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/2305.11744v1
Inference-time Re-ranker Relevance Feedback for Neural Information Retrieval
Revanth Gangi Reddy, Pradeep Dasigi, Md Arafat Sultan, Arman Cohan, Avirup Sil, Heng Ji, Hannaneh Hajishirzi
cs.IR, cs.CL
Neural information retrieval often adopts a retrieve-and-rerank framework: a bi-encoder network first retrieves K (e.g., 100) candidates that are then re-ranked using a more powerful cross-encoder model to rank the better candidates higher. The re-ranker generally produces better candidate scores than the retriever, bu...
http://arxiv.org/pdf/2305.11744v1
2023-05-19T15:30:33
2023-05-19T15:30:33
{ "title": [ 0.0020198822, 0.009780884, -0.016815186, -0.011878967, 0.009300232, -0.018981934, 0.0011167526, -0.043548584, 0.016098022, 0.052734375, -0.028930664, -0.00642395, 0.028640747, -0.0052757263, -0.054016113, 0.002614975, -0.0011634827, ...
http://arxiv.org/abs/2305.11738v2
CRITIC: Large Language Models Can Self-Correct with Tool-Interactive Critiquing
Zhibin Gou, Zhihong Shao, Yeyun Gong, Yelong Shen, Yujiu Yang, Nan Duan, Weizhu Chen
cs.CL, cs.AI
Recent developments in large language models (LLMs) have been impressive. However, these models sometimes show inconsistencies and problematic behavior, such as hallucinating facts, generating flawed code, or creating offensive and toxic content. Unlike these models, humans typically utilize external tools to cross-che...
http://arxiv.org/pdf/2305.11738v2
2023-09-30T08:35:29
2023-05-19T15:19:44
{ "title": [ -0.0051651, -0.013504028, -0.012329102, -0.0423584, -0.046783447, 0.00059223175, -0.052520752, -0.016098022, -0.0423584, 0.027252197, -0.030838013, -0.002292633, -0.020019531, -0.033203125, -0.018081665, -0.045410156, -0.009262085, 0...
http://arxiv.org/abs/2305.13331v1
A New Benchmark of Aphasia Speech Recognition and Detection Based on E-Branchformer and Multi-task Learning
Jiyang Tang, William Chen, Xuankai Chang, Shinji Watanabe, Brian MacWhinney
eess.AS, cs.CL
Aphasia is a language disorder that affects the speaking ability of millions of patients. This paper presents a new benchmark for Aphasia speech recognition and detection tasks using state-of-the-art speech recognition techniques with the AphsiaBank dataset. Specifically, we introduce two multi-task learning methods ba...
http://arxiv.org/pdf/2305.13331v1
2023-05-19T15:10:36
2023-05-19T15:10:36
{ "title": [ 0.0015974045, 0.021957397, -0.061340332, -0.048980713, 0.008163452, -0.02053833, -0.050598145, -0.009117126, -0.019515991, 0.054779053, -0.00655365, -0.0037117004, 0.0062179565, -0.033935547, -0.04736328, -0.01285553, -0.0026817322, ...
http://arxiv.org/abs/2305.11731v1
Persian Typographical Error Type Detection using Many-to-Many Deep Neural Networks on Algorithmically-Generated Misspellings
Mohammad Dehghani, Heshaam Faili
cs.CL, cs.AI
Digital technologies have led to an influx of text created daily in a variety of languages, styles, and formats. A great deal of the popularity of spell-checking systems can be attributed to this phenomenon since they are crucial to polishing the digitally conceived text. In this study, we tackle Typographical Error Ty...
http://arxiv.org/pdf/2305.11731v1
2023-05-19T15:05:39
2023-05-19T15:05:39
{ "title": [ -0.003534317, 0.0005207062, -0.022125244, -0.020996094, -0.0012674332, -0.008224487, -0.00095558167, -0.010192871, 0.02230835, 0.023986816, -0.02670288, 0.0032482147, 0.018341064, -0.026260376, -0.020263672, -0.01322937, -0.0385437, ...
http://arxiv.org/abs/2305.11725v1
S$^3$HQA: A Three-Stage Approach for Multi-hop Text-Table Hybrid Question Answering
Fangyu Lei, Xiang Li, Yifan Wei, Shizhu He, Yiming Huang, Jun Zhao, Kang Liu
cs.CL
Answering multi-hop questions over hybrid factual knowledge from the given text and table (TextTableQA) is a challenging task. Existing models mainly adopt a retriever-reader framework, which have several deficiencies, such as noisy labeling in training retriever, insufficient utilization of heterogeneous information o...
http://arxiv.org/pdf/2305.11725v1
2023-05-19T15:01:48
2023-05-19T15:01:48
{ "title": [ -0.02104187, -0.017807007, -0.029174805, -0.041412354, -0.008865356, 0.004055023, -0.034179688, 0.017700195, 0.017700195, 0.05316162, -0.06188965, 0.020584106, 0.027893066, 0.021331787, -0.06652832, -0.0073127747, 0.0035591125, 0.004...
http://arxiv.org/abs/2305.11719v2
Information Screening whilst Exploiting! Multimodal Relation Extraction with Feature Denoising and Multimodal Topic Modeling
Shengqiong Wu, Hao Fei, Yixin Cao, Lidong Bing, Tat-Seng Chua
cs.CV, cs.CL
Existing research on multimodal relation extraction (MRE) faces two co-existing challenges, internal-information over-utilization and external-information under-exploitation. To combat that, we propose a novel framework that simultaneously implements the idea of internal-information screening and external-information e...
http://arxiv.org/pdf/2305.11719v2
2023-05-25T04:08:21
2023-05-19T14:56:57
{ "title": [ -0.016403198, 0.020187378, -0.06506348, -0.04034424, -0.0006322861, -0.0211792, -0.03741455, -0.033477783, 0.035308838, 0.0068588257, 0.006778717, 0.013679504, -0.007385254, -0.047424316, -0.025680542, 0.0093688965, -0.008216858, 0.0...
http://arxiv.org/abs/2305.11707v2
What Comes Next? Evaluating Uncertainty in Neural Text Generators Against Human Production Variability
Mario Giulianelli, Joris Baan, Wilker Aziz, Raquel Fernández, Barbara Plank
cs.CL, cs.AI, cs.LG
In Natural Language Generation (NLG) tasks, for any input, multiple communicative goals are plausible, and any goal can be put into words, or produced, in multiple ways. We characterise the extent to which human production varies lexically, syntactically, and semantically across four NLG tasks, connecting human product...
http://arxiv.org/pdf/2305.11707v2
2023-10-20T14:31:21
2023-05-19T14:41:55
{ "title": [ -0.043701172, 0.013664246, -0.049346924, -0.015182495, -0.005882263, -0.021499634, -0.042022705, -0.013328552, -0.028884888, 0.03643799, -0.012275696, 0.022384644, -0.019317627, -0.011932373, -0.054504395, -0.013145447, -0.005645752, ...
http://arxiv.org/abs/2305.11694v2
QUEST: A Retrieval Dataset of Entity-Seeking Queries with Implicit Set Operations
Chaitanya Malaviya, Peter Shaw, Ming-Wei Chang, Kenton Lee, Kristina Toutanova
cs.CL
Formulating selective information needs results in queries that implicitly specify set operations, such as intersection, union, and difference. For instance, one might search for "shorebirds that are not sandpipers" or "science-fiction films shot in England". To study the ability of retrieval systems to meet such infor...
http://arxiv.org/pdf/2305.11694v2
2023-05-31T05:11:21
2023-05-19T14:19:32
{ "title": [ 0.0029850006, 0.003692627, -0.04928589, -0.04660034, 0.041046143, -0.021591187, 0.013534546, -0.008224487, 0.028305054, 0.04156494, -0.038391113, -0.0033512115, 0.012802124, 0.0036449432, -0.062805176, 0.010719299, -0.009346008, 0.02...
http://arxiv.org/abs/2305.11692v1
Surgical-VQLA: Transformer with Gated Vision-Language Embedding for Visual Question Localized-Answering in Robotic Surgery
Long Bai, Mobarakol Islam, Lalithkumar Seenivasan, Hongliang Ren
cs.CV, cs.AI, cs.CL, cs.LG, cs.RO
Despite the availability of computer-aided simulators and recorded videos of surgical procedures, junior residents still heavily rely on experts to answer their queries. However, expert surgeons are often overloaded with clinical and academic workloads and limit their time in answering. For this purpose, we develop a s...
http://arxiv.org/pdf/2305.11692v1
2023-05-19T14:13:47
2023-05-19T14:13:47
{ "title": [ -0.014709473, -0.008865356, -0.047912598, -0.04272461, -0.00491333, -0.013008118, -0.039520264, 0.0012836456, 0.02281189, 0.0051078796, -0.03463745, -0.013999939, -0.0013437271, -0.02999878, -0.026489258, -0.026321411, 0.0287323, 0.0...
http://arxiv.org/abs/2305.11685v2
Recycle-and-Distill: Universal Compression Strategy for Transformer-based Speech SSL Models with Attention Map Reusing and Masking Distillation
Kangwook Jang, Sungnyun Kim, Se-Young Yun, Hoirin Kim
eess.AS, cs.CL, cs.LG
Transformer-based speech self-supervised learning (SSL) models, such as HuBERT, show surprising performance in various speech processing tasks. However, huge number of parameters in speech SSL models necessitate the compression to a more compact model for wider usage in academia or small companies. In this study, we su...
http://arxiv.org/pdf/2305.11685v2
2023-10-26T10:43:07
2023-05-19T14:07:43
{ "title": [ -0.0017490387, 0.029388428, -0.046051025, -0.06677246, -0.0033397675, 0.012306213, -0.09527588, 0.0016851425, 0.015586853, 0.064697266, 0.021957397, -0.026382446, 0.07318115, -0.048950195, -0.0069770813, 0.024291992, 0.00141716, 0.02...
http://arxiv.org/abs/2305.11683v1
Sensing of inspiration events from speech: comparison of deep learning and linguistic methods
Aki Härmä, Ulf Grossekathöfer, Okke Ouweltjes, Venkata Srikanth Nallanthighal
cs.SD, cs.CL, cs.LG, eess.AS
Respiratory chest belt sensor can be used to measure the respiratory rate and other respiratory health parameters. Virtual Respiratory Belt, VRB, algorithms estimate the belt sensor waveform from speech audio. In this paper we compare the detection of inspiration events (IE) from respiratory belt sensor data using a no...
http://arxiv.org/pdf/2305.11683v1
2023-05-19T14:06:16
2023-05-19T14:06:16
{ "title": [ -0.0033168793, -0.0076408386, -0.023498535, -0.045532227, -0.015007019, -0.01876831, -0.036621094, -0.013366699, 0.005996704, 0.041046143, -0.009025574, -0.019180298, 0.044952393, -0.031433105, -0.021972656, 0.015167236, -0.014205933, ...
http://arxiv.org/abs/2305.11926v1
MParrotTTS: Multilingual Multi-speaker Text to Speech Synthesis in Low Resource Setting
Neil Shah, Vishal Tambrahalli, Saiteja Kosgi, Niranjan Pedanekar, Vineet Gandhi
cs.SD, cs.CL, cs.LG, eess.AS
We present MParrotTTS, a unified multilingual, multi-speaker text-to-speech (TTS) synthesis model that can produce high-quality speech. Benefiting from a modularized training paradigm exploiting self-supervised speech representations, MParrotTTS adapts to a new language with minimal supervised data and generalizes to l...
http://arxiv.org/pdf/2305.11926v1
2023-05-19T13:43:36
2023-05-19T13:43:36
{ "title": [ -0.0362854, 0.0009007454, -0.035858154, -0.07281494, 0.003194809, 0.025939941, -0.08380127, 0.018417358, -0.025375366, 0.04763794, -0.016967773, 0.01902771, 0.013694763, -0.014793396, -0.031799316, 0.002166748, -0.024841309, 0.039978...
http://arxiv.org/abs/2305.11673v1
Bias Beyond English: Counterfactual Tests for Bias in Sentiment Analysis in Four Languages
Seraphina Goldfarb-Tarrant, Adam Lopez, Roi Blanco, Diego Marcheggiani
cs.CL
Sentiment analysis (SA) systems are used in many products and hundreds of languages. Gender and racial biases are well-studied in English SA systems, but understudied in other languages, with few resources for such studies. To remedy this, we build a counterfactual evaluation corpus for gender and racial/migrant bias i...
http://arxiv.org/pdf/2305.11673v1
2023-05-19T13:38:53
2023-05-19T13:38:53
{ "title": [ -0.01461792, -0.003162384, -0.025604248, -0.029785156, 0.016464233, -0.008834839, -0.012649536, -0.033172607, 0.0135269165, 0.032562256, -0.036712646, -0.005207062, -0.0006008148, -0.011268616, -0.0647583, -0.010505676, -0.016845703, ...
http://arxiv.org/abs/2305.11663v1
Algorithmic failure as a humanities methodology: machine learning's mispredictions identify rich cases for qualitative analysis
Jill Walker Rettberg
cs.LG, cs.AI, cs.CL, cs.CY, J.5
This commentary tests a methodology proposed by Munk et al. (2022) for using failed predictions in machine learning as a method to identify ambiguous and rich cases for qualitative analysis. Using a dataset describing actions performed by fictional characters interacting with machine vision technologies in 500 artworks...
http://arxiv.org/pdf/2305.11663v1
2023-05-19T13:24:32
2023-05-19T13:24:32
{ "title": [ 0.004611969, 0.004047394, 0.017852783, 0.032562256, -0.017669678, -0.03643799, 0.053253174, -0.05935669, -0.023986816, 0.024719238, -0.04736328, -0.0065574646, 0.0005631447, -0.025756836, -0.03265381, -0.04257202, -0.019485474, 0.056...
http://arxiv.org/abs/2305.11662v2
Separating form and meaning: Using self-consistency to quantify task understanding across multiple senses
Xenia Ohmer, Elia Bruni, Dieuwke Hupkes
cs.CL, cs.AI
At the staggering pace with which the capabilities of large language models (LLMs) are increasing, creating future-proof evaluation sets to assess their understanding becomes more and more challenging. In this paper, we propose a novel paradigm for evaluating LLMs which leverages the idea that correct world understandi...
http://arxiv.org/pdf/2305.11662v2
2023-05-23T15:12:45
2023-05-19T13:23:51
{ "title": [ -0.03213501, 0.00039839745, -0.05038452, -0.03652954, 0.008132935, -0.024353027, -0.038208008, -0.015975952, 0.010444641, 0.028396606, 0.007457733, -0.014915466, -0.00008928776, -0.03274536, -0.04776001, 0.0021572113, 0.0034122467, 0...
http://arxiv.org/abs/2305.11627v3
LLM-Pruner: On the Structural Pruning of Large Language Models
Xinyin Ma, Gongfan Fang, Xinchao Wang
cs.CL
Large language models (LLMs) have shown remarkable capabilities in language understanding and generation. However, such impressive capability typically comes with a substantial model size, which presents significant challenges in both the deployment, inference, and training stages. With LLM being a general-purpose task...
http://arxiv.org/pdf/2305.11627v3
2023-09-28T03:59:27
2023-05-19T12:10:53
{ "title": [ -0.046051025, -0.020935059, -0.04345703, -0.07757568, 0.0026073456, 0.030532837, -0.07476807, 0.030090332, 0.008964539, 0.015037537, -0.0113220215, 0.0016460419, 0.007926941, -0.023742676, -0.03390503, 0.008674622, -0.014633179, 0.03...
http://arxiv.org/abs/2305.11626v1
CCT-Code: Cross-Consistency Training for Multilingual Clone Detection and Code Search
Nikita Sorokin, Dmitry Abulkhanov, Sergey Nikolenko, Valentin Malykh
cs.CL, cs.SE
We consider the clone detection and information retrieval problems for source code, well-known tasks important for any programming language. Although it is also an important and interesting problem to find code snippets that operate identically but are written in different programming languages, to the best of our know...
http://arxiv.org/pdf/2305.11626v1
2023-05-19T12:09:49
2023-05-19T12:09:49
{ "title": [ -0.002714157, -0.005088806, -0.011756897, -0.05343628, -0.0015258789, 0.0069007874, -0.044952393, -0.017578125, 0.005367279, 0.060760498, -0.013168335, -0.0047454834, 0.05706787, -0.027496338, -0.025161743, 0.020401001, -0.008453369, ...
http://arxiv.org/abs/2305.11625v1
Searching by Code: a New SearchBySnippet Dataset and SnippeR Retrieval Model for Searching by Code Snippets
Ivan Sedykh, Dmitry Abulkhanov, Nikita Sorokin, Sergey Nikolenko, Valentin Malykh
cs.CL, cs.SE
Code search is an important task that has seen many developments in recent years. However, previous attempts have mostly considered the problem of searching for code by a text query. We argue that using a code snippet (and possibly an associated traceback) as a query and looking for answers with bugfixing instructions ...
http://arxiv.org/pdf/2305.11625v1
2023-05-19T12:09:30
2023-05-19T12:09:30
{ "title": [ -0.019088745, -0.019546509, -0.038482666, -0.037994385, -0.00749588, -0.021591187, -0.030075073, -0.016799927, 0.027053833, 0.030029297, -0.040893555, -0.0056495667, 0.015266418, -0.013145447, -0.032104492, 0.008216858, 0.000024080276, ...
http://arxiv.org/abs/2305.11603v1
Attributable and Scalable Opinion Summarization
Tom Hosking, Hao Tang, Mirella Lapata
cs.CL
We propose a method for unsupervised opinion summarization that encodes sentences from customer reviews into a hierarchical discrete latent space, then identifies common opinions based on the frequency of their encodings. We are able to generate both abstractive summaries by decoding these frequent encodings, and extra...
http://arxiv.org/pdf/2305.11603v1
2023-05-19T11:30:37
2023-05-19T11:30:37
{ "title": [ 0.008460999, -0.0071105957, -0.021911621, -0.08166504, 0.0021648407, 0.011314392, -0.049041748, -0.007789612, 0.021865845, 0.045440674, -0.036895752, 0.0022468567, -0.0002837181, 0.012237549, -0.06048584, 0.00093603134, 0.001042366, ...
http://arxiv.org/abs/2305.11598v1
Introspective Tips: Large Language Model for In-Context Decision Making
Liting Chen, Lu Wang, Hang Dong, Yali Du, Jie Yan, Fangkai Yang, Shuang Li, Pu Zhao, Si Qin, Saravan Rajmohan, Qingwei Lin, Dongmei Zhang
cs.AI, cs.CL
The emergence of large language models (LLMs) has substantially influenced natural language processing, demonstrating exceptional results across various tasks. In this study, we employ ``Introspective Tips" to facilitate LLMs in self-optimizing their decision-making. By introspectively examining trajectories, LLM refin...
http://arxiv.org/pdf/2305.11598v1
2023-05-19T11:20:37
2023-05-19T11:20:37
{ "title": [ -0.010528564, -0.024093628, -0.042816162, -0.045318604, 0.0026054382, -0.013046265, -0.017852783, -0.011260986, 0.007133484, 0.035888672, -0.034606934, 0.00843811, -0.0064315796, -0.011878967, -0.03390503, -0.0087890625, -0.0064811707, ...
http://arxiv.org/abs/2305.11596v2
Mitigating Backdoor Poisoning Attacks through the Lens of Spurious Correlation
Xuanli He, Qiongkai Xu, Jun Wang, Benjamin Rubinstein, Trevor Cohn
cs.CL, cs.CR
Modern NLP models are often trained over large untrusted datasets, raising the potential for a malicious adversary to compromise model behaviour. For instance, backdoors can be implanted through crafting training instances with a specific textual trigger and a target label. This paper posits that backdoor poisoning att...
http://arxiv.org/pdf/2305.11596v2
2023-10-20T08:10:27
2023-05-19T11:18:20
{ "title": [ 0.00447464, 0.019744873, -0.030731201, -0.04269409, 0.02293396, -0.017028809, 0.003194809, -0.030914307, -0.070251465, 0.033294678, 0.019607544, 0.028274536, -0.035614014, -0.02368164, -0.017700195, -0.025100708, -0.0060539246, 0.033...
http://arxiv.org/abs/2305.11595v3
Examining Inter-Consistency of Large Language Models Collaboration: An In-depth Analysis via Debate
Kai Xiong, Xiao Ding, Yixin Cao, Ting Liu, Bing Qin
cs.CL, cs.AI
Large Language Models (LLMs) have shown impressive capabilities in various applications, but they still face various inconsistency issues. Existing works primarily focus on the inconsistency issues within a single LLM, while we complementarily explore the inter-consistency among multiple LLMs for collaboration. To exam...
http://arxiv.org/pdf/2305.11595v3
2023-10-18T06:32:15
2023-05-19T11:15:33
{ "title": [ -0.049835205, -0.005290985, -0.079833984, -0.061309814, -0.017700195, -0.007007599, -0.07366943, 0.026824951, -0.030776978, 0.02911377, -0.007858276, -0.0077590942, 0.0012273788, -0.029922485, -0.022613525, -0.030273438, -0.022537231, ...
http://arxiv.org/abs/2305.11592v1
IKDSumm: Incorporating Key-phrases into BERT for extractive Disaster Tweet Summarization
Piyush Kumar Garg, Roshni Chakraborty, Srishti Gupta, Sourav Kumar Dandapat
cs.CL, cs.SI
Online social media platforms, such as Twitter, are one of the most valuable sources of information during disaster events. Therefore, humanitarian organizations, government agencies, and volunteers rely on a summary of this information, i.e., tweets, for effective disaster management. Although there are several existi...
http://arxiv.org/pdf/2305.11592v1
2023-05-19T11:05:55
2023-05-19T11:05:55
{ "title": [ 0.001159668, -0.0039978027, -0.0050315857, -0.08746338, -0.017333984, 0.01802063, -0.06524658, 0.00014531612, 0.0036830902, 0.08239746, -0.034454346, 0.06008911, 0.03717041, 0.019165039, -0.059692383, 0.026229858, -0.0107421875, -0.0...
http://arxiv.org/abs/2305.11579v2
Speech-Text Dialog Pre-training for Spoken Dialog Understanding with Explicit Cross-Modal Alignment
Tianshu Yu, Haoyu Gao, Ting-En Lin, Min Yang, Yuchuan Wu, Wentao Ma, Chao Wang, Fei Huang, Yongbin Li
cs.CL, cs.AI
Recently, speech-text pre-training methods have shown remarkable success in many speech and natural language processing tasks. However, most previous pre-trained models are usually tailored for one or two specific tasks, but fail to conquer a wide range of speech-text tasks. In addition, existing speech-text pre-traini...
http://arxiv.org/pdf/2305.11579v2
2023-06-09T03:48:42
2023-05-19T10:37:56
{ "title": [ 0.004878998, 0.002779007, -0.050231934, -0.08618164, -0.004802704, -0.012931824, -0.0680542, 0.011383057, 0.000025868416, 0.04159546, -0.01876831, 0.010353088, 0.027023315, -0.016159058, -0.028640747, -0.004463196, -0.018447876, 0.03...
http://arxiv.org/abs/2305.11576v1
Language-universal phonetic encoder for low-resource speech recognition
Siyuan Feng, Ming Tu, Rui Xia, Chuanzeng Huang, Yuxuan Wang
eess.AS, cs.CL, cs.SD
Multilingual training is effective in improving low-resource ASR, which may partially be explained by phonetic representation sharing between languages. In end-to-end (E2E) ASR systems, graphemes are often used as basic modeling units, however graphemes may not be ideal for multilingual phonetic sharing. In this paper,...
http://arxiv.org/pdf/2305.11576v1
2023-05-19T10:24:30
2023-05-19T10:24:30
{ "title": [ -0.024002075, -0.0054244995, -0.04397583, -0.07141113, -0.0018672943, 0.0055999756, -0.05480957, 0.017730713, 0.008529663, 0.05328369, -0.012077332, 0.0069770813, 0.011657715, -0.014854431, -0.043426514, 0.021392822, 0.0017518997, 0....
http://arxiv.org/abs/2305.11569v1
Language-Universal Phonetic Representation in Multilingual Speech Pretraining for Low-Resource Speech Recognition
Siyuan Feng, Ming Tu, Rui Xia, Chuanzeng Huang, Yuxuan Wang
eess.AS, cs.CL, cs.SD
We improve low-resource ASR by integrating the ideas of multilingual training and self-supervised learning. Concretely, we leverage an International Phonetic Alphabet (IPA) multilingual model to create frame-level pseudo labels for unlabeled speech, and use these pseudo labels to guide hidden-unit BERT (HuBERT) based s...
http://arxiv.org/pdf/2305.11569v1
2023-05-19T10:15:11
2023-05-19T10:15:11
{ "title": [ -0.0011205673, -0.002407074, -0.036346436, -0.082214355, -0.0020065308, 0.004623413, -0.0713501, 0.023864746, -0.017913818, 0.057769775, 0.0038757324, 0.014465332, 0.014434814, -0.029541016, -0.027877808, 0.0067825317, 0.0027694702, ...
http://arxiv.org/abs/2305.11564v2
Decouple knowledge from parameters for plug-and-play language modeling
Xin Cheng, Yankai Lin, Xiuying Chen, Dongyan Zhao, Rui Yan
cs.CL, cs.AI
Pre-trained language models(PLM) have made impressive results in various NLP tasks. It has been revealed that one of the key factors to their success is the parameters of these models implicitly learn all kinds of knowledge during pre-training. However, encoding knowledge implicitly in the model parameters has two fund...
http://arxiv.org/pdf/2305.11564v2
2023-09-18T09:42:26
2023-05-19T10:01:55
{ "title": [ -0.029403687, -0.0151901245, -0.041137695, -0.07623291, -0.020812988, 0.007091522, -0.07165527, 0.014846802, -0.032928467, 0.013587952, -0.0077400208, 0.000089108944, -0.013877869, -0.03186035, -0.01651001, -0.0062789917, 0.0005698204, ...
http://arxiv.org/abs/2305.11558v1
Blank-regularized CTC for Frame Skipping in Neural Transducer
Yifan Yang, Xiaoyu Yang, Liyong Guo, Zengwei Yao, Wei Kang, Fangjun Kuang, Long Lin, Xie Chen, Daniel Povey
eess.AS, cs.CL
Neural Transducer and connectionist temporal classification (CTC) are popular end-to-end automatic speech recognition systems. Due to their frame-synchronous design, blank symbols are introduced to address the length mismatch between acoustic frames and output tokens, which might bring redundant computation. Previous s...
http://arxiv.org/pdf/2305.11558v1
2023-05-19T09:56:09
2023-05-19T09:56:09
{ "title": [ 0.0118255615, -0.0023860931, -0.0023231506, -0.02456665, -0.014793396, -0.0029640198, -0.0018386841, -0.02104187, 0.014472961, 0.045928955, -0.030715942, -0.019180298, 0.026428223, -0.029205322, -0.03225708, 0.0018672943, -0.01234436, ...
http://arxiv.org/abs/2305.11554v3
ToolkenGPT: Augmenting Frozen Language Models with Massive Tools via Tool Embeddings
Shibo Hao, Tianyang Liu, Zhen Wang, Zhiting Hu
cs.CL, cs.LG
Augmenting large language models (LLMs) with external tools has emerged as a promising approach to solving complex problems. However, traditional methods, which finetune LLMs with tool demonstration data, can be both costly and restricted to a predefined set of tools. Recent in-context learning paradigm alleviates thes...
http://arxiv.org/pdf/2305.11554v3
2023-10-30T21:46:30
2023-05-19T09:54:21
{ "title": [ -0.025756836, -0.019729614, -0.038330078, -0.06677246, -0.0064468384, 0.016036987, -0.06311035, 0.015617371, 0.014518738, 0.02406311, -0.008834839, 0.0040245056, 0.024978638, -0.017364502, -0.034698486, 0.026290894, -0.00957489, 0.04...
http://arxiv.org/abs/2305.11553v1
Unsupervised Scientific Abstract Segmentation with Normalized Mutual Information
Yingqiang Gao, Jessica Lam, Nianlong Gu, Richard H. R. Hahnloser
cs.CL
The abstracts of scientific papers consist of premises and conclusions. Structured abstracts explicitly highlight the conclusion sentences, whereas non-structured abstracts may have conclusion sentences at uncertain positions. This implicit nature of conclusion positions makes the automatic segmentation of scientific a...
http://arxiv.org/pdf/2305.11553v1
2023-05-19T09:53:45
2023-05-19T09:53:45
{ "title": [ -0.02885437, 0.026275635, -0.034484863, -0.036254883, 0.01033783, 0.0072288513, -0.028259277, -0.028274536, -0.006767273, 0.027954102, -0.0019779205, 0.018295288, -0.022766113, -0.017669678, -0.021850586, 0.017318726, 0.013885498, 0....
http://arxiv.org/abs/2305.11550v2
Viewing Knowledge Transfer in Multilingual Machine Translation Through a Representational Lens
David Stap, Vlad Niculae, Christof Monz
cs.CL
We argue that translation quality alone is not a sufficient metric for measuring knowledge transfer in multilingual neural machine translation. To support this claim, we introduce Representational Transfer Potential (RTP), which measures representational similarities between languages. We show that RTP can measure both...
http://arxiv.org/pdf/2305.11550v2
2023-10-23T07:29:13
2023-05-19T09:36:48
{ "title": [ -0.029449463, -0.018478394, -0.014038086, -0.032806396, -0.028381348, -0.00573349, -0.022644043, -0.004802704, -0.011726379, 0.057739258, -0.01537323, 0.014602661, 0.035369873, -0.011924744, -0.049560547, 0.00869751, 0.0019273758, 0....
http://arxiv.org/abs/2305.11543v2
Constructing Word-Context-Coupled Space Aligned with Associative Knowledge Relations for Interpretable Language Modeling
Fanyu Wang, Zhenping Xie
cs.CL
As the foundation of current natural language processing methods, pre-trained language model has achieved excellent performance. However, the black-box structure of the deep neural network in pre-trained language models seriously limits the interpretability of the language modeling process. After revisiting the coupled...
http://arxiv.org/pdf/2305.11543v2
2023-06-27T15:45:46
2023-05-19T09:26:02
{ "title": [ -0.040405273, -0.005973816, -0.031433105, -0.058929443, -0.028686523, 0.023788452, -0.07745361, 0.00054597855, 0.0056762695, 0.034973145, -0.018539429, -0.00063610077, 0.012298584, -0.010520935, -0.056274414, -0.0061149597, -0.0039100647...
http://arxiv.org/abs/2305.11541v3
Empower Large Language Model to Perform Better on Industrial Domain-Specific Question Answering
Fangkai Yang, Pu Zhao, Zezhong Wang, Lu Wang, Jue Zhang, Mohit Garg, Qingwei Lin, Saravan Rajmohan, Dongmei Zhang
cs.CL, cs.AI
Large Language Model (LLM) has gained popularity and achieved remarkable results in open-domain tasks, but its performance in real industrial domain-specific scenarios is average due to its lack of specific domain knowledge. This issue has attracted widespread attention, but there are few relevant benchmarks available....
http://arxiv.org/pdf/2305.11541v3
2023-10-16T10:48:00
2023-05-19T09:23:25
{ "title": [ -0.012374878, -0.025894165, -0.066589355, -0.06347656, -0.019134521, -0.026809692, -0.043548584, 0.047821045, -0.003742218, 0.022659302, -0.04284668, -0.005542755, 0.02507019, -0.019515991, -0.004470825, -0.006175995, -0.0069885254, ...
http://arxiv.org/abs/2305.11540v1
Efficient Cross-Lingual Transfer for Chinese Stable Diffusion with Images as Pivots
Jinyi Hu, Xu Han, Xiaoyuan Yi, Yutong Chen, Wenhao Li, Zhiyuan Liu, Maosong Sun
cs.CV, cs.CL
Diffusion models have made impressive progress in text-to-image synthesis. However, training such large-scale models (e.g. Stable Diffusion), from scratch requires high computational costs and massive high-quality text-image pairs, which becomes unaffordable in other languages. To handle this challenge, we propose IAP,...
http://arxiv.org/pdf/2305.11540v1
2023-05-19T09:20:27
2023-05-19T09:20:27
{ "title": [ 0.0008664131, -0.010284424, 0.0065994263, -0.023406982, 0.0030784607, 0.000849247, 0.00045466423, -0.035614014, 0.033111572, 0.040496826, -0.015777588, 0.0066566467, 0.017044067, -0.029022217, -0.021102905, 0.01473999, -0.00033211708, ...
http://arxiv.org/abs/2305.11536v1
PORTRAIT: a hybrid aPproach tO cReate extractive ground-TRuth summAry for dIsaster evenT
Piyush Kumar Garg, Roshni Chakraborty, Sourav Kumar Dandapat
cs.CL, cs.SI
Disaster summarization approaches provide an overview of the important information posted during disaster events on social media platforms, such as, Twitter. However, the type of information posted significantly varies across disasters depending on several factors like the location, type, severity, etc. Verification of...
http://arxiv.org/pdf/2305.11536v1
2023-05-19T09:07:52
2023-05-19T09:07:52
{ "title": [ -0.028442383, -0.00058078766, -0.057525635, -0.0904541, -0.0033187866, -0.0076065063, -0.062927246, 0.022903442, -0.034301758, 0.051208496, -0.010482788, 0.04748535, -0.0118637085, -0.013648987, -0.024658203, -0.0016994476, 0.0047073364,...
http://arxiv.org/abs/2305.11529v1
A Sequence-to-Sequence Approach for Arabic Pronoun Resolution
Hanan S. Murayshid, Hafida Benhidour, Said Kerrache
cs.CL, cs.LG
This paper proposes a sequence-to-sequence learning approach for Arabic pronoun resolution, which explores the effectiveness of using advanced natural language processing (NLP) techniques, specifically Bi-LSTM and the BERT pre-trained Language Model, in solving the pronoun resolution problem in Arabic. The proposed app...
http://arxiv.org/pdf/2305.11529v1
2023-05-19T08:53:41
2023-05-19T08:53:41
{ "title": [ -0.0033054352, -0.00083065033, -0.020141602, -0.122680664, 0.017623901, 0.018173218, -0.06756592, 0.031158447, 0.011833191, 0.025604248, 0.011566162, -0.011062622, 0.015151978, -0.04168701, -0.03314209, 0.0030841827, -0.004890442, 0....
http://arxiv.org/abs/2305.11527v1
InstructIE: A Chinese Instruction-based Information Extraction Dataset
Honghao Gui, Jintian Zhang, Hongbin Ye, Ningyu Zhang
cs.CL, cs.AI, cs.IR, cs.LG
We introduce a new Information Extraction (IE) task dubbed Instruction-based IE, which aims to ask the system to follow specific instructions or guidelines to extract information. To facilitate research in this area, we construct a dataset called InstructIE, consisting of 270,000 weakly supervised data from Chinese Wik...
http://arxiv.org/pdf/2305.11527v1
2023-05-19T08:51:11
2023-05-19T08:51:11
{ "title": [ -0.048919678, -0.010719299, -0.034210205, -0.03741455, -0.0012140274, -0.00070667267, -0.024780273, -0.003774643, 0.0071525574, 0.022979736, -0.031311035, 0.015701294, -0.005004883, -0.010505676, -0.031341553, -0.0014886856, -0.015464783...
http://arxiv.org/abs/2305.11517v1
DiffuSIA: A Spiral Interaction Architecture for Encoder-Decoder Text Diffusion
Chao-Hong Tan, Jia-Chen Gu, Zhen-Hua Ling
cs.CL, cs.AI
Diffusion models have emerged as the new state-of-the-art family of deep generative models, and their promising potentials for text generation have recently attracted increasing attention. Existing studies mostly adopt a single encoder architecture with partially noising processes for conditional text generation, but i...
http://arxiv.org/pdf/2305.11517v1
2023-05-19T08:30:11
2023-05-19T08:30:11
{ "title": [ -0.027252197, -0.00034475327, -0.02696228, -0.018692017, -0.0018291473, -0.0014324188, -0.005760193, -0.013336182, 0.036468506, 0.047454834, -0.033935547, 0.012031555, 0.011634827, 0.0026283264, -0.062927246, 0.0059509277, -0.015975952, ...
http://arxiv.org/abs/2305.11516v1
Contextualized Word Vector-based Methods for Discovering Semantic Differences with No Training nor Word Alignment
Ryo Nagata, Hiroya Takamura, Naoki Otani, Yoshifumi Kawasaki
cs.CL
In this paper, we propose methods for discovering semantic differences in words appearing in two corpora based on the norms of contextualized word vectors. The key idea is that the coverage of meanings is reflected in the norm of its mean word vector. The proposed methods do not require the assumptions concerning words...
http://arxiv.org/pdf/2305.11516v1
2023-05-19T08:27:17
2023-05-19T08:27:17
{ "title": [ -0.031173706, 0.0046577454, -0.0491333, -0.0362854, 0.01424408, -0.0074653625, -0.030166626, -0.020828247, 0.0011119843, 0.030792236, -0.019714355, 0.025527954, -0.004261017, -0.016281128, -0.021347046, 0.016937256, -0.0012979507, 0....
http://arxiv.org/abs/2305.11508v2
PlugMed: Improving Specificity in Patient-Centered Medical Dialogue Generation using In-Context Learning
Chengfeng Dou, Zhi Jin, Wenping Jiao, Haiyan Zhao, Zhenwei Tao, Yongqiang Zhao
cs.CL, cs.AI, I.2.7
The patient-centered medical dialogue systems strive to offer diagnostic interpretation services to users who are less knowledgeable about medical knowledge, through emphasizing the importance of providing responses specific to the patients. It is difficult for the large language models (LLMs) to guarantee the specific...
http://arxiv.org/pdf/2305.11508v2
2023-10-18T12:11:44
2023-05-19T08:18:24
{ "title": [ 0.005783081, -0.0022773743, -0.03994751, -0.06903076, 0.007408142, -0.00687027, -0.029632568, 0.009353638, -0.036956787, 0.0473938, 0.0031757355, 0.03112793, -0.00894928, -0.025405884, -0.03125, -0.011260986, -0.0113220215, 0.0184478...
http://arxiv.org/abs/2305.11503v1
A Topic-aware Summarization Framework with Different Modal Side Information
Xiuying Chen, Mingzhe Li, Shen Gao, Xin Cheng, Qiang Yang, Qishen Zhang, Xin Gao, Xiangliang Zhang
cs.CL
Automatic summarization plays an important role in the exponential document growth on the Web. On content websites such as CNN.com and WikiHow.com, there often exist various kinds of side information along with the main document for attention attraction and easier understanding, such as videos, images, and queries. Suc...
http://arxiv.org/pdf/2305.11503v1
2023-05-19T08:09:45
2023-05-19T08:09:45
{ "title": [ 0.005962372, 0.016342163, -0.028152466, -0.08178711, 0.0014972687, 0.007972717, -0.06414795, -0.01473999, 0.017669678, 0.05697632, -0.003616333, 0.027999878, 0.012626648, -0.009407043, -0.046722412, 0.029571533, 0.007522583, -0.00997...
http://arxiv.org/abs/2305.11501v1
From Alignment to Entailment: A Unified Textual Entailment Framework for Entity Alignment
Yu Zhao, Yike Wu, Xiangrui Cai, Ying Zhang, Haiwei Zhang, Xiaojie Yuan
cs.CL
Entity Alignment (EA) aims to find the equivalent entities between two Knowledge Graphs (KGs). Existing methods usually encode the triples of entities as embeddings and learn to align the embeddings, which prevents the direct interaction between the original information of the cross-KG entities. Moreover, they encode t...
http://arxiv.org/pdf/2305.11501v1
2023-05-19T08:06:50
2023-05-19T08:06:50
{ "title": [ -0.021072388, -0.011047363, -0.049468994, -0.0960083, 0.0340271, 0.01838684, -0.044189453, 0.009811401, 0.027526855, 0.03353882, 0.0038452148, 0.011878967, 0.03842163, -0.015457153, -0.034210205, 0.035858154, -0.025756836, 0.03390503...
http://arxiv.org/abs/2305.11499v2
RCOT: Detecting and Rectifying Factual Inconsistency in Reasoning by Reversing Chain-of-Thought
Tianci Xue, Ziqi Wang, Zhenhailong Wang, Chi Han, Pengfei Yu, Heng Ji
cs.CL
Large language Models (LLMs) have achieved promising performance on arithmetic reasoning tasks by incorporating step-by-step chain-of-thought (CoT) prompting. However, LLMs face challenges in maintaining factual consistency during reasoning, exhibiting tendencies to condition overlooking, question misinterpretation, an...
http://arxiv.org/pdf/2305.11499v2
2023-10-02T03:59:04
2023-05-19T08:02:52
{ "title": [ 0.010375977, -0.018920898, -0.02192688, -0.01777649, -0.027572632, -0.023666382, 0.009269714, -0.01222229, -0.01713562, 0.02319336, -0.026351929, -0.009468079, 0.033569336, -0.04269409, 0.0019893646, -0.030044556, -0.021408081, 0.062...
http://arxiv.org/abs/2305.11498v1
Recouple Event Field via Probabilistic Bias for Event Extraction
Xingyu Bai, Taiqiang Wu, Han Guo, Zhe Zhao, Xuefeng Yang, Jiayi Li, Weijie Liu, Qi Ju, Weigang Guo, Yujiu Yang
cs.CL, cs.AI, cs.IR
Event Extraction (EE), aiming to identify and classify event triggers and arguments from event mentions, has benefited from pre-trained language models (PLMs). However, existing PLM-based methods ignore the information of trigger/argument fields, which is crucial for understanding event schemas. To this end, we propose...
http://arxiv.org/pdf/2305.11498v1
2023-05-19T07:55:37
2023-05-19T07:55:37
{ "title": [ -0.003440857, 0.02281189, -0.05834961, -0.105895996, 0.036254883, 0.011787415, -0.06188965, -0.024795532, 0.014831543, 0.006290436, 0.015563965, -0.029052734, -0.00068569183, -0.046081543, -0.030395508, -0.0013360977, -0.015411377, -...
http://arxiv.org/abs/2305.11497v1
TreePrompt: Learning to Compose Tree Prompts for Explainable Visual Grounding
Chenchi Zhang, Jun Xiao, Lei Chen, Jian Shao, Long Chen
cs.CV, cs.AI, cs.CL, cs.MM
Prompt tuning has achieved great success in transferring the knowledge from large pretrained vision-language models into downstream tasks, and has dominated the performance on visual grounding (VG). However, almost all existing prompt tuning paradigms suffer from poor interpretability. In this paper, we argue that thei...
http://arxiv.org/pdf/2305.11497v1
2023-05-19T07:52:22
2023-05-19T07:52:22
{ "title": [ -0.00020301342, -0.0065727234, -0.03503418, -0.084228516, -0.0124053955, 0.0042762756, -0.06878662, 0.008361816, 0.022949219, 0.0005335808, -0.05432129, 0.0069618225, -0.02003479, -0.013435364, -0.03503418, -0.028137207, -0.009109497, ...
http://arxiv.org/abs/2305.11490v4
LLM-CXR: Instruction-Finetuned LLM for CXR Image Understanding and Generation
Suhyeon Lee, Won Jun Kim, Jinho Chang, Jong Chul Ye
cs.CV, cs.AI, cs.CL, cs.LG
Following the impressive development of LLMs, vision-language alignment in LLMs is actively being researched to enable multimodal reasoning and visual IO. This direction of research is particularly relevant to medical imaging because medical image analysis and generation consist of reasoning based on a combination of v...
http://arxiv.org/pdf/2305.11490v4
2023-10-17T12:16:03
2023-05-19T07:44:39
{ "title": [ -0.0446167, -0.020843506, -0.05090332, -0.027801514, -0.04119873, -0.014526367, -0.06048584, -0.010803223, -0.016693115, 0.005908966, -0.03793335, 0.031021118, -0.013038635, -0.049102783, 0.024551392, -0.011940002, 0.0064964294, 0.02...
http://arxiv.org/abs/2305.11482v1
Enhancing Personalized Dialogue Generation with Contrastive Latent Variables: Combining Sparse and Dense Persona
Yihong Tang, Bo Wang, Miao Fang, Dongming Zhao, Kun Huang, Ruifang He, Yuexian Hou
cs.CL
The personalized dialogue explores the consistent relationship between dialogue generation and personality. Existing personalized dialogue agents model persona profiles from three resources: sparse or dense persona descriptions and dialogue histories. However, sparse structured persona attributes are explicit but uninf...
http://arxiv.org/pdf/2305.11482v1
2023-05-19T07:24:27
2023-05-19T07:24:27
{ "title": [ 0.0016355515, -0.005104065, -0.041107178, -0.07122803, -0.009803772, 0.0032176971, -0.05001831, 0.002691269, 0.015701294, 0.034484863, 0.009460449, 0.028320312, -0.0018262863, -0.028656006, -0.033172607, 0.01449585, -0.02557373, 0.02...
http://arxiv.org/abs/2305.11480v1
CCGen: Explainable Complementary Concept Generation in E-Commerce
Jie Huang, Yifan Gao, Zheng Li, Jingfeng Yang, Yangqiu Song, Chao Zhang, Zining Zhu, Haoming Jiang, Kevin Chen-Chuan Chang, Bing Yin
cs.CL, cs.AI
We propose and study Complementary Concept Generation (CCGen): given a concept of interest, e.g., "Digital Cameras", generating a list of complementary concepts, e.g., 1) Camera Lenses 2) Batteries 3) Camera Cases 4) Memory Cards 5) Battery Chargers. CCGen is beneficial for various applications like query suggestion an...
http://arxiv.org/pdf/2305.11480v1
2023-05-19T07:16:04
2023-05-19T07:16:04
{ "title": [ -0.026504517, -0.017593384, -0.030090332, -0.04220581, -0.013313293, 0.0078125, -0.03253174, -0.007369995, -0.0041542053, 0.031829834, -0.027374268, 0.014083862, 0.013885498, -0.008781433, -0.03173828, -0.0017709732, -0.019729614, 0....
http://arxiv.org/abs/2305.11473v2
Graphologue: Exploring Large Language Model Responses with Interactive Diagrams
Peiling Jiang, Jude Rayan, Steven P. Dow, Haijun Xia
cs.HC, cs.AI, cs.CL
Large language models (LLMs) have recently soared in popularity due to their ease of access and the unprecedented ability to synthesize text responses to diverse user questions. However, LLMs like ChatGPT present significant limitations in supporting complex information tasks due to the insufficient affordances of the ...
http://arxiv.org/pdf/2305.11473v2
2023-08-04T18:23:55
2023-05-19T06:53:25
{ "title": [ -0.022613525, -0.035003662, -0.022354126, -0.055541992, -0.025177002, 0.0022640228, -0.044677734, -0.00881958, 0.0071525574, 0.023330688, -0.0006341934, 0.005420685, -0.02217102, -0.024337769, -0.057647705, 0.00349617, -0.008506775, ...
http://arxiv.org/abs/2305.11462v1
Extending Memory for Language Modelling
Anupiya Nugaliyadde
cs.CL
Breakthroughs in deep learning and memory networks have made major advances in natural language understanding. Language is sequential and information carried through the sequence can be captured through memory networks. Learning the sequence is one of the key aspects in learning the language. However, memory networks a...
http://arxiv.org/pdf/2305.11462v1
2023-05-19T06:30:19
2023-05-19T06:30:19
{ "title": [ -0.018463135, -0.01663208, -0.032836914, -0.085754395, -0.0074386597, 0.014503479, -0.066589355, 0.019195557, 0.018585205, 0.04446411, -0.0073013306, -0.021774292, 0.010238647, -0.015457153, -0.07281494, 0.031173706, -0.010803223, 0....
http://arxiv.org/abs/2305.11460v1
Self-Agreement: A Framework for Fine-tuning Language Models to Find Agreement among Diverse Opinions
Shiyao Ding, Takayuki Ito
cs.CL, cs.AI, cs.MA
Finding an agreement among diverse opinions is a challenging topic in multiagent systems. Recently, large language models (LLMs) have shown great potential in addressing this challenge due to their remarkable capabilities in comprehending human opinions and generating human-like text. However, they typically rely on ex...
http://arxiv.org/pdf/2305.11460v1
2023-05-19T06:27:16
2023-05-19T06:27:16
{ "title": [ -0.017410278, -0.0050582886, -0.05496216, -0.06500244, 0.024627686, -0.008964539, -0.050842285, -0.011192322, 0.0045814514, 0.029327393, -0.0044059753, -0.0020942688, -0.010765076, -0.017791748, -0.06652832, 0.004776001, -0.008460999, ...
http://arxiv.org/abs/2305.11455v1
Shattering the Agent-Environment Interface for Fine-Tuning Inclusive Language Models
Wanqiao Xu, Shi Dong, Dilip Arumugam, Benjamin Van Roy
cs.CL, cs.AI, cs.LG
A centerpiece of the ever-popular reinforcement learning from human feedback (RLHF) approach to fine-tuning autoregressive language models is the explicit training of a reward model to emulate human feedback, distinct from the language model itself. This reward model is then coupled with policy-gradient methods to dram...
http://arxiv.org/pdf/2305.11455v1
2023-05-19T06:21:15
2023-05-19T06:21:15
{ "title": [ -0.027130127, -0.008415222, -0.03463745, -0.066589355, -0.0029201508, 0.0039043427, -0.04333496, 0.016220093, 0.0062789917, 0.036590576, -0.035339355, -0.003194809, 0.019317627, -0.0031604767, -0.054016113, 0.011772156, -0.024734497, ...
http://arxiv.org/abs/2305.11449v1
Analyzing and Reducing the Performance Gap in Cross-Lingual Transfer with Fine-tuning Slow and Fast
Yiduo Guo, Yaobo Liang, Dongyan Zhao, Bing Liu, Duan Nan
cs.CL
Existing research has shown that a multilingual pre-trained language model fine-tuned with one (source) language also performs well on downstream tasks for non-source languages, even though no fine-tuning is done on these languages. However, there is a clear gap between the performance of the source language and that o...
http://arxiv.org/pdf/2305.11449v1
2023-05-19T06:04:21
2023-05-19T06:04:21
{ "title": [ -0.008628845, 0.012939453, -0.030288696, -0.034240723, 0.012672424, -0.024353027, 0.00022184849, -0.02128601, -0.01222229, 0.04611206, -0.015258789, 0.016159058, 0.028244019, -0.022979736, -0.011497498, 0.010612488, 0.0036945343, 0.0...
http://arxiv.org/abs/2305.11444v1
Arukikata Travelogue Dataset
Hiroki Ouchi, Hiroyuki Shindo, Shoko Wakamiya, Yuki Matsuda, Naoya Inoue, Shohei Higashiyama, Satoshi Nakamura, Taro Watanabe
cs.CL, cs.AI, cs.DL
We have constructed Arukikata Travelogue Dataset and released it free of charge for academic research. This dataset is a Japanese text dataset with a total of over 31 million words, comprising 4,672 Japanese domestic travelogues and 9,607 overseas travelogues. Before providing our dataset, there was a scarcity of widel...
http://arxiv.org/pdf/2305.11444v1
2023-05-19T05:53:49
2023-05-19T05:53:49
{ "title": [ 0.014099121, -0.01272583, -0.012145996, -0.021835327, -0.03967285, -0.00856781, -0.044067383, -0.007858276, -0.000828743, 0.04940796, -0.046691895, -0.028137207, 0.016708374, -0.025375366, -0.018875122, -0.038238525, 0.037963867, -0....
http://arxiv.org/abs/2305.11442v2
Zero-Shot Text Classification via Self-Supervised Tuning
Chaoqun Liu, Wenxuan Zhang, Guizhen Chen, Xiaobao Wu, Anh Tuan Luu, Chip Hong Chang, Lidong Bing
cs.CL, cs.AI, cs.LG
Existing solutions to zero-shot text classification either conduct prompting with pre-trained language models, which is sensitive to the choices of templates, or rely on large-scale annotated data of relevant tasks for meta-tuning. In this work, we propose a new paradigm based on self-supervised learning to solve zero-...
http://arxiv.org/pdf/2305.11442v2
2023-05-25T06:10:04
2023-05-19T05:47:33
{ "title": [ -0.015350342, 0.0039634705, -0.02394104, -0.04034424, -0.011062622, -0.011207581, -0.016998291, -0.0034999847, 0.0058288574, 0.037719727, -0.014862061, 0.016403198, -0.03652954, -0.0017137527, -0.061767578, -0.0053749084, -0.0030269623, ...
http://arxiv.org/abs/2305.11438v1
Phonetic and Prosody-aware Self-supervised Learning Approach for Non-native Fluency Scoring
Kaiqi Fu, Shaojun Gao, Shuju Shi, Xiaohai Tian, Wei Li, Zejun Ma
cs.CL, eess.AS
Speech fluency/disfluency can be evaluated by analyzing a range of phonetic and prosodic features. Deep neural networks are commonly trained to map fluency-related features into the human scores. However, the effectiveness of deep learning-based models is constrained by the limited amount of labeled training samples. T...
http://arxiv.org/pdf/2305.11438v1
2023-05-19T05:39:41
2023-05-19T05:39:41
{ "title": [ 0.000058412552, 0.015052795, -0.041107178, -0.05419922, 0.0016498566, -0.008041382, -0.02494812, 0.02458191, 0.0041160583, 0.010101318, -0.00034308434, -0.030838013, -0.014549255, -0.05581665, -0.029800415, -0.028320312, -0.006290436, ...
http://arxiv.org/abs/2305.11435v2
Syllable Discovery and Cross-Lingual Generalization in a Visually Grounded, Self-Supervised Speech Model
Puyuan Peng, Shang-Wen Li, Okko Räsänen, Abdelrahman Mohamed, David Harwath
eess.AS, cs.AI, cs.CL, cs.SD
In this paper, we show that representations capturing syllabic units emerge when training a self-supervised speech model with a visually-grounded training objective. We demonstrate that a nearly identical model architecture (HuBERT) trained with a masked language modeling loss does not exhibit this same ability, sugges...
http://arxiv.org/pdf/2305.11435v2
2023-07-23T05:32:05
2023-05-19T05:19:04
{ "title": [ -0.012298584, 0.0102005005, -0.028549194, -0.053222656, -0.021697998, -0.010902405, -0.0579834, 0.008041382, -0.0003399849, 0.04840088, -0.034454346, 0.014297485, 0.006340027, -0.0039596558, -0.057739258, -0.001666069, -0.003118515, ...
http://arxiv.org/abs/2305.11430v2
TELeR: A General Taxonomy of LLM Prompts for Benchmarking Complex Tasks
Shubhra Kanti Karmaker Santu, Dongji Feng
cs.AI, cs.CL, cs.IR, cs.LG, I.2.7
While LLMs have shown great success in understanding and generating text in traditional conversational settings, their potential for performing ill-defined complex tasks is largely under-studied. Indeed, we are yet to conduct comprehensive benchmarking studies with multiple LLMs that are exclusively focused on a comple...
http://arxiv.org/pdf/2305.11430v2
2023-10-24T22:50:02
2023-05-19T04:59:34
{ "title": [ -0.06237793, -0.007904053, -0.050079346, -0.014892578, -0.028442383, -0.0045051575, -0.07312012, -0.018569946, -0.009994507, 0.015838623, -0.012588501, 0.020385742, 0.0016269684, -0.045288086, 0.009010315, 0.0030918121, -0.0014533997, ...
http://arxiv.org/abs/2305.11426v2
Post Hoc Explanations of Language Models Can Improve Language Models
Satyapriya Krishna, Jiaqi Ma, Dylan Slack, Asma Ghandeharioun, Sameer Singh, Himabindu Lakkaraju
cs.CL, cs.AI
Large Language Models (LLMs) have demonstrated remarkable capabilities in performing complex tasks. Moreover, recent research has shown that incorporating human-annotated rationales (e.g., Chain-of-Thought prompting) during in-context learning can significantly enhance the performance of these models, particularly on t...
http://arxiv.org/pdf/2305.11426v2
2023-05-28T14:30:02
2023-05-19T04:46:04
{ "title": [ 0.03314209, -0.022384644, 0.013809204, -0.086242676, -0.050201416, 0.0003285408, -0.07849121, 0.0017557144, -0.027816772, 0.016342163, -0.0064888, -0.011642456, 0.017547607, -0.052856445, 0.029525757, 0.001625061, 0.0022087097, 0.028...
http://arxiv.org/abs/2305.11411v1
DUB: Discrete Unit Back-translation for Speech Translation
Dong Zhang, Rong Ye, Tom Ko, Mingxuan Wang, Yaqian Zhou
cs.CL, cs.SD, eess.AS
How can speech-to-text translation (ST) perform as well as machine translation (MT)? The key point is to bridge the modality gap between speech and text so that useful MT techniques can be applied to ST. Recently, the approach of representing speech with unsupervised discrete units yields a new way to ease the modality...
http://arxiv.org/pdf/2305.11411v1
2023-05-19T03:48:16
2023-05-19T03:48:16
{ "title": [ -0.034088135, -0.016326904, -0.03161621, -0.0635376, -0.010940552, 0.009468079, -0.057525635, 0.005874634, -0.018295288, 0.058410645, 0.006099701, 0.005607605, 0.06341553, -0.043945312, -0.022842407, 0.030456543, -0.005657196, 0.0288...
http://arxiv.org/abs/2305.11408v2
AlignAtt: Using Attention-based Audio-Translation Alignments as a Guide for Simultaneous Speech Translation
Sara Papi, Marco Turchi, Matteo Negri
cs.CL, cs.LG, cs.SD, eess.AS
Attention is the core mechanism of today's most used architectures for natural language processing and has been analyzed from many perspectives, including its effectiveness for machine translation-related tasks. Among these studies, attention resulted to be a useful source of information to get insights about word alig...
http://arxiv.org/pdf/2305.11408v2
2023-07-20T00:58:30
2023-05-19T03:31:42
{ "title": [ -0.011116028, -0.001376152, -0.034484863, -0.056671143, -0.0022468567, 0.0076904297, -0.048034668, 0.005115509, -0.009887695, 0.07080078, -0.006084442, 0.013847351, 0.052764893, -0.014472961, -0.044891357, 0.025787354, 0.0011091232, ...
http://arxiv.org/abs/2305.11398v1
Comfort Foods and Community Connectedness: Investigating Diet Change during COVID-19 Using YouTube Videos on Twitter
Yelena Mejova, Lydia Manikonda
cs.SI, cs.CL
Unprecedented lockdowns at the start of the COVID-19 pandemic have drastically changed the routines of millions of people, potentially impacting important health-related behaviors. In this study, we use YouTube videos embedded in tweets about diet, exercise and fitness posted before and during COVID-19 to investigate t...
http://arxiv.org/pdf/2305.11398v1
2023-05-19T02:51:25
2023-05-19T02:51:25
{ "title": [ 0.029006958, 0.0056915283, -0.035583496, -0.055908203, 0.021530151, -0.027328491, 0.007434845, 0.01977539, -0.02748108, 0.037017822, -0.015838623, 0.018112183, 0.037353516, -0.023452759, 0.010322571, -0.020950317, 0.006111145, 0.0219...
http://arxiv.org/abs/2305.11392v1
Fast-StrucTexT: An Efficient Hourglass Transformer with Modality-guided Dynamic Token Merge for Document Understanding
Mingliang Zhai, Yulin Li, Xiameng Qin, Chen Yi, Qunyi Xie, Chengquan Zhang, Kun Yao, Yuwei Wu, Yunde Jia
cs.CV, cs.CL
Transformers achieve promising performance in document understanding because of their high effectiveness and still suffer from quadratic computational complexity dependency on the sequence length. General efficient transformers are challenging to be directly adapted to model document. They are unable to handle the layo...
http://arxiv.org/pdf/2305.11392v1
2023-05-19T02:42:35
2023-05-19T02:42:35
{ "title": [ -0.020233154, -0.000834465, -0.061920166, -0.080322266, 0.009346008, 0.006969452, -0.054840088, 0.006248474, -0.0014657974, 0.026535034, -0.032287598, 0.005592346, 0.046661377, -0.0046806335, -0.022781372, -0.005645752, 0.017089844, ...
http://arxiv.org/abs/2306.05552v1
ChatGPT for Us: Preserving Data Privacy in ChatGPT via Dialogue Text Ambiguation to Expand Mental Health Care Delivery
Anaelia Ovalle, Mehrab Beikzadeh, Parshan Teimouri, Kai-Wei Chang, Majid Sarrafzadeh
cs.CL, I.2; I.7
Large language models have been useful in expanding mental health care delivery. ChatGPT, in particular, has gained popularity for its ability to generate human-like dialogue. However, data-sensitive domains -- including but not limited to healthcare -- face challenges in using ChatGPT due to privacy and data-ownership...
http://arxiv.org/pdf/2306.05552v1
2023-05-19T02:09:52
2023-05-19T02:09:52
{ "title": [ 0.0018825531, -0.01939392, -0.03265381, -0.024032593, -0.019607544, -0.014785767, -0.030532837, -0.0061950684, -0.037994385, 0.039276123, -0.0109939575, 0.014205933, 0.012969971, -0.04071045, -0.0012397766, -0.014785767, -0.024017334, ...
http://arxiv.org/abs/2305.13330v2
Unsupervised ASR via Cross-Lingual Pseudo-Labeling
Tatiana Likhomanenko, Loren Lugosch, Ronan Collobert
eess.AS, cs.CL, cs.LG, cs.SD
Recent work has shown that it is possible to train an $\textit{unsupervised}$ automatic speech recognition (ASR) system using only unpaired audio and text. Existing unsupervised ASR methods assume that no labeled data can be used for training. We argue that even if one does not have any labeled audio for a given langua...
http://arxiv.org/pdf/2305.13330v2
2023-09-30T02:16:31
2023-05-19T01:59:20
{ "title": [ -0.01626587, 0.00592041, -0.04989624, -0.08294678, 0.00415802, 0.0031204224, -0.06222534, 0.013458252, -0.017578125, 0.036590576, -0.00089263916, -0.01939392, 0.024368286, -0.04345703, -0.0061912537, 0.014175415, 0.010765076, 0.04913...
http://arxiv.org/abs/2305.11374v1
Characterizing tradeoffs between teaching via language and demonstrations in multi-agent systems
Dhara Yu, Noah D. Goodman, Jesse Mu
cs.CL
Humans teach others about the world through language and demonstration. When might one of these modalities be more effective than the other? In this work, we study the factors that modulate the effectiveness of language vs. demonstration using multi-agent systems to model human communication. Specifically, we train neu...
http://arxiv.org/pdf/2305.11374v1
2023-05-19T01:27:29
2023-05-19T01:27:29
{ "title": [ -0.013458252, -0.002412796, -0.013504028, -0.021621704, 0.010070801, -0.0124053955, -0.01876831, -0.030471802, -0.0085372925, 0.017318726, -0.023513794, 0.014434814, -0.021621704, -0.008216858, -0.011810303, -0.01713562, -0.015792847, ...
http://arxiv.org/abs/2305.11366v2
AutoTrial: Prompting Language Models for Clinical Trial Design
Zifeng Wang, Cao Xiao, Jimeng Sun
cs.CL
Clinical trials are critical for drug development. Constructing the appropriate eligibility criteria (i.e., the inclusion/exclusion criteria for patient recruitment) is essential for the trial's success. Proper design of clinical trial protocols should consider similar precedent trials and their eligibility criteria to...
http://arxiv.org/pdf/2305.11366v2
2023-10-08T03:44:35
2023-05-19T01:04:16
{ "title": [ -0.04888916, -0.0065612793, -0.050476074, -0.033996582, -0.01751709, 0.012168884, -0.08190918, -0.025024414, -0.02015686, 0.034454346, -0.015205383, 0.028793335, -0.0018405914, -0.012756348, -0.027404785, 0.009727478, -0.003660202, 0...
http://arxiv.org/abs/2305.11364v2
Visualizing Linguistic Diversity of Text Datasets Synthesized by Large Language Models
Emily Reif, Minsuk Kahng, Savvas Petridis
cs.CL, cs.AI
Large language models (LLMs) can be used to generate smaller, more refined datasets via few-shot prompting for benchmarking, fine-tuning or other use cases. However, understanding and evaluating these datasets is difficult, and the failure modes of LLM-generated data are still not well understood. Specifically, the dat...
http://arxiv.org/pdf/2305.11364v2
2023-09-27T22:08:13
2023-05-19T00:53:45
{ "title": [ -0.021118164, 0.010871887, -0.04434204, -0.041870117, 0.010826111, -0.008918762, -0.040771484, -0.01525116, 0.001953125, 0.047790527, -0.022232056, 0.0068855286, -0.008392334, 0.0023880005, -0.070739746, -0.0045814514, -0.009147644, ...
http://arxiv.org/abs/2305.11355v1
MD3: The Multi-Dialect Dataset of Dialogues
Jacob Eisenstein, Vinodkumar Prabhakaran, Clara Rivera, Dorottya Demszky, Devyani Sharma
cs.CL
We introduce a new dataset of conversational speech representing English from India, Nigeria, and the United States. The Multi-Dialect Dataset of Dialogues (MD3) strikes a new balance between open-ended conversational speech and task-oriented dialogue by prompting participants to perform a series of short information-s...
http://arxiv.org/pdf/2305.11355v1
2023-05-19T00:14:10
2023-05-19T00:14:10
{ "title": [ -0.024276733, -0.0132369995, -0.056396484, -0.0927124, 0.023712158, 0.0057525635, -0.068603516, 0.025543213, 0.018844604, 0.035949707, 0.0030555725, -0.0069847107, 0.003868103, -0.024841309, -0.056549072, 0.021911621, -0.0019397736, ...
http://arxiv.org/abs/2305.11351v1
Data Redaction from Conditional Generative Models
Zhifeng Kong, Kamalika Chaudhuri
cs.LG, cs.CL, cs.CV
Deep generative models are known to produce undesirable samples such as harmful content. Traditional mitigation methods include re-training from scratch, filtering, or editing; however, these are either computationally expensive or can be circumvented by third parties. In this paper, we take a different approach and st...
http://arxiv.org/pdf/2305.11351v1
2023-05-18T23:58:53
2023-05-18T23:58:53
{ "title": [ -0.03213501, 0.008125305, -0.06555176, -0.08605957, 0.010955811, 0.014282227, -0.08306885, -0.006477356, 0.009918213, 0.018066406, -0.01398468, -0.0051879883, -0.018081665, -0.03186035, -0.033172607, 0.006767273, 0.001660347, 0.02114...
http://arxiv.org/abs/2305.11349v1
Unsupervised Domain-agnostic Fake News Detection using Multi-modal Weak Signals
Amila Silva, Ling Luo, Shanika Karunasekera, Christopher Leckie
cs.LG, cs.CL
The emergence of social media as one of the main platforms for people to access news has enabled the wide dissemination of fake news. This has motivated numerous studies on automating fake news detection. Although there have been limited attempts at unsupervised fake news detection, their performance suffers due to not...
http://arxiv.org/pdf/2305.11349v1
2023-05-18T23:49:31
2023-05-18T23:49:31
{ "title": [ -0.0014219284, 0.032287598, -0.023269653, -0.01084137, 0.021835327, -0.004180908, -0.009887695, -0.059661865, -0.032073975, 0.0395813, -0.0039253235, 0.014343262, 0.026290894, -0.025802612, -0.013191223, -0.0046577454, -0.03277588, 0...
http://arxiv.org/abs/2305.11348v1
In the Name of Fairness: Assessing the Bias in Clinical Record De-identification
Yuxin Xiao, Shulammite Lim, Tom Joseph Pollard, Marzyeh Ghassemi
cs.LG, cs.CL, cs.CR, cs.CY
Data sharing is crucial for open science and reproducible research, but the legal sharing of clinical data requires the removal of protected health information from electronic health records. This process, known as de-identification, is often achieved through the use of machine learning algorithms by many commercial an...
http://arxiv.org/pdf/2305.11348v1
2023-05-18T23:47:00
2023-05-18T23:47:00
{ "title": [ 0.021011353, 0.025924683, -0.05038452, -0.06921387, 0.038391113, -0.017532349, -0.028579712, -0.022476196, -0.02456665, 0.020187378, -0.027404785, 0.0061836243, -0.009468079, -0.023345947, -0.0151901245, -0.021316528, -0.02079773, 0....
http://arxiv.org/abs/2305.11334v1
Writing your own book: A method for going from closed to open book QA to improve robustness and performance of smaller LLMs
Giorgi Kokaia, Pratyush Sinha, Yutong Jiang, Nozha Boujemaa
cs.CL, cs.AI
We introduce two novel methods, Tree-Search and Self-contextualizing QA, designed to enhance the performance of large language models (LLMs) in question-answering tasks. Tree-Search is a sampling technique specifically created to extract diverse information from an LLM for a given prompt. Self-contextualizing QA levera...
http://arxiv.org/pdf/2305.11334v1
2023-05-18T22:47:06
2023-05-18T22:47:06
{ "title": [ -0.06768799, -0.008125305, -0.028961182, 0.02154541, -0.059265137, 0.0077590942, -0.07946777, -0.025482178, -0.037506104, -0.016189575, -0.010169983, 0.050811768, -0.070373535, -0.040527344, 0.037506104, -0.06008911, 0.032043457, -0....
http://arxiv.org/abs/2305.11326v2
Towards the Automatic Generation of Conversational Interfaces to Facilitate the Exploration of Tabular Data
Marcos Gomez, Jordi Cabot, Robert Clarisó
cs.CL
Tabular data is the most common format to publish and exchange structured data online. A clear example is the growing number of open data portals published by all types of public administrations. However, exploitation of these data sources is currently limited to technical people able to programmatically manipulate and...
http://arxiv.org/pdf/2305.11326v2
2023-05-24T11:55:16
2023-05-18T22:23:40
{ "title": [ -0.01335907, -0.013259888, -0.04611206, -0.03982544, -0.012390137, -0.017242432, -0.046142578, 0.0037002563, 0.020080566, 0.04159546, -0.029144287, -0.009208679, 0.0032081604, -0.011672974, -0.06750488, -0.017242432, -0.019943237, 0....
http://arxiv.org/abs/2305.11317v2
Collaborative Generative AI: Integrating GPT-k for Efficient Editing in Text-to-Image Generation
Wanrong Zhu, Xinyi Wang, Yujie Lu, Tsu-Jui Fu, Xin Eric Wang, Miguel Eckstein, William Yang Wang
cs.CL
The field of text-to-image (T2I) generation has garnered significant attention both within the research community and among everyday users. Despite the advancements of T2I models, a common issue encountered by users is the need for repetitive editing of input prompts in order to receive a satisfactory image, which is t...
http://arxiv.org/pdf/2305.11317v2
2023-10-28T04:13:44
2023-05-18T21:53:58
{ "title": [ -0.031951904, 0.00021827221, -0.027130127, -0.027496338, -0.015853882, 0.0056495667, -0.035858154, -0.02507019, 0.019165039, 0.035583496, -0.015808105, 0.017929077, -0.0009832382, -0.015106201, -0.043029785, 0.008125305, -0.011512756, ...
http://arxiv.org/abs/2305.11315v1
Improving Toponym Resolution with Better Candidate Generation, Transformer-based Reranking, and Two-Stage Resolution
Zeyu Zhang, Steven Bethard
cs.CL
Geocoding is the task of converting location mentions in text into structured data that encodes the geospatial semantics. We propose a new architecture for geocoding, GeoNorm. GeoNorm first uses information retrieval techniques to generate a list of candidate entries from the geospatial ontology. Then it reranks the ca...
http://arxiv.org/pdf/2305.11315v1
2023-05-18T21:52:48
2023-05-18T21:52:48
{ "title": [ 0.012649536, 0.008102417, 0.008163452, -0.026687622, -0.01133728, -0.0065193176, -0.02368164, -0.049316406, -0.011360168, 0.029144287, -0.011489868, 0.026123047, 0.029373169, -0.028442383, 0.014724731, 0.0158844, -0.002664566, 0.0075...
http://arxiv.org/abs/2305.11271v1
Towards Collaborative Plan Acquisition through Theory of Mind Modeling in Situated Dialogue
Cristian-Paul Bara, Ziqiao Ma, Yingzhuo Yu, Julie Shah, Joyce Chai
cs.AI, cs.CL, cs.CV, cs.LG
Collaborative tasks often begin with partial task knowledge and incomplete initial plans from each partner. To complete these tasks, agents need to engage in situated communication with their partners and coordinate their partial plans towards a complete plan to achieve a joint task goal. While such collaboration seems...
http://arxiv.org/pdf/2305.11271v1
2023-05-18T19:42:04
2023-05-18T19:42:04
{ "title": [ -0.0132369995, -0.013763428, -0.02645874, -0.03717041, -0.011299133, 0.00035119057, -0.014953613, 0.01071167, -0.03967285, 0.01361084, -0.0049934387, 0.0036010742, -0.032165527, -0.028442383, -0.03677368, -0.029815674, -0.01914978, 0...
http://arxiv.org/abs/2305.11262v1
CHBias: Bias Evaluation and Mitigation of Chinese Conversational Language Models
Jiaxu Zhao, Meng Fang, Zijing Shi, Yitong Li, Ling Chen, Mykola Pechenizkiy
cs.CL
\textit{\textbf{\textcolor{red}{Warning}:} This paper contains content that may be offensive or upsetting.} Pretrained conversational agents have been exposed to safety issues, exhibiting a range of stereotypical human biases such as gender bias. However, there are still limited bias categories in current research, and...
http://arxiv.org/pdf/2305.11262v1
2023-05-18T18:58:30
2023-05-18T18:58:30
{ "title": [ -0.017242432, -0.00082969666, -0.031021118, -0.035491943, -0.006465912, -0.0062026978, -0.05078125, -0.018600464, 0.008598328, 0.036590576, -0.016296387, -0.009803772, 0.03768921, -0.028381348, -0.04171753, 0.0105896, -0.03527832, 0....
http://arxiv.org/abs/2305.11255v4
Reasoning Implicit Sentiment with Chain-of-Thought Prompting
Hao Fei, Bobo Li, Qian Liu, Lidong Bing, Fei Li, Tat-Seng Chua
cs.CL
While sentiment analysis systems try to determine the sentiment polarities of given targets based on the key opinion expressions in input texts, in implicit sentiment analysis (ISA) the opinion cues come in an implicit and obscure manner. Thus detecting implicit sentiment requires the common-sense and multi-hop reasoni...
http://arxiv.org/pdf/2305.11255v4
2023-06-09T01:27:58
2023-05-18T18:38:32
{ "title": [ 0.017623901, -0.008857727, 0.010139465, -0.062194824, -0.0052719116, 0.0087509155, -0.020233154, -0.013786316, 0.032073975, 0.0036621094, -0.060150146, -0.0154800415, -0.022277832, -0.012702942, -0.041900635, -0.026641846, -0.0051078796,...
http://arxiv.org/abs/2305.11251v1
Computational thematics: Comparing algorithms for clustering the genres of literary fiction
Oleg Sobchuk, Artjoms Šeļa
cs.CL
What are the best methods of capturing thematic similarity between literary texts? Knowing the answer to this question would be useful for automatic clustering of book genres, or any other thematic grouping. This paper compares a variety of algorithms for unsupervised learning of thematic similarities between texts, wh...
http://arxiv.org/pdf/2305.11251v1
2023-05-18T18:32:03
2023-05-18T18:32:03
{ "title": [ -0.025146484, 0.010345459, -0.030090332, -0.0047912598, 0.010772705, -0.008590698, 0.010101318, -0.02420044, -0.013938904, 0.048828125, 0.0050811768, 0.01449585, 0.020553589, -0.024642944, -0.029220581, 0.016906738, -0.020950317, 0.0...
http://arxiv.org/abs/2305.11244v2
A Parameter-Efficient Learning Approach to Arabic Dialect Identification with Pre-Trained General-Purpose Speech Model
Srijith Radhakrishnan, Chao-Han Huck Yang, Sumeer Ahmad Khan, Narsis A. Kiani, David Gomez-Cabrero, Jesper N. Tegner
cs.CL, cs.AI, cs.LG, cs.NE, eess.AS
In this work, we explore Parameter-Efficient-Learning (PEL) techniques to repurpose a General-Purpose-Speech (GSM) model for Arabic dialect identification (ADI). Specifically, we investigate different setups to incorporate trainable features into a multi-layer encoder-decoder GSM formulation under frozen pre-trained se...
http://arxiv.org/pdf/2305.11244v2
2023-10-03T19:45:25
2023-05-18T18:15:53
{ "title": [ 0.00819397, 0.0072135925, -0.031051636, -0.07989502, 0.0066184998, -0.00017035007, -0.05239868, 0.009933472, -0.02571106, 0.032592773, 0.025924683, -0.01953125, -0.010284424, -0.057159424, -0.038635254, -0.00034499168, 0.018112183, 0...
http://arxiv.org/abs/2305.11243v2
Comparing Machines and Children: Using Developmental Psychology Experiments to Assess the Strengths and Weaknesses of LaMDA Responses
Eliza Kosoy, Emily Rose Reagan, Leslie Lai, Alison Gopnik, Danielle Krettek Cobb
cs.CL, cs.AI
Developmental psychologists have spent decades devising experiments to test the intelligence and knowledge of infants and children, tracing the origin of crucial concepts and capacities. Moreover, experimental techniques in developmental psychology have been carefully designed to discriminate the cognitive capacities t...
http://arxiv.org/pdf/2305.11243v2
2023-11-07T20:20:47
2023-05-18T18:15:43
{ "title": [ -0.005657196, -0.008766174, -0.0014858246, 0.03250122, -0.01776123, -0.029907227, 0.035736084, -0.03552246, -0.046142578, 0.014694214, -0.02822876, 0.01146698, -0.036743164, -0.020812988, -0.022903442, -0.06185913, -0.0025157928, 0.0...
http://arxiv.org/abs/2305.11242v1
Comparing Biases and the Impact of Multilingual Training across Multiple Languages
Sharon Levy, Neha Anna John, Ling Liu, Yogarshi Vyas, Jie Ma, Yoshinari Fujinuma, Miguel Ballesteros, Vittorio Castelli, Dan Roth
cs.CL
Studies in bias and fairness in natural language processing have primarily examined social biases within a single language and/or across few attributes (e.g. gender, race). However, biases can manifest differently across various languages for individual attributes. As a result, it is critical to examine biases within e...
http://arxiv.org/pdf/2305.11242v1
2023-05-18T18:15:07
2023-05-18T18:15:07
{ "title": [ -0.009963989, -0.0020999908, -0.044036865, -0.027770996, -0.028396606, -0.035614014, -0.032928467, 0.0018386841, -0.0519104, 0.040618896, -0.0015325546, 0.0135269165, -0.0018148422, -0.043548584, 0.018844604, -0.015617371, -0.020507812, ...
http://arxiv.org/abs/2305.11231v1
Recent Trends in Unsupervised Summarization
Mohammad Khosravani, Amine Trabelsi
cs.CL
Unsupervised summarization is a powerful technique that enables training summarizing models without requiring labeled datasets. This survey covers different recent techniques and models used for unsupervised summarization. We cover extractive, abstractive, and hybrid models and strategies used to achieve unsupervised s...
http://arxiv.org/pdf/2305.11231v1
2023-05-18T18:00:44
2023-05-18T18:00:44
{ "title": [ -0.0059661865, 0.010986328, -0.038360596, -0.080078125, 0.0121536255, 0.0067481995, -0.05319214, -0.013053894, 0.0062789917, 0.041931152, -0.016906738, 0.024978638, -0.009803772, 0.0035514832, -0.04574585, 0.018615723, 0.003124237, 0...
http://arxiv.org/abs/2305.11172v1
ONE-PEACE: Exploring One General Representation Model Toward Unlimited Modalities
Peng Wang, Shijie Wang, Junyang Lin, Shuai Bai, Xiaohuan Zhou, Jingren Zhou, Xinggang Wang, Chang Zhou
cs.CV, cs.CL, cs.SD, eess.AS
In this work, we explore a scalable way for building a general representation model toward unlimited modalities. We release ONE-PEACE, a highly extensible model with 4B parameters that can seamlessly align and integrate representations across vision, audio, and language modalities. The architecture of ONE-PEACE compris...
http://arxiv.org/pdf/2305.11172v1
2023-05-18T17:59:06
2023-05-18T17:59:06
{ "title": [ 0.020568848, 0.011405945, -0.0368042, -0.084472656, 0.008857727, -0.0048599243, -0.044433594, -0.02418518, 0.0070533752, 0.033172607, 0.0047073364, -0.015594482, 0.024856567, -0.051879883, -0.036895752, 0.01689148, -0.00258255, 0.004...
http://arxiv.org/abs/2305.11171v3
TrueTeacher: Learning Factual Consistency Evaluation with Large Language Models
Zorik Gekhman, Jonathan Herzig, Roee Aharoni, Chen Elkind, Idan Szpektor
cs.CL
Factual consistency evaluation is often conducted using Natural Language Inference (NLI) models, yet these models exhibit limited success in evaluating summaries. Previous work improved such models with synthetic training data. However, the data is typically based on perturbed human-written summaries, which often diffe...
http://arxiv.org/pdf/2305.11171v3
2023-10-18T19:16:18
2023-05-18T17:58:35
{ "title": [ -0.028762817, -0.0070724487, -0.050811768, -0.06262207, 0.009590149, -0.0046691895, -0.0440979, 0.014282227, -0.0038528442, 0.03164673, -0.010894775, -0.01096344, 0.0093688965, -0.010696411, -0.05911255, 0.0008215904, -0.027450562, 0...
http://arxiv.org/abs/2305.11170v1
Efficient Prompting via Dynamic In-Context Learning
Wangchunshu Zhou, Yuchen Eleanor Jiang, Ryan Cotterell, Mrinmaya Sachan
cs.CL, cs.AI, cs.LG
The primary way of building AI applications is shifting from training specialist models to prompting generalist models. A common practice for prompting generalist models, often referred to as in-context learning, is to append a few examples (demonstrations) to the prompt to help the model better understand the task. Wh...
http://arxiv.org/pdf/2305.11170v1
2023-05-18T17:58:31
2023-05-18T17:58:31
{ "title": [ 0.030258179, -0.0014734268, -0.013015747, -0.072631836, 0.012886047, -0.0074806213, -0.016662598, -0.021347046, -0.007484436, 0.00894165, -0.03955078, 0.013137817, -0.031234741, -0.015281677, -0.013748169, -0.021591187, 0.015449524, ...
http://arxiv.org/abs/2305.11169v2
Evidence of Meaning in Language Models Trained on Programs
Charles Jin, Martin Rinard
cs.LG, cs.AI, cs.CL, cs.PL
We present evidence that language models can learn meaning despite being trained only to perform next token prediction on text, specifically a corpus of programs. Each program is preceded by a specification in the form of (textual) input-output examples. Working with programs enables us to precisely define concepts rel...
http://arxiv.org/pdf/2305.11169v2
2023-05-24T11:52:13
2023-05-18T17:58:08
{ "title": [ -0.010543823, -0.0033168793, -0.017684937, -0.013633728, -0.07293701, -0.021484375, -0.057495117, -0.021057129, -0.021896362, 0.011016846, 0.009757996, -0.016906738, 0.010002136, -0.07733154, 0.022567749, -0.004802704, -0.0013275146, ...
http://arxiv.org/abs/2305.11159v1
Aligning Instruction Tasks Unlocks Large Language Models as Zero-Shot Relation Extractors
Kai Zhang, Bernal Jiménez Gutiérrez, Yu Su
cs.CL
Recent work has shown that fine-tuning large language models (LLMs) on large-scale instruction-following datasets substantially improves their performance on a wide range of NLP tasks, especially in the zero-shot setting. However, even advanced instruction-tuned LLMs still fail to outperform small LMs on relation extra...
http://arxiv.org/pdf/2305.11159v1
2023-05-18T17:48:03
2023-05-18T17:48:03
{ "title": [ -0.01927185, -0.011543274, -0.04940796, -0.06512451, -0.017349243, -0.0005016327, -0.07318115, 0.011917114, 0.024658203, 0.021697998, -0.007904053, 0.009094238, -0.012458801, -0.009086609, -0.042816162, 0.017669678, 0.0014572144, 0.0...
http://arxiv.org/abs/2305.11206v1
LIMA: Less Is More for Alignment
Chunting Zhou, Pengfei Liu, Puxin Xu, Srini Iyer, Jiao Sun, Yuning Mao, Xuezhe Ma, Avia Efrat, Ping Yu, Lili Yu, Susan Zhang, Gargi Ghosh, Mike Lewis, Luke Zettlemoyer, Omer Levy
cs.CL, cs.AI, cs.LG
Large language models are trained in two stages: (1) unsupervised pretraining from raw text, to learn general-purpose representations, and (2) large scale instruction tuning and reinforcement learning, to better align to end tasks and user preferences. We measure the relative importance of these two stages by training ...
http://arxiv.org/pdf/2305.11206v1
2023-05-18T17:45:22
2023-05-18T17:45:22
{ "title": [ -0.016021729, -0.04309082, -0.05206299, -0.014091492, -0.049560547, -0.012382507, -0.024276733, 0.014526367, -0.0033493042, 0.022125244, -0.024230957, 0.0023708344, -0.017669678, -0.024459839, -0.013465881, -0.03024292, 0.0211792, 0....
http://arxiv.org/abs/2305.11142v1
Discourse Centric Evaluation of Machine Translation with a Densely Annotated Parallel Corpus
Yuchen Eleanor Jiang, Tianyu Liu, Shuming Ma, Dongdong Zhang, Mrinmaya Sachan, Ryan Cotterell
cs.CL
Several recent papers claim human parity at sentence-level Machine Translation (MT), especially in high-resource languages. Thus, in response, the MT community has, in part, shifted its focus to document-level translation. Translating documents requires a deeper understanding of the structure and meaning of text, which...
http://arxiv.org/pdf/2305.11142v1
2023-05-18T17:36:41
2023-05-18T17:36:41
{ "title": [ -0.04083252, 0.0026760101, -0.010635376, -0.04083252, -0.020019531, 0.014579773, -0.040771484, 0.006290436, 0.0007634163, 0.049591064, 0.00013816357, 0.0043525696, 0.016799927, -0.018066406, -0.06323242, 0.013023376, -0.018295288, 0....
http://arxiv.org/abs/2305.11140v1
Exploiting Biased Models to De-bias Text: A Gender-Fair Rewriting Model
Chantal Amrhein, Florian Schottmann, Rico Sennrich, Samuel Läubli
cs.CL, I.2.7
Natural language generation models reproduce and often amplify the biases present in their training data. Previous research explored using sequence-to-sequence rewriting models to transform biased model outputs (or original texts) into more gender-fair language by creating pseudo training data through linguistic rules....
http://arxiv.org/pdf/2305.11140v1
2023-05-18T17:35:28
2023-05-18T17:35:28
{ "title": [ -0.04525757, -0.0016384125, -0.05279541, -0.057800293, -0.0066337585, 0.00459671, -0.08026123, -0.011123657, 0.015007019, 0.040374756, -0.026184082, -0.0062446594, 0.038879395, -0.00046014786, -0.057373047, -0.009986877, -0.047180176, ...
http://arxiv.org/abs/2305.11129v2
mLongT5: A Multilingual and Efficient Text-To-Text Transformer for Longer Sequences
David Uthus, Santiago Ontañón, Joshua Ainslie, Mandy Guo
cs.CL
We present our work on developing a multilingual, efficient text-to-text transformer that is suitable for handling long inputs. This model, called mLongT5, builds upon the architecture of LongT5, while leveraging the multilingual datasets used for pretraining mT5 and the pretraining tasks of UL2. We evaluate this model...
http://arxiv.org/pdf/2305.11129v2
2023-10-26T22:43:26
2023-05-18T17:22:53
{ "title": [ -0.012123108, -0.008293152, -0.030014038, -0.07952881, 0.011817932, 0.012527466, -0.059326172, 0.006175995, 0.010154724, 0.04208374, -0.009880066, 0.009681702, 0.04458618, -0.012527466, -0.021636963, 0.025772095, -0.0035037994, 0.036...
http://arxiv.org/abs/2305.11116v1
LLMScore: Unveiling the Power of Large Language Models in Text-to-Image Synthesis Evaluation
Yujie Lu, Xianjun Yang, Xiujun Li, Xin Eric Wang, William Yang Wang
cs.CV, cs.CL
Existing automatic evaluation on text-to-image synthesis can only provide an image-text matching score, without considering the object-level compositionality, which results in poor correlation with human judgments. In this work, we propose LLMScore, a new framework that offers evaluation scores with multi-granularity c...
http://arxiv.org/pdf/2305.11116v1
2023-05-18T16:57:57
2023-05-18T16:57:57
{ "title": [ -0.040802002, -0.0084991455, -0.015182495, -0.024734497, -0.01725769, 0.0042762756, -0.03668213, -0.015594482, 0.01058197, 0.026473999, -0.037963867, 0.014671326, -0.017196655, -0.012687683, -0.03945923, -0.010101318, -0.007255554, 0...
http://arxiv.org/abs/2305.11096v3
Cross-modality Data Augmentation for End-to-End Sign Language Translation
Jinhui Ye, Wenxiang Jiao, Xing Wang, Zhaopeng Tu, Hui Xiong
cs.CL
End-to-end sign language translation (SLT) aims to convert sign language videos into spoken language texts directly without intermediate representations. It has been a challenging task due to the modality gap between sign videos and texts and the data scarcity of labeled data. To tackle these challenges, we propose a n...
http://arxiv.org/pdf/2305.11096v3
2023-10-18T11:59:57
2023-05-18T16:34:18
{ "title": [ -0.033691406, -0.001745224, -0.049194336, -0.052703857, -0.008071899, -0.0001013875, -0.04626465, 0.00009000301, 0.024139404, 0.03390503, -0.035491943, 0.01020813, 0.022155762, -0.021072388, -0.037750244, 0.011108398, -0.010131836, 0...
http://arxiv.org/abs/2305.11095v3
Prompting the Hidden Talent of Web-Scale Speech Models for Zero-Shot Task Generalization
Puyuan Peng, Brian Yan, Shinji Watanabe, David Harwath
eess.AS, cs.AI, cs.CL, cs.LG, cs.SD
We investigate the emergent abilities of the recently proposed web-scale speech model Whisper, by adapting it to unseen tasks with prompt engineering. We selected three tasks: audio-visual speech recognition (AVSR), code-switched speech recognition (CS-ASR), and speech translation (ST) on unseen language pairs. We desi...
http://arxiv.org/pdf/2305.11095v3
2023-08-16T00:57:34
2023-05-18T16:32:58
{ "title": [ 0.01499939, 0.004447937, -0.04547119, -0.0748291, -0.00315094, -0.020568848, -0.03805542, 0.027954102, -0.0018539429, 0.055664062, -0.008033752, 0.0052757263, 0.013641357, -0.008270264, -0.058044434, 0.007972717, 0.004211426, 0.04840...
http://arxiv.org/abs/2305.11080v1
Inspecting the Geographical Representativeness of Images from Text-to-Image Models
Abhipsa Basu, R. Venkatesh Babu, Danish Pruthi
cs.CV, cs.CL
Recent progress in generative models has resulted in models that produce both realistic as well as relevant images for most textual inputs. These models are being used to generate millions of images everyday, and hold the potential to drastically impact areas such as generative art, digital marketing and data augmentat...
http://arxiv.org/pdf/2305.11080v1
2023-05-18T16:08:11
2023-05-18T16:08:11
{ "title": [ 0.010070801, 0.017868042, 0.00806427, -0.032287598, 0.006942749, -0.0040626526, -0.01108551, -0.052703857, 0.024856567, 0.038848877, -0.05392456, 0.0029392242, 0.009094238, -0.002960205, -0.050598145, -0.0018529892, 0.017303467, 0.00...
http://arxiv.org/abs/2305.11073v1
A Comparative Study on E-Branchformer vs Conformer in Speech Recognition, Translation, and Understanding Tasks
Yifan Peng, Kwangyoun Kim, Felix Wu, Brian Yan, Siddhant Arora, William Chen, Jiyang Tang, Suwon Shon, Prashant Sridhar, Shinji Watanabe
cs.CL, cs.SD, eess.AS
Conformer, a convolution-augmented Transformer variant, has become the de facto encoder architecture for speech processing due to its superior performance in various tasks, including automatic speech recognition (ASR), speech translation (ST) and spoken language understanding (SLU). Recently, a new encoder called E-Bra...
http://arxiv.org/pdf/2305.11073v1
2023-05-18T16:00:48
2023-05-18T16:00:48
{ "title": [ -0.0143585205, -0.0051460266, -0.056152344, -0.072387695, -0.019485474, -0.002632141, -0.090148926, 0.026763916, -0.021499634, 0.057861328, -0.023330688, -0.023254395, 0.044921875, -0.026229858, -0.038146973, -0.029052734, 0.0036735535, ...
http://arxiv.org/abs/2305.11072v1
Self-supervised Fine-tuning for Improved Content Representations by Speaker-invariant Clustering
Heng-Jui Chang, Alexander H. Liu, James Glass
cs.CL, eess.AS
Self-supervised speech representation models have succeeded in various tasks, but improving them for content-related problems using unlabeled data is challenging. We propose speaker-invariant clustering (Spin), a novel self-supervised learning method that clusters speech representations and performs swapped prediction ...
http://arxiv.org/pdf/2305.11072v1
2023-05-18T15:59:36
2023-05-18T15:59:36
{ "title": [ 0.011421204, 0.022583008, -0.042785645, -0.04324341, 0.013328552, -0.021835327, -0.026443481, -0.016403198, -0.026489258, 0.049713135, 0.005558014, 0.019195557, -0.036071777, -0.0107040405, -0.052825928, 0.008392334, 0.0069122314, 0....
http://arxiv.org/abs/2305.11052v1
BERM: Training the Balanced and Extractable Representation for Matching to Improve Generalization Ability of Dense Retrieval
Shicheng Xu, Liang Pang, Huawei Shen, Xueqi Cheng
cs.IR, cs.CL
Dense retrieval has shown promise in the first-stage retrieval process when trained on in-domain labeled datasets. However, previous studies have found that dense retrieval is hard to generalize to unseen domains due to its weak modeling of domain-invariant and interpretable feature (i.e., matching signal between two t...
http://arxiv.org/pdf/2305.11052v1
2023-05-18T15:43:09
2023-05-18T15:43:09
{ "title": [ 0.020721436, 0.014091492, -0.03213501, -0.039764404, 0.01777649, -0.01108551, -0.0040626526, -0.030700684, 0.037139893, 0.036102295, -0.021835327, 0.01939392, -0.0005145073, -0.007671356, -0.052215576, 0.0132369995, -0.010887146, 0.0...