 0.015426636, -0.027694702, 0.026046753, -0.018478394, 0.006717682, -0.0030918121, -0.02407837, -0.013153076, -0.007736206, -0.00039482117, ...
http://arxiv.org/abs/2310.05688v1
Larth: Dataset and Machine Translation for Etruscan
Gianluca Vico, Gerasimos Spanakis
cs.CL
Etruscan is an ancient language spoken in Italy from the 7th century BC to the 1st century AD. There are no native speakers of the language at the present day, and its resources are scarce, as there exist only around 12,000 known inscriptions. To the best of our knowledge, there are no publicly available Etruscan corpo...
http://arxiv.org/pdf/2310.05688v1
2023-10-09T12:56:08
2023-10-09T12:56:08
{ "title": [ -0.042266846, -0.011528015, -0.04916382, -0.02558899, -0.011871338, -0.0002644062, -0.04168701, 0.0027675629, 0.022750854, 0.040161133, -0.029937744, -0.0085372925, 0.0496521, -0.032714844, -0.02583313, 0.021118164, 0.008659363, 0.04...
http://arxiv.org/abs/2310.05686v1
The potential of large language models for improving probability learning: A study on ChatGPT3.5 and first-year computer engineering students
Angel Udias, Antonio Alonso-Ayuso, Ignacio Sanchez, Sonia Hernandez, Maria Eugenia Castellanos, Raquel Montes Diez, Emilio Lopez Cano
cs.CL, cs.AI, I2, I.2
In this paper, we assess the efficacy of ChatGPT (version Feb 2023), a large-scale language model, in solving probability problems typically presented in introductory computer engineering exams. Our study comprised a set of 23 probability exercises administered to students at Rey Juan Carlos University (URJC) in Madrid...
http://arxiv.org/pdf/2310.05686v1
2023-10-09T12:54:58
2023-10-09T12:54:58
{ "title": [ -0.010276794, -0.012619019, -0.015197754, -0.012916565, -0.017303467, -0.009742737, -0.014656067, -0.00042438507, 0.0012769699, 0.023025513, -0.01576233, 0.00012922287, 0.0135650635, -0.030151367, -0.0024299622, 0.011436462, -0.0262146, ...
http://arxiv.org/abs/2310.05657v1
A Closer Look into Automatic Evaluation Using Large Language Models
Cheng-Han Chiang, Hung-yi Lee
cs.CL
Using large language models (LLMs) to evaluate text quality has recently gained popularity. Some prior works explore the idea of using LLMs for evaluation, while they differ in some details of the evaluation process. In this paper, we analyze LLM evaluation (Chiang and Lee, 2023) and G-Eval (Liu et al., 2023), and we d...
http://arxiv.org/pdf/2310.05657v1
2023-10-09T12:12:55
2023-10-09T12:12:55
{ "title": [ -0.026641846, 0.0019445419, -0.020004272, -0.04940796, 0.00566864, 0.00289917, -0.048858643, -0.0206604, 0.0017967224, 0.023025513, -0.018127441, -0.014602661, -0.01638794, -0.022109985, -0.047058105, -0.0025405884, -0.0234375, 0.038...
http://arxiv.org/abs/2310.05650v1
RAUCG: Retrieval-Augmented Unsupervised Counter Narrative Generation for Hate Speech
Shuyu Jiang, Wenyi Tang, Xingshu Chen, Rui Tanga, Haizhou Wang, Wenxian Wang
cs.CL
The Counter Narrative (CN) is a promising approach to combat online hate speech (HS) without infringing on freedom of speech. In recent years, there has been a growing interest in automatically generating CNs using natural language generation techniques. However, current automatic CN generation methods mainly rely on e...
http://arxiv.org/pdf/2310.05650v1
2023-10-09T12:01:26
2023-10-09T12:01:26
{ "title": [ 0.026855469, -0.009315491, -0.0061149597, -0.043823242, -0.005844116, -0.012840271, 0.0031471252, -0.020950317, -0.010688782, 0.04058838, -0.022125244, 0.034973145, 0.007297516, -0.005016327, -0.044921875, -0.0026664734, -0.02822876, ...
http://arxiv.org/abs/2310.05634v1
Towards Verifiable Generation: A Benchmark for Knowledge-aware Language Model Attribution
Xinze Li, Yixin Cao2, Liangming Pan, Yubo Ma, Aixin Sun
cs.CL
Although achieving great success, Large Language Models (LLMs) usually suffer from unreliable hallucinations. In this paper, we define a new task of Knowledge-aware Language Model Attribution (KaLMA) that improves upon three core concerns on conventional attributed LMs. First, we extend attribution source from unstruct...
http://arxiv.org/pdf/2310.05634v1
2023-10-09T11:45:59
2023-10-09T11:45:59
{ "title": [ -0.027038574, -0.02671814, -0.047058105, -0.072143555, -0.012001038, 0.0059432983, -0.09588623, 0.0076942444, -0.00042295456, 0.025848389, -0.0019025803, 0.008552551, 0.013389587, -0.024368286, -0.031463623, 0.011650085, -0.027038574, ...
http://arxiv.org/abs/2310.05628v2
Glitter or Gold? Deriving Structured Insights from Sustainability Reports via Large Language Models
Marco Bronzini, Carlo Nicolini, Bruno Lepri, Andrea Passerini, Jacopo Staiano
cs.CL, cs.CE, cs.CY
Over the last decade, several regulatory bodies have started requiring the disclosure of non-financial information from publicly listed companies, in light of the investors' increasing attention to Environmental, Social, and Governance (ESG) issues. Such information is publicly released in a variety of non-structured a...
http://arxiv.org/pdf/2310.05628v2
2023-11-03T10:43:06
2023-10-09T11:34:41
{ "title": [ -0.04486084, -0.022872925, -0.06762695, -0.09246826, -0.02357483, 0.0028705597, -0.09631348, 0.02067566, -0.0024604797, 0.010414124, -0.046844482, -0.006816864, -0.034362793, -0.008460999, -0.045959473, -0.0362854, -0.007286072, 0.03...
http://arxiv.org/abs/2310.05627v1
Integrating Stock Features and Global Information via Large Language Models for Enhanced Stock Return Prediction
Yujie Ding, Shuai Jia, Tianyi Ma, Bingcheng Mao, Xiuze Zhou, Liuliu Li, Dongming Han
cs.CL, cs.LG, q-fin.ST
The remarkable achievements and rapid advancements of Large Language Models (LLMs) such as ChatGPT and GPT-4 have showcased their immense potential in quantitative investment. Traders can effectively leverage these LLMs to analyze financial news and predict stock returns accurately. However, integrating LLMs into exist...
http://arxiv.org/pdf/2310.05627v1
2023-10-09T11:34:18
2023-10-09T11:34:18
{ "title": [ -0.018951416, -0.009155273, -0.058929443, -0.057769775, 0.0030918121, -0.009460449, -0.038391113, 0.006729126, 0.016143799, 0.03213501, -0.0056495667, -0.018218994, 0.0017080307, -0.011177063, -0.048187256, 0.0065689087, 0.0074691772, ...
http://arxiv.org/abs/2310.05620v1
LAiW: A Chinese Legal Large Language Models Benchmark (A Technical Report)
Yongfu Dai, Duanyu Feng, Jimin Huang, Haochen Jia, Qianqian Xie, Yifang Zhang, Weiguang Han, Wei Tian, Hao Wang
cs.CL
With the emergence of numerous legal LLMs, there is currently a lack of a comprehensive benchmark for evaluating their legal abilities. In this paper, we propose the first Chinese Legal LLMs benchmark based on legal capabilities. Through the collaborative efforts of legal and artificial intelligence experts, we divide ...
http://arxiv.org/pdf/2310.05620v1
2023-10-09T11:19:55
2023-10-09T11:19:55
{ "title": [ -0.053222656, -0.018096924, -0.04473877, -0.044281006, -0.036010742, 0.006706238, -0.06347656, 0.01927185, -0.020599365, 0.031280518, -0.021942139, -0.003074646, 0.031341553, -0.045166016, 0.002614975, -0.010322571, -0.015464783, 0.0...
http://arxiv.org/abs/2310.05619v2
Dynamic Top-k Estimation Consolidates Disagreement between Feature Attribution Methods
Jonathan Kamp, Lisa Beinborn, Antske Fokkens
cs.CL, cs.AI
Feature attribution scores are used for explaining the prediction of a text classifier to users by highlighting a k number of tokens. In this work, we propose a way to determine the number of optimal k tokens that should be displayed from sequential properties of the attribution scores. Our approach is dynamic across s...
http://arxiv.org/pdf/2310.05619v2
2023-11-03T12:11:17
2023-10-09T11:19:33
{ "title": [ 0.0003335476, -0.017440796, -0.009925842, -0.004436493, -0.02394104, -0.014877319, 0.014373779, -0.03479004, -0.034576416, -0.0042419434, -0.0087890625, 0.009033203, -0.041900635, -0.045318604, 0.010772705, -0.01576233, 0.006248474, ...