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/2310.07968v1
Think, Act, and Ask: Open-World Interactive Personalized Robot Navigation
Yinpei Dai, Run Peng, Sikai Li, Joyce Chai
cs.RO, cs.CL, cs.HC
Zero-Shot Object Navigation (ZSON) enables agents to navigate towards open-vocabulary objects in unknown environments. The existing works of ZSON mainly focus on following individual instructions to find generic object classes, neglecting the utilization of natural language interaction and the complexities of identifyi...
http://arxiv.org/pdf/2310.07968v1
2023-10-12T01:17:56
2023-10-12T01:17:56
{ "title": [ -0.017959595, -0.0043525696, -0.0035057068, -0.01423645, -0.0003437996, -0.002111435, 0.0065307617, -0.030578613, -0.031311035, 0.01737976, -0.02230835, 0.010726929, -0.041503906, -0.016601562, -0.024627686, -0.028030396, 0.010894775, ...
http://arxiv.org/abs/2310.07962v1
Clustering of Spell Variations for Proper Nouns Transliterated from the other languages
Prathamesh Pawar
cs.CL
One of the prominent problems with processing and operating on text data is the non uniformity of it. Due to the change in the dialects and languages, the caliber of translation is low. This creates a unique problem while using NLP in text data; which is the spell variation arising from the inconsistent translations an...
http://arxiv.org/pdf/2310.07962v1
2023-10-12T00:57:32
2023-10-12T00:57:32
{ "title": [ 0.0039043427, 0.00033187866, -0.036071777, 0.0028057098, -0.038146973, -0.037841797, -0.012138367, -0.019805908, -0.013809204, 0.051086426, -0.023025513, 0.015213013, 0.050476074, -0.029144287, 0.0033130646, 0.00548172, 0.009819031, ...
http://arxiv.org/abs/2310.07957v2
A New Approach Towards Autoformalization
Nilay Patel, Rahul Saha, Jeffrey Flanigan
cs.CL, cs.AI
Verifying mathematical proofs is difficult, but can be automated with the assistance of a computer. Autoformalization is the task of automatically translating natural language mathematics into a formal language that can be verified by a program. This is a challenging task, and especially for higher-level mathematics fo...
http://arxiv.org/pdf/2310.07957v2
2023-10-19T18:08:20
2023-10-12T00:50:24
{ "title": [ 0.019073486, 0.014511108, -0.014625549, -0.10699463, -0.0026626587, 0.012580872, -0.097229004, 0.017730713, -0.020874023, 0.053131104, 0.008728027, -0.03488159, 0.0065574646, -0.050994873, -0.0015792847, 0.007888794, 0.03062439, -0.0...
http://arxiv.org/abs/2310.07931v1
D2 Pruning: Message Passing for Balancing Diversity and Difficulty in Data Pruning
Adyasha Maharana, Prateek Yadav, Mohit Bansal
cs.LG, cs.AI, cs.CL, cs.CV
Analytical theories suggest that higher-quality data can lead to lower test errors in models trained on a fixed data budget. Moreover, a model can be trained on a lower compute budget without compromising performance if a dataset can be stripped of its redundancies. Coreset selection (or data pruning) seeks to select a...
http://arxiv.org/pdf/2310.07931v1
2023-10-11T23:01:29
2023-10-11T23:01:29
{ "title": [ -0.040863037, 0.00083351135, -0.052093506, -0.0657959, 0.01386261, -0.00009584427, -0.044555664, -0.0004043579, 0.0096206665, 0.024734497, -0.01625061, 0.0036525726, 0.009002686, -0.028167725, -0.037994385, -0.003545761, 0.0015420914, ...
http://arxiv.org/abs/2310.07929v1
Crosslingual Structural Priming and the Pre-Training Dynamics of Bilingual Language Models
Catherine Arnett, Tyler A. Chang, James A. Michaelov, Benjamin K. Bergen
cs.CL
Do multilingual language models share abstract grammatical representations across languages, and if so, when do these develop? Following Sinclair et al. (2022), we use structural priming to test for abstract grammatical representations with causal effects on model outputs. We extend the approach to a Dutch-English bili...
http://arxiv.org/pdf/2310.07929v1
2023-10-11T22:57:03
2023-10-11T22:57:03
{ "title": [ -0.023361206, -0.0073242188, 0.004383087, -0.0491333, -0.020629883, 0.017593384, -0.051940918, -0.013290405, -0.022232056, 0.03552246, 0.0057868958, 0.026443481, 0.00024342537, -0.027160645, -0.028182983, 0.014389038, -0.010368347, 0...
http://arxiv.org/abs/2310.07923v3
The Expressive Power of Transformers with Chain of Thought
William Merrill, Ashish Sabharwal
cs.LG, cs.CC, cs.CL, cs.LO
Recent theoretical work has identified surprisingly simple reasoning problems, such as checking if two nodes in a graph are connected or simulating finite-state machines, that are provably unsolvable by standard transformers that answer immediately after reading their input. However, in practice, transformers' reasonin...
http://arxiv.org/pdf/2310.07923v3
2023-10-18T02:38:02
2023-10-11T22:35:18
{ "title": [ 0.058135986, 0.0010213852, -0.015060425, -0.026138306, -0.010314941, -0.032196045, 0.022613525, -0.0104599, -0.015472412, 0.018295288, -0.059570312, -0.015113831, -0.001750946, -0.044189453, 0.0063667297, -0.030227661, 0.014053345, 0...
http://arxiv.org/abs/2310.07911v1
Pit One Against Many: Leveraging Attention-head Embeddings for Parameter-efficient Multi-head Attention
Huiyin Xue, Nikolaos Aletras
cs.CL
Scaling pre-trained language models has resulted in large performance gains in various natural language processing tasks but comes with a large cost in memory requirements. Inspired by the position embeddings in transformers, we aim to simplify and reduce the memory footprint of the multi-head attention (MHA) mechanism...
http://arxiv.org/pdf/2310.07911v1
2023-10-11T21:38:40
2023-10-11T21:38:40
{ "title": [ 0.02508545, 0.01902771, -0.033569336, -0.07623291, 0.031555176, -0.0135650635, -0.013504028, -0.012008667, 0.020935059, 0.024291992, -0.016433716, 0.022491455, 0.0067214966, -0.005302429, -0.02217102, 0.022109985, -0.015853882, 0.048...
http://arxiv.org/abs/2310.07889v1
LangNav: Language as a Perceptual Representation for Navigation
Bowen Pan, Rameswar Panda, SouYoung Jin, Rogerio Feris, Aude Oliva, Phillip Isola, Yoon Kim
cs.CV, cs.AI, cs.CL, cs.RO
We explore the use of language as a perceptual representation for vision-and-language navigation. Our approach uses off-the-shelf vision systems (for image captioning and object detection) to convert an agent's egocentric panoramic view at each time step into natural language descriptions. We then finetune a pretrained...
http://arxiv.org/pdf/2310.07889v1
2023-10-11T20:52:30
2023-10-11T20:52:30
{ "title": [ -0.018295288, -0.015792847, -0.021881104, -0.020019531, -0.0033855438, -0.0178833, 0.00712204, -0.024871826, 0.0044898987, 0.03125, -0.018508911, -0.0087890625, -0.009895325, -0.020324707, -0.04949951, -0.005340576, -0.0001847744, 0....
http://arxiv.org/abs/2310.07875v1
TabLib: A Dataset of 627M Tables with Context
Gus Eggert, Kevin Huo, Mike Biven, Justin Waugh
cs.CL, cs.AI, cs.DB, cs.LG
It is well-established that large, diverse datasets play a pivotal role in the performance of modern AI systems for text and image modalities. However, there are no datasets for tabular data of comparable size and diversity to those available for text and images. Thus we present "TabLib'', a compilation of 627 million ...
http://arxiv.org/pdf/2310.07875v1
2023-10-11T20:34:42
2023-10-11T20:34:42
{ "title": [ 0.015930176, 0.004470825, -0.08709717, -0.05886841, -0.0014562607, -0.026763916, -0.05255127, -0.0061683655, -0.034729004, 0.047821045, -0.04333496, 0.010681152, -0.009117126, -0.017211914, 0.002374649, -0.03793335, 0.030563354, -0.0...
http://arxiv.org/abs/2310.07856v1
Assessing Evaluation Metrics for Neural Test Oracle Generation
Jiho Shin, Hadi Hemmati, Moshi Wei, Song Wang
cs.CL, cs.SE
In this work, we revisit existing oracle generation studies plus ChatGPT to empirically investigate the current standing of their performance in both NLG-based and test adequacy metrics. Specifically, we train and run four state-of-the-art test oracle generation models on five NLG-based and two test adequacy metrics fo...
http://arxiv.org/pdf/2310.07856v1
2023-10-11T19:58:07
2023-10-11T19:58:07
{ "title": [ -0.030548096, -0.009429932, -0.033599854, -0.044281006, -0.004459381, 0.0010938644, -0.029937744, 0.012313843, 0.0048599243, 0.032928467, -0.02180481, -0.006969452, -0.010307312, -0.015350342, -0.06750488, -0.0069885254, -0.012275696, ...
http://arxiv.org/abs/2310.07849v2
Synthetic Data Generation with Large Language Models for Text Classification: Potential and Limitations
Zhuoyan Li, Hangxiao Zhu, Zhuoran Lu, Ming Yin
cs.CL, cs.AI
The collection and curation of high-quality training data is crucial for developing text classification models with superior performance, but it is often associated with significant costs and time investment. Researchers have recently explored using large language models (LLMs) to generate synthetic datasets as an alte...
http://arxiv.org/pdf/2310.07849v2
2023-10-13T01:31:59
2023-10-11T19:51:13
{ "title": [ -0.036102295, -0.0071792603, -0.034210205, -0.03656006, -0.011741638, 0.0055885315, -0.04159546, -0.0069389343, -0.0046157837, 0.02470398, -0.021484375, 0.0051651, -0.021972656, -0.007713318, -0.042877197, -0.019729614, -0.02760315, ...
http://arxiv.org/abs/2310.07848v1
Framework for Question-Answering in Sanskrit through Automated Construction of Knowledge Graphs
Hrishikesh Terdalkar, Arnab Bhattacharya
cs.CL
Sanskrit (sa\d{m}sk\d{r}ta) enjoys one of the largest and most varied literature in the whole world. Extracting the knowledge from it, however, is a challenging task due to multiple reasons including complexity of the language and paucity of standard natural language processing tools. In this paper, we target the probl...
http://arxiv.org/pdf/2310.07848v1
2023-10-11T19:50:59
2023-10-11T19:50:59
{ "title": [ -0.018859863, -0.011260986, -0.016860962, -0.05517578, -0.02822876, 0.014450073, -0.04373169, -0.00066900253, -0.0014629364, 0.016830444, -0.029067993, 0.008026123, -0.004814148, -0.013923645, -0.046295166, -0.017318726, 0.009346008, ...
http://arxiv.org/abs/2310.07830v1
Does Synthetic Data Make Large Language Models More Efficient?
Sia Gholami, Marwan Omar
cs.CL, cs.AI, cs.LG
Natural Language Processing (NLP) has undergone transformative changes with the advent of deep learning methodologies. One challenge persistently confronting researchers is the scarcity of high-quality, annotated datasets that drive these models. This paper explores the nuances of synthetic data generation in NLP, with...
http://arxiv.org/pdf/2310.07830v1
2023-10-11T19:16:09
2023-10-11T19:16:09
{ "title": [ -0.022994995, -0.0028953552, -0.050811768, -0.056488037, -0.032165527, -0.019515991, -0.052581787, 0.016571045, -0.025466919, 0.0082473755, -0.044067383, -0.0034732819, -0.017456055, -0.016586304, 0.010620117, -0.031066895, -0.015075684,...
http://arxiv.org/abs/2310.07826v1
Antarlekhaka: A Comprehensive Tool for Multi-task Natural Language Annotation
Hrishikesh Terdalkar, Arnab Bhattacharya
cs.CL
One of the primary obstacles in the advancement of Natural Language Processing (NLP) technologies for low-resource languages is the lack of annotated datasets for training and testing machine learning models. In this paper, we present Antarlekhaka, a tool for manual annotation of a comprehensive set of tasks relevant t...
http://arxiv.org/pdf/2310.07826v1
2023-10-11T19:09:07
2023-10-11T19:09:07
{ "title": [ -0.025024414, -0.004562378, -0.032287598, -0.087402344, 0.0040283203, 0.019454956, -0.07019043, -0.004901886, -0.0028247833, 0.03262329, 0.0018796921, 0.023025513, -0.015670776, -0.010772705, -0.03857422, 0.020080566, -0.013587952, 0...
http://arxiv.org/abs/2310.07821v1
Non-autoregressive Text Editing with Copy-aware Latent Alignments
Yu Zhang, Yue Zhang, Leyang Cui, Guohong Fu
cs.CL
Recent work has witnessed a paradigm shift from Seq2Seq to Seq2Edit in the field of text editing, with the aim of addressing the slow autoregressive inference problem posed by the former. Despite promising results, Seq2Edit approaches still face several challenges such as inflexibility in generation and difficulty in g...
http://arxiv.org/pdf/2310.07821v1
2023-10-11T19:02:57
2023-10-11T19:02:57
{ "title": [ -0.03152466, 0.005695343, -0.04586792, -0.030288696, -0.0043907166, -0.0033168793, -0.041137695, -0.009529114, 0.011665344, 0.05114746, 0.0031108856, 0.007888794, 0.04534912, -0.016937256, -0.04748535, 0.021728516, -0.0149383545, 0.0...
http://arxiv.org/abs/2310.07819v1
Faithfulness Measurable Masked Language Models
Andreas Madsen, Siva Reddy, Sarath Chandar
cs.CL, cs.LG
A common approach to explain NLP models, is to use importance measures that express which tokens are important for a prediction. Unfortunately, such explanations are often wrong despite being persuasive. Therefore, it is essential to measure their faithfulness. One such metric is if tokens are truly important, then mas...
http://arxiv.org/pdf/2310.07819v1
2023-10-11T19:00:40
2023-10-11T19:00:40
{ "title": [ -0.061462402, 0.01928711, -0.084350586, -0.045043945, -0.026794434, -0.016601562, -0.097351074, 0.000051140785, -0.0033893585, 0.04147339, -0.0016222, -0.008773804, 0.038269043, -0.03967285, -0.0074882507, 0.032287598, -0.019836426, ...
http://arxiv.org/abs/2310.07818v2
Exploring the Relationship between Analogy Identification and Sentence Structure Encoding in Large Language Models
Thilini Wijesiriwardene, Ruwan Wickramarachchi, Aishwarya Naresh Reganti, Vinija Jain, Aman Chadha, Amit Sheth, Amitava Das
cs.CL, cs.AI
Identifying analogies plays a pivotal role in human cognition and language proficiency. In the last decade, there has been extensive research on word analogies in the form of ``A is to B as C is to D.'' However, there is a growing interest in analogies that involve longer text, such as sentences and collections of sent...
http://arxiv.org/pdf/2310.07818v2
2023-10-13T19:14:48
2023-10-11T18:59:48
{ "title": [ -0.03164673, -0.00843811, -0.004096985, -0.022415161, -0.018341064, 0.009613037, -0.04260254, -0.021621704, 0.0037612915, 0.037139893, 0.0034275055, 0.003238678, 0.04550171, -0.020507812, -0.037597656, 0.027130127, -0.016738892, 0.05...
http://arxiv.org/abs/2310.07815v1
Language Models As Semantic Indexers
Bowen Jin, Hansi Zeng, Guoyin Wang, Xiusi Chen, Tianxin Wei, Ruirui Li, Zhengyang Wang, Zheng Li, Yang Li, Hanqing Lu, Suhang Wang, Jiawei Han, Xianfeng Tang
cs.IR, cs.CL, cs.LG
Semantic identifier (ID) is an important concept in information retrieval that aims to preserve the semantics of objects such as documents and items inside their IDs. Previous studies typically adopt a two-stage pipeline to learn semantic IDs by first procuring embeddings using off-the-shelf text encoders and then deri...
http://arxiv.org/pdf/2310.07815v1
2023-10-11T18:56:15
2023-10-11T18:56:15
{ "title": [ -0.028564453, -0.00894165, -0.022628784, -0.052734375, -0.0026283264, 0.017700195, -0.05142212, -0.012458801, 0.027175903, 0.04135132, -0.026885986, -0.019577026, 0.011665344, -0.0082473755, -0.066467285, 0.018157959, -0.01322937, 0....
http://arxiv.org/abs/2310.07803v1
A general mechanism of humor: reformulating the semantic overlap
Javier Martínez
cs.CL, cs.AI, I.2.m; J.5
This article proposes a cognitive mechanism of humour of general applicability, not restricted to verbal communication. It is indebted to Raskin's concept of script overlap, and conforms to the incongruity-resolution theoretical framework, but it is built on the notion of constraint, an abstract correspondence between ...
http://arxiv.org/pdf/2310.07803v1
2023-10-11T18:36:13
2023-10-11T18:36:13
{ "title": [ 0.008865356, 0.00083351135, -0.02003479, -0.047943115, 0.022323608, -0.0042419434, -0.016738892, -0.021362305, -0.010070801, 0.036132812, -0.0041122437, -0.011512756, 0.0051879883, -0.023773193, -0.041625977, 0.015838623, -0.0060768127, ...
http://arxiv.org/abs/2310.07795v1
Ontology Enrichment for Effective Fine-grained Entity Typing
Siru Ouyang, Jiaxin Huang, Pranav Pillai, Yunyi Zhang, Yu Zhang, Jiawei Han
cs.CL
Fine-grained entity typing (FET) is the task of identifying specific entity types at a fine-grained level for entity mentions based on their contextual information. Conventional methods for FET require extensive human annotation, which is time-consuming and costly. Recent studies have been developing weakly supervised ...
http://arxiv.org/pdf/2310.07795v1
2023-10-11T18:30:37
2023-10-11T18:30:37
{ "title": [ -0.0041770935, -0.0051574707, -0.057617188, -0.09222412, 0.038513184, -0.0021648407, -0.034423828, 0.01234436, 0.008041382, 0.034332275, -0.014801025, 0.009880066, 0.031341553, -0.014556885, -0.027648926, 0.012504578, -0.028503418, 0...
http://arxiv.org/abs/2310.07793v2
GenTKG: Generative Forecasting on Temporal Knowledge Graph
Ruotong Liao, Xu Jia, Yunpu Ma, Volker Tresp
cs.CL, cs.AI, cs.LG
The rapid advancements in large language models (LLMs) have ignited interest in the temporal knowledge graph (tKG) domain, where conventional carefully designed embedding-based and rule-based models dominate. The question remains open of whether pre-trained LLMs can understand structured temporal relational data and re...
http://arxiv.org/pdf/2310.07793v2
2023-11-14T15:51:18
2023-10-11T18:27:12
{ "title": [ -0.012283325, -0.023971558, -0.07336426, -0.112976074, -0.005054474, -0.0043907166, -0.07312012, 0.029388428, 0.01939392, 0.03933716, -0.014465332, 0.01789856, -0.028137207, -0.004558563, -0.07312012, -0.0026435852, 0.009979248, 0.03...
http://arxiv.org/abs/2310.07715v1
To Build Our Future, We Must Know Our Past: Contextualizing Paradigm Shifts in Natural Language Processing
Sireesh Gururaja, Amanda Bertsch, Clara Na, David Gray Widder, Emma Strubell
cs.CL, cs.CY
NLP is in a period of disruptive change that is impacting our methodologies, funding sources, and public perception. In this work, we seek to understand how to shape our future by better understanding our past. We study factors that shape NLP as a field, including culture, incentives, and infrastructure by conducting l...
http://arxiv.org/pdf/2310.07715v1
2023-10-11T17:59:36
2023-10-11T17:59:36
{ "title": [ -0.017211914, -0.0055351257, -0.057128906, -0.056762695, 0.02229309, -0.020599365, -0.017242432, 0.008087158, -0.0006170273, 0.027252197, -0.027770996, 0.014541626, 0.010017395, -0.0075798035, -0.049438477, -0.002336502, -0.017059326, ...
http://arxiv.org/abs/2310.07713v1
InstructRetro: Instruction Tuning post Retrieval-Augmented Pretraining
Boxin Wang, Wei Ping, Lawrence McAfee, Peng Xu, Bo Li, Mohammad Shoeybi, Bryan Catanzaro
cs.CL, cs.AI, cs.IR, cs.LG
Pretraining auto-regressive large language models (LLMs) with retrieval demonstrates better perplexity and factual accuracy by leveraging external databases. However, the size of existing pretrained retrieval-augmented LLM is still limited (e.g., Retro has 7.5B parameters), which limits the effectiveness of instruction...
http://arxiv.org/pdf/2310.07713v1
2023-10-11T17:59:05
2023-10-11T17:59:05
{ "title": [ -0.019729614, -0.019805908, -0.016296387, -0.027252197, -0.030853271, -0.015266418, -0.0236969, -0.011459351, -0.033691406, 0.036468506, -0.015113831, 0.016906738, -0.003627777, -0.025619507, 0.0065956116, -0.014274597, 0.0037288666, ...
http://arxiv.org/abs/2310.07712v1
Found in the Middle: Permutation Self-Consistency Improves Listwise Ranking in Large Language Models
Raphael Tang, Xinyu Zhang, Xueguang Ma, Jimmy Lin, Ferhan Ture
cs.CL, cs.LG
Large language models (LLMs) exhibit positional bias in how they use context, which especially complicates listwise ranking. To address this, we propose permutation self-consistency, a form of self-consistency over ranking list outputs of black-box LLMs. Our key idea is to marginalize out different list orders in the p...
http://arxiv.org/pdf/2310.07712v1
2023-10-11T17:59:02
2023-10-11T17:59:02
{ "title": [ -0.0016040802, -0.00091791153, -0.03213501, -0.039520264, -0.0025672913, -0.019332886, -0.040496826, -0.0017757416, -0.009933472, 0.043121338, -0.010665894, 0.013320923, 0.0016450882, -0.022827148, -0.040527344, -0.0018835068, -0.0221710...
http://arxiv.org/abs/2310.07710v1
DiPmark: A Stealthy, Efficient and Resilient Watermark for Large Language Models
Yihan Wu, Zhengmian Hu, Hongyang Zhang, Heng Huang
cs.CR, cs.CL, cs.LG
Watermarking techniques offer a promising way to secure data via embedding covert information into the data. A paramount challenge in the domain lies in preserving the distribution of original data during watermarking. Our research extends and refines existing watermarking framework, placing emphasis on the importance ...
http://arxiv.org/pdf/2310.07710v1
2023-10-11T17:57:35
2023-10-11T17:57:35
{ "title": [ -0.0115737915, -0.002729416, -0.04876709, -0.07775879, 0.030014038, -0.0007238388, -0.047454834, -0.0045394897, -0.003326416, 0.03353882, 0.0039901733, -0.00035357475, 0.025726318, -0.025909424, -0.03414917, 0.022094727, -0.022384644, ...
http://arxiv.org/abs/2310.07707v1
MatFormer: Nested Transformer for Elastic Inference
Devvrit, Sneha Kudugunta, Aditya Kusupati, Tim Dettmers, Kaifeng Chen, Inderjit Dhillon, Yulia Tsvetkov, Hannaneh Hajishirzi, Sham Kakade, Ali Farhadi, Prateek Jain
cs.LG, cs.CL, cs.CV
Transformer models are deployed in a wide range of settings, from multi-accelerator clusters to standalone mobile phones. The diverse inference constraints in these scenarios necessitate practitioners to train foundation models such as PaLM 2, Llama, & ViTs as a series of models of varying sizes. Due to significant tra...
http://arxiv.org/pdf/2310.07707v1
2023-10-11T17:57:14
2023-10-11T17:57:14
{ "title": [ -0.019302368, -0.013343811, -0.058288574, -0.07946777, 0.01083374, -0.0014429092, -0.055633545, -0.0034446716, -0.024856567, 0.018325806, -0.014091492, 0.008926392, 0.0052108765, -0.023330688, -0.006401062, 0.000016450882, 0.017410278, ...
http://arxiv.org/abs/2310.07704v1
Ferret: Refer and Ground Anything Anywhere at Any Granularity
Haoxuan You, Haotian Zhang, Zhe Gan, Xianzhi Du, Bowen Zhang, Zirui Wang, Liangliang Cao, Shih-Fu Chang, Yinfei Yang
cs.CV, cs.CL
We introduce Ferret, a new Multimodal Large Language Model (MLLM) capable of understanding spatial referring of any shape or granularity within an image and accurately grounding open-vocabulary descriptions. To unify referring and grounding in the LLM paradigm, Ferret employs a novel and powerful hybrid region represen...
http://arxiv.org/pdf/2310.07704v1
2023-10-11T17:55:15
2023-10-11T17:55:15
{ "title": [ -0.015853882, 0.00034070015, -0.046661377, -0.02897644, -0.012008667, -0.0030460358, -0.020431519, 0.018737793, 0.0068893433, 0.040496826, -0.026489258, -0.009033203, 0.05819702, -0.028274536, -0.00819397, -0.0027103424, -0.005970001, ...
http://arxiv.org/abs/2310.07700v1
Knowledge-enhanced Memory Model for Emotional Support Conversation
Mengzhao Jia, Qianglong Chen, Liqiang Jing, Dawei Fu, Renyu Li
cs.CL
The prevalence of mental disorders has become a significant issue, leading to the increased focus on Emotional Support Conversation as an effective supplement for mental health support. Existing methods have achieved compelling results, however, they still face three challenges: 1) variability of emotions, 2) practical...
http://arxiv.org/pdf/2310.07700v1
2023-10-11T17:51:28
2023-10-11T17:51:28
{ "title": [ -0.010749817, -0.0014257431, -0.041046143, -0.051361084, -0.00548172, -0.008804321, -0.039886475, 0.005302429, 0.017044067, 0.05029297, -0.011756897, -0.004722595, 0.041168213, -0.03692627, -0.042419434, 0.021316528, 0.0110321045, -0...
http://arxiv.org/abs/2310.07676v1
Composite Backdoor Attacks Against Large Language Models
Hai Huang, Zhengyu Zhao, Michael Backes, Yun Shen, Yang Zhang
cs.CR, cs.CL, cs.LG
Large language models (LLMs) have demonstrated superior performance compared to previous methods on various tasks, and often serve as the foundation models for many researches and services. However, the untrustworthy third-party LLMs may covertly introduce vulnerabilities for downstream tasks. In this paper, we explore...
http://arxiv.org/pdf/2310.07676v1
2023-10-11T17:21:03
2023-10-11T17:21:03
{ "title": [ -0.028244019, -0.008338928, -0.022201538, -0.06768799, -0.0030593872, 0.010375977, -0.047424316, 0.010421753, -0.049346924, 0.029281616, 0.009773254, 0.01852417, -0.01864624, -0.027511597, -0.021850586, -0.007457733, -0.015235901, 0....
http://arxiv.org/abs/2310.07659v3
Well Begun is Half Done: Generator-agnostic Knowledge Pre-Selection for Knowledge-Grounded Dialogue
Lang Qin, Yao Zhang, Hongru Liang, Jun Wang, Zhenglu Yang
cs.CL
Accurate knowledge selection is critical in knowledge-grounded dialogue systems. Towards a closer look at it, we offer a novel perspective to organize existing literature, i.e., knowledge selection coupled with, after, and before generation. We focus on the third under-explored category of study, which can not only sel...
http://arxiv.org/pdf/2310.07659v3
2023-10-20T13:43:35
2023-10-11T17:00:29
{ "title": [ -0.010482788, -0.005592346, -0.066223145, -0.08117676, 0.009170532, -0.01727295, -0.039031982, 0.019180298, -0.0088272095, 0.013381958, -0.0037879944, 0.028945923, -0.026123047, -0.025787354, -0.037017822, -0.0058670044, 0.0024700165, ...
http://arxiv.org/abs/2310.07654v1
Audio-Visual Neural Syntax Acquisition
Cheng-I Jeff Lai, Freda Shi, Puyuan Peng, Yoon Kim, Kevin Gimpel, Shiyu Chang, Yung-Sung Chuang, Saurabhchand Bhati, David Cox, David Harwath, Yang Zhang, Karen Livescu, James Glass
cs.CL, cs.LG, cs.SD, eess.AS
We study phrase structure induction from visually-grounded speech. The core idea is to first segment the speech waveform into sequences of word segments, and subsequently induce phrase structure using the inferred segment-level continuous representations. We present the Audio-Visual Neural Syntax Learner (AV-NSL) that ...
http://arxiv.org/pdf/2310.07654v1
2023-10-11T16:54:57
2023-10-11T16:54:57
{ "title": [ -0.0090408325, -0.008979797, -0.026535034, -0.034698486, -0.041748047, -0.017593384, -0.03878784, 0.008155823, -0.008666992, 0.045532227, -0.03930664, -0.010650635, 0.015350342, -0.02722168, -0.03741455, 0.000037789345, 0.010299683, ...
http://arxiv.org/abs/2310.07652v2
LLM4Vis: Explainable Visualization Recommendation using ChatGPT
Lei Wang, Songheng Zhang, Yun Wang, Ee-Peng Lim, Yong Wang
cs.HC, cs.CL
Data visualization is a powerful tool for exploring and communicating insights in various domains. To automate visualization choice for datasets, a task known as visualization recommendation has been proposed. Various machine-learning-based approaches have been developed for this purpose, but they often require a large...
http://arxiv.org/pdf/2310.07652v2
2023-10-16T03:34:47
2023-10-11T16:51:46
{ "title": [ -0.0066375732, -0.01335144, -0.012519836, -0.027282715, -0.014862061, 0.0076789856, -0.012886047, -0.02470398, -0.013221741, 0.00819397, -0.01751709, 0.040222168, -0.03643799, -0.021713257, -0.01586914, 0.0012130737, -0.006980896, 0....
http://arxiv.org/abs/2310.07644v2
Rethinking the BERT-like Pretraining for DNA Sequences
Chaoqi Liang, Weiqiang Bai, Lifeng Qiao, Yuchen Ren, Jianle Sun, Peng Ye, Hongliang Yan, Xinzhu Ma, Wangmeng Zuo, Wanli Ouyang
cs.AI, cs.CL, cs.LG
With the success of large-scale pretraining in NLP, there is an increasing trend of applying it to the domain of life sciences. In particular, pretraining methods based on DNA sequences have garnered growing attention due to their potential to capture generic information about genes. However, existing pretraining metho...
http://arxiv.org/pdf/2310.07644v2
2023-10-12T03:32:32
2023-10-11T16:40:57
{ "title": [ -0.009407043, -0.017440796, -0.03189087, -0.06518555, -0.0071792603, -0.0023918152, -0.03262329, 0.025131226, 0.00856781, 0.044921875, -0.028137207, 0.004081726, 0.023788452, -0.015899658, -0.029220581, 0.01852417, 0.0046920776, 0.04...
http://arxiv.org/abs/2310.07641v1
Evaluating Large Language Models at Evaluating Instruction Following
Zhiyuan Zeng, Jiatong Yu, Tianyu Gao, Yu Meng, Tanya Goyal, Danqi Chen
cs.CL, cs.LG
As research in large language models (LLMs) continues to accelerate, LLM-based evaluation has emerged as a scalable and cost-effective alternative to human evaluations for comparing the ever increasing list of models. This paper investigates the efficacy of these "LLM evaluators", particularly in using them to assess i...
http://arxiv.org/pdf/2310.07641v1
2023-10-11T16:38:11
2023-10-11T16:38:11
{ "title": [ -0.028152466, -0.007949829, -0.029190063, -0.040924072, -0.0088272095, -0.0051002502, -0.033996582, -0.012245178, -0.0101623535, 0.019683838, -0.019073486, -0.0011987686, -0.028030396, -0.016647339, -0.04284668, -0.013694763, -0.03109741...
http://arxiv.org/abs/2310.07629v1
The Past, Present and Better Future of Feedback Learning in Large Language Models for Subjective Human Preferences and Values
Hannah Rose Kirk, Andrew M. Bean, Bertie Vidgen, Paul Röttger, Scott A. Hale
cs.CL, cs.CY
Human feedback is increasingly used to steer the behaviours of Large Language Models (LLMs). However, it is unclear how to collect and incorporate feedback in a way that is efficient, effective and unbiased, especially for highly subjective human preferences and values. In this paper, we survey existing approaches for ...
http://arxiv.org/pdf/2310.07629v1
2023-10-11T16:18:13
2023-10-11T16:18:13
{ "title": [ -0.012969971, 0.0015125275, -0.018554688, -0.033843994, 0.002658844, -0.018203735, -0.021697998, -0.014213562, 0.012428284, 0.023086548, -0.034698486, -0.010917664, 0.013328552, -0.016174316, -0.035980225, -0.007270813, -0.012123108, ...
http://arxiv.org/abs/2310.07611v2
Democratizing LLMs: An Exploration of Cost-Performance Trade-offs in Self-Refined Open-Source Models
Sumuk Shashidhar, Abhinav Chinta, Vaibhav Sahai, Zhenhailong Wang, Heng Ji
cs.CL, cs.AI, cs.PF, 68T50 (Primary), I.2.7; A.2; H.3.4; K.4.1; C.4
The dominance of proprietary LLMs has led to restricted access and raised information privacy concerns. High-performing open-source alternatives are crucial for information-sensitive and high-volume applications but often lag behind in performance. To address this gap, we propose (1) A untargeted variant of iterative s...
http://arxiv.org/pdf/2310.07611v2
2023-10-22T00:37:06
2023-10-11T15:56:00
{ "title": [ -0.06604004, 0.01020813, -0.041534424, -0.05227661, -0.035980225, 0.016235352, -0.09320068, -0.015945435, -0.04159546, 0.04751587, -0.01838684, 0.024597168, -0.00944519, -0.020065308, -0.02444458, -0.0017051697, -0.0039367676, 0.0029...
http://arxiv.org/abs/2310.07609v1
QACHECK: A Demonstration System for Question-Guided Multi-Hop Fact-Checking
Liangming Pan, Xinyuan Lu, Min-Yen Kan, Preslav Nakov
cs.CL
Fact-checking real-world claims often requires complex, multi-step reasoning due to the absence of direct evidence to support or refute them. However, existing fact-checking systems often lack transparency in their decision-making, making it challenging for users to comprehend their reasoning process. To address this, ...
http://arxiv.org/pdf/2310.07609v1
2023-10-11T15:51:53
2023-10-11T15:51:53
{ "title": [ 0.008430481, -0.0037670135, -0.029190063, -0.0670166, 0.006652832, -0.0020313263, -0.015090942, 0.007408142, -0.020217896, 0.022064209, -0.033111572, 0.011222839, -0.0004720688, -0.017105103, -0.035125732, -0.030014038, -0.01927185, ...
http://arxiv.org/abs/2310.07588v1
Accurate Use of Label Dependency in Multi-Label Text Classification Through the Lens of Causality
Caoyun Fan, Wenqing Chen, Jidong Tian, Yitian Li, Hao He, Yaohui Jin
cs.CL, cs.AI, cs.LG
Multi-Label Text Classification (MLTC) aims to assign the most relevant labels to each given text. Existing methods demonstrate that label dependency can help to improve the model's performance. However, the introduction of label dependency may cause the model to suffer from unwanted prediction bias. In this study, we ...
http://arxiv.org/pdf/2310.07588v1
2023-10-11T15:28:44
2023-10-11T15:28:44
{ "title": [ -0.0072517395, -0.0044784546, -0.0541687, -0.0803833, 0.0028858185, -0.016479492, -0.038848877, 0.005432129, -0.008926392, 0.005962372, -0.0047950745, 0.021194458, -0.011726379, -0.027770996, -0.022766113, -0.009315491, -0.02607727, ...
http://arxiv.org/abs/2310.07521v2
Survey on Factuality in Large Language Models: Knowledge, Retrieval and Domain-Specificity
Cunxiang Wang, Xiaoze Liu, Yuanhao Yue, Xiangru Tang, Tianhang Zhang, Cheng Jiayang, Yunzhi Yao, Wenyang Gao, Xuming Hu, Zehan Qi, Yidong Wang, Linyi Yang, Jindong Wang, Xing Xie, Zheng Zhang, Yue Zhang
cs.CL
This survey addresses the crucial issue of factuality in Large Language Models (LLMs). As LLMs find applications across diverse domains, the reliability and accuracy of their outputs become vital. We define the Factuality Issue as the probability of LLMs to produce content inconsistent with established facts. We first ...
http://arxiv.org/pdf/2310.07521v2
2023-10-18T14:09:19
2023-10-11T14:18:03
{ "title": [ -0.0181427, -0.003063202, -0.05154419, -0.07696533, -0.0022850037, -0.019729614, -0.061462402, -0.0032367706, 0.017242432, 0.044952393, -0.009262085, 0.0016918182, 0.057159424, -0.017349243, -0.03552246, 0.040496826, -0.029510498, 0....
http://arxiv.org/abs/2310.07488v2
KwaiYiiMath: Technical Report
Jiayi Fu, Lei Lin, Xiaoyang Gao, Pengli Liu, Zhengzong Chen, Zhirui Yang, Shengnan Zhang, Xue Zheng, Yan Li, Yuliang Liu, Xucheng Ye, Yiqiao Liao, Chao Liao, Bin Chen, Chengru Song, Junchen Wan, Zijia Lin, Fuzheng Zhang, Zhongyuan Wang, Di Zhang, Kun Gai
cs.CL, cs.AI, cs.LG
Recent advancements in large language models (LLMs) have demonstrated remarkable abilities in handling a variety of natural language processing (NLP) downstream tasks, even on mathematical tasks requiring multi-step reasoning. In this report, we introduce the KwaiYiiMath which enhances the mathematical reasoning abilit...
http://arxiv.org/pdf/2310.07488v2
2023-10-19T12:34:32
2023-10-11T13:35:05
{ "title": [ -0.0008792877, -0.011810303, -0.009521484, -0.011001587, -0.047973633, 0.0026512146, -0.05493164, -0.028717041, -0.045776367, 0.07244873, 0.007461548, -0.011039734, 0.054656982, -0.034423828, -0.0025959015, -0.01725769, 0.038116455, ...
http://arxiv.org/abs/2310.07487v2
Cognate Transformer for Automated Phonological Reconstruction and Cognate Reflex Prediction
V. S. D. S. Mahesh Akavarapu, Arnab Bhattacharya
cs.CL, I.2.7
Phonological reconstruction is one of the central problems in historical linguistics where a proto-word of an ancestral language is determined from the observed cognate words of daughter languages. Computational approaches to historical linguistics attempt to automate the task by learning models on available linguistic...
http://arxiv.org/pdf/2310.07487v2
2023-10-18T06:48:37
2023-10-11T13:34:22
{ "title": [ 0.0018157959, 0.0043029785, 0.0053100586, -0.03302002, -0.031585693, 0.010620117, -0.03213501, -0.013549805, -0.016403198, 0.05834961, -0.028884888, 0.026382446, 0.068847656, -0.024002075, 0.0000308156, 0.025848389, 0.00011909008, 0....
http://arxiv.org/abs/2310.07423v1
Adapting the adapters for code-switching in multilingual ASR
Atharva Kulkarni, Ajinkya Kulkarni, Miguel Couceiro, Hanan Aldarmaki
cs.CL, cs.SD, eess.AS
Recently, large pre-trained multilingual speech models have shown potential in scaling Automatic Speech Recognition (ASR) to many low-resource languages. Some of these models employ language adapters in their formulation, which helps to improve monolingual performance and avoids some of the drawbacks of multi-lingual m...
http://arxiv.org/pdf/2310.07423v1
2023-10-11T12:15:24
2023-10-11T12:15:24
{ "title": [ -0.0012598038, -0.0018815994, -0.02911377, -0.031982422, -0.023895264, -0.008773804, -0.026123047, 0.0023517609, -0.021331787, 0.06488037, -0.0046463013, -0.028808594, 0.04901123, -0.046020508, -0.016616821, 0.009223938, 0.008354187, ...
http://arxiv.org/abs/2310.07403v1
DASpeech: Directed Acyclic Transformer for Fast and High-quality Speech-to-Speech Translation
Qingkai Fang, Yan Zhou, Yang Feng
cs.CL, cs.AI, cs.SD, eess.AS, I.2.7
Direct speech-to-speech translation (S2ST) translates speech from one language into another using a single model. However, due to the presence of linguistic and acoustic diversity, the target speech follows a complex multimodal distribution, posing challenges to achieving both high-quality translations and fast decodin...
http://arxiv.org/pdf/2310.07403v1
2023-10-11T11:39:36
2023-10-11T11:39:36
{ "title": [ -0.017364502, 0.0007429123, -0.038330078, -0.08026123, 0.011749268, 0.00026369095, -0.061157227, 0.011039734, -0.022521973, 0.040008545, -0.014884949, 0.0039787292, 0.0181427, -0.034088135, -0.03982544, -0.002729416, 0.0084991455, 0....
http://arxiv.org/abs/2310.07397v2
Target-oriented Proactive Dialogue Systems with Personalization: Problem Formulation and Dataset Curation
Jian Wang, Yi Cheng, Dongding Lin, Chak Tou Leong, Wenjie Li
cs.CL, cs.AI
Target-oriented dialogue systems, designed to proactively steer conversations toward predefined targets or accomplish specific system-side goals, are an exciting area in conversational AI. In this work, by formulating a <dialogue act, topic> pair as the conversation target, we explore a novel problem of personalized ta...
http://arxiv.org/pdf/2310.07397v2
2023-10-13T11:16:58
2023-10-11T11:32:57
{ "title": [ -0.0018091202, 0.0069503784, -0.04812622, -0.08544922, 0.023971558, 0.0043792725, -0.039855957, 0.019821167, -0.014457703, 0.019897461, 0.031982422, -0.00014066696, -0.040130615, -0.036865234, -0.04107666, -0.0046539307, -0.003288269, ...
http://arxiv.org/abs/2310.07387v1
Linguistic laws in biology
Stuart Semple, Ramon Ferrer-i-Cancho, Morgan L. Gustison
cs.CL, physics.bio-ph
Linguistic laws, the common statistical patterns of human language, have been investigated by quantitative linguists for nearly a century. Recently, biologists from a range of disciplines have started to explore the prevalence of these laws beyond language, finding patterns consistent with linguistic laws across multip...
http://arxiv.org/pdf/2310.07387v1
2023-10-11T11:08:20
2023-10-11T11:08:20
{ "title": [ -0.033966064, -0.045074463, -0.020767212, -0.0039024353, -0.07287598, -0.010696411, -0.043273926, 0.020523071, 0.01083374, 0.05267334, -0.021728516, -0.035705566, 0.0079193115, -0.031066895, -0.024765015, 0.011749268, 0.01146698, 0.0...
http://arxiv.org/abs/2310.07347v1
Fast-ELECTRA for Efficient Pre-training
Chengyu Dong, Liyuan Liu, Hao Cheng, Jingbo Shang, Jianfeng Gao, Xiaodong Liu
cs.CL, cs.AI, cs.LG
ELECTRA pre-trains language models by detecting tokens in a sequence that have been replaced by an auxiliary model. Although ELECTRA offers a significant boost in efficiency, its potential is constrained by the training cost brought by the auxiliary model. Notably, this model, which is jointly trained with the main mod...
http://arxiv.org/pdf/2310.07347v1
2023-10-11T09:55:46
2023-10-11T09:55:46
{ "title": [ 0.018676758, -0.016708374, -0.052337646, -0.08288574, -0.036071777, -0.013969421, -0.0368042, 0.02230835, -0.030929565, 0.029937744, -0.012718201, 0.016326904, 0.012763977, -0.04171753, 0.04046631, -0.001162529, -0.00021898746, 0.029...
http://arxiv.org/abs/2310.07345v1
Investigating the Effect of Language Models in Sequence Discriminative Training for Neural Transducers
Zijian Yang, Wei Zhou, Ralf Schlüter, Hermann Ney
cs.CL, cs.SD, eess.AS
In this work, we investigate the effect of language models (LMs) with different context lengths and label units (phoneme vs. word) used in sequence discriminative training for phoneme-based neural transducers. Both lattice-free and N-best-list approaches are examined. For lattice-free methods with phoneme-level LMs, we...
http://arxiv.org/pdf/2310.07345v1
2023-10-11T09:53:17
2023-10-11T09:53:17
{ "title": [ -0.002008438, 0.009185791, -0.027908325, -0.027709961, -0.0012130737, -0.018295288, -0.016998291, -0.016296387, -0.0021133423, 0.02381897, -0.010292053, -0.022964478, 0.0057525635, -0.04449463, -0.031799316, -0.008720398, -0.019714355, ...
http://arxiv.org/abs/2310.07343v1
How Do Large Language Models Capture the Ever-changing World Knowledge? A Review of Recent Advances
Zihan Zhang, Meng Fang, Ling Chen, Mohammad-Reza Namazi-Rad, Jun Wang
cs.CL
Although large language models (LLMs) are impressive in solving various tasks, they can quickly be outdated after deployment. Maintaining their up-to-date status is a pressing concern in the current era. This paper provides a comprehensive review of recent advances in aligning LLMs with the ever-changing world knowledg...
http://arxiv.org/pdf/2310.07343v1
2023-10-11T09:46:32
2023-10-11T09:46:32
{ "title": [ -0.035949707, -0.014839172, -0.05203247, -0.07525635, -0.0022792816, -0.0020160675, -0.05697632, 0.018188477, 0.009307861, 0.027938843, -0.0209198, 0.0035591125, 0.01612854, -0.015823364, -0.04309082, 0.020233154, -0.0030727386, 0.02...
http://arxiv.org/abs/2310.07328v2
An Empirical Study of Instruction-tuning Large Language Models in Chinese
Qingyi Si, Tong Wang, Zheng Lin, Xu Zhang, Yanan Cao, Weiping Wang
cs.CL, cs.AI
The success of ChatGPT validates the potential of large language models (LLMs) in artificial general intelligence (AGI). Subsequently, the release of LLMs has sparked the open-source community's interest in instruction-tuning, which is deemed to accelerate ChatGPT's replication process. However, research on instruction...
http://arxiv.org/pdf/2310.07328v2
2023-10-20T08:02:09
2023-10-11T09:18:09
{ "title": [ -0.047454834, -0.012619019, -0.015548706, -0.028961182, -0.022277832, 0.0021858215, -0.047943115, -0.004890442, -0.018295288, 0.030731201, -0.025802612, 0.009513855, 0.0006337166, -0.015792847, -0.031799316, -0.0077209473, -0.024993896, ...
http://arxiv.org/abs/2310.07321v2
On the Impact of Cross-Domain Data on German Language Models
Amin Dada, Aokun Chen, Cheng Peng, Kaleb E Smith, Ahmad Idrissi-Yaghir, Constantin Marc Seibold, Jianning Li, Lars Heiliger, Xi Yang, Christoph M. Friedrich, Daniel Truhn, Jan Egger, Jiang Bian, Jens Kleesiek, Yonghui Wu
cs.CL, cs.AI, cs.LG
Traditionally, large language models have been either trained on general web crawls or domain-specific data. However, recent successes of generative large language models, have shed light on the benefits of cross-domain datasets. To examine the significance of prioritizing data diversity over quality, we present a Germ...
http://arxiv.org/pdf/2310.07321v2
2023-10-13T14:24:31
2023-10-11T09:09:55
{ "title": [ -0.02279663, -0.00078868866, -0.052459717, -0.037597656, -0.00032114983, -0.02330017, -0.031951904, -0.008651733, -0.010665894, 0.04421997, 0.006313324, -0.011100769, 0.058135986, -0.03942871, -0.018966675, 0.03463745, -0.016448975, ...
http://arxiv.org/abs/2310.07306v1
SNOiC: Soft Labeling and Noisy Mixup based Open Intent Classification Model
Aditi Kanwar, Aditi Seetha, Satyendra Singh Chouhan, Rajdeep Niyogi
cs.LG, cs.CL, ACM-class: F.2.2, I.2.7
This paper presents a Soft Labeling and Noisy Mixup-based open intent classification model (SNOiC). Most of the previous works have used threshold-based methods to identify open intents, which are prone to overfitting and may produce biased predictions. Additionally, the need for more available data for an open intent ...
http://arxiv.org/pdf/2310.07306v1
2023-10-11T08:40:06
2023-10-11T08:40:06
{ "title": [ -0.017730713, 0.01864624, -0.047424316, -0.04510498, 0.026062012, -0.010238647, -0.01878357, -0.029846191, -0.014877319, 0.017288208, -0.023742676, 0.024932861, -0.011657715, -0.023910522, -0.007572174, -0.011390686, -0.009857178, 0....
http://arxiv.org/abs/2310.07301v1
Parrot: Enhancing Multi-Turn Chat Models by Learning to Ask Questions
Yuchong Sun, Che Liu, Jinwen Huang, Ruihua Song, Fuzheng Zhang, Di Zhang, Zhongyuan Wang, Kun Gai
cs.CL
Impressive progress has been made on chat models based on Large Language Models (LLMs) recently; however, there is a noticeable lag in multi-turn conversations between open-source chat models (e.g., Alpaca and Vicuna) and the leading chat models (e.g., ChatGPT and GPT-4). Through a series of analyses, we attribute the ...
http://arxiv.org/pdf/2310.07301v1
2023-10-11T08:36:43
2023-10-11T08:36:43
{ "title": [ 0.004600525, -0.00024354458, -0.021392822, -0.07232666, 0.0015707016, 0.010383606, -0.04736328, 0.013664246, 0.008659363, 0.026367188, -0.039001465, 0.014320374, -0.025527954, 0.0004711151, -0.058258057, -0.012336731, -0.016860962, 0...
http://arxiv.org/abs/2310.07299v1
RobustGEC: Robust Grammatical Error Correction Against Subtle Context Perturbation
Yue Zhang, Leyang Cui, Enbo Zhao, Wei Bi, Shuming Shi
cs.CL, cs.AI
Grammatical Error Correction (GEC) systems play a vital role in assisting people with their daily writing tasks. However, users may sometimes come across a GEC system that initially performs well but fails to correct errors when the inputs are slightly modified. To ensure an ideal user experience, a reliable GEC system...
http://arxiv.org/pdf/2310.07299v1
2023-10-11T08:33:23
2023-10-11T08:33:23
{ "title": [ 0.004852295, 0.0040740967, -0.028427124, -0.061279297, 0.0143966675, -0.004535675, -0.030212402, -0.018081665, 0.009613037, 0.030700684, -0.035064697, 0.016921997, 0.021820068, -0.010215759, -0.043518066, 0.0055656433, -0.03262329, 0...
http://arxiv.org/abs/2310.07289v1
Beyond Factuality: A Comprehensive Evaluation of Large Language Models as Knowledge Generators
Liang Chen, Yang Deng, Yatao Bian, Zeyu Qin, Bingzhe Wu, Tat-Seng Chua, Kam-Fai Wong
cs.CL
Large language models (LLMs) outperform information retrieval techniques for downstream knowledge-intensive tasks when being prompted to generate world knowledge. However, community concerns abound regarding the factuality and potential implications of using this uncensored knowledge. In light of this, we introduce CON...
http://arxiv.org/pdf/2310.07289v1
2023-10-11T08:22:37
2023-10-11T08:22:37
{ "title": [ -0.029006958, -0.022903442, -0.05569458, -0.085632324, 0.0020961761, 0.009651184, -0.073791504, 0.011047363, 0.0034484863, 0.037872314, -0.012138367, -0.0051002502, 0.012611389, -0.013092041, -0.05722046, 0.011314392, -0.0066337585, ...
http://arxiv.org/abs/2310.07284v3
Typing to Listen at the Cocktail Party: Text-Guided Target Speaker Extraction
Xiang Hao, Jibin Wu, Jianwei Yu, Chenglin Xu, Kay Chen Tan
eess.AS, cs.CL
Humans possess an extraordinary ability to selectively focus on the sound source of interest amidst complex acoustic environments, commonly referred to as cocktail party scenarios. In an attempt to replicate this remarkable auditory attention capability in machines, target speaker extraction (TSE) models have been deve...
http://arxiv.org/pdf/2310.07284v3
2023-10-15T03:58:29
2023-10-11T08:17:54
{ "title": [ -0.00944519, 0.019515991, -0.03062439, -0.043395996, -0.0026283264, 0.005428314, -0.04421997, -0.017623901, -0.0007452965, 0.044128418, -0.008285522, 0.025253296, 0.011421204, -0.013595581, -0.036712646, 0.014076233, -0.013420105, 0....
http://arxiv.org/abs/2310.07282v2
An Analysis on Large Language Models in Healthcare: A Case Study of BioBERT
Shyni Sharaf, V. S. Anoop
cs.AI, cs.CL
This paper conducts a comprehensive investigation into applying large language models, particularly on BioBERT, in healthcare. It begins with thoroughly examining previous natural language processing (NLP) approaches in healthcare, shedding light on the limitations and challenges these methods face. Following that, thi...
http://arxiv.org/pdf/2310.07282v2
2023-10-12T07:53:53
2023-10-11T08:16:35
{ "title": [ -0.017669678, -0.006626129, -0.042877197, -0.07897949, -0.0072631836, 0.0020122528, -0.079956055, 0.0010986328, -0.039245605, 0.040039062, -0.0067825317, 0.0027923584, 0.00082683563, -0.027023315, -0.011680603, -0.014434814, -0.001422882...
http://arxiv.org/abs/2310.07279v1
Enhancing expressivity transfer in textless speech-to-speech translation
Jarod Duret, Benjamin O'Brien, Yannick Estève, Titouan Parcollet
cs.SD, cs.CL, eess.AS
Textless speech-to-speech translation systems are rapidly advancing, thanks to the integration of self-supervised learning techniques. However, existing state-of-the-art systems fall short when it comes to capturing and transferring expressivity accurately across different languages. Expressivity plays a vital role in ...
http://arxiv.org/pdf/2310.07279v1
2023-10-11T08:07:22
2023-10-11T08:07:22
{ "title": [ -0.009788513, -0.0066947937, -0.021514893, -0.06567383, -0.0017633438, 0.009239197, -0.00944519, 0.016174316, 0.00031352043, 0.050567627, -0.0021629333, 0.004558563, 0.0385437, -0.03012085, -0.04449463, 0.017990112, -0.0021076202, 0....
http://arxiv.org/abs/2310.07276v2
BioT5: Enriching Cross-modal Integration in Biology with Chemical Knowledge and Natural Language Associations
Qizhi Pei, Wei Zhang, Jinhua Zhu, Kehan Wu, Kaiyuan Gao, Lijun Wu, Yingce Xia, Rui Yan
cs.CL, cs.AI, cs.LG, q-bio.BM
Recent advancements in biological research leverage the integration of molecules, proteins, and natural language to enhance drug discovery. However, current models exhibit several limitations, such as the generation of invalid molecular SMILES, underutilization of contextual information, and equal treatment of structur...
http://arxiv.org/pdf/2310.07276v2
2023-10-17T14:55:58
2023-10-11T07:57:08
{ "title": [ -0.021728516, -0.015007019, -0.048675537, -0.07086182, 0.0028381348, 0.0026664734, -0.048675537, -0.005115509, 0.027954102, 0.00674057, -0.00932312, 0.0026836395, 0.008087158, -0.04168701, -0.0049057007, 0.014801025, -0.008720398, 0....
http://arxiv.org/abs/2310.07251v1
Ethical Reasoning over Moral Alignment: A Case and Framework for In-Context Ethical Policies in LLMs
Abhinav Rao, Aditi Khandelwal, Kumar Tanmay, Utkarsh Agarwal, Monojit Choudhury
cs.CL, cs.AI
In this position paper, we argue that instead of morally aligning LLMs to specific set of ethical principles, we should infuse generic ethical reasoning capabilities into them so that they can handle value pluralism at a global scale. When provided with an ethical policy, an LLM should be capable of making decisions th...
http://arxiv.org/pdf/2310.07251v1
2023-10-11T07:27:34
2023-10-11T07:27:34
{ "title": [ -0.010665894, 0.0063591003, -0.038757324, -0.030670166, -0.008865356, -0.023361206, -0.010795593, -0.008728027, -0.0395813, 0.02671814, -0.040893555, 0.01802063, -0.022659302, -0.020019531, -0.022842407, -0.027420044, -0.03213501, 0....
http://arxiv.org/abs/2310.07225v1
Exploring the Landscape of Large Language Models In Medical Question Answering: Observations and Open Questions
Karolina Korgul, Andrew M. Bean, Felix Krones, Robert McCraith, Adam Mahdi
cs.CL
Large Language Models (LLMs) have shown promise in medical question answering by achieving passing scores in standardised exams and have been suggested as tools for supporting healthcare workers. Deploying LLMs into such a high-risk context requires a clear understanding of the limitations of these models. With the rap...
http://arxiv.org/pdf/2310.07225v1
2023-10-11T06:26:19
2023-10-11T06:26:19
{ "title": [ -0.015914917, -0.006679535, -0.030853271, -0.055725098, -0.0010890961, 0.005756378, -0.03729248, -0.002067566, -0.0038433075, 0.027770996, -0.008407593, -0.0039482117, 0.011543274, -0.017303467, -0.042510986, 0.012428284, 0.0036067963, ...
http://arxiv.org/abs/2310.07188v1
Adaptive Gating in Mixture-of-Experts based Language Models
Jiamin Li, Qiang Su, Yitao Yang, Yimin Jiang, Cong Wang, Hong Xu
cs.CL, cs.AI
Large language models, such as OpenAI's ChatGPT, have demonstrated exceptional language understanding capabilities in various NLP tasks. Sparsely activated mixture-of-experts (MoE) has emerged as a promising solution for scaling models while maintaining a constant number of computational operations. Existing MoE model ...
http://arxiv.org/pdf/2310.07188v1
2023-10-11T04:30:18
2023-10-11T04:30:18
{ "title": [ -0.038604736, -0.0016517639, -0.032165527, -0.044891357, -0.008850098, -0.0035877228, -0.052124023, 0.00034046173, -0.004989624, 0.049072266, 0.013046265, 0.0017881393, 0.017486572, -0.015106201, -0.05606079, 0.03010559, -0.0031776428, ...
http://arxiv.org/abs/2310.07177v2
Online Speculative Decoding
Xiaoxuan Liu, Lanxiang Hu, Peter Bailis, Ion Stoica, Zhijie Deng, Alvin Cheung, Hao Zhang
cs.AI, cs.CL, cs.LG
Speculative decoding is a pivotal technique to accelerate the inference of large language models (LLMs) by employing a smaller draft model to predict the target model's outputs. However, its efficacy can be limited due to the low predictive accuracy of the draft model, particularly when faced with diverse text inputs a...
http://arxiv.org/pdf/2310.07177v2
2023-10-17T18:02:19
2023-10-11T04:03:42
{ "title": [ 0.007347107, -0.016052246, 0.015296936, -0.036010742, -0.070007324, 0.001709938, -0.047302246, -0.014312744, 0.009429932, 0.05581665, -0.012123108, -0.044189453, 0.066833496, -0.035095215, -0.003982544, 0.009719849, 0.047973633, 0.03...
http://arxiv.org/abs/2310.07170v1
PHALM: Building a Knowledge Graph from Scratch by Prompting Humans and a Language Model
Tatsuya Ide, Eiki Murata, Daisuke Kawahara, Takato Yamazaki, Shengzhe Li, Kenta Shinzato, Toshinori Sato
cs.CL
Despite the remarkable progress in natural language understanding with pretrained Transformers, neural language models often do not handle commonsense knowledge well. Toward commonsense-aware models, there have been attempts to obtain knowledge, ranging from automatic acquisition to crowdsourcing. However, it is diffic...
http://arxiv.org/pdf/2310.07170v1
2023-10-11T03:39:46
2023-10-11T03:39:46
{ "title": [ -0.0385437, -0.020690918, -0.05670166, -0.07232666, -0.02192688, 0.0107421875, -0.09899902, 0.005428314, -0.007396698, 0.00894928, -0.031982422, 0.013153076, 0.005886078, -0.017196655, -0.009780884, -0.009315491, 0.008140564, 0.00993...
http://arxiv.org/abs/2310.07161v1
Psychoacoustic Challenges Of Speech Enhancement On VoIP Platforms
Joseph Konan, Ojas Bhargave, Shikhar Agnihotri, Shuo Han, Yunyang Zeng, Ankit Shah, Bhiksha Raj
cs.SD, cs.CL, eess.AS
Within the ambit of VoIP (Voice over Internet Protocol) telecommunications, the complexities introduced by acoustic transformations merit rigorous analysis. This research, rooted in the exploration of proprietary sender-side denoising effects, meticulously evaluates platforms such as Google Meets and Zoom. The study dr...
http://arxiv.org/pdf/2310.07161v1
2023-10-11T03:19:22
2023-10-11T03:19:22
{ "title": [ 0.02281189, 0.014228821, -0.027450562, -0.054534912, 0.016403198, -0.01638794, 0.00042271614, -0.004852295, -0.023834229, 0.06744385, 0.004825592, 0.0082473755, 0.037261963, -0.020095825, -0.028305054, 0.024154663, -0.023666382, 0.05...
http://arxiv.org/abs/2310.07155v2
"A Tale of Two Movements": Identifying and Comparing Perspectives in #BlackLivesMatter and #BlueLivesMatter Movements-related Tweets using Weakly Supervised Graph-based Structured Prediction
Shamik Roy, Dan Goldwasser
cs.CL
Social media has become a major driver of social change, by facilitating the formation of online social movements. Automatically understanding the perspectives driving the movement and the voices opposing it, is a challenging task as annotated data is difficult to obtain. We propose a weakly supervised graph-based appr...
http://arxiv.org/pdf/2310.07155v2
2023-10-21T16:28:07
2023-10-11T03:01:42
{ "title": [ 0.041625977, 0.0069084167, -0.012931824, -0.08074951, -0.02407837, -0.0011301041, -0.07531738, -0.00059366226, -0.02468872, 0.023010254, -0.033477783, 0.040740967, -0.006038666, -0.022537231, 0.005622864, -0.03302002, -0.009384155, 0...
http://arxiv.org/abs/2310.07147v1
QFT: Quantized Full-parameter Tuning of LLMs with Affordable Resources
Zhikai Li, Xiaoxuan Liu, Banghua Zhu, Zhen Dong, Qingyi Gu, Kurt Keutzer
cs.CL, cs.LG
Large Language Models (LLMs) have showcased remarkable impacts across a wide spectrum of natural language processing tasks. Fine-tuning these pre-trained models on downstream datasets provides further significant performance gains, but this process has been challenging due to its extraordinary resource requirements. To...
http://arxiv.org/pdf/2310.07147v1
2023-10-11T02:47:40
2023-10-11T02:47:40
{ "title": [ -0.06817627, -0.017608643, -0.044128418, -0.06567383, -0.034240723, 0.013114929, -0.09088135, 0.008285522, -0.025924683, 0.039855957, 0.0011463165, 0.02267456, 0.007572174, -0.027496338, -0.018615723, 0.0149002075, -0.0069122314, 0.0...
http://arxiv.org/abs/2310.07146v1
Empowering Psychotherapy with Large Language Models: Cognitive Distortion Detection through Diagnosis of Thought Prompting
Zhiyu Chen, Yujie Lu, William Yang Wang
cs.CL
Mental illness remains one of the most critical public health issues of our time, due to the severe scarcity and accessibility limit of professionals. Psychotherapy requires high-level expertise to conduct deep, complex reasoning and analysis on the cognition modeling of the patients. In the era of Large Language Model...
http://arxiv.org/pdf/2310.07146v1
2023-10-11T02:47:21
2023-10-11T02:47:21
{ "title": [ -0.010719299, -0.007610321, -0.017318726, -0.025527954, -0.02986145, -0.0044403076, -0.0647583, -0.026809692, -0.002609253, 0.0037384033, 0.0026817322, -0.016464233, 0.053588867, -0.06085205, 0.014457703, 0.00019967556, 0.009887695, ...
http://arxiv.org/abs/2310.07137v1
AE-smnsMLC: Multi-Label Classification with Semantic Matching and Negative Label Sampling for Product Attribute Value Extraction
Zhongfen Deng, Wei-Te Chen, Lei Chen, Philip S. Yu
cs.IR, cs.CL, cs.LG
Product attribute value extraction plays an important role for many real-world applications in e-Commerce such as product search and recommendation. Previous methods treat it as a sequence labeling task that needs more annotation for position of values in the product text. This limits their application to real-world sc...
http://arxiv.org/pdf/2310.07137v1
2023-10-11T02:22:28
2023-10-11T02:22:28
{ "title": [ -0.021759033, -0.012336731, -0.041992188, -0.03552246, 0.0026721954, -0.019851685, -0.012802124, -0.0104522705, 0.009338379, 0.015304565, -0.0026474, 0.0066337585, -0.010147095, -0.03137207, -0.041137695, 0.0026931763, -0.016571045, ...
http://arxiv.org/abs/2310.07135v1
Comparing Styles across Languages
Shreya Havaldar, Matthew Pressimone, Eric Wong, Lyle Ungar
cs.CL
Understanding how styles differ across languages is advantageous for training both humans and computers to generate culturally appropriate text. We introduce an explanation framework to extract stylistic differences from multilingual LMs and compare styles across languages. Our framework (1) generates comprehensive sty...
http://arxiv.org/pdf/2310.07135v1
2023-10-11T02:16:12
2023-10-11T02:16:12
{ "title": [ -0.026062012, -0.0059661865, -0.022598267, -0.017532349, -0.05596924, -0.013008118, -0.061035156, -0.002248764, -0.020721436, 0.052642822, -0.0070114136, -0.006717682, 0.019821167, -0.038726807, 0.018447876, 0.019821167, 0.039611816, ...
http://arxiv.org/abs/2310.07106v1
The Temporal Structure of Language Processing in the Human Brain Corresponds to The Layered Hierarchy of Deep Language Models
Ariel Goldstein, Eric Ham, Mariano Schain, Samuel Nastase, Zaid Zada, Avigail Dabush, Bobbi Aubrey, Harshvardhan Gazula, Amir Feder, Werner K Doyle, Sasha Devore, Patricia Dugan, Daniel Friedman, Roi Reichart, Michael Brenner, Avinatan Hassidim, Orrin Devinsky, Adeen Flinker, Omer Levy, Uri Hasson
cs.CL, cs.AI, cs.LG, q-bio.NC
Deep Language Models (DLMs) provide a novel computational paradigm for understanding the mechanisms of natural language processing in the human brain. Unlike traditional psycholinguistic models, DLMs use layered sequences of continuous numerical vectors to represent words and context, allowing a plethora of emerging ap...
http://arxiv.org/pdf/2310.07106v1
2023-10-11T01:03:42
2023-10-11T01:03:42
{ "title": [ 0.025314331, 0.0052986145, -0.025848389, -0.04244995, -0.023010254, -0.025970459, -0.03942871, -0.016204834, -0.024307251, 0.051727295, 0.030273438, -0.008598328, 0.03869629, -0.048034668, -0.008460999, 0.007221222, 0.0022411346, 0.0...
http://arxiv.org/abs/2310.07096v1
Sparse Universal Transformer
Shawn Tan, Yikang Shen, Zhenfang Chen, Aaron Courville, Chuang Gan
cs.CL, cs.AI
The Universal Transformer (UT) is a variant of the Transformer that shares parameters across its layers. Empirical evidence shows that UTs have better compositional generalization than Vanilla Transformers (VTs) in formal language tasks. The parameter-sharing also affords it better parameter efficiency than VTs. Despit...
http://arxiv.org/pdf/2310.07096v1
2023-10-11T00:38:57
2023-10-11T00:38:57
{ "title": [ -0.0049095154, 0.018875122, -0.070617676, -0.052734375, -0.027999878, -0.0061454773, -0.08306885, -0.0031223297, 0.008911133, 0.05822754, 0.0061073303, -0.021865845, 0.058563232, -0.05050659, 0.00440979, 0.042053223, 0.047790527, 0.0...
http://arxiv.org/abs/2310.07093v1
Argumentative Stance Prediction: An Exploratory Study on Multimodality and Few-Shot Learning
Arushi Sharma, Abhibha Gupta, Maneesh Bilalpur
cs.CL
To advance argumentative stance prediction as a multimodal problem, the First Shared Task in Multimodal Argument Mining hosted stance prediction in crucial social topics of gun control and abortion. Our exploratory study attempts to evaluate the necessity of images for stance prediction in tweets and compare out-of-the...
http://arxiv.org/pdf/2310.07093v1
2023-10-11T00:18:29
2023-10-11T00:18:29
{ "title": [ 0.0256958, -0.016220093, -0.034851074, -0.071899414, -0.028869629, -0.02003479, -0.05090332, 0.012313843, -0.00074005127, 0.012908936, 0.00945282, -0.007133484, -0.028259277, -0.039489746, -0.035491943, -0.036865234, -0.0015640259, 0...
http://arxiv.org/abs/2310.07091v2
Jaeger: A Concatenation-Based Multi-Transformer VQA Model
Jieting Long, Zewei Shi, Penghao Jiang, Yidong Gan
cs.CL, cs.AI
Document-based Visual Question Answering poses a challenging task between linguistic sense disambiguation and fine-grained multimodal retrieval. Although there has been encouraging progress in document-based question answering due to the utilization of large language and open-world prior models\cite{1}, several challen...
http://arxiv.org/pdf/2310.07091v2
2023-10-19T04:03:08
2023-10-11T00:14:40
{ "title": [ -0.011306763, -0.0043258667, -0.07543945, -0.064331055, -0.0005993843, -0.012741089, -0.05569458, 0.006855011, -0.026031494, 0.051757812, 0.03778076, -0.030975342, 0.01713562, -0.05657959, -0.03387451, 0.006454468, -0.010818481, 0.05...
http://arxiv.org/abs/2310.07088v1
Diversity of Thought Improves Reasoning Abilities of Large Language Models
Ranjita Naik, Varun Chandrasekaran, Mert Yuksekgonul, Hamid Palangi, Besmira Nushi
cs.CL, cs.AI
Large language models (LLMs) are documented to struggle in settings that require complex reasoning. Nevertheless, instructing the model to break down the problem into smaller reasoning steps (Wei et al., 2022), or ensembling various generations through modifying decoding steps (Wang et al., 2023) boosts performance. Cu...
http://arxiv.org/pdf/2310.07088v1
2023-10-11T00:01:41
2023-10-11T00:01:41
{ "title": [ -0.0074501038, -0.008728027, -0.061553955, -0.05657959, -0.022232056, -0.03656006, -0.05682373, 0.010017395, -0.021148682, 0.014045715, -0.016540527, -0.01789856, 0.009941101, -0.047058105, -0.008460999, -0.009788513, -0.00945282, 0....
http://arxiv.org/abs/2310.07081v2
Crossing the Threshold: Idiomatic Machine Translation through Retrieval Augmentation and Loss Weighting
Emmy Liu, Aditi Chaudhary, Graham Neubig
cs.CL
Idioms are common in everyday language, but often pose a challenge to translators because their meanings do not follow from the meanings of their parts. Despite significant advances, machine translation systems still struggle to translate idiomatic expressions. We provide a simple characterization of idiomatic translat...
http://arxiv.org/pdf/2310.07081v2
2023-10-20T19:24:14
2023-10-10T23:47:25
{ "title": [ -0.016204834, -0.013092041, 0.010276794, -0.035003662, -0.009300232, 0.024093628, -0.006793976, -0.008796692, 0.010818481, 0.06982422, -0.01725769, 0.020568848, 0.041870117, 0.0010566711, -0.06500244, 0.027648926, -0.011924744, 0.010...
http://arxiv.org/abs/2310.07075v1
Syntax Error-Free and Generalizable Tool Use for LLMs via Finite-State Decoding
Kexun Zhang, Hongqiao Chen, Lei Li, William Wang
cs.CL, cs.AI
Large language models (LLMs) have shown promising capabilities in using external tools to solve complex problems. However, existing approaches either involve fine-tuning on tool demonstrations, which do not generalize to new tools without additional training, or providing tool documentation in context, limiting the num...
http://arxiv.org/pdf/2310.07075v1
2023-10-10T23:37:53
2023-10-10T23:37:53
{ "title": [ -0.013000488, -0.011558533, -0.046325684, -0.08441162, -0.0362854, -0.002193451, -0.09338379, 0.025772095, 0.018051147, 0.02571106, -0.03656006, -0.0041923523, 0.008598328, -0.028564453, -0.004825592, -0.0060310364, -0.000310421, 0.0...
http://arxiv.org/abs/2310.07064v1
Large Language Models can Learn Rules
Zhaocheng Zhu, Yuan Xue, Xinyun Chen, Denny Zhou, Jian Tang, Dale Schuurmans, Hanjun Dai
cs.AI, cs.CL
When prompted with a few examples and intermediate steps, large language models (LLMs) have demonstrated impressive performance in various reasoning tasks. However, prompting methods that rely on implicit knowledge in an LLM often hallucinate incorrect answers when the implicit knowledge is wrong or inconsistent with t...
http://arxiv.org/pdf/2310.07064v1
2023-10-10T23:07:01
2023-10-10T23:07:01
{ "title": [ 0.013664246, -0.021697998, -0.031921387, -0.060791016, -0.05923462, -0.016326904, -0.06762695, 0.029418945, -0.024475098, 0.027618408, -0.05114746, -0.013511658, -0.010673523, -0.028778076, 0.02355957, -0.029327393, 0.02519226, 0.048...
http://arxiv.org/abs/2310.07059v1
DKEC: Domain Knowledge Enhanced Multi-Label Classification for Electronic Health Records
Xueren Ge, Ronald Dean Williams, John A. Stankovic, Homa Alemzadeh
cs.CL, cs.AI, cs.LG
Multi-label text classification (MLTC) tasks in the medical domain often face long-tail label distribution, where rare classes have fewer training samples than frequent classes. Although previous works have explored different model architectures and hierarchical label structures to find important features, most of them...
http://arxiv.org/pdf/2310.07059v1
2023-10-10T22:53:15
2023-10-10T22:53:15
{ "title": [ -0.01889038, 0.0033740997, -0.074523926, -0.08001709, 0.03189087, -0.010932922, -0.027862549, 0.010353088, -0.029541016, 0.017974854, -0.011909485, 0.020980835, -0.009979248, -0.022338867, -0.019714355, -0.012512207, -0.011054993, 0....
http://arxiv.org/abs/2310.07023v1
Automatic Macro Mining from Interaction Traces at Scale
Forrest Huang, Gang Li, Tao Li, Yang Li
cs.HC, cs.CL, cs.LG
Macros are building block tasks of our everyday smartphone activity (e.g., "login", or "booking a flight"). Effectively extracting macros is important for understanding mobile interaction and enabling task automation. These macros are however difficult to extract at scale as they can be comprised of multiple steps yet ...
http://arxiv.org/pdf/2310.07023v1
2023-10-10T21:23:47
2023-10-10T21:23:47
{ "title": [ -0.012397766, -0.02796936, -0.016662598, -0.07788086, -0.0021820068, 0.02420044, -0.057037354, -0.004245758, -0.0014896393, 0.013137817, -0.022598267, 0.014183044, -0.025726318, -0.01826477, -0.031051636, -0.0046806335, -0.018539429, ...
http://arxiv.org/abs/2310.10677v1
LLMs as Potential Brainstorming Partners for Math and Science Problems
Sophia Gu
cs.CL
With the recent rise of widely successful deep learning models, there is emerging interest among professionals in various math and science communities to see and evaluate the state-of-the-art models' abilities to collaborate on finding or solving problems that often require creativity and thus brainstorming. While a si...
http://arxiv.org/pdf/2310.10677v1
2023-10-10T21:16:35
2023-10-10T21:16:35
{ "title": [ -0.011909485, -0.02746582, -0.02180481, 0.035095215, -0.06530762, -0.025314331, -0.0077705383, -0.014602661, -0.031951904, 0.01776123, -0.014381409, 0.012481689, -0.026550293, -0.040405273, 0.020599365, -0.037200928, -0.000688076, 0....
http://arxiv.org/abs/2310.07018v1
NEWTON: Are Large Language Models Capable of Physical Reasoning?
Yi Ru Wang, Jiafei Duan, Dieter Fox, Siddhartha Srinivasa
cs.CL, cs.AI, cs.RO
Large Language Models (LLMs), through their contextualized representations, have been empirically proven to encapsulate syntactic, semantic, word sense, and common-sense knowledge. However, there has been limited exploration of their physical reasoning abilities, specifically concerning the crucial attributes for compr...
http://arxiv.org/pdf/2310.07018v1
2023-10-10T21:08:51
2023-10-10T21:08:51
{ "title": [ -0.038208008, -0.0015668869, -0.043029785, -0.04562378, -0.038391113, -0.0027370453, -0.06286621, -0.0045318604, -0.02407837, 0.019104004, -0.011001587, -0.026016235, 0.027954102, -0.047454834, -0.0047302246, -0.019714355, -0.0011787415,...
http://arxiv.org/abs/2310.07008v1
Answer Candidate Type Selection: Text-to-Text Language Model for Closed Book Question Answering Meets Knowledge Graphs
Mikhail Salnikov, Maria Lysyuk, Pavel Braslavski, Anton Razzhigaev, Valentin Malykh, Alexander Panchenko
cs.CL, cs.AI, cs.IR, cs.LG
Pre-trained Text-to-Text Language Models (LMs), such as T5 or BART yield promising results in the Knowledge Graph Question Answering (KGQA) task. However, the capacity of the models is limited and the quality decreases for questions with less popular entities. In this paper, we present a novel approach which works on t...
http://arxiv.org/pdf/2310.07008v1
2023-10-10T20:49:43
2023-10-10T20:49:43
{ "title": [ -0.03665161, -0.01864624, -0.030410767, -0.04827881, -0.026367188, 0.010002136, -0.038513184, 0.008934021, 0.014579773, 0.028762817, -0.029434204, 0.022369385, 0.0020542145, -0.004333496, -0.04840088, -0.00015628338, 0.003276825, 0.0...
http://arxiv.org/abs/2310.06987v1
Catastrophic Jailbreak of Open-source LLMs via Exploiting Generation
Yangsibo Huang, Samyak Gupta, Mengzhou Xia, Kai Li, Danqi Chen
cs.CL, cs.AI, cs.CR
The rapid progress in open-source large language models (LLMs) is significantly advancing AI development. Extensive efforts have been made before model release to align their behavior with human values, with the primary goal of ensuring their helpfulness and harmlessness. However, even carefully aligned models can be m...
http://arxiv.org/pdf/2310.06987v1
2023-10-10T20:15:54
2023-10-10T20:15:54
{ "title": [ -0.028533936, 0.00945282, -0.040802002, -0.03024292, -0.020736694, -0.012870789, -0.034057617, -0.0003490448, -0.06008911, 0.034240723, 0.0044136047, 0.028778076, 0.010177612, -0.053527832, 0.018554688, -0.0014190674, -0.0021266937, ...
http://arxiv.org/abs/2310.06983v1
Violation of Expectation via Metacognitive Prompting Reduces Theory of Mind Prediction Error in Large Language Models
Courtland Leer, Vincent Trost, Vineeth Voruganti
cs.CL, cs.LG
Recent research shows that Large Language Models (LLMs) exhibit a compelling level of proficiency in Theory of Mind (ToM) tasks. This ability to impute unobservable mental states to others is vital to human social cognition and may prove equally important in principal-agent relations between individual humans and Artif...
http://arxiv.org/pdf/2310.06983v1
2023-10-10T20:05:13
2023-10-10T20:05:13
{ "title": [ -0.0019931793, -0.04296875, 0.017181396, -0.019592285, -0.051452637, -0.01159668, -0.020233154, -0.01953125, -0.010871887, 0.012878418, -0.01612854, -0.008476257, 0.027770996, -0.041290283, -0.02027893, -0.012023926, -0.008232117, 0....
http://arxiv.org/abs/2310.06977v1
Why bother with geometry? On the relevance of linear decompositions of Transformer embeddings
Timothee Mickus, Raúl Vázquez
cs.CL
A recent body of work has demonstrated that Transformer embeddings can be linearly decomposed into well-defined sums of factors, that can in turn be related to specific network inputs or components. There is however still a dearth of work studying whether these mathematical reformulations are empirically meaningful. In...
http://arxiv.org/pdf/2310.06977v1
2023-10-10T19:56:10
2023-10-10T19:56:10
{ "title": [ -0.009231567, 0.0211792, -0.020248413, -0.04324341, 0.010726929, -0.012275696, -0.0289917, -0.033203125, -0.017623901, 0.014419556, 0.008506775, 0.011695862, -0.00491333, -0.042144775, -0.0016527176, 0.016113281, 0.0054244995, 0.0370...
http://arxiv.org/abs/2310.06960v1
Jaynes Machine: The universal microstructure of deep neural networks
Venkat Venkatasubramanian, N. Sanjeevrajan, Manasi Khandekar
cond-mat.stat-mech, cs.CL
We present a novel theory of the microstructure of deep neural networks. Using a theoretical framework called statistical teleodynamics, which is a conceptual synthesis of statistical thermodynamics and potential game theory, we predict that all highly connected layers of deep neural networks have a universal microstru...
http://arxiv.org/pdf/2310.06960v1
2023-10-10T19:22:01
2023-10-10T19:22:01
{ "title": [ -0.0132751465, -0.0009160042, -0.045196533, -0.030838013, 0.030059814, -0.013679504, 0.0002553463, -0.018127441, -0.00004005432, 0.036132812, -0.004245758, 0.005760193, -0.00028014183, -0.019744873, -0.049041748, 0.01461792, -0.002048492...
http://arxiv.org/abs/2310.10675v1
Creation Of A ChatBot Based On Natural Language Proccesing For Whatsapp
Valderrama Jonatan, Aguilar-Alonso Igor
cs.CL
In the era of digital transformation, customer service is of paramount importance to the success of organizations, and to meet the growing demand for immediate responses and personalized assistance 24 hours a day, chatbots have become a promising tool to solve these problems. Currently, there are many companies that ne...
http://arxiv.org/pdf/2310.10675v1
2023-10-10T18:54:15
2023-10-10T18:54:15
{ "title": [ -0.025909424, -0.012962341, -0.05078125, -0.036895752, -0.034576416, -0.0030651093, -0.032928467, 0.042510986, -0.025604248, 0.045043945, 0.008895874, -0.008720398, 0.037719727, -0.032592773, 0.000790596, -0.023147583, 0.0054473877, ...
http://arxiv.org/abs/2310.06940v1
Document-Level Supervision for Multi-Aspect Sentiment Analysis Without Fine-grained Labels
Kasturi Bhattacharjee, Rashmi Gangadharaiah
cs.CL
Aspect-based sentiment analysis (ABSA) is a widely studied topic, most often trained through supervision from human annotations of opinionated texts. These fine-grained annotations include identifying aspects towards which a user expresses their sentiment, and their associated polarities (aspect-based sentiments). Such...
http://arxiv.org/pdf/2310.06940v1
2023-10-10T18:53:21
2023-10-10T18:53:21
{ "title": [ 0.007797241, -0.013877869, -0.017318726, -0.0690918, -0.0023231506, 0.0019874573, -0.03262329, 0.000058710575, 0.0335083, 0.026748657, -0.031555176, -0.015541077, 0.0021743774, -0.007194519, -0.05871582, -0.0070991516, -0.008850098, ...
http://arxiv.org/abs/2310.06927v2
Sparse Fine-tuning for Inference Acceleration of Large Language Models
Eldar Kurtic, Denis Kuznedelev, Elias Frantar, Michael Goin, Dan Alistarh
cs.CL, cs.AI
We consider the problem of accurate sparse fine-tuning of large language models (LLMs), that is, fine-tuning pretrained LLMs on specialized tasks, while inducing sparsity in their weights. On the accuracy side, we observe that standard loss-based fine-tuning may fail to recover accuracy, especially at high sparsities. ...
http://arxiv.org/pdf/2310.06927v2
2023-10-13T13:47:44
2023-10-10T18:28:38
{ "title": [ -0.042663574, -0.038757324, -0.04989624, -0.08526611, -0.014854431, 0.0141067505, -0.07739258, 0.05633545, 0.007762909, 0.027069092, -0.011436462, -0.006752014, 0.013954163, -0.016860962, -0.042999268, 0.01725769, 0.012588501, 0.0089...
http://arxiv.org/abs/2310.06918v2
Improving Contrastive Learning of Sentence Embeddings with Focal-InfoNCE
Pengyue Hou, Xingyu Li
cs.CL, cs.LG
The recent success of SimCSE has greatly advanced state-of-the-art sentence representations. However, the original formulation of SimCSE does not fully exploit the potential of hard negative samples in contrastive learning. This study introduces an unsupervised contrastive learning framework that combines SimCSE with h...
http://arxiv.org/pdf/2310.06918v2
2023-10-20T19:39:40
2023-10-10T18:15:24
{ "title": [ 0.018371582, 0.012519836, -0.0071907043, -0.04385376, -0.0013866425, -0.007888794, -0.017120361, -0.025665283, 0.026626587, 0.024383545, -0.03692627, 0.026016235, 0.008262634, 0.00045132637, -0.04144287, 0.009788513, -0.022125244, 0....
http://arxiv.org/abs/2310.06913v1
A Comparative Study of Transformer-based Neural Text Representation Techniques on Bug Triaging
Atish Kumar Dipongkor, Kevin Moran
cs.SE, cs.CL, cs.IR
Often, the first step in managing bug reports is related to triaging a bug to the appropriate developer who is best suited to understand, localize, and fix the target bug. Additionally, assigning a given bug to a particular part of a software project can help to expedite the fixing process. However, despite the importa...
http://arxiv.org/pdf/2310.06913v1
2023-10-10T18:09:32
2023-10-10T18:09:32
{ "title": [ 0.00699234, -0.017074585, -0.01576233, -0.04156494, -0.010429382, -0.0048217773, -0.03173828, 0.011177063, -0.0066833496, 0.06173706, -0.03100586, 0.014457703, 0.024551392, -0.0062294006, -0.05001831, 0.0022888184, -0.01876831, 0.063...
http://arxiv.org/abs/2310.06839v1
LongLLMLingua: Accelerating and Enhancing LLMs in Long Context Scenarios via Prompt Compression
Huiqiang Jiang, Qianhui Wu, Xufang Luo, Dongsheng Li, Chin-Yew Lin, Yuqing Yang, Lili Qiu
cs.CL, cs.LG
In long context scenarios, large language models (LLMs) face three main challenges: higher computational/financial cost, longer latency, and inferior performance. Some studies reveal that the performance of LLMs depends on both the density and the position of the key information (question relevant) in the input prompt....
http://arxiv.org/pdf/2310.06839v1
2023-10-10T17:59:58
2023-10-10T17:59:58
{ "title": [ -0.049621582, -0.026733398, -0.056396484, -0.07067871, -0.018829346, 0.012886047, -0.07330322, 0.028686523, -0.0006928444, 0.028884888, -0.016052246, 0.01448822, -0.0018997192, -0.03955078, -0.0034370422, 0.0015630722, -0.0016813278, ...
http://arxiv.org/abs/2310.06837v1
Generating and Evaluating Tests for K-12 Students with Language Model Simulations: A Case Study on Sentence Reading Efficiency
Eric Zelikman, Wanjing Anya Ma, Jasmine E. Tran, Diyi Yang, Jason D. Yeatman, Nick Haber
cs.CL, cs.LG
Developing an educational test can be expensive and time-consuming, as each item must be written by experts and then evaluated by collecting hundreds of student responses. Moreover, many tests require multiple distinct sets of questions administered throughout the school year to closely monitor students' progress, know...
http://arxiv.org/pdf/2310.06837v1
2023-10-10T17:59:51
2023-10-10T17:59:51
{ "title": [ -0.04296875, -0.007633209, -0.016540527, -0.021560669, -0.028915405, 0.01159668, -0.048431396, -0.015914917, -0.0029888153, 0.036193848, -0.0053901672, 0.0070114136, -0.004337311, -0.008483887, -0.058624268, -0.00447464, -0.048095703, ...
http://arxiv.org/abs/2310.06830v1
Lemur: Harmonizing Natural Language and Code for Language Agents
Yiheng Xu, Hongjin Su, Chen Xing, Boyu Mi, Qian Liu, Weijia Shi, Binyuan Hui, Fan Zhou, Yitao Liu, Tianbao Xie, Zhoujun Cheng, Siheng Zhao, Lingpeng Kong, Bailin Wang, Caiming Xiong, Tao Yu
cs.CL
We introduce Lemur and Lemur-Chat, openly accessible language models optimized for both natural language and coding capabilities to serve as the backbone of versatile language agents. The evolution from language chat models to functional language agents demands that models not only master human interaction, reasoning, ...
http://arxiv.org/pdf/2310.06830v1
2023-10-10T17:57:45
2023-10-10T17:57:45
{ "title": [ -0.03982544, -0.033325195, -0.034240723, -0.0670166, 0.0021343231, 0.01637268, -0.04055786, 0.022598267, 0.0060310364, 0.027420044, -0.03488159, 0.007209778, 0.025390625, -0.005619049, -0.0345459, 0.007129669, -0.0206604, 0.0362854, ...
http://arxiv.org/abs/2310.06827v3
Teaching Language Models to Hallucinate Less with Synthetic Tasks
Erik Jones, Hamid Palangi, Clarisse Simões, Varun Chandrasekaran, Subhabrata Mukherjee, Arindam Mitra, Ahmed Awadallah, Ece Kamar
cs.CL, cs.LG
Large language models (LLMs) frequently hallucinate on abstractive summarization tasks such as document-based question-answering, meeting summarization, and clinical report generation, even though all necessary information is included in context. However, optimizing LLMs to hallucinate less on these tasks is challengin...
http://arxiv.org/pdf/2310.06827v3
2023-11-07T05:11:46
2023-10-10T17:57:00
{ "title": [ -0.0018072128, 0.009567261, -0.026977539, -0.03744507, -0.021270752, -0.013381958, -0.04434204, -0.016418457, -0.019073486, 0.031280518, -0.018341064, 0.017440796, 0.0046806335, -0.021148682, -0.02571106, 0.0027828217, -0.008262634, ...
http://arxiv.org/abs/2310.06825v1
Mistral 7B
Albert Q. Jiang, Alexandre Sablayrolles, Arthur Mensch, Chris Bamford, Devendra Singh Chaplot, Diego de las Casas, Florian Bressand, Gianna Lengyel, Guillaume Lample, Lucile Saulnier, Lélio Renard Lavaud, Marie-Anne Lachaux, Pierre Stock, Teven Le Scao, Thibaut Lavril, Thomas Wang, Timothée Lacroix, William El Sayed
cs.CL, cs.AI, cs.LG
We introduce Mistral 7B v0.1, a 7-billion-parameter language model engineered for superior performance and efficiency. Mistral 7B outperforms Llama 2 13B across all evaluated benchmarks, and Llama 1 34B in reasoning, mathematics, and code generation. Our model leverages grouped-query attention (GQA) for faster inferenc...
http://arxiv.org/pdf/2310.06825v1
2023-10-10T17:54:58
2023-10-10T17:54:58
{ "title": [ -0.049560547, 0.006652832, -0.06585693, -0.0026779175, -0.043029785, -0.013946533, -0.054260254, 0.0019989014, 0.009933472, 0.035888672, -0.04550171, 0.0038757324, 0.027175903, -0.014770508, 0.061767578, -0.001414299, 0.015205383, 0....
http://arxiv.org/abs/2310.06816v1
Text Embeddings Reveal (Almost) As Much As Text
John X. Morris, Volodymyr Kuleshov, Vitaly Shmatikov, Alexander M. Rush
cs.CL, cs.LG
How much private information do text embeddings reveal about the original text? We investigate the problem of embedding \textit{inversion}, reconstructing the full text represented in dense text embeddings. We frame the problem as controlled generation: generating text that, when reembedded, is close to a fixed point i...
http://arxiv.org/pdf/2310.06816v1
2023-10-10T17:39:03
2023-10-10T17:39:03
{ "title": [ 0.018218994, 0.013442993, -0.021972656, -0.019714355, -0.02822876, -0.04232788, -0.02684021, -0.046691895, 0.0077552795, 0.025100708, -0.06286621, -0.008026123, 0.005241394, -0.024398804, -0.0021915436, -0.015541077, 0.013557434, 0.0...
http://arxiv.org/abs/2310.06803v1
Advancing Transformer's Capabilities in Commonsense Reasoning
Yu Zhou, Yunqiu Han, Hanyu Zhou, Yulun Wu
cs.CL, cs.AI, cs.LG
Recent advances in general purpose pre-trained language models have shown great potential in commonsense reasoning. However, current works still perform poorly on standard commonsense reasoning benchmarks including the Com2Sense Dataset. We argue that this is due to a disconnect with current cutting-edge machine learni...
http://arxiv.org/pdf/2310.06803v1
2023-10-10T17:21:03
2023-10-10T17:21:03
{ "title": [ 0.05633545, -0.022140503, -0.015640259, -0.055786133, -0.016235352, -0.01612854, -0.015914917, -0.0014324188, -0.0073051453, 0.029541016, -0.020263672, -0.007331848, 0.021911621, -0.04144287, 0.027801514, -0.0014467239, 0.010108948, ...
http://arxiv.org/abs/2310.06786v1
OpenWebMath: An Open Dataset of High-Quality Mathematical Web Text
Keiran Paster, Marco Dos Santos, Zhangir Azerbayev, Jimmy Ba
cs.AI, cs.CL, cs.LG
There is growing evidence that pretraining on high quality, carefully thought-out tokens such as code or mathematics plays an important role in improving the reasoning abilities of large language models. For example, Minerva, a PaLM model finetuned on billions of tokens of mathematical documents from arXiv and the web,...
http://arxiv.org/pdf/2310.06786v1
2023-10-10T16:57:28
2023-10-10T16:57:28
{ "title": [ -0.008956909, 0.020019531, -0.04611206, -0.05105591, 0.032409668, -0.02418518, -0.00793457, -0.015388489, 0.0068130493, 0.0390625, 0.0007953644, 0.02255249, 0.0072288513, -0.009208679, -0.033843994, 0.029525757, 0.011245728, 0.008140...
http://arxiv.org/abs/2310.06773v1
Uni3D: Exploring Unified 3D Representation at Scale
Junsheng Zhou, Jinsheng Wang, Baorui Ma, Yu-Shen Liu, Tiejun Huang, Xinlong Wang
cs.CV, cs.CL
Scaling up representations for images or text has been extensively investigated in the past few years and has led to revolutions in learning vision and language. However, scalable representation for 3D objects and scenes is relatively unexplored. In this work, we present Uni3D, a 3D foundation model to explore the unif...
http://arxiv.org/pdf/2310.06773v1
2023-10-10T16:49:21
2023-10-10T16:49:21
{ "title": [ -0.0463562, 0.0023593903, -0.051208496, -0.052947998, 0.010543823, 0.004638672, -0.034698486, -0.01927185, 0.002073288, 0.036743164, -0.009414673, -0.00051403046, -0.008293152, -0.02508545, -0.035308838, 0.012741089, 0.018234253, 0.0...