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/2305.12918v1
Improving Metrics for Speech Translation
Claudio Paonessa, Dominik Frefel, Manfred Vogel
cs.CL, cs.AI
We introduce Parallel Paraphrasing ($\text{Para}_\text{both}$), an augmentation method for translation metrics making use of automatic paraphrasing of both the reference and hypothesis. This method counteracts the typically misleading results of speech translation metrics such as WER, CER, and BLEU if only a single ref...
http://arxiv.org/pdf/2305.12918v1
2023-05-22T11:01:38
2023-05-22T11:01:38
{ "title": [ -0.045288086, -0.014015198, -0.016403198, -0.02909851, -0.025253296, 0.0059280396, -0.047546387, 0.0034065247, -0.016143799, 0.060058594, -0.0073623657, -0.0015525818, 0.021392822, -0.021011353, -0.0368042, 0.017196655, 0.01626587, 0...
http://arxiv.org/abs/2305.12908v1
Language Models for German Text Simplification: Overcoming Parallel Data Scarcity through Style-specific Pre-training
Miriam Anschütz, Joshua Oehms, Thomas Wimmer, Bartłomiej Jezierski, Georg Groh
cs.CL
Automatic text simplification systems help to reduce textual information barriers on the internet. However, for languages other than English, only few parallel data to train these systems exists. We propose a two-step approach to overcome this data scarcity issue. First, we fine-tuned language models on a corpus of Ger...
http://arxiv.org/pdf/2305.12908v1
2023-05-22T10:41:30
2023-05-22T10:41:30
{ "title": [ -0.051208496, -0.004776001, -0.05822754, -0.053497314, -0.022872925, 0.01512146, -0.09283447, -0.004383087, -0.014709473, 0.07501221, 0.026031494, 0.015182495, 0.06506348, -0.02923584, -0.03894043, 0.029037476, 0.0056877136, 0.028533...
http://arxiv.org/abs/2305.12907v1
Meta-in-context learning in large language models
Julian Coda-Forno, Marcel Binz, Zeynep Akata, Matthew Botvinick, Jane X. Wang, Eric Schulz
cs.CL, cs.AI, cs.LG
Large language models have shown tremendous performance in a variety of tasks. In-context learning -- the ability to improve at a task after being provided with a number of demonstrations -- is seen as one of the main contributors to their success. In the present paper, we demonstrate that the in-context learning abili...
http://arxiv.org/pdf/2305.12907v1
2023-05-22T10:40:36
2023-05-22T10:40:36
{ "title": [ -0.011787415, -0.011047363, -0.035461426, -0.06890869, -0.0021400452, 0.0031414032, -0.045532227, 0.0041885376, 0.014335632, 0.048339844, -0.032806396, 0.012687683, 0.0073165894, -0.009086609, -0.044525146, 0.012794495, -0.009132385, ...
http://arxiv.org/abs/2305.12900v2
Evaluating Prompt-based Question Answering for Object Prediction in the Open Research Knowledge Graph
Jennifer D'Souza, Moussab Hrou, Sören Auer
cs.CL, cs.AI, cs.DL, cs.IT, math.IT
There have been many recent investigations into prompt-based training of transformer language models for new text genres in low-resource settings. The prompt-based training approach has been found to be effective in generalizing pre-trained or fine-tuned models for transfer to resource-scarce settings. This work, for t...
http://arxiv.org/pdf/2305.12900v2
2023-06-11T08:23:23
2023-05-22T10:35:18
{ "title": [ -0.019302368, -0.00031352043, -0.036621094, -0.06311035, -0.01424408, 0.007827759, -0.04849243, -0.010635376, -0.008399963, 0.011581421, -0.020355225, 0.008049011, -0.024414062, -0.015304565, -0.03237915, -0.019073486, 0.015579224, 0...
http://arxiv.org/abs/2305.12894v2
Leveraging Human Feedback to Scale Educational Datasets: Combining Crowdworkers and Comparative Judgement
Owen Henkel, Libby Hills
cs.CL
Machine Learning models have many potentially beneficial applications in education settings, but a key barrier to their development is securing enough data to train these models. Labelling educational data has traditionally relied on highly skilled raters using complex, multi-class rubrics, making the process expensive...
http://arxiv.org/pdf/2305.12894v2
2023-11-09T18:02:58
2023-05-22T10:22:14
{ "title": [ -0.0077781677, 0.005596161, -0.03778076, -0.0018539429, 0.0052452087, -0.03274536, 0.0032138824, -0.020828247, -0.015716553, 0.03463745, -0.05819702, 0.016479492, -0.032348633, 0.022079468, -0.06137085, -0.05050659, -0.0067214966, 0....
http://arxiv.org/abs/2305.12880v1
Yes, this Way! Learning to Ground Referring Expressions into Actions with Intra-episodic Feedback from Supportive Teachers
Philipp Sadler, Sherzod Hakimov, David Schlangen
cs.CV, cs.CL
The ability to pick up on language signals in an ongoing interaction is crucial for future machine learning models to collaborate and interact with humans naturally. In this paper, we present an initial study that evaluates intra-episodic feedback given in a collaborative setting. We use a referential language game as ...
http://arxiv.org/pdf/2305.12880v1
2023-05-22T10:01:15
2023-05-22T10:01:15
{ "title": [ 0.020889282, -0.016647339, 0.0019111633, -0.019729614, -0.009315491, -0.017608643, 0.016342163, 0.001578331, -0.0015983582, 0.009536743, -0.022506714, 0.016342163, 0.018554688, -0.03225708, 0.010261536, -0.04232788, -0.0146102905, 0....
http://arxiv.org/abs/2305.12878v2
Non-Autoregressive Document-Level Machine Translation
Guangsheng Bao, Zhiyang Teng, Hao Zhou, Jianhao Yan, Yue Zhang
cs.CL
Non-autoregressive translation (NAT) models achieve comparable performance and superior speed compared to auto-regressive translation (AT) models in the context of sentence-level machine translation (MT). However, their abilities are unexplored in document-level MT, hindering their usage in real scenarios. In this pape...
http://arxiv.org/pdf/2305.12878v2
2023-10-08T08:03:11
2023-05-22T09:59:59
{ "title": [ -0.03466797, -0.0055389404, -0.022949219, -0.054992676, -0.008758545, 0.005493164, -0.055267334, 0.005268097, 0.018707275, 0.057922363, -0.016479492, 0.001282692, 0.04055786, -0.00573349, -0.045410156, 0.025543213, 0.00071525574, 0.0...
http://arxiv.org/abs/2305.12876v2
Gloss-Free End-to-End Sign Language Translation
Kezhou Lin, Xiaohan Wang, Linchao Zhu, Ke Sun, Bang Zhang, Yi Yang
cs.CV, cs.CL
In this paper, we tackle the problem of sign language translation (SLT) without gloss annotations. Although intermediate representation like gloss has been proven effective, gloss annotations are hard to acquire, especially in large quantities. This limits the domain coverage of translation datasets, thus handicapping ...
http://arxiv.org/pdf/2305.12876v2
2023-05-27T16:43:18
2023-05-22T09:57:43
{ "title": [ -0.018615723, -0.015930176, -0.03729248, -0.04800415, -0.040100098, -0.002117157, -0.04928589, 0.016616821, -0.002878189, 0.06124878, 0.00083589554, -0.011413574, 0.026290894, -0.04156494, -0.025238037, 0.0014381409, 0.015945435, 0.0...
http://arxiv.org/abs/2305.12870v2
Lion: Adversarial Distillation of Proprietary Large Language Models
Yuxin Jiang, Chunkit Chan, Mingyang Chen, Wei Wang
cs.CL
The practice of transferring knowledge from a sophisticated, proprietary large language model (LLM) to a compact, open-source LLM has garnered considerable attention. Previous works have focused on a unidirectional knowledge distillation way by aligning the responses of the student model with those of the teacher model...
http://arxiv.org/pdf/2305.12870v2
2023-10-14T02:21:24
2023-05-22T09:49:16
{ "title": [ -0.029418945, -0.013839722, -0.042236328, -0.06524658, 0.0013074875, 0.012916565, -0.06378174, 0.005619049, -0.013000488, 0.0284729, -0.004219055, 0.008415222, -0.0035114288, -0.017623901, -0.033721924, -0.000019133091, -0.009246826, ...
http://arxiv.org/abs/2305.12868v1
NAS-FM: Neural Architecture Search for Tunable and Interpretable Sound Synthesis based on Frequency Modulation
Zhen Ye, Wei Xue, Xu Tan, Qifeng Liu, Yike Guo
cs.SD, cs.AI, cs.CL, cs.LG, cs.MM, eess.AS
Developing digital sound synthesizers is crucial to the music industry as it provides a low-cost way to produce high-quality sounds with rich timbres. Existing traditional synthesizers often require substantial expertise to determine the overall framework of a synthesizer and the parameters of submodules. Since expert ...
http://arxiv.org/pdf/2305.12868v1
2023-05-22T09:46:10
2023-05-22T09:46:10
{ "title": [ -0.039001465, 0.0064315796, -0.027648926, -0.017471313, -0.010253906, -0.004535675, -0.04324341, -0.036071777, -0.019088745, 0.050964355, -0.00730896, 0.0066337585, 0.0103302, -0.019638062, -0.053710938, -0.0025749207, -0.0013389587, ...
http://arxiv.org/abs/2305.12851v2
Enhancing Coherence of Extractive Summarization with Multitask Learning
Renlong Jie, Xiaojun Meng, Lifeng Shang, Xin Jiang, Qun Liu
cs.CL, cs.AI
This study proposes a multitask learning architecture for extractive summarization with coherence boosting. The architecture contains an extractive summarizer and coherent discriminator module. The coherent discriminator is trained online on the sentence vectors of the augmented textual input, thus improving its genera...
http://arxiv.org/pdf/2305.12851v2
2023-07-21T10:22:53
2023-05-22T09:20:58
{ "title": [ -0.0063209534, 0.02116394, -0.056762695, -0.07397461, 0.00737381, -0.013938904, -0.07775879, -0.01512146, -0.0036907196, 0.05581665, 0.0049324036, 0.035491943, -0.015419006, -0.001294136, -0.052947998, 0.0060653687, -0.010116577, 0.0...
http://arxiv.org/abs/2305.12839v1
CopyNE: Better Contextual ASR by Copying Named Entities
Shilin Zhou, Zhenghua Li, Yu Hong, Min Zhang, Zhefeng Wang, Baoxing Huai
cs.CL
Recent years have seen remarkable progress in automatic speech recognition (ASR). However, traditional token-level ASR models have struggled with accurately transcribing entities due to the problem of homophonic and near-homophonic tokens. This paper introduces a novel approach called CopyNE, which uses a span-level co...
http://arxiv.org/pdf/2305.12839v1
2023-05-22T09:03:11
2023-05-22T09:03:11
{ "title": [ -0.0005774498, -0.0063591003, -0.05618286, -0.075927734, 0.011665344, -0.006965637, -0.039276123, -0.0028915405, -0.019714355, 0.02835083, -0.005306244, 0.007549286, 0.04727173, -0.034698486, 0.013130188, 0.011619568, 0.0154800415, 0...
http://arxiv.org/abs/2305.12835v1
Open-Domain Event Graph Induction for Mitigating Framing Bias
Siyi Liu, Hongming Zhang, Hongwei Wang, Kaiqiang Song, Dan Roth, Dong Yu
cs.CL, cs.AI
Researchers have proposed various information extraction (IE) techniques to convert news articles into structured knowledge for news understanding. However, none of the existing methods have explicitly addressed the issue of framing bias that is inherent in news articles. We argue that studying and identifying framing ...
http://arxiv.org/pdf/2305.12835v1
2023-05-22T08:57:42
2023-05-22T08:57:42
{ "title": [ 0.023590088, 0.012313843, -0.04168701, -0.074279785, 0.022140503, -0.0107421875, -0.043548584, -0.030044556, -0.026763916, 0.011306763, 0.003156662, 0.017684937, 0.0014724731, -0.036468506, 0.0070152283, -0.0026359558, -0.022766113, ...
http://arxiv.org/abs/2305.12829v2
On Bias and Fairness in NLP: How to have a fairer text classification?
Fatma Elsafoury, Stamos Katsigiannis, Naeem Ramzan
cs.CL
In this paper, we provide a holistic analysis of the different sources of bias, Upstream, Sample and Overampflication biases, in NLP models. We investigate how they impact the fairness of the task of text classification. We also investigate the impact of removing these biases using different debiasing techniques on the...
http://arxiv.org/pdf/2305.12829v2
2023-05-31T10:13:26
2023-05-22T08:44:00
{ "title": [ -0.02381897, 0.004753113, -0.05441284, -0.036071777, 0.022903442, -0.016723633, -0.00982666, -0.0110321045, -0.006324768, 0.008666992, -0.052520752, 0.024276733, -0.0015792847, -0.012413025, -0.02406311, -0.01272583, -0.03225708, 0.0...
http://arxiv.org/abs/2306.05523v2
FACTIFY3M: A Benchmark for Multimodal Fact Verification with Explainability through 5W Question-Answering
Megha Chakraborty, Khushbu Pahwa, Anku Rani, Shreyas Chatterjee, Dwip Dalal, Harshit Dave, Ritvik G, Preethi Gurumurthy, Adarsh Mahor, Samahriti Mukherjee, Aditya Pakala, Ishan Paul, Janvita Reddy, Arghya Sarkar, Kinjal Sensharma, Aman Chadha, Amit P. Sheth, Amitava Das
cs.CL, cs.AI, cs.CV, cs.MM
Combating disinformation is one of the burning societal crises -- about 67% of the American population believes that disinformation produces a lot of uncertainty, and 10% of them knowingly propagate disinformation. Evidence shows that disinformation can manipulate democratic processes and public opinion, causing disrup...
http://arxiv.org/pdf/2306.05523v2
2023-10-31T01:51:56
2023-05-22T08:29:47
{ "title": [ -0.016998291, 0.0006427765, -0.032165527, -0.092041016, 0.00036478043, 0.026672363, -0.0680542, 0.012039185, -0.004032135, 0.015991211, -0.014770508, 0.024337769, 0.01398468, -0.023376465, -0.016616821, -0.0064353943, -0.028244019, 0...
http://arxiv.org/abs/2305.12820v2
MultiTabQA: Generating Tabular Answers for Multi-Table Question Answering
Vaishali Pal, Andrew Yates, Evangelos Kanoulas, Maarten de Rijke
cs.CL, cs.AI
Recent advances in tabular question answering (QA) with large language models are constrained in their coverage and only answer questions over a single table. However, real-world queries are complex in nature, often over multiple tables in a relational database or web page. Single table questions do not involve common ...
http://arxiv.org/pdf/2305.12820v2
2023-05-24T17:13:47
2023-05-22T08:25:15
{ "title": [ -0.019805908, -0.019882202, -0.037597656, -0.07720947, -0.0015306473, 0.013397217, -0.048797607, 0.00818634, 0.02116394, 0.033477783, -0.022140503, 0.006072998, 0.004371643, -0.003862381, -0.066345215, 0.0014772415, 0.0025081635, 0.0...
http://arxiv.org/abs/2305.12818v2
Crosslingual Transfer Learning for Low-Resource Languages Based on Multilingual Colexification Graphs
Yihong Liu, Haotian Ye, Leonie Weissweiler, Renhao Pei, Hinrich Schütze
cs.CL, cs.AI
In comparative linguistics, colexification refers to the phenomenon of a lexical form conveying two or more distinct meanings. Existing work on colexification patterns relies on annotated word lists, limiting scalability and usefulness in NLP. In contrast, we identify colexification patterns of more than 2,000 concepts...
http://arxiv.org/pdf/2305.12818v2
2023-10-19T14:38:17
2023-05-22T08:20:23
{ "title": [ -0.028778076, -0.027023315, -0.04397583, -0.038269043, -0.023086548, -0.0023880005, -0.046569824, 0.0060272217, -0.003967285, 0.03375244, -0.0037212372, 0.019363403, -0.0023880005, -0.025405884, -0.026016235, -0.0012369156, -0.01725769, ...
http://arxiv.org/abs/2305.12816v1
Farewell to Aimless Large-scale Pretraining: Influential Subset Selection for Language Model
Xiao Wang, Weikang Zhou, Qi Zhang, Jie Zhou, Songyang Gao, Junzhe Wang, Menghan Zhang, Xiang Gao, Yunwen Chen, Tao Gui
cs.CL
Pretrained language models have achieved remarkable success in various natural language processing tasks. However, pretraining has recently shifted toward larger models and larger data, and this has resulted in significant computational and energy costs. In this paper, we propose Influence Subset Selection (ISS) for la...
http://arxiv.org/pdf/2305.12816v1
2023-05-22T08:18:58
2023-05-22T08:18:58
{ "title": [ -0.01828003, 0.006034851, -0.05340576, -0.044891357, -0.006011963, -0.017318726, -0.038604736, 0.006351471, -0.03274536, 0.022033691, -0.017196655, 0.022903442, -0.009468079, -0.02494812, -0.020339966, -0.02029419, -0.017837524, 0.05...
http://arxiv.org/abs/2305.12815v1
Towards Dialogue Systems with Agency in Human-AI Collaboration Tasks
Ashish Sharma, Sudha Rao, Chris Brockett, Akanksha Malhotra, Nebojsa Jojic, Bill Dolan
cs.CL
Agency, the capacity to proactively shape events, is crucial to how humans interact and collaborate with other humans. In this paper, we investigate Agency as a potentially desirable function of dialogue agents, and how it can be measured and controlled. We build upon the social-cognitive theory of Bandura (2001) to de...
http://arxiv.org/pdf/2305.12815v1
2023-05-22T08:17:14
2023-05-22T08:17:14
{ "title": [ -0.034820557, -0.0059394836, -0.055480957, -0.037841797, 0.009902954, -0.016113281, -0.018051147, 0.0022392273, 0.0022068024, 0.024414062, -0.016159058, -0.0051956177, -0.020721436, -0.016906738, -0.0579834, -0.014709473, -0.030471802, ...
http://arxiv.org/abs/2305.12802v1
Ultra-Fine Entity Typing with Prior Knowledge about Labels: A Simple Clustering Based Strategy
Na Li, Zied Bouraoui, Steven Schockaert
cs.CL, cs.AI
Ultra-fine entity typing (UFET) is the task of inferring the semantic types, from a large set of fine-grained candidates, that apply to a given entity mention. This task is especially challenging because we only have a small number of training examples for many of the types, even with distant supervision strategies. St...
http://arxiv.org/pdf/2305.12802v1
2023-05-22T08:00:56
2023-05-22T08:00:56
{ "title": [ -0.006778717, -0.0006866455, -0.041290283, -0.075683594, 0.027572632, -0.00605011, -0.033721924, 0.002614975, -0.009765625, 0.035369873, 0.0053901672, 0.022079468, 0.015525818, -0.02532959, -0.029205322, 0.011268616, -0.01776123, 0.0...
http://arxiv.org/abs/2305.12798v1
LM-Switch: Lightweight Language Model Conditioning in Word Embedding Space
Chi Han, Jialiang Xu, Manling Li, Yi Fung, Chenkai Sun, Nan Jiang, Tarek Abdelzaher, Heng Ji
cs.CL, cs.AI, cs.LG
In recent years, large language models (LMs) have achieved remarkable progress across various natural language processing tasks. As pre-training and fine-tuning are costly and might negatively impact model performance, it is desired to efficiently adapt an existing model to different conditions such as styles, sentimen...
http://arxiv.org/pdf/2305.12798v1
2023-05-22T07:52:04
2023-05-22T07:52:04
{ "title": [ -0.017669678, -0.01335144, -0.04360962, -0.036193848, -0.004673004, -0.008651733, -0.033569336, -0.0005726814, 0.0044403076, 0.022827148, -0.012939453, 0.010818481, 0.014198303, -0.01689148, -0.023132324, 0.009643555, -0.013381958, 0...
http://arxiv.org/abs/2305.12793v1
Zero-Shot End-to-End Spoken Language Understanding via Cross-Modal Selective Self-Training
Jianfeng He, Julian Salazar, Kaisheng Yao, Haoqi Li, Jinglun Cai
eess.AS, cs.CL, cs.MM, cs.SD
End-to-end (E2E) spoken language understanding (SLU) is constrained by the cost of collecting speech-semantics pairs, especially when label domains change. Hence, we explore \textit{zero-shot} E2E SLU, which learns E2E SLU without speech-semantics pairs, instead using only speech-text and text-semantics pairs. Previous...
http://arxiv.org/pdf/2305.12793v1
2023-05-22T07:42:52
2023-05-22T07:42:52
{ "title": [ 0.002363205, -0.0027580261, -0.009307861, -0.08270264, -0.015609741, 0.007286072, -0.058563232, 0.013793945, -0.0018224716, 0.011657715, 0.002544403, 0.0028133392, -0.0057640076, -0.03591919, -0.029754639, -0.000582695, -0.0050735474, ...
http://arxiv.org/abs/2305.12792v1
Semantic Structure Enhanced Event Causality Identification
Zhilei Hu, Zixuan Li, Xiaolong Jin, Long Bai, Saiping Guan, Jiafeng Guo, Xueqi Cheng
cs.CL, cs.AI
Event Causality Identification (ECI) aims to identify causal relations between events in unstructured texts. This is a very challenging task, because causal relations are usually expressed by implicit associations between events. Existing methods usually capture such associations by directly modeling the texts with pre...
http://arxiv.org/pdf/2305.12792v1
2023-05-22T07:42:35
2023-05-22T07:42:35
{ "title": [ -0.0017280579, 0.0011463165, -0.049560547, -0.099731445, 0.016555786, -0.00022053719, -0.066101074, -0.008285522, -0.0056419373, 0.0132751465, 0.029327393, -0.0017309189, 0.014251709, -0.051239014, -0.0054512024, 0.021850586, -0.00969696...
http://arxiv.org/abs/2305.12786v2
Mitigating Data Imbalance and Representation Degeneration in Multilingual Machine Translation
Wen Lai, Alexandra Chronopoulou, Alexander Fraser
cs.CL
Despite advances in multilingual neural machine translation (MNMT), we argue that there are still two major challenges in this area: data imbalance and representation degeneration. The data imbalance problem refers to the imbalance in the amount of parallel corpora for all language pairs, especially for long-tail langu...
http://arxiv.org/pdf/2305.12786v2
2023-10-24T20:10:04
2023-05-22T07:31:08
{ "title": [ -0.014022827, 0.0014266968, -0.016799927, -0.059020996, 0.0007286072, 0.010292053, -0.036865234, -0.006477356, 0.010612488, 0.06713867, -0.011314392, 0.002500534, 0.049102783, -0.0038337708, -0.051574707, 0.022247314, -0.015449524, 0...
http://arxiv.org/abs/2305.12785v2
MacLaSa: Multi-Aspect Controllable Text Generation via Efficient Sampling from Compact Latent Space
Hanxing Ding, Liang Pang, Zihao Wei, Huawei Shen, Xueqi Cheng, Tat-Seng Chua
cs.CL
Multi-aspect controllable text generation aims to generate fluent sentences that possess multiple desired attributes simultaneously. Traditional methods either combine many operators in the decoding stage, often with costly iteration or search in the discrete text space, or train separate controllers for each aspect, r...
http://arxiv.org/pdf/2305.12785v2
2023-10-17T15:48:15
2023-05-22T07:30:35
{ "title": [ -0.035247803, -0.015144348, -0.035614014, -0.048034668, -0.027938843, 0.008483887, -0.06008911, -0.009475708, 0.00006681681, 0.042633057, 0.0012397766, 0.03955078, 0.0069389343, -0.005470276, -0.044158936, 0.022109985, -0.011802673, ...
http://arxiv.org/abs/2305.12782v1
Towards Robust Personalized Dialogue Generation via Order-Insensitive Representation Regularization
Liang Chen, Hongru Wang, Yang Deng, Wai-Chung Kwan, Zezhong Wang, Kam-Fai Wong
cs.CL, cs.AI
Generating persona consistent dialogue response is important for developing an intelligent conversational agent. Recent works typically fine-tune large-scale pre-trained models on this task by concatenating persona texts and dialogue history as a single input sequence to generate the target response. While simple and e...
http://arxiv.org/pdf/2305.12782v1
2023-05-22T07:24:29
2023-05-22T07:24:29
{ "title": [ 0.016067505, 0.006500244, -0.048339844, -0.0871582, 0.018234253, 0.00447464, -0.05355835, 0.014602661, -0.0030269623, 0.040649414, 0.018081665, 0.005886078, -0.0004286766, -0.021362305, -0.05581665, 0.0018692017, -0.030075073, 0.0490...
http://arxiv.org/abs/2305.12777v1
Evaluating Pragmatic Abilities of Image Captioners on A3DS
Polina Tsvilodub, Michael Franke
cs.CL
Evaluating grounded neural language model performance with respect to pragmatic qualities like the trade off between truthfulness, contrastivity and overinformativity of generated utterances remains a challenge in absence of data collected from humans. To enable such evaluation, we present a novel open source image-tex...
http://arxiv.org/pdf/2305.12777v1
2023-05-22T07:15:33
2023-05-22T07:15:33
{ "title": [ 0.00030732155, 0.018875122, -0.009185791, -0.014556885, 0.0046653748, -0.014640808, 0.010787964, -0.022567749, 0.01763916, 0.048309326, -0.035461426, 0.0028934479, -0.010437012, -0.006137848, -0.060302734, -0.008323669, -0.0038318634, ...
http://arxiv.org/abs/2305.12774v3
Learning Optimal Policy for Simultaneous Machine Translation via Binary Search
Shoutao Guo, Shaolei Zhang, Yang Feng
cs.CL
Simultaneous machine translation (SiMT) starts to output translation while reading the source sentence and needs a precise policy to decide when to output the generated translation. Therefore, the policy determines the number of source tokens read during the translation of each target token. However, it is difficult to...
http://arxiv.org/pdf/2305.12774v3
2023-05-27T15:27:40
2023-05-22T07:03:06
{ "title": [ -0.021636963, -0.02407837, 0.014251709, -0.06518555, -0.02746582, 0.020828247, -0.04473877, 0.00630188, -0.044708252, 0.048706055, 0.004142761, 0.027404785, -0.027816772, -0.014152527, -0.040130615, -0.0055236816, -0.012168884, 0.030...
http://arxiv.org/abs/2305.12767v2
D$^2$TV: Dual Knowledge Distillation and Target-oriented Vision Modeling for Many-to-Many Multimodal Summarization
Yunlong Liang, Fandong Meng, Jiaan Wang, Jinan Xu, Yufeng Chen, Jie Zhou
cs.CV, cs.AI, cs.CL
Many-to-many multimodal summarization (M$^3$S) task aims to generate summaries in any language with document inputs in any language and the corresponding image sequence, which essentially comprises multimodal monolingual summarization (MMS) and multimodal cross-lingual summarization (MXLS) tasks. Although much work has...
http://arxiv.org/pdf/2305.12767v2
2023-10-16T10:04:37
2023-05-22T06:47:35
{ "title": [ -0.0020599365, 0.0037956238, -0.05960083, -0.10430908, 0.00031876564, 0.006088257, -0.07458496, 0.015701294, 0.012374878, 0.049591064, -0.015426636, 0.0070991516, 0.016555786, -0.009643555, -0.040039062, -0.0128479, 0.016235352, 0.00...
http://arxiv.org/abs/2305.12766v2
Explaining Emergent In-Context Learning as Kernel Regression
Chi Han, Ziqi Wang, Han Zhao, Heng Ji
cs.CL, cs.AI, cs.LG
Large language models (LLMs) have initiated a paradigm shift in transfer learning. In contrast to the classic pretraining-then-finetuning procedure, in order to use LLMs for downstream prediction tasks, one only needs to provide a few demonstrations, known as in-context examples, without adding more or updating existin...
http://arxiv.org/pdf/2305.12766v2
2023-10-05T16:04:43
2023-05-22T06:45:02
{ "title": [ -0.0029621124, -0.0024528503, -0.02104187, -0.02671814, -0.0028572083, -0.022018433, 0.005592346, -0.023254395, -0.0028266907, 0.025985718, -0.0435791, -0.0031833649, -0.02041626, -0.006839752, -0.05480957, -0.025527954, -0.0021686554, ...
http://arxiv.org/abs/2305.12761v1
Enhancing Cross-lingual Natural Language Inference by Soft Prompting with Multilingual Verbalizer
Shuang Li, Xuming Hu, Aiwei Liu, Yawen Yang, Fukun Ma, Philip S. Yu, Lijie Wen
cs.CL, cs.AI
Cross-lingual natural language inference is a fundamental problem in cross-lingual language understanding. Many recent works have used prompt learning to address the lack of annotated parallel corpora in XNLI. However, these methods adopt discrete prompting by simply translating the templates to the target language and...
http://arxiv.org/pdf/2305.12761v1
2023-05-22T06:31:29
2023-05-22T06:31:29
{ "title": [ -0.027145386, -0.011474609, -0.029846191, -0.09185791, -0.009567261, 0.019485474, -0.09899902, 0.0041618347, -0.009239197, 0.012168884, -0.02911377, 0.02923584, -0.00018167496, -0.0141067505, -0.00046420097, -0.0005760193, 0.003572464, ...
http://arxiv.org/abs/2305.12759v1
Kanbun-LM: Reading and Translating Classical Chinese in Japanese Methods by Language Models
Hao Wang, Hirofumi Shimizu, Daisuke Kawahara
cs.CL
Recent studies in natural language processing (NLP) have focused on modern languages and achieved state-of-the-art results in many tasks. Meanwhile, little attention has been paid to ancient texts and related tasks. Classical Chinese first came to Japan approximately 2,000 years ago. It was gradually adapted to a Japan...
http://arxiv.org/pdf/2305.12759v1
2023-05-22T06:30:02
2023-05-22T06:30:02
{ "title": [ -0.064208984, -0.0014028549, -0.038146973, 0.01979065, -0.027008057, -0.010543823, -0.0413208, -0.038146973, 0.0008006096, 0.040527344, -0.0009841919, -0.015777588, 0.0670166, -0.057037354, -0.0309906, 0.018218994, 0.0028438568, 0.02...
http://arxiv.org/abs/2305.12757v1
This Prompt is Measuring <MASK>: Evaluating Bias Evaluation in Language Models
Seraphina Goldfarb-Tarrant, Eddie Ungless, Esma Balkir, Su Lin Blodgett
cs.CL
Bias research in NLP seeks to analyse models for social biases, thus helping NLP practitioners uncover, measure, and mitigate social harms. We analyse the body of work that uses prompts and templates to assess bias in language models. We draw on a measurement modelling framework to create a taxonomy of attributes that ...
http://arxiv.org/pdf/2305.12757v1
2023-05-22T06:28:48
2023-05-22T06:28:48
{ "title": [ -0.059570312, -0.007713318, -0.06530762, 0.048461914, -0.029785156, -0.016830444, -0.020141602, -0.0059890747, -0.024810791, 0.011398315, -0.04724121, 0.0032024384, 0.018310547, -0.03564453, 0.030685425, -0.03955078, -0.024230957, 0....
http://arxiv.org/abs/2305.12755v1
GNCformer Enhanced Self-attention for Automatic Speech Recognition
J. Li, Z. Duan, S. Li, X. Yu, G. Yang
cs.SD, cs.CL, eess.AS
In this paper,an Enhanced Self-Attention (ESA) mechanism has been put forward for robust feature extraction.The proposed ESA is integrated with the recursive gated convolution and self-attention mechanism.In particular, the former is used to capture multi-order feature interaction and the latter is for global feature e...
http://arxiv.org/pdf/2305.12755v1
2023-05-22T06:26:05
2023-05-22T06:26:05
{ "title": [ 0.016540527, 0.010314941, -0.050354004, -0.06896973, 0.011154175, -0.012062073, -0.036376953, 0.010848999, -0.012207031, 0.049438477, -0.018203735, -0.025665283, 0.044769287, -0.028762817, -0.038330078, 0.00021636486, -0.006980896, 0...
http://arxiv.org/abs/2305.12753v1
Learning to Rank Utterances for Query-Focused Meeting Summarization
Xingxian Liu, Yajing Xu
cs.CL, cs.AI
Query-focused meeting summarization(QFMS) aims to generate a specific summary for the given query according to the meeting transcripts. Due to the conflict between long meetings and limited input size, previous works mainly adopt extract-then-summarize methods, which use extractors to simulate binary labels or ROUGE sc...
http://arxiv.org/pdf/2305.12753v1
2023-05-22T06:25:09
2023-05-22T06:25:09
{ "title": [ 0.01802063, 0.016967773, -0.04272461, -0.10406494, 0.017608643, 0.0017814636, -0.07293701, 0.002325058, 0.0059776306, 0.04815674, -0.019042969, 0.013900757, 0.012939453, 0.013809204, -0.06518555, -0.0056114197, -0.008354187, 0.005016...
http://arxiv.org/abs/2305.12749v1
A Benchmark on Extremely Weakly Supervised Text Classification: Reconcile Seed Matching and Prompting Approaches
Zihan Wang, Tianle Wang, Dheeraj Mekala, Jingbo Shang
cs.CL
Etremely Weakly Supervised Text Classification (XWS-TC) refers to text classification based on minimal high-level human guidance, such as a few label-indicative seed words or classification instructions. There are two mainstream approaches for XWS-TC, however, never being rigorously compared: (1) training classifiers b...
http://arxiv.org/pdf/2305.12749v1
2023-05-22T06:18:23
2023-05-22T06:18:23
{ "title": [ -0.004962921, 0.010353088, -0.013282776, -0.045898438, 0.005214691, -0.0078125, -0.019363403, -0.023361206, -0.003917694, 0.013069153, -0.044006348, 0.028137207, -0.009338379, -0.0077209473, -0.03137207, -0.012924194, -0.0035972595, ...
http://arxiv.org/abs/2305.12744v1
Fact-Checking Complex Claims with Program-Guided Reasoning
Liangming Pan, Xiaobao Wu, Xinyuan Lu, Anh Tuan Luu, William Yang Wang, Min-Yen Kan, Preslav Nakov
cs.CL, cs.AI
Fact-checking real-world claims often requires collecting multiple pieces of evidence and applying complex multi-step reasoning. In this paper, we present Program-Guided Fact-Checking (ProgramFC), a novel fact-checking model that decomposes complex claims into simpler sub-tasks that can be solved using a shared library...
http://arxiv.org/pdf/2305.12744v1
2023-05-22T06:11:15
2023-05-22T06:11:15
{ "title": [ -0.0065574646, -0.015914917, -0.015609741, -0.088378906, -0.0074310303, 0.011413574, -0.039764404, 0.011062622, -0.037597656, 0.024337769, -0.020629883, 0.018554688, -0.01423645, -0.026168823, -0.014282227, -0.011657715, -0.017303467, ...
http://arxiv.org/abs/2305.12740v1
Can We Edit Factual Knowledge by In-Context Learning?
Ce Zheng, Lei Li, Qingxiu Dong, Yuxuan Fan, Zhiyong Wu, Jingjing Xu, Baobao Chang
cs.CL
Previous studies have shown that large language models (LLMs) like GPTs store massive factual knowledge in their parameters. However, the stored knowledge could be false or out-dated. Traditional knowledge editing methods refine LLMs via fine-tuning on texts containing specific knowledge. However, with the increasing s...
http://arxiv.org/pdf/2305.12740v1
2023-05-22T06:07:58
2023-05-22T06:07:58
{ "title": [ 0.017547607, -0.0005784035, -0.048034668, -0.01777649, -0.009902954, -0.033599854, 0.02861023, 0.0032405853, -0.015007019, 0.03857422, -0.032196045, -0.022567749, -0.010353088, -0.013496399, -0.0149383545, -0.04537964, 0.005836487, 0...
http://arxiv.org/abs/2305.12737v1
The Best of Both Worlds: Combining Human and Machine Translations for Multilingual Semantic Parsing with Active Learning
Zhuang Li, Lizhen Qu, Philip R. Cohen, Raj V. Tumuluri, Gholamreza Haffari
cs.CL, cs.AI
Multilingual semantic parsing aims to leverage the knowledge from the high-resource languages to improve low-resource semantic parsing, yet commonly suffers from the data imbalance problem. Prior works propose to utilize the translations by either humans or machines to alleviate such issues. However, human translations...
http://arxiv.org/pdf/2305.12737v1
2023-05-22T05:57:47
2023-05-22T05:57:47
{ "title": [ -0.021942139, -0.00198555, 0.0063552856, -0.064331055, -0.0036201477, 0.01689148, -0.046417236, -0.009597778, -0.009414673, 0.051757812, -0.0024318695, 0.017318726, 0.00504303, -0.015083313, -0.047546387, 0.010688782, -0.0022716522, ...
http://arxiv.org/abs/2305.12733v2
MADNet: Maximizing Addressee Deduction Expectation for Multi-Party Conversation Generation
Jia-Chen Gu, Chao-Hong Tan, Caiyuan Chu, Zhen-Hua Ling, Chongyang Tao, Quan Liu, Cong Liu
cs.CL
Modeling multi-party conversations (MPCs) with graph neural networks has been proven effective at capturing complicated and graphical information flows. However, existing methods rely heavily on the necessary addressee labels and can only be applied to an ideal setting where each utterance must be tagged with an addres...
http://arxiv.org/pdf/2305.12733v2
2023-10-18T02:09:24
2023-05-22T05:50:11
{ "title": [ -0.005630493, -0.0059661865, -0.046722412, -0.07141113, -0.012283325, 0.00024843216, -0.077941895, -0.001704216, -0.0043144226, 0.018829346, 0.02041626, 0.019424438, 0.004787445, -0.04815674, -0.023162842, -0.0025787354, -0.03289795, ...
http://arxiv.org/abs/2305.12726v2
Towards Explainable In-the-Wild Video Quality Assessment: A Database and a Language-Prompted Approach
Haoning Wu, Erli Zhang, Liang Liao, Chaofeng Chen, Jingwen Hou, Annan Wang, Wenxiu Sun, Qiong Yan, Weisi Lin
cs.CV, cs.CL, cs.MM
The proliferation of in-the-wild videos has greatly expanded the Video Quality Assessment (VQA) problem. Unlike early definitions that usually focus on limited distortion types, VQA on in-the-wild videos is especially challenging as it could be affected by complicated factors, including various distortions and diverse ...
http://arxiv.org/pdf/2305.12726v2
2023-08-03T09:26:36
2023-05-22T05:20:23
{ "title": [ -0.010482788, 0.004760742, -0.015449524, -0.030792236, 0.011680603, -0.0071754456, -0.013198853, -0.017913818, -0.004714966, 0.025115967, -0.017837524, -0.004940033, -0.022888184, -0.010231018, -0.05734253, -0.015266418, -0.0033359528, ...
http://arxiv.org/abs/2305.12723v1
Enhancing Small Medical Learners with Privacy-preserving Contextual Prompting
Xinlu Zhang, Shiyang Li, Xianjun Yang, Chenxin Tian, Yao Qin, Linda Ruth Petzold
cs.CL, cs.AI
Large language models (LLMs) demonstrate remarkable medical expertise, but data privacy concerns impede their direct use in healthcare environments. Although offering improved data privacy protection, domain-specific small language models (SLMs) often underperform LLMs, emphasizing the need for methods that reduce this...
http://arxiv.org/pdf/2305.12723v1
2023-05-22T05:14:38
2023-05-22T05:14:38
{ "title": [ 0.0087509155, 0.0039482117, -0.02822876, -0.041015625, 0.010429382, -0.006690979, -0.017822266, -0.030044556, -0.054504395, 0.023925781, -0.024795532, 0.018554688, -0.003824234, -0.023101807, -0.0026721954, -0.03086853, 0.007583618, ...
http://arxiv.org/abs/2305.12720v1
llm-japanese-dataset v0: Construction of Japanese Chat Dataset for Large Language Models and its Methodology
Masanori Hirano, Masahiro Suzuki, Hiroki Sakaji
cs.CL, cs.AI
This study constructed a Japanese chat dataset for tuning large language models (LLMs), which consist of about 8.4 million records. Recently, LLMs have been developed and gaining popularity. However, high-performing LLMs are usually mainly for English. There are two ways to support languages other than English by those...
http://arxiv.org/pdf/2305.12720v1
2023-05-22T04:59:33
2023-05-22T04:59:33
{ "title": [ -0.036132812, -0.011123657, -0.04626465, -0.055145264, -0.0154800415, -0.00012338161, -0.08392334, 0.0066223145, -0.002866745, 0.037109375, -0.03366089, -0.0095825195, 0.037078857, -0.036346436, -0.02861023, -0.010627747, -0.006008148, ...
http://arxiv.org/abs/2305.12717v1
TADA: Efficient Task-Agnostic Domain Adaptation for Transformers
Chia-Chien Hung, Lukas Lange, Jannik Strötgen
cs.CL, cs.LG
Intermediate training of pre-trained transformer-based language models on domain-specific data leads to substantial gains for downstream tasks. To increase efficiency and prevent catastrophic forgetting alleviated from full domain-adaptive pre-training, approaches such as adapters have been developed. However, these re...
http://arxiv.org/pdf/2305.12717v1
2023-05-22T04:53:59
2023-05-22T04:53:59
{ "title": [ -0.013931274, -0.0010595322, -0.04458618, -0.064819336, 0.018203735, -0.0110321045, -0.036102295, -0.005302429, 0.006542206, 0.031234741, -0.0066452026, 0.008216858, 0.016555786, -0.019088745, -0.028137207, 0.020462036, -0.012458801, ...
http://arxiv.org/abs/2305.12710v2
Beyond Labels: Empowering Human Annotators with Natural Language Explanations through a Novel Active-Learning Architecture
Bingsheng Yao, Ishan Jindal, Lucian Popa, Yannis Katsis, Sayan Ghosh, Lihong He, Yuxuan Lu, Shashank Srivastava, Yunyao Li, James Hendler, Dakuo Wang
cs.CL
Real-world domain experts (e.g., doctors) rarely annotate only a decision label in their day-to-day workflow without providing explanations. Yet, existing low-resource learning techniques, such as Active Learning (AL), that aim to support human annotators mostly focus on the label while neglecting the natural language ...
http://arxiv.org/pdf/2305.12710v2
2023-10-23T16:44:59
2023-05-22T04:38:10
{ "title": [ -0.015991211, -0.005672455, -0.0010919571, -0.07006836, -0.013298035, 0.0024700165, -0.039855957, -0.014419556, 0.0055999756, 0.021987915, 0.011131287, 0.0017938614, -0.01272583, -0.032440186, -0.026275635, 0.007045746, 0.0056266785, ...
http://arxiv.org/abs/2305.12709v1
Cross-lingual Transfer Can Worsen Bias in Sentiment Analysis
Seraphina Goldfarb-Tarrant, Björn Ross, Adam Lopez
cs.CL
Sentiment analysis (SA) systems are widely deployed in many of the world's languages, and there is well-documented evidence of demographic bias in these systems. In languages beyond English, scarcer training data is often supplemented with transfer learning using pre-trained models, including multilingual models traine...
http://arxiv.org/pdf/2305.12709v1
2023-05-22T04:37:49
2023-05-22T04:37:49
{ "title": [ 0.03137207, 0.0028896332, -0.0054359436, -0.010902405, -0.035491943, -0.02659607, 0.0054244995, -0.028839111, 0.001832962, 0.058044434, -0.048797607, -0.017974854, 0.044769287, -0.016021729, -0.01701355, -0.010810852, 0.0052871704, 0...
http://arxiv.org/abs/2305.12707v1
Quantifying Association Capabilities of Large Language Models and Its Implications on Privacy Leakage
Hanyin Shao, Jie Huang, Shen Zheng, Kevin Chen-Chuan Chang
cs.CL, cs.AI, cs.CR
The advancement of large language models (LLMs) brings notable improvements across various applications, while simultaneously raising concerns about potential private data exposure. One notable capability of LLMs is their ability to form associations between different pieces of information, but this raises concerns whe...
http://arxiv.org/pdf/2305.12707v1
2023-05-22T04:30:35
2023-05-22T04:30:35
{ "title": [ -0.022720337, -0.006828308, -0.0519104, -0.057037354, 0.0067214966, -0.0132369995, -0.03829956, 0.01600647, -0.021621704, 0.04296875, 0.0030841827, 0.0105896, 0.009429932, -0.01209259, -0.035736084, 0.005340576, -0.02017212, 0.038574...
http://arxiv.org/abs/2305.12696v2
Learning Interpretable Style Embeddings via Prompting LLMs
Ajay Patel, Delip Rao, Ansh Kothary, Kathleen McKeown, Chris Callison-Burch
cs.CL
Style representation learning builds content-independent representations of author style in text. Stylometry, the analysis of style in text, is often performed by expert forensic linguists and no large dataset of stylometric annotations exists for training. Current style representation learning uses neural methods to d...
http://arxiv.org/pdf/2305.12696v2
2023-10-09T19:20:32
2023-05-22T04:07:54
{ "title": [ -0.045715332, -0.0035648346, -0.048065186, -0.040039062, -0.01902771, 0.004627228, -0.084350586, 0.0015287399, 0.004673004, -0.010864258, -0.034118652, 0.025115967, -0.01687622, -0.02470398, 0.021102905, 0.0006380081, 0.014778137, 0....
http://arxiv.org/abs/2305.12694v1
Automatic Spell Checker and Correction for Under-represented Spoken Languages: Case Study on Wolof
Thierno Ibrahima Cissé, Fatiha Sadat
cs.CL, cs.AI
This paper presents a spell checker and correction tool specifically designed for Wolof, an under-represented spoken language in Africa. The proposed spell checker leverages a combination of a trie data structure, dynamic programming, and the weighted Levenshtein distance to generate suggestions for misspelled words. W...
http://arxiv.org/pdf/2305.12694v1
2023-05-22T04:03:20
2023-05-22T04:03:20
{ "title": [ -0.0067825317, 0.005935669, -0.0066490173, -0.010864258, -0.012786865, -0.0071372986, -0.02670288, -0.037963867, 0.0024032593, 0.059631348, -0.027008057, 0.009605408, 0.05911255, -0.018539429, -0.042785645, 0.018295288, -0.023269653, ...
http://arxiv.org/abs/2305.12692v1
MetaAdapt: Domain Adaptive Few-Shot Misinformation Detection via Meta Learning
Zhenrui Yue, Huimin Zeng, Yang Zhang, Lanyu Shang, Dong Wang
cs.CL, cs.AI
With emerging topics (e.g., COVID-19) on social media as a source for the spreading misinformation, overcoming the distributional shifts between the original training domain (i.e., source domain) and such target domains remains a non-trivial task for misinformation detection. This presents an elusive challenge for earl...
http://arxiv.org/pdf/2305.12692v1
2023-05-22T04:00:38
2023-05-22T04:00:38
{ "title": [ 0.015716553, -0.002046585, -0.030273438, -0.026275635, 0.007144928, -0.020263672, 0.0044822693, -0.022979736, -0.01940918, 0.029220581, 0.0031204224, 0.016586304, -0.014160156, -0.023788452, -0.03564453, -0.009384155, -0.021575928, 0...
http://arxiv.org/abs/2305.12689v2
FIT: Far-reaching Interleaved Transformers
Ting Chen, Lala Li
cs.LG, cs.AI, cs.CL, cs.CV
We present FIT: a transformer-based architecture with efficient self-attention and adaptive computation. Unlike original transformers, which operate on a single sequence of data tokens, we divide the data tokens into groups, with each group being a shorter sequence of tokens. We employ two types of transformer layers: ...
http://arxiv.org/pdf/2305.12689v2
2023-05-25T16:27:30
2023-05-22T03:56:44
{ "title": [ 0.0066833496, -0.0017433167, -0.032226562, -0.056610107, -0.028503418, -0.007965088, -0.040740967, -0.0071792603, -0.06262207, 0.0146865845, -0.025543213, 0.015289307, 0.007133484, -0.044067383, 0.033843994, -0.038269043, 0.021148682, ...
http://arxiv.org/abs/2305.14380v2
Finding the Pillars of Strength for Multi-Head Attention
Jinjie Ni, Rui Mao, Zonglin Yang, Han Lei, Erik Cambria
cs.LG, cs.CL, I.2.0; I.2.7
Recent studies have revealed some issues of Multi-Head Attention (MHA), e.g., redundancy and over-parameterization. Specifically, the heads of MHA were originally designed to attend to information from different representation subspaces, whereas prior studies found that some attention heads likely learn similar feature...
http://arxiv.org/pdf/2305.14380v2
2023-10-15T04:10:13
2023-05-22T03:44:44
{ "title": [ 0.015159607, 0.021438599, -0.047424316, -0.05050659, 0.004421234, -0.02319336, -0.016830444, -0.022094727, -0.020599365, 0.048614502, -0.011558533, 0.010643005, 0.014205933, -0.012870789, -0.03857422, 0.0055160522, 0.014030457, 0.017...
http://arxiv.org/abs/2305.12680v2
G3Detector: General GPT-Generated Text Detector
Haolan Zhan, Xuanli He, Qiongkai Xu, Yuxiang Wu, Pontus Stenetorp
cs.CL
The burgeoning progress in the field of Large Language Models (LLMs) heralds significant benefits due to their unparalleled capacities. However, it is critical to acknowledge the potential misuse of these models, which could give rise to a spectrum of social and ethical dilemmas. Despite numerous preceding efforts cent...
http://arxiv.org/pdf/2305.12680v2
2023-08-04T06:07:49
2023-05-22T03:35:00
{ "title": [ -0.030181885, 0.006389618, -0.03555298, -0.026748657, 0.0019512177, -0.008079529, -0.028137207, -0.03488159, -0.014907837, 0.033447266, -0.02999878, 0.010795593, 0.004497528, -0.011512756, -0.031585693, 0.0001989603, -0.018325806, 0....
http://arxiv.org/abs/2305.12678v2
Gradient-Boosted Decision Tree for Listwise Context Model in Multimodal Review Helpfulness Prediction
Thong Nguyen, Xiaobao Wu, Xinshuai Dong, Anh Tuan Luu, Cong-Duy Nguyen, Zhen Hai, Lidong Bing
cs.CL
Multimodal Review Helpfulness Prediction (MRHP) aims to rank product reviews based on predicted helpfulness scores and has been widely applied in e-commerce via presenting customers with useful reviews. Previous studies commonly employ fully-connected neural networks (FCNNs) as the final score predictor and pairwise lo...
http://arxiv.org/pdf/2305.12678v2
2023-05-25T04:51:43
2023-05-22T03:31:00
{ "title": [ 0.004611969, 0.0054969788, -0.024414062, -0.050476074, -0.013298035, -0.004425049, -0.045196533, -0.011505127, 0.0070266724, 0.044036865, -0.06793213, 0.017150879, 0.027130127, 0.0038871765, -0.03564453, -0.01663208, -0.009559631, 0....
http://arxiv.org/abs/2305.12676v3
Exploring Energy-based Language Models with Different Architectures and Training Methods for Speech Recognition
Hong Liu, Zhaobiao Lv, Zhijian Ou, Wenbo Zhao, Qing Xiao
cs.CL
Energy-based language models (ELMs) parameterize an unnormalized distribution for natural sentences and are radically different from popular autoregressive language models (ALMs). As an important application, ELMs have been successfully used as a means for calculating sentence scores in speech recognition, but they all...
http://arxiv.org/pdf/2305.12676v3
2023-05-29T06:38:21
2023-05-22T03:28:48
{ "title": [ -0.019088745, 0.010734558, -0.042144775, -0.04727173, 0.00472641, -0.010902405, -0.03970337, 0.011512756, -0.023101807, 0.048980713, -0.02748108, 0.009880066, 0.01612854, -0.024032593, -0.025482178, 0.004070282, -0.0052986145, 0.0372...
http://arxiv.org/abs/2305.12675v1
A Frustratingly Simple Decoding Method for Neural Text Generation
Haoran Yang, Deng Cai, Huayang Li, Wei Bi, Wai Lam, Shuming Shi
cs.CL
We introduce a frustratingly simple, super efficient and surprisingly effective decoding method, which we call Frustratingly Simple Decoding (FSD), for neural text generation. The idea behind FSD is straightforward: we build an anti-LM based on previously generated text and use this anti-LM to penalize future generatio...
http://arxiv.org/pdf/2305.12675v1
2023-05-22T03:28:47
2023-05-22T03:28:47
{ "title": [ -0.034698486, -0.010284424, -0.027786255, -0.031143188, -0.011375427, 0.006061554, -0.031585693, -0.011817932, 0.0038757324, 0.051635742, -0.01008606, 0.0077819824, 0.012237549, -0.0070343018, -0.05630493, 0.01852417, -0.006439209, 0...
http://arxiv.org/abs/2305.12660v2
Beneath Surface Similarity: Large Language Models Make Reasonable Scientific Analogies after Structure Abduction
Siyu Yuan, Jiangjie Chen, Xuyang Ge, Yanghua Xiao, Deqing Yang
cs.CL, cs.AI
The vital role of analogical reasoning in human cognition allows us to grasp novel concepts by linking them with familiar ones through shared relational structures. Despite the attention previous research has given to word analogies, this work suggests that Large Language Models (LLMs) often overlook the structures tha...
http://arxiv.org/pdf/2305.12660v2
2023-10-10T11:36:08
2023-05-22T03:04:06
{ "title": [ -0.039215088, -0.0079956055, -0.007007599, -0.034301758, -0.025787354, 0.008178711, -0.04043579, -0.028625488, -0.005367279, 0.019439697, -0.0018482208, 0.010643005, 0.0068893433, -0.029876709, -0.013633728, 0.011672974, -0.015731812, ...
http://arxiv.org/abs/2305.12647v1
Reflective Linguistic Programming (RLP): A Stepping Stone in Socially-Aware AGI (SocialAGI)
Kevin A. Fischer
cs.AI, cs.CL, cs.HC, cs.LG
This paper presents Reflective Linguistic Programming (RLP), a unique approach to conversational AI that emphasizes self-awareness and strategic planning. RLP encourages models to introspect on their own predefined personality traits, emotional responses to incoming messages, and planned strategies, enabling contextual...
http://arxiv.org/pdf/2305.12647v1
2023-05-22T02:43:15
2023-05-22T02:43:15
{ "title": [ -0.05480957, -0.0063819885, -0.03302002, -0.02357483, -0.03656006, 0.015670776, -0.07330322, -0.008956909, -0.035980225, 0.03579712, 0.01436615, -0.03173828, 0.010498047, -0.05670166, -0.026641846, -0.021835327, 0.0073661804, -0.0290...
http://arxiv.org/abs/2305.12641v2
Beyond Words: A Comprehensive Survey of Sentence Representations
Abhinav Ramesh Kashyap, Thanh-Tung Nguyen, Viktor Schlegel, Stefan Winkler, See-Kiong Ng, Soujanya Poria
cs.CL
Sentence representations have become a critical component in natural language processing applications, such as retrieval, question answering, and text classification. They capture the semantics and meaning of a sentence, enabling machines to understand and reason over human language. In recent years, significant progre...
http://arxiv.org/pdf/2305.12641v2
2023-05-23T15:33:07
2023-05-22T02:31:15
{ "title": [ 0.0067253113, -0.004219055, -0.014511108, -0.06707764, 0.01927185, 0.01398468, -0.052581787, -0.015640259, 0.029190063, 0.0335083, -0.016662598, 0.008071899, 0.008125305, -0.0028629303, -0.049987793, 0.018493652, -0.013900757, 0.0396...
http://arxiv.org/abs/2305.12634v2
Data-efficient Active Learning for Structured Prediction with Partial Annotation and Self-Training
Zhisong Zhang, Emma Strubell, Eduard Hovy
cs.CL
In this work we propose a pragmatic method that reduces the annotation cost for structured label spaces using active learning. Our approach leverages partial annotation, which reduces labeling costs for structured outputs by selecting only the most informative sub-structures for annotation. We also utilize self-trainin...
http://arxiv.org/pdf/2305.12634v2
2023-10-19T02:13:51
2023-05-22T01:58:42
{ "title": [ -0.005130768, -0.009689331, -0.023223877, -0.10437012, 0.008674622, 0.00075149536, -0.038360596, 0.011268616, 0.015075684, 0.021392822, 0.005340576, 0.0077209473, -0.030899048, -0.032043457, -0.023376465, 0.025497437, 0.030319214, -0...
http://arxiv.org/abs/2305.12628v1
Duplex Diffusion Models Improve Speech-to-Speech Translation
Xianchao Wu
cs.CL, cs.LG, cs.SD, eess.AS
Speech-to-speech translation is a typical sequence-to-sequence learning task that naturally has two directions. How to effectively leverage bidirectional supervision signals to produce high-fidelity audio for both directions? Existing approaches either train two separate models or a multitask-learned model with low eff...
http://arxiv.org/pdf/2305.12628v1
2023-05-22T01:39:40
2023-05-22T01:39:40
{ "title": [ -0.008117676, 0.006175995, -0.018692017, -0.049682617, -0.027694702, -0.009361267, -0.06274414, 0.00092077255, -0.035308838, 0.062194824, -0.012641907, 0.0007009506, 0.05041504, -0.04336548, -0.0009784698, 0.0053253174, 0.0015678406, ...
http://arxiv.org/abs/2305.12627v1
MvP: Multi-view Prompting Improves Aspect Sentiment Tuple Prediction
Zhibin Gou, Qingyan Guo, Yujiu Yang
cs.CL, cs.AI
Generative methods greatly promote aspect-based sentiment analysis via generating a sequence of sentiment elements in a specified format. However, existing studies usually predict sentiment elements in a fixed order, which ignores the effect of the interdependence of the elements in a sentiment tuple and the diversity ...
http://arxiv.org/pdf/2305.12627v1
2023-05-22T01:32:50
2023-05-22T01:32:50
{ "title": [ 0.011138916, -0.019927979, -0.029083252, -0.106933594, -0.023162842, -0.00030851364, -0.0680542, 0.025146484, 0.036468506, 0.03515625, -0.030227661, -0.0027770996, 0.015808105, -0.013473511, -0.057922363, 0.008880615, 0.00027537346, ...
http://arxiv.org/abs/2305.12620v1
Keeping Up with the Language Models: Robustness-Bias Interplay in NLI Data and Models
Ioana Baldini, Chhavi Yadav, Payel Das, Kush R. Varshney
cs.CL
Auditing unwanted social bias in language models (LMs) is inherently hard due to the multidisciplinary nature of the work. In addition, the rapid evolution of LMs can make benchmarks irrelevant in no time. Bias auditing is further complicated by LM brittleness: when a presumably biased outcome is observed, is it due to...
http://arxiv.org/pdf/2305.12620v1
2023-05-22T01:02:45
2023-05-22T01:02:45
{ "title": [ -0.020233154, 0.003643036, -0.04046631, -0.042633057, -0.004207611, -0.011207581, -0.053771973, -0.026779175, 0.0014572144, 0.018661499, -0.036895752, 0.0012893677, 0.009429932, -0.013542175, -0.040405273, -0.0067710876, -0.025512695, ...
http://arxiv.org/abs/2305.12612v1
PrOnto: Language Model Evaluations for 859 Languages
Luke Gessler
cs.CL
Evaluation datasets are critical resources for measuring the quality of pretrained language models. However, due to the high cost of dataset annotation, these resources are scarce for most languages other than English, making it difficult to assess the quality of language models. In this work, we present a new method f...
http://arxiv.org/pdf/2305.12612v1
2023-05-22T00:33:52
2023-05-22T00:33:52
{ "title": [ -0.026641846, -0.012145996, -0.05041504, -0.06665039, 0.012008667, -0.014587402, -0.06298828, 0.00030851364, -0.0058403015, 0.03543091, 0.013442993, -0.017959595, 0.032073975, -0.058624268, -0.023986816, 0.022613525, -0.0020923615, 0...
http://arxiv.org/abs/2305.12606v2
Comparison of Multilingual Self-Supervised and Weakly-Supervised Speech Pre-Training for Adaptation to Unseen Languages
Andrew Rouditchenko, Sameer Khurana, Samuel Thomas, Rogerio Feris, Leonid Karlinsky, Hilde Kuehne, David Harwath, Brian Kingsbury, James Glass
cs.CL, cs.SD, eess.AS
Recent models such as XLS-R and Whisper have made multilingual speech technologies more accessible by pre-training on audio from around 100 spoken languages each. However, there are thousands of spoken languages worldwide, and adapting to new languages is an important problem. In this work, we aim to understand which m...
http://arxiv.org/pdf/2305.12606v2
2023-05-31T01:27:41
2023-05-21T23:53:12
{ "title": [ 0.0068244934, 0.0014696121, -0.0262146, -0.06555176, -0.01033783, -0.005607605, -0.05114746, 0.012458801, -0.050689697, 0.042907715, 0.005935669, 0.011734009, 0.0019454956, -0.040985107, -0.006095886, -0.009506226, 0.0056533813, 0.03...
http://arxiv.org/abs/2305.12599v2
Enhancing Logical Reasoning of Large Language Models through Logic-Driven Data Augmentation
Qiming Bao, Alex Yuxuan Peng, Zhenyun Deng, Wanjun Zhong, Gael Gendron, Timothy Pistotti, Neset Tan, Nathan Young, Yang Chen, Yonghua Zhu, Paul Denny, Michael Witbrock, Jiamou Liu
cs.CL, cs.AI
Combining large language models with logical reasoning enhance their capacity to address problems in a robust and reliable manner. Nevertheless, the intricate nature of logical reasoning poses challenges to gathering reliable data from web for building comprehensive training datasets, subsequently affecting the perform...
http://arxiv.org/pdf/2305.12599v2
2023-10-14T09:44:32
2023-05-21T23:16:26
{ "title": [ -0.013832092, -0.017807007, -0.07470703, -0.10913086, 0.0073127747, -0.0006146431, -0.079956055, 0.043792725, 0.005218506, 0.019454956, -0.034210205, 0.013633728, -0.0009860992, -0.008056641, -0.02571106, 0.0051460266, -0.020126343, ...
http://arxiv.org/abs/2305.12594v1
Modeling User Satisfaction Dynamics in Dialogue via Hawkes Process
Fanghua Ye, Zhiyuan Hu, Emine Yilmaz
cs.CL
Dialogue systems have received increasing attention while automatically evaluating their performance remains challenging. User satisfaction estimation (USE) has been proposed as an alternative. It assumes that the performance of a dialogue system can be measured by user satisfaction and uses an estimator to simulate us...
http://arxiv.org/pdf/2305.12594v1
2023-05-21T23:04:14
2023-05-21T23:04:14
{ "title": [ 0.01222229, 0.010025024, -0.011978149, -0.013931274, -0.0064888, -0.013801575, -0.022399902, -0.02835083, 0.01260376, 0.022491455, -0.009925842, -0.015342712, -0.008598328, -0.026779175, -0.06964111, 0.0004146099, -0.029922485, 0.043...
http://arxiv.org/abs/2305.13342v1
On the Limitations of Simulating Active Learning
Katerina Margatina, Nikolaos Aletras
cs.LG, cs.CL
Active learning (AL) is a human-and-model-in-the-loop paradigm that iteratively selects informative unlabeled data for human annotation, aiming to improve over random sampling. However, performing AL experiments with human annotations on-the-fly is a laborious and expensive process, thus unrealistic for academic resear...
http://arxiv.org/pdf/2305.13342v1
2023-05-21T22:52:13
2023-05-21T22:52:13
{ "title": [ 0.02168274, 0.0013847351, 0.015792847, -0.010437012, -0.0006046295, -0.023406982, 0.036254883, -0.039001465, -0.022140503, 0.03765869, -0.005809784, -0.005302429, -0.044799805, -0.03164673, -0.02973938, 0.010726929, 0.012298584, 0.02...
http://arxiv.org/abs/2305.12586v1
Enhancing Few-shot Text-to-SQL Capabilities of Large Language Models: A Study on Prompt Design Strategies
Linyong Nan, Yilun Zhao, Weijin Zou, Narutatsu Ri, Jaesung Tae, Ellen Zhang, Arman Cohan, Dragomir Radev
cs.CL
In-context learning (ICL) has emerged as a new approach to various natural language processing tasks, utilizing large language models (LLMs) to make predictions based on context that has been supplemented with a few examples or task-specific instructions. In this paper, we aim to extend this method to question answerin...
http://arxiv.org/pdf/2305.12586v1
2023-05-21T22:44:25
2023-05-21T22:44:25
{ "title": [ -0.011497498, -0.03552246, -0.04232788, -0.052520752, -0.011291504, -0.004840851, -0.04522705, 0.022491455, 0.027328491, 0.044921875, -0.05078125, 0.007583618, 0.024673462, 0.00075912476, -0.066589355, -0.007633209, 0.0070991516, 0.0...
http://arxiv.org/abs/2305.12580v2
A Framework for Bidirectional Decoding: Case Study in Morphological Inflection
Marc E. Canby, Julia Hockenmaier
cs.CL
Transformer-based encoder-decoder models that generate outputs in a left-to-right fashion have become standard for sequence-to-sequence tasks. In this paper, we propose a framework for decoding that produces sequences from the "outside-in": at each step, the model chooses to generate a token on the left, on the right, ...
http://arxiv.org/pdf/2305.12580v2
2023-10-30T05:51:34
2023-05-21T22:08:31
{ "title": [ -0.015396118, -0.008338928, -0.026885986, -0.07385254, -0.009567261, 0.00422287, -0.07122803, 0.012786865, 0.004737854, 0.058776855, -0.014976501, -0.0055274963, 0.02671814, -0.025314331, -0.057769775, 0.009407043, -0.025177002, 0.06...
http://arxiv.org/abs/2305.12579v1
Hystoc: Obtaining word confidences for fusion of end-to-end ASR systems
Karel Beneš, Martin Kocour, Lukáš Burget
cs.CL, cs.SD, eess.AS
End-to-end (e2e) systems have recently gained wide popularity in automatic speech recognition. However, these systems do generally not provide well-calibrated word-level confidences. In this paper, we propose Hystoc, a simple method for obtaining word-level confidences from hypothesis-level scores. Hystoc is an iterati...
http://arxiv.org/pdf/2305.12579v1
2023-05-21T22:06:14
2023-05-21T22:06:14
{ "title": [ -0.032928467, -0.0024662018, -0.064331055, -0.070495605, 0.008132935, 0.0031795502, -0.05670166, 0.010665894, 0.0013828278, 0.03427124, -0.03540039, -0.0018482208, 0.03314209, -0.023956299, -0.031051636, 0.006706238, -0.013023376, 0....
http://arxiv.org/abs/2305.12576v2
Automated Few-shot Classification with Instruction-Finetuned Language Models
Rami Aly, Xingjian Shi, Kaixiang Lin, Aston Zhang, Andrew Gordon Wilson
cs.CL
A particularly successful class of approaches for few-shot learning combines language models with prompts -- hand-crafted task descriptions that complement data samples. However, designing prompts by hand for each task commonly requires domain knowledge and substantial guesswork. We observe, in the context of classific...
http://arxiv.org/pdf/2305.12576v2
2023-10-21T06:46:51
2023-05-21T21:50:27
{ "title": [ 0.0017652512, -0.025497437, -0.015914917, -0.038269043, -0.023025513, -0.006374359, -0.029327393, -0.0043258667, -0.0028419495, 0.023147583, -0.03616333, 0.0211792, -0.02583313, -0.004676819, -0.049346924, -0.026062012, 0.00037527084, ...
http://arxiv.org/abs/2305.12567v1
Model-Generated Pretraining Signals Improves Zero-Shot Generalization of Text-to-Text Transformers
Linyuan Gong, Chenyan Xiong, Xiaodong Liu, Payal Bajaj, Yiqing Xie, Alvin Cheung, Jianfeng Gao, Xia Song
cs.CL
This paper explores the effectiveness of model-generated signals in improving zero-shot generalization of text-to-text Transformers such as T5. We study various designs to pretrain T5 using an auxiliary model to construct more challenging token replacements for the main model to denoise. Key aspects under study include...
http://arxiv.org/pdf/2305.12567v1
2023-05-21T21:06:23
2023-05-21T21:06:23
{ "title": [ 0.02319336, 0.00970459, 0.0005145073, -0.057250977, -0.015823364, 0.0024662018, -0.04208374, -0.02268982, 0.00674057, 0.05630493, -0.018920898, 0.029693604, 0.017044067, -0.0127334595, -0.033172607, 0.011711121, 0.0073013306, 0.04678...
http://arxiv.org/abs/2305.12565v1
Understanding the Effect of Data Augmentation on Knowledge Distillation
Ziqi Wang, Chi Han, Wenxuan Bao, Heng Ji
cs.CL
Knowledge distillation (KD) requires sufficient data to transfer knowledge from large-scale teacher models to small-scale student models. Therefore, data augmentation has been widely used to mitigate the shortage of data under specific scenarios. Classic data augmentation techniques, such as synonym replacement and k-n...
http://arxiv.org/pdf/2305.12565v1
2023-05-21T21:02:55
2023-05-21T21:02:55
{ "title": [ 0.013084412, 0.013641357, -0.09210205, -0.08514404, 0.026687622, -0.037902832, -0.042114258, 0.011177063, 0.014953613, 0.010650635, -0.06567383, -0.008758545, -0.010032654, -0.010215759, -0.029830933, -0.02949524, 0.006008148, 0.0147...
http://arxiv.org/abs/2305.12564v1
ChatGPT Is More Likely to Be Perceived as Male Than Female
Jared Wong, Jin Kim
cs.HC, cs.AI, cs.CL, cs.LG
We investigate how people perceive ChatGPT, and, in particular, how they assign human-like attributes such as gender to the chatbot. Across five pre-registered studies (N = 1,552), we find that people are more likely to perceive ChatGPT to be male than female. Specifically, people perceive male gender identity (1) foll...
http://arxiv.org/pdf/2305.12564v1
2023-05-21T20:57:12
2023-05-21T20:57:12
{ "title": [ -0.002544403, 0.01486969, -0.04159546, 0.003419876, -0.01776123, -0.022476196, -0.011169434, -0.0085372925, -0.0067214966, 0.063964844, -0.034454346, -0.012321472, 0.041229248, -0.012863159, -0.002664566, 0.0146102905, -0.037109375, ...
http://arxiv.org/abs/2305.12563v1
A Symbolic Framework for Systematic Evaluation of Mathematical Reasoning with Transformers
Jordan Meadows, Marco Valentino, Damien Teney, Andre Freitas
cs.CL, cs.LG
Whether Transformers can learn to apply symbolic rules and generalise to out-of-distribution examples is an open research question. In this paper, we devise a data generation method for producing intricate mathematical derivations, and systematically perturb them with respect to syntax, structure, and semantics. Our ta...
http://arxiv.org/pdf/2305.12563v1
2023-05-21T20:40:37
2023-05-21T20:40:37
{ "title": [ 0.014633179, 0.012939453, -0.009063721, -0.056793213, 0.021148682, -0.00037670135, -0.0109939575, -0.019729614, -0.007858276, 0.014289856, -0.027374268, -0.0074005127, -0.015274048, -0.014518738, -0.04006958, -0.017486572, -0.00705719, ...
http://arxiv.org/abs/2305.12552v1
Wav2SQL: Direct Generalizable Speech-To-SQL Parsing
Huadai Liu, Rongjie Huang, Jinzheng He, Gang Sun, Ran Shen, Xize Cheng, Zhou Zhao
cs.CL, cs.SD, eess.AS
Speech-to-SQL (S2SQL) aims to convert spoken questions into SQL queries given relational databases, which has been traditionally implemented in a cascaded manner while facing the following challenges: 1) model training is faced with the major issue of data scarcity, where limited parallel data is available; and 2) the ...
http://arxiv.org/pdf/2305.12552v1
2023-05-21T19:26:46
2023-05-21T19:26:46
{ "title": [ 0.002796173, -0.0016050339, -0.0637207, -0.105285645, 0.03579712, 0.0017166138, -0.05496216, 0.031707764, 0.01008606, 0.049743652, -0.0284729, -0.005554199, 0.021606445, -0.015319824, -0.061401367, -0.010543823, -0.002620697, 0.02526...
http://arxiv.org/abs/2305.12544v1
A PhD Student's Perspective on Research in NLP in the Era of Very Large Language Models
Oana Ignat, Zhijing Jin, Artem Abzaliev, Laura Biester, Santiago Castro, Naihao Deng, Xinyi Gao, Aylin Gunal, Jacky He, Ashkan Kazemi, Muhammad Khalifa, Namho Koh, Andrew Lee, Siyang Liu, Do June Min, Shinka Mori, Joan Nwatu, Veronica Perez-Rosas, Siqi Shen, Zekun Wang, Winston Wu, Rada Mihalcea
cs.CL, cs.AI
Recent progress in large language models has enabled the deployment of many generative NLP applications. At the same time, it has also led to a misleading public discourse that ``it's all been solved.'' Not surprisingly, this has in turn made many NLP researchers -- especially those at the beginning of their career -- ...
http://arxiv.org/pdf/2305.12544v1
2023-05-21T19:06:30
2023-05-21T19:06:30
{ "title": [ -0.008079529, -0.0060920715, -0.071777344, -0.044403076, -0.019210815, -0.045043945, -0.017944336, 0.042633057, 0.012962341, 0.014854431, -0.045928955, -0.029006958, 0.009552002, -0.018432617, -0.017059326, -0.03741455, 0.001127243, ...
http://arxiv.org/abs/2305.12542v1
ToxBuster: In-game Chat Toxicity Buster with BERT
Zachary Yang, Yasmine Maricar, MohammadReza Davari, Nicolas Grenon-Godbout, Reihaneh Rabbany
cs.CL, cs.CY
Detecting toxicity in online spaces is challenging and an ever more pressing problem given the increase in social media and gaming consumption. We introduce ToxBuster, a simple and scalable model trained on a relatively large dataset of 194k lines of game chat from Rainbow Six Siege and For Honor, carefully annotated f...
http://arxiv.org/pdf/2305.12542v1
2023-05-21T18:53:26
2023-05-21T18:53:26
{ "title": [ 0.030944824, -0.02229309, 0.008804321, -0.039398193, -0.058563232, -0.008346558, -0.044769287, -0.022064209, 0.010948181, 0.039794922, 0.011817932, 0.010536194, 0.07122803, -0.04751587, 0.025421143, 0.008300781, 0.0035972595, 0.00736...
http://arxiv.org/abs/2305.12535v1
Explaining How Transformers Use Context to Build Predictions
Javier Ferrando, Gerard I. Gállego, Ioannis Tsiamas, Marta R. Costa-jussà
cs.CL, cs.AI, cs.LG
Language Generation Models produce words based on the previous context. Although existing methods offer input attributions as explanations for a model's prediction, it is still unclear how prior words affect the model's decision throughout the layers. In this work, we leverage recent advances in explainability of the T...
http://arxiv.org/pdf/2305.12535v1
2023-05-21T18:29:10
2023-05-21T18:29:10
{ "title": [ 0.05126953, -0.0076065063, -0.009292603, -0.06036377, -0.0051956177, -0.0146102905, 0.0019798279, 0.007713318, -0.041290283, 0.044799805, -0.05734253, 0.011695862, 0.0018167496, -0.009155273, 0.03793335, -0.04031372, 0.026779175, 0.0...
http://arxiv.org/abs/2305.12532v4
Multilingual Simplification of Medical Texts
Sebastian Joseph, Kathryn Kazanas, Keziah Reina, Vishnesh J. Ramanathan, Wei Xu, Byron C. Wallace, Junyi Jessy Li
cs.CL
Automated text simplification aims to produce simple versions of complex texts. This task is especially useful in the medical domain, where the latest medical findings are typically communicated via complex and technical articles. This creates barriers for laypeople seeking access to up-to-date medical findings, conseq...
http://arxiv.org/pdf/2305.12532v4
2023-10-18T04:20:05
2023-05-21T18:25:07
{ "title": [ -0.027328491, -0.0035572052, -0.037139893, -0.077819824, -0.022491455, 0.01927185, -0.080444336, 0.013153076, -0.0062713623, 0.075683594, 0.023086548, 0.011108398, 0.039520264, -0.033935547, -0.011543274, 0.021820068, 0.032958984, 0....
http://arxiv.org/abs/2305.12524v2
TheoremQA: A Theorem-driven Question Answering dataset
Wenhu Chen, Ming Yin, Max Ku, Pan Lu, Yixin Wan, Xueguang Ma, Jianyu Xu, Xinyi Wang, Tony Xia
cs.CL, cs.AI
The recent LLMs like GPT-4 and PaLM-2 have made tremendous progress in solving fundamental math problems like GSM8K by achieving over 90% accuracy. However, their capabilities to solve more challenging math problems which require domain-specific knowledge (i.e. theorem) have yet to be investigated. In this paper, we in...
http://arxiv.org/pdf/2305.12524v2
2023-05-23T22:35:20
2023-05-21T17:51:35
{ "title": [ -0.025238037, -0.005832672, -0.050811768, -0.06036377, 0.0098724365, -0.0028095245, -0.023971558, 0.01159668, 0.026611328, 0.012924194, -0.043395996, -0.00084495544, -0.017150879, 0.0071411133, -0.061767578, -0.011039734, -0.004627228, ...
http://arxiv.org/abs/2305.12519v1
GPT Paternity Test: GPT Generated Text Detection with GPT Genetic Inheritance
Xiao Yu, Yuang Qi, Kejiang Chen, Guoqiang Chen, Xi Yang, Pengyuan Zhu, Weiming Zhang, Nenghai Yu
cs.CL, cs.AI, cs.LG
Large Language Models (LLMs) can generate texts that carry the risk of various misuses, including plagiarism, planting fake reviews on e-commerce platforms, or creating fake social media postings that can sway election results. Detecting whether a text is machine-generated has thus become increasingly important. While ...
http://arxiv.org/pdf/2305.12519v1
2023-05-21T17:26:16
2023-05-21T17:26:16
{ "title": [ -0.027816772, 0.010948181, -0.039276123, -0.011550903, -0.021118164, -0.010681152, -0.03753662, -0.016494751, -0.011955261, 0.02482605, -0.025756836, -0.024551392, 0.01727295, -0.038513184, -0.018478394, -0.027267456, -0.021713257, 0...
http://arxiv.org/abs/2305.12518v1
VAKTA-SETU: A Speech-to-Speech Machine Translation Service in Select Indic Languages
Shivam Mhaskar, Vineet Bhat, Akshay Batheja, Sourabh Deoghare, Paramveer Choudhary, Pushpak Bhattacharyya
cs.CL
In this work, we present our deployment-ready Speech-to-Speech Machine Translation (SSMT) system for English-Hindi, English-Marathi, and Hindi-Marathi language pairs. We develop the SSMT system by cascading Automatic Speech Recognition (ASR), Disfluency Correction (DC), Machine Translation (MT), and Text-to-Speech Synt...
http://arxiv.org/pdf/2305.12518v1
2023-05-21T17:23:54
2023-05-21T17:23:54
{ "title": [ -0.029663086, -0.01436615, -0.03994751, -0.041778564, -0.008598328, -0.014053345, -0.029159546, 0.010131836, -0.026885986, 0.05899048, -0.006290436, -0.0037021637, 0.029144287, -0.024246216, -0.058380127, -0.013343811, -0.0040130615, ...
http://arxiv.org/abs/2305.12517v2
Retrieving Texts based on Abstract Descriptions
Shauli Ravfogel, Valentina Pyatkin, Amir DN Cohen, Avshalom Manevich, Yoav Goldberg
cs.CL, cs.IR, cs.LG
While instruction-tuned Large Language Models (LLMs) excel at extracting information from text, they are not suitable for locating texts conforming to a given description in a large document collection (semantic retrieval). Similarity search over embedding vectors does allow to perform retrieval by query, but the simil...
http://arxiv.org/pdf/2305.12517v2
2023-10-22T17:38:42
2023-05-21T17:14:31
{ "title": [ -0.0022525787, 0.0056915283, -0.0141067505, -0.05529785, -0.014053345, -0.0033779144, -0.04446411, -0.02760315, 0.012886047, 0.06555176, -0.03152466, 0.00040221214, 0.01751709, -0.0024337769, -0.064819336, 0.0030326843, 0.0082092285, ...
http://arxiv.org/abs/2305.12510v1
A Deeper (Autoregressive) Approach to Non-Convergent Discourse Parsing
Yoav Tulpan, Oren Tsur
cs.CL, cs.SI
Online social platforms provide a bustling arena for information-sharing and for multi-party discussions. Various frameworks for dialogic discourse parsing were developed and used for the processing of discussions and for predicting the productivity of a dialogue. However, most of these frameworks are not suitable for ...
http://arxiv.org/pdf/2305.12510v1
2023-05-21T17:04:21
2023-05-21T17:04:21
{ "title": [ -0.00052547455, 0.012840271, -0.026824951, -0.08959961, 0.016342163, 0.00749588, -0.06573486, 0.004787445, 0.007247925, 0.015533447, -0.016021729, 0.0027542114, -0.020507812, -0.0062789917, -0.06384277, -0.010818481, -0.023483276, 0....
http://arxiv.org/abs/2305.12501v1
Exploring How Generative Adversarial Networks Learn Phonological Representations
Jingyi Chen, Micha Elsner
cs.CL, cs.SD, eess.AS
This paper explores how Generative Adversarial Networks (GANs) learn representations of phonological phenomena. We analyze how GANs encode contrastive and non-contrastive nasality in French and English vowels by applying the ciwGAN architecture (Begus 2021a). Begus claims that ciwGAN encodes linguistically meaningful r...
http://arxiv.org/pdf/2305.12501v1
2023-05-21T16:37:21
2023-05-21T16:37:21
{ "title": [ -0.021697998, -0.0038414001, -0.025665283, -0.02897644, -0.011444092, -0.0049438477, -0.02003479, 0.01524353, -0.022476196, 0.041748047, -0.010009766, 0.01525116, 0.00083732605, -0.02746582, -0.02861023, 0.003396988, -0.013450623, 0....
http://arxiv.org/abs/2305.12498v2
Multi-Head State Space Model for Speech Recognition
Yassir Fathullah, Chunyang Wu, Yuan Shangguan, Junteng Jia, Wenhan Xiong, Jay Mahadeokar, Chunxi Liu, Yangyang Shi, Ozlem Kalinli, Mike Seltzer, Mark J. F. Gales
eess.AS, cs.AI, cs.CL, cs.LG, cs.SD
State space models (SSMs) have recently shown promising results on small-scale sequence and language modelling tasks, rivalling and outperforming many attention-based approaches. In this paper, we propose a multi-head state space (MH-SSM) architecture equipped with special gating mechanisms, where parallel heads are ta...
http://arxiv.org/pdf/2305.12498v2
2023-05-25T21:55:58
2023-05-21T16:28:57
{ "title": [ -0.01739502, 0.012496948, -0.04534912, -0.06488037, 0.006706238, -0.00667572, -0.055236816, 0.005908966, -0.0020008087, 0.038513184, -0.034301758, 0.009025574, -0.016098022, -0.0126571655, -0.041900635, -0.00082206726, -0.0032081604, ...
http://arxiv.org/abs/2305.12495v2
Fair Without Leveling Down: A New Intersectional Fairness Definition
Gaurav Maheshwari, Aurélien Bellet, Pascal Denis, Mikaela Keller
cs.LG, cs.CL, cs.CY
In this work, we consider the problem of intersectional group fairness in the classification setting, where the objective is to learn discrimination-free models in the presence of several intersecting sensitive groups. First, we illustrate various shortcomings of existing fairness measures commonly used to capture inte...
http://arxiv.org/pdf/2305.12495v2
2023-11-07T10:19:02
2023-05-21T16:15:12
{ "title": [ -0.0005578995, 0.017501831, -0.04434204, -0.060180664, 0.024230957, -0.021408081, -0.01876831, -0.018447876, -0.012672424, 0.034942627, -0.03375244, 0.019515991, 0.008232117, -0.0155181885, -0.03564453, 0.013885498, -0.01701355, 0.04...
http://arxiv.org/abs/2305.12493v5
Contextualized End-to-End Speech Recognition with Contextual Phrase Prediction Network
Kaixun Huang, Ao Zhang, Zhanheng Yang, Pengcheng Guo, Bingshen Mu, Tianyi Xu, Lei Xie
eess.AS, cs.CL, cs.SD
Contextual information plays a crucial role in speech recognition technologies and incorporating it into the end-to-end speech recognition models has drawn immense interest recently. However, previous deep bias methods lacked explicit supervision for bias tasks. In this study, we introduce a contextual phrase predictio...
http://arxiv.org/pdf/2305.12493v5
2023-07-12T17:41:53
2023-05-21T16:08:04
{ "title": [ -0.00021672249, 0.014533997, -0.051239014, -0.09442139, 0.028900146, 0.0009622574, -0.053741455, 0.023544312, -0.024810791, 0.05014038, -0.009132385, 0.005748749, 0.019622803, -0.017547607, -0.02696228, 0.0069274902, -0.011672974, 0....
http://arxiv.org/abs/2305.12487v1
Augmenting Autotelic Agents with Large Language Models
Cédric Colas, Laetitia Teodorescu, Pierre-Yves Oudeyer, Xingdi Yuan, Marc-Alexandre Côté
cs.AI, cs.CL, cs.LG
Humans learn to master open-ended repertoires of skills by imagining and practicing their own goals. This autotelic learning process, literally the pursuit of self-generated (auto) goals (telos), becomes more and more open-ended as the goals become more diverse, abstract and creative. The resulting exploration of the s...
http://arxiv.org/pdf/2305.12487v1
2023-05-21T15:42:41
2023-05-21T15:42:41
{ "title": [ -0.016220093, -0.00029969215, -0.024414062, -0.060577393, 0.01335907, 0.012924194, -0.046966553, -0.008483887, 0.0039024353, 0.03778076, -0.02709961, 0.000084102154, 0.022521973, 0.0004312992, -0.03543091, 0.008529663, -0.014389038, ...
http://arxiv.org/abs/2305.12485v2
A Confidence-based Partial Label Learning Model for Crowd-Annotated Named Entity Recognition
Limao Xiong, Jie Zhou, Qunxi Zhu, Xiao Wang, Yuanbin Wu, Qi Zhang, Tao Gui, Xuanjing Huang, Jin Ma, Ying Shan
cs.CL, cs.AI
Existing models for named entity recognition (NER) are mainly based on large-scale labeled datasets, which always obtain using crowdsourcing. However, it is hard to obtain a unified and correct label via majority voting from multiple annotators for NER due to the large labeling space and complexity of this task. To add...
http://arxiv.org/pdf/2305.12485v2
2023-07-27T10:06:49
2023-05-21T15:31:23
{ "title": [ 0.001247406, 0.020507812, -0.042388916, -0.07562256, 0.037322998, -0.0051727295, -0.015899658, 0.01461792, -0.017349243, 0.037078857, -0.04034424, 0.04397583, -0.0034427643, -0.00034499168, -0.032989502, -0.0016508102, 0.0013561249, ...
http://arxiv.org/abs/2305.12483v1
Model Analysis & Evaluation for Ambiguous Question Answering
Konstantinos Papakostas, Irene Papadopoulou
cs.CL, cs.AI, cs.LG
Ambiguous questions are a challenge for Question Answering models, as they require answers that cover multiple interpretations of the original query. To this end, these models are required to generate long-form answers that often combine conflicting pieces of information. Although recent advances in the field have show...
http://arxiv.org/pdf/2305.12483v1
2023-05-21T15:20:20
2023-05-21T15:20:20
{ "title": [ -0.028762817, 0.010223389, -0.04525757, -0.050048828, 0.008506775, 0.0034370422, -0.038391113, -0.0141067505, 0.006416321, 0.024612427, -0.03289795, -0.01676941, -0.012168884, -0.0015859604, -0.072509766, -0.018508911, -0.013511658, ...
http://arxiv.org/abs/2305.12480v1
Is Translation Helpful? An Empirical Analysis of Cross-Lingual Transfer in Low-Resource Dialog Generation
Lei Shen, Shuai Yu, Xiaoyu Shen
cs.CL
Cross-lingual transfer is important for developing high-quality chatbots in multiple languages due to the strongly imbalanced distribution of language resources. A typical approach is to leverage off-the-shelf machine translation (MT) systems to utilize either the training corpus or developed models from high-resource ...
http://arxiv.org/pdf/2305.12480v1
2023-05-21T15:07:04
2023-05-21T15:07:04
{ "title": [ -0.033996582, -0.020584106, -0.03366089, -0.03717041, -0.018615723, -0.007217407, -0.05859375, -0.009147644, -0.012397766, 0.049682617, 0.026107788, 0.025238037, 0.01776123, -0.024749756, -0.04373169, 0.013267517, -0.011634827, 0.039...
http://arxiv.org/abs/2305.12477v2
GPT-3.5, GPT-4, or BARD? Evaluating LLMs Reasoning Ability in Zero-Shot Setting and Performance Boosting Through Prompts
Jessica López Espejel, El Hassane Ettifouri, Mahaman Sanoussi Yahaya Alassan, El Mehdi Chouham, Walid Dahhane
cs.CL, cs.AI
Large Language Models (LLMs) have exhibited remarkable performance on various Natural Language Processing (NLP) tasks. However, there is a current hot debate regarding their reasoning capacity. In this paper, we examine the performance of GPT-3.5, GPT-4, and BARD models, by performing a thorough technical evaluation on...
http://arxiv.org/pdf/2305.12477v2
2023-09-19T21:34:51
2023-05-21T14:45:17
{ "title": [ 0.0045928955, -0.0071640015, -0.035736084, -0.006778717, -0.037628174, -0.022338867, -0.04095459, 0.00016415119, -0.0005259514, 0.0037078857, -0.039093018, 0.01876831, -0.0039367676, -0.031234741, 0.025466919, -0.022216797, -0.02268982, ...
http://arxiv.org/abs/2305.12474v2
Evaluating the Performance of Large Language Models on GAOKAO Benchmark
Xiaotian Zhang, Chunyang Li, Yi Zong, Zhengyu Ying, Liang He, Xipeng Qiu
cs.CL, cs.AI
Large language models have demonstrated remarkable performance across various natural language processing tasks; however, their efficacy in more challenging and domain-specific tasks remains less explored. This paper introduces the GAOKAO-Benchmark (GAOKAO-Bench), an intuitive benchmark that employs questions from the ...
http://arxiv.org/pdf/2305.12474v2
2023-05-23T01:02:11
2023-05-21T14:39:28
{ "title": [ -0.06439209, -0.016021729, -0.039642334, -0.054840088, -0.011062622, 0.010391235, -0.07727051, 0.0032157898, 0.015068054, 0.047729492, 0.0021362305, 0.00015950203, 0.0079193115, -0.012428284, -0.0715332, 0.027114868, -0.010093689, 0....
http://arxiv.org/abs/2305.12473v2
Continually Improving Extractive QA via Human Feedback
Ge Gao, Hung-Ting Chen, Yoav Artzi, Eunsol Choi
cs.CL, cs.AI, cs.LG
We study continually improving an extractive question answering (QA) system via human user feedback. We design and deploy an iterative approach, where information-seeking users ask questions, receive model-predicted answers, and provide feedback. We conduct experiments involving thousands of user interactions under div...
http://arxiv.org/pdf/2305.12473v2
2023-11-03T18:23:58
2023-05-21T14:35:32
{ "title": [ -0.009399414, 0.007446289, -0.02709961, -0.031707764, -0.009346008, -0.02444458, -0.021896362, -0.030456543, 0.013153076, 0.0385437, -0.039855957, -0.005783081, -0.00756073, -0.0011701584, -0.048706055, -0.004360199, 0.010475159, 0.0...
http://arxiv.org/abs/2305.12464v2
Self-supervised Predictive Coding Models Encode Speaker and Phonetic Information in Orthogonal Subspaces
Oli Liu, Hao Tang, Sharon Goldwater
cs.CL, cs.SD
Self-supervised speech representations are known to encode both speaker and phonetic information, but how they are distributed in the high-dimensional space remains largely unexplored. We hypothesize that they are encoded in orthogonal subspaces, a property that lends itself to simple disentanglement. Applying principa...
http://arxiv.org/pdf/2305.12464v2
2023-06-06T10:49:29
2023-05-21T14:03:54
{ "title": [ -0.012748718, 0.010604858, -0.038604736, 0.0074310303, -0.04437256, -0.019439697, -0.028045654, -0.01965332, -0.01626587, 0.046691895, -0.0107803345, -0.0051651, 0.02178955, -0.047943115, 0.009811401, 0.0041503906, 0.019973755, 0.013...
http://arxiv.org/abs/2305.12463v1
Teaching the Pre-trained Model to Generate Simple Texts for Text Simplification
Renliang Sun, Wei Xu, Xiaojun Wan
cs.CL, cs.AI
Randomly masking text spans in ordinary texts in the pre-training stage hardly allows models to acquire the ability to generate simple texts. It can hurt the performance of pre-trained models on text simplification tasks. In this paper, we propose a new continued pre-training strategy to teach the pre-trained model to ...
http://arxiv.org/pdf/2305.12463v1
2023-05-21T14:03:49
2023-05-21T14:03:49
{ "title": [ 0.0019187927, -0.0057144165, -0.038726807, -0.043273926, -0.037506104, -0.00522995, -0.06311035, -0.022079468, -0.0068511963, 0.07043457, -0.003704071, 0.011657715, 0.045196533, -0.02420044, -0.02079773, -0.0028743744, 0.014945984, 0...
http://arxiv.org/abs/2305.12458v1
Infor-Coef: Information Bottleneck-based Dynamic Token Downsampling for Compact and Efficient language model
Wenxi Tan
cs.CL
The prevalence of Transformer-based pre-trained language models (PLMs) has led to their wide adoption for various natural language processing tasks. However, their excessive overhead leads to large latency and computational costs. The statically compression methods allocate fixed computation to different samples, resul...
http://arxiv.org/pdf/2305.12458v1
2023-05-21T13:30:56
2023-05-21T13:30:56
{ "title": [ -0.035491943, -0.010253906, -0.051940918, -0.062469482, 0.00089645386, 0.006149292, -0.0440979, 0.017837524, 0.02796936, 0.04058838, -0.03488159, 0.014518738, 0.03237915, -0.0045318604, -0.050689697, 0.026428223, -0.0115737915, 0.059...