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/2305.13808v2
Asking Clarification Questions to Handle Ambiguity in Open-Domain QA
Dongryeol Lee, Segwang Kim, Minwoo Lee, Hwanhee Lee, Joonsuk Park, Sang-Woo Lee, Kyomin Jung
cs.CL
Ambiguous questions persist in open-domain question answering, because formulating a precise question with a unique answer is often challenging. Previously, Min et al. (2020) have tackled this issue by generating disambiguated questions for all possible interpretations of the ambiguous question. This can be effective, ...
http://arxiv.org/pdf/2305.13808v2
2023-10-25T06:24:16
2023-05-23T08:20:01
{ "title": [ -0.028930664, 0.002702713, -0.052886963, -0.025177002, -0.0042304993, -0.01826477, -0.010055542, -0.014099121, -0.010025024, 0.016036987, -0.005039215, -0.016906738, -0.015136719, -0.038879395, -0.03744507, -0.027694702, -0.0065078735, ...
http://arxiv.org/abs/2305.13805v1
Towards Zero-shot Relation Extraction in Web Mining: A Multimodal Approach with Relative XML Path
Zilong Wang, Jingbo Shang
cs.CL
The rapid growth of web pages and the increasing complexity of their structure poses a challenge for web mining models. Web mining models are required to understand the semi-structured web pages, particularly when little is known about the subject or template of a new page. Current methods migrate language models to th...
http://arxiv.org/pdf/2305.13805v1
2023-05-23T08:16:52
2023-05-23T08:16:52
{ "title": [ 0.004432678, -0.01928711, -0.05508423, -0.05279541, -0.0016555786, -0.02142334, -0.019836426, 0.02128601, 0.035125732, 0.020339966, -0.015029907, -0.006542206, -0.013420105, -0.022216797, -0.056365967, -0.0038108826, -0.02079773, 0.0...
http://arxiv.org/abs/2305.13794v1
Personalized Predictive ASR for Latency Reduction in Voice Assistants
Andreas Schwarz, Di He, Maarten Van Segbroeck, Mohammed Hethnawi, Ariya Rastrow
cs.CL, eess.AS
Streaming Automatic Speech Recognition (ASR) in voice assistants can utilize prefetching to partially hide the latency of response generation. Prefetching involves passing a preliminary ASR hypothesis to downstream systems in order to prefetch and cache a response. If the final ASR hypothesis after endpoint detection m...
http://arxiv.org/pdf/2305.13794v1
2023-05-23T08:05:43
2023-05-23T08:05:43
{ "title": [ 0.0035133362, -0.013847351, -0.040161133, -0.08148193, 0.0011301041, -0.007041931, -0.023086548, 0.034606934, -0.042816162, 0.06781006, 0.024627686, -0.006099701, 0.035125732, -0.03829956, -0.008102417, 0.015419006, 0.0021762848, 0.0...
http://arxiv.org/abs/2305.13788v2
Can Large Language Models Capture Dissenting Human Voices?
Noah Lee, Na Min An, James Thorne
cs.CL, cs.AI
Large language models (LLMs) have shown impressive achievements in solving a broad range of tasks. Augmented by instruction fine-tuning, LLMs have also been shown to generalize in zero-shot settings as well. However, whether LLMs closely align with the human disagreement distribution has not been well-studied, especial...
http://arxiv.org/pdf/2305.13788v2
2023-10-27T11:25:00
2023-05-23T07:55:34
{ "title": [ 0.0038223267, 0.0055274963, -0.025543213, -0.022842407, -0.039886475, -0.020690918, -0.023895264, -0.013801575, -0.0032596588, 0.02961731, -0.03668213, -0.024856567, 0.0113220215, -0.025863647, -0.028656006, -0.023117065, -0.007751465, ...
http://arxiv.org/abs/2305.13785v2
Enhancing Black-Box Few-Shot Text Classification with Prompt-Based Data Augmentation
Danqing Luo, Chen Zhang, Jiahui Xu, Bin Wang, Yiming Chen, Yan Zhang, Haizhou Li
cs.CL
Training or finetuning large-scale language models (LLMs) such as GPT-3 requires substantial computation resources, motivating recent efforts to explore parameter-efficient adaptation to downstream tasks. One practical area of research is to treat these models as black boxes and interact with them through their inferen...
http://arxiv.org/pdf/2305.13785v2
2023-10-20T08:44:21
2023-05-23T07:54:34
{ "title": [ 0.023620605, 0.0021705627, -0.04336548, -0.04837036, 0.006046295, -0.027954102, -0.009506226, -0.00023663044, 0.005886078, 0.014381409, -0.06210327, 0.030380249, -0.023208618, 0.011314392, -0.042175293, -0.029037476, -0.010467529, 0....
http://arxiv.org/abs/2305.13782v1
Images in Language Space: Exploring the Suitability of Large Language Models for Vision & Language Tasks
Sherzod Hakimov, David Schlangen
cs.CL
Large language models have demonstrated robust performance on various language tasks using zero-shot or few-shot learning paradigms. While being actively researched, multimodal models that can additionally handle images as input have yet to catch up in size and generality with language-only models. In this work, we ask...
http://arxiv.org/pdf/2305.13782v1
2023-05-23T07:50:36
2023-05-23T07:50:36
{ "title": [ -0.025054932, -0.009117126, -0.019210815, -0.03253174, -0.006893158, -0.010375977, -0.03237915, -0.019104004, 0.029296875, 0.022262573, -0.039978027, -0.0043907166, 0.005607605, -0.01386261, -0.039123535, 0.0009522438, 0.0063591003, ...
http://arxiv.org/abs/2305.13776v1
Counterspeeches up my sleeve! Intent Distribution Learning and Persistent Fusion for Intent-Conditioned Counterspeech Generation
Rishabh Gupta, Shaily Desai, Manvi Goel, Anil Bandhakavi, Tanmoy Chakraborty, Md. Shad Akhtar
cs.CL, cs.AI
Counterspeech has been demonstrated to be an efficacious approach for combating hate speech. While various conventional and controlled approaches have been studied in recent years to generate counterspeech, a counterspeech with a certain intent may not be sufficient in every scenario. Due to the complex and multifacete...
http://arxiv.org/pdf/2305.13776v1
2023-05-23T07:45:17
2023-05-23T07:45:17
{ "title": [ 0.00869751, 0.0023555756, -0.02949524, -0.03753662, -0.007587433, -0.018203735, -0.014816284, -0.027999878, 0.023422241, 0.04095459, -0.0076408386, 0.009963989, 0.0046920776, -0.023025513, -0.059661865, 0.00073194504, -0.019836426, 0...
http://arxiv.org/abs/2305.13775v1
Concept-aware Training Improves In-context Learning Ability of Language Models
Michal Štefánik, Marek Kadlčík
cs.CL, cs.AI
Many recent language models (LMs) of Transformers family exhibit so-called in-context learning (ICL) ability, manifested in the LMs' ability to modulate their function by a task described in a natural language input. Previous work curating these models assumes that ICL emerges from vast over-parametrization or the scal...
http://arxiv.org/pdf/2305.13775v1
2023-05-23T07:44:52
2023-05-23T07:44:52
{ "title": [ 0.01675415, -0.005683899, -0.02532959, -0.048828125, -0.013214111, -0.020874023, -0.034423828, 0.0027160645, 0.0069618225, 0.020629883, -0.025421143, 0.0060272217, 0.014450073, -0.023880005, -0.030227661, 0.00051498413, -0.015975952, ...
http://arxiv.org/abs/2305.13755v1
Topic-driven Distant Supervision Framework for Macro-level Discourse Parsing
Feng Jiang, Longwang He, Peifeng Li, Qiaoming Zhu, Haizhou Li
cs.CL, cs.AI
Discourse parsing, the task of analyzing the internal rhetorical structure of texts, is a challenging problem in natural language processing. Despite the recent advances in neural models, the lack of large-scale, high-quality corpora for training remains a major obstacle. Recent studies have attempted to overcome this ...
http://arxiv.org/pdf/2305.13755v1
2023-05-23T07:13:51
2023-05-23T07:13:51
{ "title": [ -0.0138549805, -0.0041923523, -0.024871826, -0.09161377, 0.0017004013, 0.015052795, -0.07421875, 0.02067566, 0.005355835, 0.004966736, -0.039733887, 0.016418457, -0.009399414, -0.004070282, -0.020980835, -0.016281128, -0.032562256, 0...
http://arxiv.org/abs/2305.13751v2
Challenges in Context-Aware Neural Machine Translation
Linghao Jin, Jacqueline He, Jonathan May, Xuezhe Ma
cs.CL
Context-aware neural machine translation involves leveraging information beyond sentence-level context to resolve inter-sentential discourse dependencies and improve document-level translation quality, and has given rise to a number of recent techniques. However, despite well-reasoned intuitions, most context-aware tra...
http://arxiv.org/pdf/2305.13751v2
2023-10-23T21:01:26
2023-05-23T07:08:18
{ "title": [ -0.015579224, 0.0016231537, -0.015602112, -0.044158936, 0.009986877, 0.008728027, -0.018539429, 0.0007252693, -0.0018491745, 0.06286621, -0.015060425, -0.0051002502, 0.03604126, -0.009208679, -0.04650879, 0.030532837, -0.007423401, 0...
http://arxiv.org/abs/2305.13749v2
Goal-Driven Explainable Clustering via Language Descriptions
Zihan Wang, Jingbo Shang, Ruiqi Zhong
cs.CL
Unsupervised clustering is widely used to explore large corpora, but existing formulations neither consider the users' goals nor explain clusters' meanings. We propose a new task formulation, "Goal-Driven Clustering with Explanations" (GoalEx), which represents both the goal and the explanations as free-form language d...
http://arxiv.org/pdf/2305.13749v2
2023-11-13T18:27:21
2023-05-23T07:05:50
{ "title": [ -0.008476257, -0.010955811, -0.02861023, -0.06536865, -0.0011177063, 0.009292603, -0.048065186, 0.0014238358, -0.013671875, 0.007987976, 0.0084991455, 0.014305115, -0.02609253, -0.016159058, -0.009864807, 0.0053520203, 0.0035037994, ...
http://arxiv.org/abs/2305.13740v1
TeCS: A Dataset and Benchmark for Tense Consistency of Machine Translation
Yiming Ai, Zhiwei He, Kai Yu, Rui Wang
cs.CL
Tense inconsistency frequently occurs in machine translation. However, there are few criteria to assess the model's mastery of tense prediction from a linguistic perspective. In this paper, we present a parallel tense test set, containing French-English 552 utterances. We also introduce a corresponding benchmark, tense...
http://arxiv.org/pdf/2305.13740v1
2023-05-23T06:51:48
2023-05-23T06:51:48
{ "title": [ -0.038482666, -0.022369385, -0.00920105, -0.05960083, -0.015777588, 0.019943237, -0.057861328, 0.0071792603, 0.0047454834, 0.056915283, -0.008117676, 0.017913818, 0.049072266, -0.00085401535, -0.06021118, 0.030059814, -0.016326904, 0...
http://arxiv.org/abs/2305.13738v1
i-Code Studio: A Configurable and Composable Framework for Integrative AI
Yuwei Fang, Mahmoud Khademi, Chenguang Zhu, Ziyi Yang, Reid Pryzant, Yichong Xu, Yao Qian, Takuya Yoshioka, Lu Yuan, Michael Zeng, Xuedong Huang
cs.CL, cs.AI, cs.CV
Artificial General Intelligence (AGI) requires comprehensive understanding and generation capabilities for a variety of tasks spanning different modalities and functionalities. Integrative AI is one important direction to approach AGI, through combining multiple models to tackle complex multimodal tasks. However, there...
http://arxiv.org/pdf/2305.13738v1
2023-05-23T06:45:55
2023-05-23T06:45:55
{ "title": [ -0.045043945, -0.033325195, -0.053466797, -0.036834717, -0.009811401, -0.010902405, -0.033233643, 0.0025310516, 0.003736496, 0.016983032, -0.030014038, -0.0071907043, -0.0023727417, -0.023132324, -0.029418945, -0.020187378, -0.010093689,...
http://arxiv.org/abs/2305.13735v2
Aligning Large Language Models through Synthetic Feedback
Sungdong Kim, Sanghwan Bae, Jamin Shin, Soyoung Kang, Donghyun Kwak, Kang Min Yoo, Minjoon Seo
cs.CL, cs.AI, cs.LG
Aligning large language models (LLMs) to human values has become increasingly important as it enables sophisticated steering of LLMs. However, it requires significant human demonstrations and feedback or distillation from proprietary LLMs such as ChatGPT. In this work, we propose a novel alignment learning framework wi...
http://arxiv.org/pdf/2305.13735v2
2023-10-21T01:50:54
2023-05-23T06:41:16
{ "title": [ -0.058746338, -0.001373291, -0.049102783, -0.05847168, -0.009391785, 0.020629883, -0.08959961, 0.0012874603, 0.013870239, 0.036346436, -0.030929565, -0.018188477, 0.005962372, -0.0019569397, -0.060180664, -0.0002837181, -0.0056114197, ...
http://arxiv.org/abs/2305.13733v1
Self-Critique Prompting with Large Language Models for Inductive Instructions
Rui Wang, Hongru Wang, Fei Mi, Yi Chen, Ruifeng Xu, Kam-Fai Wong
cs.CL
Numerous works are proposed to improve or evaluate the capabilities of Large language models (LLMs) to fulfill user instructions. However, they neglect the possibility that user inputs may inherently contain incorrect information due to users' false beliefs or malicious intents. In this way, blindly adhering to users' ...
http://arxiv.org/pdf/2305.13733v1
2023-05-23T06:38:20
2023-05-23T06:38:20
{ "title": [ -0.027328491, -0.03048706, -0.019042969, -0.031036377, -0.029876709, -0.012519836, -0.037963867, -0.013999939, -0.025970459, 0.011375427, -0.025344849, 0.008674622, -0.019805908, -0.023071289, -0.01676941, -0.024139404, 0.0047035217, ...
http://arxiv.org/abs/2305.13729v1
Discrete Prompt Optimization via Constrained Generation for Zero-shot Re-ranker
Sukmin Cho, Soyeong Jeong, Jeongyeon Seo, Jong C. Park
cs.IR, cs.AI, cs.CL
Re-rankers, which order retrieved documents with respect to the relevance score on the given query, have gained attention for the information retrieval (IR) task. Rather than fine-tuning the pre-trained language model (PLM), the large-scale language model (LLM) is utilized as a zero-shot re-ranker with excellent result...
http://arxiv.org/pdf/2305.13729v1
2023-05-23T06:35:33
2023-05-23T06:35:33
{ "title": [ 0.0035209656, 0.010757446, -0.030151367, -0.00856781, -0.020126343, -0.009544373, -0.040405273, -0.051574707, 0.019241333, 0.024551392, -0.010566711, 0.006149292, 0.013671875, -0.0368042, -0.012458801, 0.008056641, 0.005592346, -0.01...
http://arxiv.org/abs/2305.16876v1
CombLM: Adapting Black-Box Language Models through Small Fine-Tuned Models
Aitor Ormazabal, Mikel Artetxe, Eneko Agirre
cs.CL
Methods for adapting language models (LMs) to new tasks and domains have traditionally assumed white-box access to the model, and work by modifying its parameters. However, this is incompatible with a recent trend in the field, where the highest quality models are only available as black-boxes through inference APIs. E...
http://arxiv.org/pdf/2305.16876v1
2023-05-23T06:32:55
2023-05-23T06:32:55
{ "title": [ -0.030441284, -0.03765869, -0.033935547, -0.07623291, -0.020126343, 0.0013837814, -0.07128906, 0.013008118, -0.00730896, 0.029022217, -0.012161255, 0.0028877258, -0.0017499924, -0.011871338, -0.05319214, 0.0049476624, -0.016860962, 0...
http://arxiv.org/abs/2305.13725v1
Conversational Recommendation as Retrieval: A Simple, Strong Baseline
Raghav Gupta, Renat Aksitov, Samrat Phatale, Simral Chaudhary, Harrison Lee, Abhinav Rastogi
cs.CL, cs.IR
Conversational recommendation systems (CRS) aim to recommend suitable items to users through natural language conversation. However, most CRS approaches do not effectively utilize the signal provided by these conversations. They rely heavily on explicit external knowledge e.g., knowledge graphs to augment the models' u...
http://arxiv.org/pdf/2305.13725v1
2023-05-23T06:21:31
2023-05-23T06:21:31
{ "title": [ 0.026275635, -0.00039219856, -0.017730713, -0.054382324, 0.012557983, -0.010322571, -0.017410278, -0.0107803345, 0.0046653748, 0.041259766, -0.010810852, 0.017990112, 0.0033016205, -0.01713562, -0.04510498, 0.014816284, -0.010635376, ...
http://arxiv.org/abs/2305.13724v1
ChatGPT-EDSS: Empathetic Dialogue Speech Synthesis Trained from ChatGPT-derived Context Word Embeddings
Yuki Saito, Shinnosuke Takamichi, Eiji Iimori, Kentaro Tachibana, Hiroshi Saruwatari
cs.SD, cs.CL, cs.LG, eess.AS
We propose ChatGPT-EDSS, an empathetic dialogue speech synthesis (EDSS) method using ChatGPT for extracting dialogue context. ChatGPT is a chatbot that can deeply understand the content and purpose of an input prompt and appropriately respond to the user's request. We focus on ChatGPT's reading comprehension and introd...
http://arxiv.org/pdf/2305.13724v1
2023-05-23T06:19:37
2023-05-23T06:19:37
{ "title": [ 0.0012149811, -0.015991211, -0.03955078, -0.02305603, -0.020568848, -0.019607544, -0.027740479, 0.012886047, -0.019042969, 0.028640747, -0.022232056, 0.0036849976, 0.037628174, -0.033813477, 0.0027160645, -0.002620697, -0.026626587, ...
http://arxiv.org/abs/2305.13723v2
PIEClass: Weakly-Supervised Text Classification with Prompting and Noise-Robust Iterative Ensemble Training
Yunyi Zhang, Minhao Jiang, Yu Meng, Yu Zhang, Jiawei Han
cs.CL
Weakly-supervised text classification trains a classifier using the label name of each target class as the only supervision, which largely reduces human annotation efforts. Most existing methods first use the label names as static keyword-based features to generate pseudo labels, which are then used for final classifie...
http://arxiv.org/pdf/2305.13723v2
2023-10-20T15:14:34
2023-05-23T06:19:14
{ "title": [ -0.025619507, 0.005706787, -0.04840088, -0.050323486, 0.017929077, -0.005466461, -0.029830933, -0.015113831, 0.010131836, 0.018112183, -0.0519104, 0.04925537, -0.008285522, 0.008811951, -0.033599854, -0.00030469894, -0.013717651, 0.0...
http://arxiv.org/abs/2305.13721v1
Continual Dialogue State Tracking via Example-Guided Question Answering
Hyundong Cho, Andrea Madotto, Zhaojiang Lin, Khyathi Raghavi Chandu, Satwik Kottur, Jing Xu, Jonathan May, Chinnadhurai Sankar
cs.CL, cs.AI
Dialogue systems are frequently updated to accommodate new services, but naively updating them by continually training with data for new services in diminishing performance on previously learnt services. Motivated by the insight that dialogue state tracking (DST), a crucial component of dialogue systems that estimates ...
http://arxiv.org/pdf/2305.13721v1
2023-05-23T06:15:43
2023-05-23T06:15:43
{ "title": [ 0.03186035, -0.0178833, -0.008041382, -0.083984375, -0.0031356812, -0.0023498535, -0.041900635, 0.003238678, 0.022369385, 0.01184845, -0.011276245, 0.008361816, 0.00012153387, -0.027999878, -0.031982422, -0.001499176, 0.0010967255, 0...
http://arxiv.org/abs/2305.13718v2
LogicLLM: Exploring Self-supervised Logic-enhanced Training for Large Language Models
Fangkai Jiao, Zhiyang Teng, Shafiq Joty, Bosheng Ding, Aixin Sun, Zhengyuan Liu, Nancy F. Chen
cs.CL
Existing efforts to improve logical reasoning ability of language models have predominantly relied on supervised fine-tuning, hindering generalization to new domains and/or tasks. The development of Large Langauge Models (LLMs) has demonstrated the capacity of compressing abundant knowledge into a single proxy, enablin...
http://arxiv.org/pdf/2305.13718v2
2023-05-24T06:38:41
2023-05-23T06:13:10
{ "title": [ -0.024139404, -0.004627228, -0.062805176, -0.083740234, -0.0029621124, -0.0051879883, -0.07678223, 0.03189087, -0.022659302, 0.02281189, -0.0059127808, 0.022476196, -0.011009216, -0.017501831, -0.02178955, 0.011802673, -0.028259277, ...
http://arxiv.org/abs/2305.13716v3
BA-SOT: Boundary-Aware Serialized Output Training for Multi-Talker ASR
Yuhao Liang, Fan Yu, Yangze Li, Pengcheng Guo, Shiliang Zhang, Qian Chen, Lei Xie
cs.SD, cs.CL, eess.AS
The recently proposed serialized output training (SOT) simplifies multi-talker automatic speech recognition (ASR) by generating speaker transcriptions separated by a special token. However, frequent speaker changes can make speaker change prediction difficult. To address this, we propose boundary-aware serialized outpu...
http://arxiv.org/pdf/2305.13716v3
2023-10-05T11:44:39
2023-05-23T06:08:13
{ "title": [ 0.003993988, -0.0060272217, -0.047698975, -0.107910156, 0.0093688965, 0.0023975372, -0.057556152, 0.03414917, -0.030395508, 0.029327393, 0.00818634, 0.0158844, -0.007003784, -0.040496826, -0.009567261, 0.0023231506, 0.00094509125, 0....
http://arxiv.org/abs/2305.13713v1
CALLS: Japanese Empathetic Dialogue Speech Corpus of Complaint Handling and Attentive Listening in Customer Center
Yuki Saito, Eiji Iimori, Shinnosuke Takamichi, Kentaro Tachibana, Hiroshi Saruwatari
cs.SD, cs.CL, cs.LG, eess.AS
We present CALLS, a Japanese speech corpus that considers phone calls in a customer center as a new domain of empathetic spoken dialogue. The existing STUDIES corpus covers only empathetic dialogue between a teacher and student in a school. To extend the application range of empathetic dialogue speech synthesis (EDSS),...
http://arxiv.org/pdf/2305.13713v1
2023-05-23T06:04:50
2023-05-23T06:04:50
{ "title": [ 0.032806396, -0.023406982, -0.046813965, -0.06274414, -0.020141602, -0.030548096, -0.041503906, 0.024627686, 0.026031494, 0.032440186, -0.029968262, -0.027038574, 0.062072754, -0.035308838, -0.001991272, -0.026885986, -0.010307312, 0...
http://arxiv.org/abs/2305.13712v1
Knowledge of Knowledge: Exploring Known-Unknowns Uncertainty with Large Language Models
Alfonso Amayuelas, Liangming Pan, Wenhu Chen, William Wang
cs.CL, cs.AI
This paper investigates the capabilities of Large Language Models (LLMs) in the context of understanding their own knowledge and measuring their uncertainty. We argue this is an important feature for mitigating hallucinations. Specifically, we focus on addressing \textit{known-unknown} questions, characterized by high ...
http://arxiv.org/pdf/2305.13712v1
2023-05-23T05:59:21
2023-05-23T05:59:21
{ "title": [ -0.054870605, -0.023391724, -0.05279541, -0.057403564, -0.010635376, -0.011146545, -0.04888916, -0.0054626465, 0.00605011, 0.015213013, -0.030014038, -0.0017900467, -0.005962372, -0.022918701, -0.0423584, -0.0021457672, 0.0022964478, ...
http://arxiv.org/abs/2305.13711v1
LLM-Eval: Unified Multi-Dimensional Automatic Evaluation for Open-Domain Conversations with Large Language Models
Yen-Ting Lin, Yun-Nung Chen
cs.CL, cs.AI
We propose LLM-Eval, a unified multi-dimensional automatic evaluation method for open-domain conversations with large language models (LLMs). Existing evaluation methods often rely on human annotations, ground-truth responses, or multiple LLM prompts, which can be expensive and time-consuming. To address these issues, ...
http://arxiv.org/pdf/2305.13711v1
2023-05-23T05:57:09
2023-05-23T05:57:09
{ "title": [ -0.029541016, 0.0101623535, -0.038909912, -0.07678223, 0.008087158, 0.012954712, -0.08154297, -0.013656616, -0.001906395, 0.025131226, -0.0030822754, -0.0028934479, 0.0041503906, -0.025726318, -0.035186768, -0.0059661865, -0.030395508, ...
http://arxiv.org/abs/2305.13710v1
Using Textual Interface to Align External Knowledge for End-to-End Task-Oriented Dialogue Systems
Qingyang Wu, Deema Alnuhait, Derek Chen, Zhou Yu
cs.CL
Traditional end-to-end task-oriented dialogue systems have been built with a modularized design. However, such design often causes misalignment between the agent response and external knowledge, due to inadequate representation of information. Furthermore, its evaluation metrics emphasize assessing the agent's pre-lexi...
http://arxiv.org/pdf/2305.13710v1
2023-05-23T05:48:21
2023-05-23T05:48:21
{ "title": [ -0.038238525, -0.020690918, -0.06304932, -0.054229736, 0.004047394, -0.007911682, -0.035858154, 0.022583008, 0.0029201508, 0.04458618, -0.0058784485, -0.016235352, -0.0026416779, -0.01689148, -0.063964844, -0.015838623, 0.0026054382, ...
http://arxiv.org/abs/2305.13707v1
Do All Languages Cost the Same? Tokenization in the Era of Commercial Language Models
Orevaoghene Ahia, Sachin Kumar, Hila Gonen, Jungo Kasai, David R. Mortensen, Noah A. Smith, Yulia Tsvetkov
cs.CL
Language models have graduated from being research prototypes to commercialized products offered as web APIs, and recent works have highlighted the multilingual capabilities of these products. The API vendors charge their users based on usage, more specifically on the number of ``tokens'' processed or generated by the ...
http://arxiv.org/pdf/2305.13707v1
2023-05-23T05:46:45
2023-05-23T05:46:45
{ "title": [ -0.030288696, -0.013450623, -0.04800415, -0.06530762, -0.017471313, -0.009277344, -0.049987793, 0.012107849, -0.059417725, 0.025054932, -0.028518677, 0.022735596, -0.016418457, -0.014122009, -0.00831604, -0.025344849, -0.0076141357, ...
http://arxiv.org/abs/2305.13703v1
MemeCap: A Dataset for Captioning and Interpreting Memes
EunJeong Hwang, Vered Shwartz
cs.CL
Memes are a widely popular tool for web users to express their thoughts using visual metaphors. Understanding memes requires recognizing and interpreting visual metaphors with respect to the text inside or around the meme, often while employing background knowledge and reasoning abilities. We present the task of meme c...
http://arxiv.org/pdf/2305.13703v1
2023-05-23T05:41:18
2023-05-23T05:41:18
{ "title": [ 0.0023651123, 0.00053453445, -0.03186035, -0.030014038, -0.0008368492, -0.021270752, -0.018325806, -0.028793335, 0.024475098, 0.052368164, -0.012863159, 0.01777649, 0.017440796, -0.009719849, -0.04272461, 0.02456665, -0.0088272095, 0...
http://arxiv.org/abs/2305.13698v1
Exploring Large Language Models for Classical Philology
Frederick Riemenschneider, Anette Frank
cs.CL, I.2.7
Recent advances in NLP have led to the creation of powerful language models for many languages including Ancient Greek and Latin. While prior work on Classical languages unanimously uses BERT, in this work we create four language models for Ancient Greek that vary along two dimensions to study their versatility for tas...
http://arxiv.org/pdf/2305.13698v1
2023-05-23T05:21:02
2023-05-23T05:21:02
{ "title": [ -0.037475586, 0.006713867, -0.027923584, -0.036499023, 0.00027370453, 0.0054130554, -0.03262329, 0.0002965927, 0.017547607, 0.034240723, -0.019241333, -0.0231781, 0.026611328, -0.024642944, -0.04714966, 0.013671875, -0.0056495667, 0....
http://arxiv.org/abs/2305.13697v1
UNIMO-3: Multi-granularity Interaction for Vision-Language Representation Learning
Hao Yang, Can Gao, Hao Líu, Xinyan Xiao, Yanyan Zhao, Bing Qin
cs.CL
Vision-and-language (VL) pre-training, which aims to learn a general representation of image-text pairs that can be transferred to various vision-and-language tasks. Compared with modeling uni-modal data, the main challenge of the VL model is: how to learn the cross-modal interaction from multimodal data, especially th...
http://arxiv.org/pdf/2305.13697v1
2023-05-23T05:11:34
2023-05-23T05:11:34
{ "title": [ -0.009613037, 0.0056533813, -0.035858154, -0.0635376, -0.0067100525, 0.0019779205, -0.06719971, -0.0060043335, 0.03439331, 0.022491455, -0.023254395, 0.0017242432, 0.010322571, -0.012641907, -0.041412354, -0.001832962, 0.0129776, 0.0...
http://arxiv.org/abs/2305.13696v1
Abstractive Text Summarization Using the BRIO Training Paradigm
Khang Nhut Lam, Thieu Gia Doan, Khang Thua Pham, Jugal Kalita
cs.CL
Summary sentences produced by abstractive summarization models may be coherent and comprehensive, but they lack control and rely heavily on reference summaries. The BRIO training paradigm assumes a non-deterministic distribution to reduce the model's dependence on reference summaries, and improve model performance duri...
http://arxiv.org/pdf/2305.13696v1
2023-05-23T05:09:53
2023-05-23T05:09:53
{ "title": [ 0.0019321442, 0.009384155, -0.026275635, -0.08526611, -0.0036811829, 0.010604858, -0.056243896, -0.0050315857, 0.02293396, 0.051635742, -0.019592285, 0.017730713, 0.0016727448, -0.008270264, -0.06188965, 0.003627777, -0.008338928, 0....
http://arxiv.org/abs/2305.13693v1
Automated Metrics for Medical Multi-Document Summarization Disagree with Human Evaluations
Lucy Lu Wang, Yulia Otmakhova, Jay DeYoung, Thinh Hung Truong, Bailey E. Kuehl, Erin Bransom, Byron C. Wallace
cs.CL
Evaluating multi-document summarization (MDS) quality is difficult. This is especially true in the case of MDS for biomedical literature reviews, where models must synthesize contradicting evidence reported across different documents. Prior work has shown that rather than performing the task, models may exploit shortcu...
http://arxiv.org/pdf/2305.13693v1
2023-05-23T05:00:59
2023-05-23T05:00:59
{ "title": [ -0.014816284, -0.004135132, -0.013626099, 0.016403198, -0.061523438, -0.017990112, 0.0011844635, -0.025161743, -0.037231445, 0.053771973, -0.04248047, -0.002670288, 0.01751709, -0.015731812, -0.024459839, -0.039215088, -0.0011110306, ...
http://arxiv.org/abs/2305.13691v1
Efficient Open Domain Multi-Hop Question Answering with Few-Shot Data Synthesis
Mingda Chen, Xilun Chen, Wen-tau Yih
cs.CL
Few-shot learning for open domain multi-hop question answering typically relies on large language models (LLMs). While powerful, LLMs are inefficient at the inference time. We propose a data synthesis framework for multi-hop question answering that allows for improving smaller language models with less than 10 human-an...
http://arxiv.org/pdf/2305.13691v1
2023-05-23T04:57:31
2023-05-23T04:57:31
{ "title": [ -0.0025997162, -0.014335632, -0.02267456, -0.052764893, -0.017440796, 0.0047683716, -0.03744507, 0.016998291, 0.01625061, 0.027175903, -0.02998352, 0.018692017, -0.0061187744, -0.0023765564, -0.041290283, -0.006450653, 0.0023479462, ...
http://arxiv.org/abs/2305.13690v1
Towards Asking Clarification Questions for Information Seeking on Task-Oriented Dialogues
Yue Feng, Hossein A. Rahmani, Aldo Lipani, Emine Yilmaz
cs.CL, cs.IR
Task-oriented dialogue systems aim at providing users with task-specific services. Users of such systems often do not know all the information about the task they are trying to accomplish, requiring them to seek information about the task. To provide accurate and personalized task-oriented information seeking results, ...
http://arxiv.org/pdf/2305.13690v1
2023-05-23T04:56:04
2023-05-23T04:56:04
{ "title": [ -0.022567749, -0.022369385, -0.029968262, -0.035125732, -0.012275696, -0.00080537796, -0.02658081, 0.0044822693, -0.0004324913, 0.019439697, -0.0020503998, -0.007896423, -0.020462036, -0.026107788, -0.06774902, -0.018051147, -0.012954712...
http://arxiv.org/abs/2306.01755v2
Training Priors Predict Text-To-Image Model Performance
Charles Lovering, Ellie Pavlick
cs.CV, cs.AI, cs.CL
Text-to-image models can often generate some relations, i.e., "astronaut riding horse", but fail to generate other relations composed of the same basic parts, i.e., "horse riding astronaut". These failures are often taken as evidence that models rely on training priors rather than constructing novel images compositiona...
http://arxiv.org/pdf/2306.01755v2
2023-10-25T02:56:03
2023-05-23T04:54:26
{ "title": [ -0.011070251, 0.01626587, -0.009216309, -0.029266357, -0.020050049, -0.00831604, -0.040496826, -0.03125, 0.006290436, 0.040893555, -0.023513794, 0.0044403076, 0.0046539307, -0.01914978, -0.034454346, -0.003850937, 0.003189087, 0.0357...
http://arxiv.org/abs/2305.13685v1
Causal Intervention for Abstractive Related Work Generation
Jiachang Liu, Qi Zhang, Chongyang Shi, Usman Naseem, Shoujin Wang, Ivor Tsang
cs.CL
Abstractive related work generation has attracted increasing attention in generating coherent related work that better helps readers grasp the background in the current research. However, most existing abstractive models ignore the inherent causality of related work generation, leading to low quality of generated relat...
http://arxiv.org/pdf/2305.13685v1
2023-05-23T04:48:30
2023-05-23T04:48:30
{ "title": [ 0.007675171, -0.0062675476, -0.053253174, -0.08508301, -0.019378662, -0.0115356445, -0.07470703, 0.017471313, -0.03277588, 0.03286743, 0.011543274, 0.008811951, -0.008895874, -0.052368164, 0.015670776, -0.00598526, -0.022232056, 0.04...
http://arxiv.org/abs/2305.13684v1
mPLM-Sim: Unveiling Better Cross-Lingual Similarity and Transfer in Multilingual Pretrained Language Models
Peiqin Lin, Chengzhi Hu, Zheyu Zhang, André F. T. Martins, Hinrich Schütze
cs.CL
Recent multilingual pretrained language models (mPLMs) have been shown to encode strong language-specific signals, which are not explicitly provided during pretraining. It remains an open question whether it is feasible to employ mPLMs to measure language similarity, and subsequently use the similarity results to selec...
http://arxiv.org/pdf/2305.13684v1
2023-05-23T04:44:26
2023-05-23T04:44:26
{ "title": [ -0.00907135, -0.0073051453, -0.025878906, -0.0715332, 0.0025787354, 0.0033168793, -0.058654785, -0.0033798218, -0.002922058, 0.033416748, 0.015357971, 0.024658203, 0.028457642, -0.028686523, -0.022979736, 0.030960083, -0.01890564, 0....
http://arxiv.org/abs/2305.13683v1
Error Detection for Text-to-SQL Semantic Parsing
Shijie Chen, Ziru Chen, Huan Sun, Yu Su
cs.CL
Despite remarkable progress in text-to-SQL semantic parsing in recent years, the performance of existing parsers is still far from perfect. At the same time, modern deep learning based text-to-SQL parsers are often over-confident and thus casting doubt on their trustworthiness when deployed for real use. To that end, w...
http://arxiv.org/pdf/2305.13683v1
2023-05-23T04:44:22
2023-05-23T04:44:22
{ "title": [ -0.006427765, 0.010307312, -0.05316162, -0.082336426, 0.030532837, 0.0008330345, -0.05218506, 0.0015459061, 0.024337769, 0.03652954, -0.0569458, 0.007762909, 0.021499634, -0.003376007, -0.053253174, -0.01876831, -0.01260376, 0.027221...
http://arxiv.org/abs/2305.13677v2
Towards Legally Enforceable Hate Speech Detection for Public Forums
Chu Fei Luo, Rohan Bhambhoria, Xiaodan Zhu, Samuel Dahan
cs.CL
Hate speech causes widespread and deep-seated societal issues. Proper enforcement of hate speech laws is key for protecting groups of people against harmful and discriminatory language. However, determining what constitutes hate speech is a complex task that is highly open to subjective interpretations. Existing works ...
http://arxiv.org/pdf/2305.13677v2
2023-11-01T18:30:21
2023-05-23T04:34:41
{ "title": [ 0.03540039, 0.0076026917, -0.030273438, -0.048034668, 0.028533936, -0.02470398, 0.007896423, -0.01789856, -0.0014867783, 0.028961182, -0.018829346, 0.008087158, 0.0017557144, -0.014816284, -0.059692383, -0.002500534, -0.017578125, 0....
http://arxiv.org/abs/2305.13675v1
Polyglot or Not? Measuring Multilingual Encyclopedic Knowledge Retrieval from Foundation Language Models
Tim Schott, Daniel Furman, Shreshta Bhat
cs.CL
In this work, we evaluate the capacity for foundation models to retrieve encyclopedic knowledge across a wide range of languages, topics, and contexts. To support this effort, we 1) produce a new dataset containing 303k factual associations in 20 different languages, 2) formulate a new counterfactual knowledge assessme...
http://arxiv.org/pdf/2305.13675v1
2023-05-23T04:31:39
2023-05-23T04:31:39
{ "title": [ -0.046722412, -0.007411957, -0.05319214, -0.031280518, -0.013580322, -0.011062622, -0.043426514, 0.0006785393, 0.013305664, 0.047851562, -0.01612854, 0.018600464, -0.006958008, -0.004043579, -0.066711426, -0.0056266785, -0.013916016, ...
http://arxiv.org/abs/2305.13673v2
Physics of Language Models: Part 1, Context-Free Grammar
Zeyuan Allen-Zhu, Yuanzhi Li
cs.CL, cs.AI, cs.LG
We design controlled experiments to study HOW generative language models, like GPT, learn context-free grammars (CFGs) -- diverse language systems with a tree-like structure capturing many aspects of natural languages, programs, and logics. CFGs are as hard as pushdown automata, and can be ambiguous so that verifying i...
http://arxiv.org/pdf/2305.13673v2
2023-10-05T01:43:23
2023-05-23T04:28:16
{ "title": [ -0.03149414, -0.014404297, -0.03378296, -0.07342529, 0.0047721863, 0.014450073, -0.040771484, 0.021316528, -0.023330688, 0.048706055, 0.015930176, 0.00035429, 0.014213562, -0.026260376, -0.03878784, 0.024291992, -0.016311646, 0.06079...
http://arxiv.org/abs/2305.13669v2
The Knowledge Alignment Problem: Bridging Human and External Knowledge for Large Language Models
Shuo Zhang, Liangming Pan, Junzhou Zhao, William Yang Wang
cs.CL, cs.AI
Large language models often necessitate grounding on external knowledge to generate faithful and reliable answers. Yet even with the correct groundings in the reference, they can ignore them and rely on wrong groundings or their inherent biases to hallucinate when users, being largely unaware of the specifics of the st...
http://arxiv.org/pdf/2305.13669v2
2023-11-03T17:40:21
2023-05-23T04:22:50
{ "title": [ -0.044311523, -0.015090942, -0.06414795, -0.07220459, 0.0073928833, 0.0011444092, -0.06530762, 0.015563965, 0.009681702, 0.031341553, -0.02168274, -0.012756348, 0.015548706, -0.0061302185, -0.047790527, -0.009353638, -0.003604889, 0....
http://arxiv.org/abs/2305.13668v1
Grounding and Distinguishing Conceptual Vocabulary Through Similarity Learning in Embodied Simulations
Sadaf Ghaffari, Nikhil Krishnaswamy
cs.CL, cs.LG
We present a novel method for using agent experiences gathered through an embodied simulation to ground contextualized word vectors to object representations. We use similarity learning to make comparisons between different object types based on their properties when interacted with, and to extract common features pert...
http://arxiv.org/pdf/2305.13668v1
2023-05-23T04:22:00
2023-05-23T04:22:00
{ "title": [ -0.01838684, 0.011985779, -0.026504517, -0.019989014, 0.0013093948, -0.00667572, -0.014625549, -0.024871826, -0.030822754, 0.028625488, -0.03555298, 0.044555664, -0.012962341, -0.014663696, -0.0039367676, -0.019851685, -0.0072288513, ...
http://arxiv.org/abs/2305.13667v2
Optimizing Non-Autoregressive Transformers with Contrastive Learning
Chenxin An, Jiangtao Feng, Fei Huang, Xipeng Qiu, Lingpeng Kong
cs.CL
Non-autoregressive Transformers (NATs) reduce the inference latency of Autoregressive Transformers (ATs) by predicting words all at once rather than in sequential order. They have achieved remarkable progress in machine translation as well as many other applications. However, a long-standing challenge for NATs is the l...
http://arxiv.org/pdf/2305.13667v2
2023-06-02T10:48:41
2023-05-23T04:20:13
{ "title": [ 0.015991211, 0.021728516, -0.013885498, -0.040100098, 0.018463135, -0.004787445, -0.025924683, -0.052001953, 0.014953613, 0.03552246, -0.045043945, 0.022216797, -0.016784668, -0.005317688, -0.055633545, -0.015129089, 0.00050640106, 0...
http://arxiv.org/abs/2305.13661v2
On the Risk of Misinformation Pollution with Large Language Models
Yikang Pan, Liangming Pan, Wenhu Chen, Preslav Nakov, Min-Yen Kan, William Yang Wang
cs.CL, cs.AI
In this paper, we comprehensively investigate the potential misuse of modern Large Language Models (LLMs) for generating credible-sounding misinformation and its subsequent impact on information-intensive applications, particularly Open-Domain Question Answering (ODQA) systems. We establish a threat model and simulate ...
http://arxiv.org/pdf/2305.13661v2
2023-10-26T20:45:39
2023-05-23T04:10:26
{ "title": [ 0.00035429, 0.0077590942, -0.04699707, -0.066345215, 0.022354126, -0.0074386597, -0.030960083, -0.0025749207, -0.021057129, 0.038146973, -0.016464233, 0.008155823, 0.004261017, -0.011917114, -0.037261963, -0.0019483566, -0.024917603, ...
http://arxiv.org/abs/2305.13660v2
Prompt-Based Monte-Carlo Tree Search for Goal-Oriented Dialogue Policy Planning
Xiao Yu, Maximillian Chen, Zhou Yu
cs.CL
Planning for goal-oriented dialogue often requires simulating future dialogue interactions and estimating task progress. Many approaches thus consider training neural networks to perform look-ahead search algorithms such as A* search and Monte Carlo Tree Search (MCTS). However, this training often requires abundant ann...
http://arxiv.org/pdf/2305.13660v2
2023-10-19T22:31:18
2023-05-23T04:07:03
{ "title": [ -0.011116028, -0.017501831, -0.020004272, -0.07043457, -0.02003479, 0.006866455, -0.0440979, 0.021987915, -0.038116455, -0.0017061234, -0.008506775, 0.032806396, -0.028808594, -0.036865234, -0.009063721, -0.026824951, -0.028533936, 0...
http://arxiv.org/abs/2305.13658v2
Understanding Compositional Data Augmentation in Typologically Diverse Morphological Inflection
Farhan Samir, Miikka Silfverberg
cs.CL
Data augmentation techniques are widely used in low-resource automatic morphological inflection to overcome data sparsity. However, the full implications of these techniques remain poorly understood. In this study, we aim to shed light on the theoretical aspects of the prominent data augmentation strategy StemCorrupt (...
http://arxiv.org/pdf/2305.13658v2
2023-10-23T18:07:29
2023-05-23T04:02:54
{ "title": [ -0.0042648315, 0.0005617142, -0.05142212, -0.09326172, 0.0051345825, -0.0038585663, -0.07019043, 0.02508545, -0.0021572113, 0.027709961, -0.016082764, 0.020248413, -0.0084991455, -0.022659302, -0.033294678, 0.0036888123, -0.029083252, ...
http://arxiv.org/abs/2305.13657v1
ChatGPT as your Personal Data Scientist
Md Mahadi Hassan, Alex Knipper, Shubhra Kanti Karmaker Santu
cs.CL
The rise of big data has amplified the need for efficient, user-friendly automated machine learning (AutoML) tools. However, the intricacy of understanding domain-specific data and defining prediction tasks necessitates human intervention making the process time-consuming while preventing full automation. Instead, envi...
http://arxiv.org/pdf/2305.13657v1
2023-05-23T04:00:16
2023-05-23T04:00:16
{ "title": [ 0.008224487, -0.013931274, -0.044128418, -0.05593872, -0.037872314, -0.01184845, -0.03540039, -0.0019779205, -0.017700195, 0.028015137, -0.015701294, 0.0034885406, -0.030761719, -0.052124023, 0.0087890625, 0.0004711151, 0.016204834, ...
http://arxiv.org/abs/2305.13654v2
Understanding and Mitigating Spurious Correlations in Text Classification with Neighborhood Analysis
Oscar Chew, Hsuan-Tien Lin, Kai-Wei Chang, Kuan-Hao Huang
cs.CL
Recent research has revealed that deep learning models have a tendency to leverage spurious correlations that exist in the training set but may not hold true in general circumstances. For instance, a sentiment classifier may erroneously learn that the token performances is commonly associated with positive movie review...
http://arxiv.org/pdf/2305.13654v2
2023-07-16T04:00:01
2023-05-23T03:55:50
{ "title": [ 0.0023403168, 0.012825012, -0.04928589, -0.028640747, 0.029632568, -0.020111084, -0.0060195923, -0.035888672, -0.017990112, 0.019821167, -0.02053833, 0.026519775, -0.022125244, -0.0008969307, -0.030029297, -0.026123047, -0.022445679, ...
http://arxiv.org/abs/2305.13652v1
Cross-lingual Knowledge Transfer and Iterative Pseudo-labeling for Low-Resource Speech Recognition with Transducers
Jan Silovsky, Liuhui Deng, Arturo Argueta, Tresi Arvizo, Roger Hsiao, Sasha Kuznietsov, Yiu-Chang Lin, Xiaoqiang Xiao, Yuanyuan Zhang
cs.CL, eess.AS
Voice technology has become ubiquitous recently. However, the accuracy, and hence experience, in different languages varies significantly, which makes the technology not equally inclusive. The availability of data for different languages is one of the key factors affecting accuracy, especially in training of all-neural...
http://arxiv.org/pdf/2305.13652v1
2023-05-23T03:50:35
2023-05-23T03:50:35
{ "title": [ -0.01576233, 0.0017147064, -0.0541687, -0.070251465, 0.0030784607, -0.008659363, -0.05899048, 0.009109497, -0.019439697, 0.038269043, -0.0049819946, 0.008628845, 0.005760193, -0.026550293, -0.030303955, -0.0044136047, 0.006679535, 0....
http://arxiv.org/abs/2305.13648v1
Non-parametric, Nearest-neighbor-assisted Fine-tuning for Neural Machine Translation
Jiayi Wang, Ke Wang, Yuqi Zhang, Yu Zhao, Pontus Stenetorp
cs.CL, cs.AI
Non-parametric, k-nearest-neighbor algorithms have recently made inroads to assist generative models such as language models and machine translation decoders. We explore whether such non-parametric models can improve machine translation models at the fine-tuning stage by incorporating statistics from the kNN prediction...
http://arxiv.org/pdf/2305.13648v1
2023-05-23T03:44:06
2023-05-23T03:44:06
{ "title": [ -0.020431519, -0.009796143, 0.017074585, -0.045410156, -0.012069702, 0.020141602, -0.024658203, -0.0058135986, -0.023483276, 0.0637207, -0.009231567, 0.014694214, 0.01979065, 0.0012016296, -0.057525635, 0.007827759, 0.0029067993, 0.0...
http://arxiv.org/abs/2305.13645v1
mPMR: A Multilingual Pre-trained Machine Reader at Scale
Weiwen Xu, Xin Li, Wai Lam, Lidong Bing
cs.CL
We present multilingual Pre-trained Machine Reader (mPMR), a novel method for multilingual machine reading comprehension (MRC)-style pre-training. mPMR aims to guide multilingual pre-trained language models (mPLMs) to perform natural language understanding (NLU) including both sequence classification and span extractio...
http://arxiv.org/pdf/2305.13645v1
2023-05-23T03:40:36
2023-05-23T03:40:36
{ "title": [ -0.05041504, -0.024307251, -0.033477783, -0.053375244, -0.019058228, 0.0138168335, -0.06060791, -0.009231567, -0.003107071, 0.041625977, -0.009811401, -0.00026917458, 0.014007568, -0.014717102, -0.04055786, 0.01411438, -0.0008940697, ...
http://arxiv.org/abs/2305.18324v1
Regex-augmented Domain Transfer Topic Classification based on a Pre-trained Language Model: An application in Financial Domain
Vanessa Liao, Syed Shariyar Murtaza, Yifan Nie, Jimmy Lin
cs.CL, cs.AI
A common way to use large pre-trained language models for downstream tasks is to fine tune them using additional layers. This may not work well if downstream domain is a specialized domain whereas the large language model has been pre-trained on a generic corpus. In this paper, we discuss the use of regular expression ...
http://arxiv.org/pdf/2305.18324v1
2023-05-23T03:26:32
2023-05-23T03:26:32
{ "title": [ -0.012557983, 0.0068969727, -0.05593872, -0.03086853, -0.0051231384, -0.02609253, -0.03479004, -0.019241333, -0.01272583, 0.021697998, 0.0023097992, 0.013198853, 0.011039734, -0.0309906, -0.017593384, -0.008621216, -0.021072388, 0.01...
http://arxiv.org/abs/2305.13641v1
AxomiyaBERTa: A Phonologically-aware Transformer Model for Assamese
Abhijnan Nath, Sheikh Mannan, Nikhil Krishnaswamy
cs.CL
Despite their successes in NLP, Transformer-based language models still require extensive computing resources and suffer in low-resource or low-compute settings. In this paper, we present AxomiyaBERTa, a novel BERT model for Assamese, a morphologically-rich low-resource language (LRL) of Eastern India. AxomiyaBERTa is ...
http://arxiv.org/pdf/2305.13641v1
2023-05-23T03:19:21
2023-05-23T03:19:21
{ "title": [ -0.027648926, 0.0005760193, -0.02305603, -0.03567505, -0.025344849, 0.019561768, -0.058288574, 0.011657715, 0.0032634735, 0.022888184, -0.0317688, -0.009674072, 0.03942871, -0.033233643, -0.018753052, -0.0023384094, -0.03793335, 0.08...
http://arxiv.org/abs/2305.13637v2
IdEALS: Idiomatic Expressions for Advancement of Language Skills
Narutatsu Ri, Bill Sun, Sam Davidson, Zhou Yu
cs.CL
Although significant progress has been made in developing methods for Grammatical Error Correction (GEC), addressing word choice improvements has been notably lacking and enhancing sentence expressivity by replacing phrases with advanced expressions is an understudied aspect. In this paper, we focus on this area and pr...
http://arxiv.org/pdf/2305.13637v2
2023-05-24T17:25:55
2023-05-23T03:11:41
{ "title": [ 0.021713257, -0.0027866364, -0.027893066, -0.014976501, -0.018066406, -0.019927979, -0.0042648315, -0.007095337, -0.024261475, 0.045318604, -0.002298355, 0.0000017881393, 0.044006348, -0.042633057, -0.0022792816, 0.0073051453, -0.0129547...
http://arxiv.org/abs/2305.13632v2
Detecting and Mitigating Hallucinations in Multilingual Summarisation
Yifu Qiu, Yftah Ziser, Anna Korhonen, Edoardo M. Ponti, Shay B. Cohen
cs.CL, cs.AI, cs.LG
Hallucinations pose a significant challenge to the reliability of neural models for abstractive summarisation. While automatically generated summaries may be fluent, they often lack faithfulness to the original document. This issue becomes even more pronounced in low-resource settings, such as cross-lingual transfer. W...
http://arxiv.org/pdf/2305.13632v2
2023-10-26T17:02:56
2023-05-23T02:59:25
{ "title": [ 0.01083374, 0.009170532, -0.01953125, -0.061767578, -0.0025959015, 0.00019991398, -0.053222656, -0.026443481, 0.00008767843, 0.06347656, 0.00995636, 0.010536194, 0.021636963, 0.0013904572, -0.06542969, 0.026062012, -0.0030517578, 0.0...
http://arxiv.org/abs/2305.13631v2
EDIS: Entity-Driven Image Search over Multimodal Web Content
Siqi Liu, Weixi Feng, Tsu-jui Fu, Wenhu Chen, William Yang Wang
cs.CL, cs.CV, cs.IR
Making image retrieval methods practical for real-world search applications requires significant progress in dataset scales, entity comprehension, and multimodal information fusion. In this work, we introduce \textbf{E}ntity-\textbf{D}riven \textbf{I}mage \textbf{S}earch (EDIS), a challenging dataset for cross-modal im...
http://arxiv.org/pdf/2305.13631v2
2023-10-23T05:42:51
2023-05-23T02:59:19
{ "title": [ -0.0045661926, 0.0021572113, -0.03665161, -0.036834717, 0.03375244, -0.020721436, 0.016983032, -0.0158844, 0.029190063, 0.042510986, -0.022109985, 0.00894928, 0.02746582, -0.007827759, -0.05038452, 0.014808655, -0.010276794, 0.020812...
http://arxiv.org/abs/2305.13628v2
Improving Self-training for Cross-lingual Named Entity Recognition with Contrastive and Prototype Learning
Ran Zhou, Xin Li, Lidong Bing, Erik Cambria, Chunyan Miao
cs.CL
In cross-lingual named entity recognition (NER), self-training is commonly used to bridge the linguistic gap by training on pseudo-labeled target-language data. However, due to sub-optimal performance on target languages, the pseudo labels are often noisy and limit the overall performance. In this work, we aim to impro...
http://arxiv.org/pdf/2305.13628v2
2023-06-04T16:32:41
2023-05-23T02:52:16
{ "title": [ -0.003282547, 0.008094788, -0.025039673, -0.050689697, 0.010948181, -0.005138397, -0.019821167, -0.0072631836, 0.00036096573, 0.025512695, -0.02330017, 0.035827637, -0.013145447, -0.016983032, -0.019454956, 0.0075302124, -0.007144928, ...
http://arxiv.org/abs/2305.13627v2
InstructAlign: High-and-Low Resource Language Alignment via Continual Crosslingual Instruction Tuning
Samuel Cahyawijaya, Holy Lovenia, Tiezheng Yu, Willy Chung, Pascale Fung
cs.CL, cs.AI
Large language models (LLMs) that are tuned with instructions have demonstrated remarkable capabilities in various tasks and languages. However, their ability to generalize to underrepresented languages is limited due to the scarcity of available data. Additionally, directly adapting new languages to instruction-tuned ...
http://arxiv.org/pdf/2305.13627v2
2023-10-24T08:08:33
2023-05-23T02:51:34
{ "title": [ -0.06173706, -0.019317627, -0.028060913, -0.049316406, -0.017547607, 0.010810852, -0.05960083, 0.00096035004, -0.012901306, 0.04550171, -0.009338379, 0.02027893, 0.00630188, -0.020263672, -0.028533936, 0.009307861, -0.008300781, 0.03...
http://arxiv.org/abs/2305.13626v2
Prompting and Evaluating Large Language Models for Proactive Dialogues: Clarification, Target-guided, and Non-collaboration
Yang Deng, Lizi Liao, Liang Chen, Hongru Wang, Wenqiang Lei, Tat-Seng Chua
cs.CL
Conversational systems based on Large Language Models (LLMs), such as ChatGPT, show exceptional proficiency in context understanding and response generation. However, despite their impressive capabilities, they still possess limitations, such as providing randomly-guessed answers to ambiguous queries or failing to refu...
http://arxiv.org/pdf/2305.13626v2
2023-10-14T06:36:05
2023-05-23T02:49:35
{ "title": [ -0.024902344, 0.00096559525, -0.048706055, -0.070739746, 0.002576828, 0.0022735596, -0.082458496, 0.0025024414, -0.000156641, 0.009590149, 0.0046424866, -0.009414673, -0.013458252, -0.038879395, -0.036010742, -0.006664276, -0.0031929016,...
http://arxiv.org/abs/2305.13623v1
Validating Multimedia Content Moderation Software via Semantic Fusion
Wenxuan Wang, Jingyuan Huang, Chang Chen, Jiazhen Gu, Jianping Zhang, Weibin Wu, Pinjia He, Michael Lyu
cs.SE, cs.AI, cs.CL, cs.CV, cs.MM
The exponential growth of social media platforms, such as Facebook and TikTok, has revolutionized communication and content publication in human society. Users on these platforms can publish multimedia content that delivers information via the combination of text, audio, images, and video. Meanwhile, the multimedia con...
http://arxiv.org/pdf/2305.13623v1
2023-05-23T02:44:15
2023-05-23T02:44:15
{ "title": [ 0.024154663, 0.022125244, -0.017852783, -0.06829834, 0.025146484, 0.0056915283, -0.029891968, -0.021621704, -0.014282227, 0.05307007, -0.010398865, 0.0013256073, 0.047943115, -0.02607727, -0.028503418, 0.004814148, -0.038024902, 0.06...
http://arxiv.org/abs/2305.13617v3
SPEECH: Structured Prediction with Energy-Based Event-Centric Hyperspheres
Shumin Deng, Shengyu Mao, Ningyu Zhang, Bryan Hooi
cs.CL, cs.AI, cs.LG
Event-centric structured prediction involves predicting structured outputs of events. In most NLP cases, event structures are complex with manifold dependency, and it is challenging to effectively represent these complicated structured events. To address these issues, we propose Structured Prediction with Energy-based ...
http://arxiv.org/pdf/2305.13617v3
2023-09-18T10:19:10
2023-05-23T02:28:09
{ "title": [ -0.010749817, 0.0028533936, -0.041809082, -0.11859131, -0.0048561096, 0.01777649, -0.083618164, 0.006828308, 0.0012645721, 0.04940796, 0.00085401535, 0.008865356, 0.03793335, -0.023773193, -0.020889282, 0.02357483, 0.017059326, 0.014...
http://arxiv.org/abs/2305.13614v1
LLM-empowered Chatbots for Psychiatrist and Patient Simulation: Application and Evaluation
Siyuan Chen, Mengyue Wu, Kenny Q. Zhu, Kunyao Lan, Zhiling Zhang, Lyuchun Cui
cs.CL
Empowering chatbots in the field of mental health is receiving increasing amount of attention, while there still lacks exploration in developing and evaluating chatbots in psychiatric outpatient scenarios. In this work, we focus on exploring the potential of ChatGPT in powering chatbots for psychiatrist and patient sim...
http://arxiv.org/pdf/2305.13614v1
2023-05-23T02:25:01
2023-05-23T02:25:01
{ "title": [ -0.0029182434, -0.024169922, -0.006259918, -0.041229248, -0.02784729, 0.0068969727, -0.040222168, -0.012825012, -0.024871826, 0.029663086, -0.007926941, 0.020980835, 0.015617371, -0.040527344, -0.0069885254, 0.0000029802322, -0.016174316...
http://arxiv.org/abs/2305.13602v2
ReSee: Responding through Seeing Fine-grained Visual Knowledge in Open-domain Dialogue
Haoqin Tu, Yitong Li, Fei Mi, Zhongliang Yang
cs.CL
Incorporating visual knowledge into text-only dialogue systems has become a potential direction to imitate the way humans think, imagine, and communicate. However, existing multimodal dialogue systems are either confined by the scale and quality of available datasets or the coarse concept of visual knowledge. To addres...
http://arxiv.org/pdf/2305.13602v2
2023-10-20T04:59:10
2023-05-23T02:08:56
{ "title": [ -0.0040779114, -0.025222778, -0.030929565, -0.0395813, 0.000062704086, -0.014625549, 0.0019311905, -0.0015077591, 0.01864624, 0.018371582, -0.016860962, 0.0020484924, -0.022888184, -0.01826477, -0.06744385, -0.020568848, 0.012794495, ...
http://arxiv.org/abs/2305.13599v3
Decoupled Rationalization with Asymmetric Learning Rates: A Flexible Lipschitz Restraint
Wei Liu, Jun Wang, Haozhao Wang, Ruixuan Li, Yang Qiu, YuanKai Zhang, Jie Han, Yixiong Zou
cs.LG, cs.CL
A self-explaining rationalization model is generally constructed by a cooperative game where a generator selects the most human-intelligible pieces from the input text as rationales, followed by a predictor that makes predictions based on the selected rationales. However, such a cooperative game may incur the degenerat...
http://arxiv.org/pdf/2305.13599v3
2023-06-24T08:54:12
2023-05-23T02:01:13
{ "title": [ -0.02104187, 0.003856659, -0.015777588, -0.01576233, -0.00831604, -0.0034160614, 0.0073013306, -0.031021118, -0.042938232, 0.013969421, -0.04248047, 0.02319336, -0.04840088, -0.010360718, -0.012962341, -0.040802002, 0.012825012, 0.02...
http://arxiv.org/abs/2305.13592v2
Understanding Programs by Exploiting (Fuzzing) Test Cases
Jianyu Zhao, Yuyang Rong, Yiwen Guo, Yifeng He, Hao Chen
cs.LG, cs.AI, cs.CL, cs.CR, cs.SE
Semantic understanding of programs has attracted great attention in the community. Inspired by recent successes of large language models (LLMs) in natural language understanding, tremendous progress has been made by treating programming language as another sort of natural language and training LLMs on corpora of progra...
http://arxiv.org/pdf/2305.13592v2
2023-06-12T00:56:21
2023-05-23T01:51:46
{ "title": [ 0.00051403046, -0.0051498413, -0.011100769, -0.058776855, -0.002916336, -0.005870819, -0.015365601, -0.015106201, -0.008277893, 0.020736694, -0.026687622, -0.011375427, -0.017562866, -0.024810791, -0.030853271, -0.012924194, 0.003982544,...
http://arxiv.org/abs/2305.13589v1
BiasX: "Thinking Slow" in Toxic Content Moderation with Explanations of Implied Social Biases
Yiming Zhang, Sravani Nanduri, Liwei Jiang, Tongshuang Wu, Maarten Sap
cs.CL
Toxicity annotators and content moderators often default to mental shortcuts when making decisions. This can lead to subtle toxicity being missed, and seemingly toxic but harmless content being over-detected. We introduce BiasX, a framework that enhances content moderation setups with free-text explanations of statemen...
http://arxiv.org/pdf/2305.13589v1
2023-05-23T01:45:18
2023-05-23T01:45:18
{ "title": [ 0.013832092, 0.009063721, -0.036193848, -0.010314941, -0.011505127, -0.030715942, 0.019256592, -0.034851074, -0.012313843, 0.023101807, -0.04196167, 0.010284424, 0.043182373, -0.035583496, 0.0063591003, -0.029937744, -0.023605347, 0....
http://arxiv.org/abs/2305.13585v1
Query Structure Modeling for Inductive Logical Reasoning Over Knowledge Graphs
Siyuan Wang, Zhongyu Wei, Meng Han, Zhihao Fan, Haijun Shan, Qi Zhang, Xuanjing Huang
cs.CL
Logical reasoning over incomplete knowledge graphs to answer complex logical queries is a challenging task. With the emergence of new entities and relations in constantly evolving KGs, inductive logical reasoning over KGs has become a crucial problem. However, previous PLMs-based methods struggle to model the logical s...
http://arxiv.org/pdf/2305.13585v1
2023-05-23T01:25:29
2023-05-23T01:25:29
{ "title": [ -0.03262329, -0.0077705383, -0.060638428, -0.07489014, 0.0059280396, 0.006160736, -0.05050659, 0.024093628, -0.012374878, 0.009635925, -0.04373169, 0.0054855347, -0.02217102, -0.012390137, -0.018875122, -0.026168823, -0.0031108856, 0...
http://arxiv.org/abs/2305.13583v4
Cross-Attention is Not Enough: Incongruity-Aware Dynamic Hierarchical Fusion for Multimodal Affect Recognition
Yaoting Wang, Yuanchao Li, Paul Pu Liang, Louis-Philippe Morency, Peter Bell, Catherine Lai
cs.CL, cs.MM, eess.AS, eess.IV
Fusing multiple modalities has proven effective for multimodal information processing. However, the incongruity between modalities poses a challenge for multimodal fusion, especially in affect recognition. In this study, we first analyze how the salient affective information in one modality can be affected by the other...
http://arxiv.org/pdf/2305.13583v4
2023-11-13T00:09:47
2023-05-23T01:24:15
{ "title": [ 0.008552551, 0.011711121, -0.06427002, -0.089538574, 0.0078086853, -0.024459839, -0.03036499, 0.01197052, 0.023513794, 0.045959473, -0.020751953, 0.0026817322, 0.033843994, -0.008880615, -0.05444336, 0.0038661957, -0.011734009, 0.008...
http://arxiv.org/abs/2305.13582v2
Better Low-Resource Entity Recognition Through Translation and Annotation Fusion
Yang Chen, Vedaant Shah, Alan Ritter
cs.CL
Pre-trained multilingual language models have enabled significant advancements in cross-lingual transfer. However, these models often exhibit a performance disparity when transferring from high-resource languages to low-resource languages, especially for languages that are underrepresented or not in the pre-training da...
http://arxiv.org/pdf/2305.13582v2
2023-05-24T04:20:10
2023-05-23T01:23:22
{ "title": [ -0.017868042, -0.0028095245, -0.039916992, -0.075927734, 0.019195557, 0.00019133091, -0.03994751, 0.0061569214, 0.002790451, 0.045074463, -0.017105103, 0.008285522, 0.023986816, -0.006904602, -0.042022705, 0.010032654, -0.0013923645, ...
http://arxiv.org/abs/2305.13571v1
Latent Positional Information is in the Self-Attention Variance of Transformer Language Models Without Positional Embeddings
Ta-Chung Chi, Ting-Han Fan, Li-Wei Chen, Alexander I. Rudnicky, Peter J. Ramadge
cs.CL
The use of positional embeddings in transformer language models is widely accepted. However, recent research has called into question the necessity of such embeddings. We further extend this inquiry by demonstrating that a randomly initialized and frozen transformer language model, devoid of positional embeddings, inhe...
http://arxiv.org/pdf/2305.13571v1
2023-05-23T01:03:40
2023-05-23T01:03:40
{ "title": [ 0.01889038, 0.010864258, -0.036834717, 0.010063171, -0.013648987, -0.046081543, -0.0039024353, -0.033233643, 0.02432251, 0.05166626, -0.013244629, -0.00491333, 0.02949524, -0.027816772, -0.015014648, 0.01876831, 0.0073890686, 0.03448...
http://arxiv.org/abs/2305.18323v1
ReWOO: Decoupling Reasoning from Observations for Efficient Augmented Language Models
Binfeng Xu, Zhiyuan Peng, Bowen Lei, Subhabrata Mukherjee, Yuchen Liu, Dongkuan Xu
cs.CL, cs.AI
Augmented Language Models (ALMs) blend the reasoning capabilities of Large Language Models (LLMs) with tools that allow for knowledge retrieval and action execution. Existing ALM systems trigger LLM thought processes while pulling observations from these tools in an interleaved fashion. Specifically, an LLM reasons to ...
http://arxiv.org/pdf/2305.18323v1
2023-05-23T00:16:48
2023-05-23T00:16:48
{ "title": [ -0.02734375, -0.032043457, -0.021652222, -0.08874512, -0.018463135, 0.024307251, -0.07281494, 0.016159058, 0.018707275, 0.014328003, -0.014282227, 0.006160736, 0.011581421, -0.0178833, -0.0256958, 0.014160156, -0.02180481, 0.04193115...
http://arxiv.org/abs/2305.13551v2
How Fragile is Relation Extraction under Entity Replacements?
Yiwei Wang, Bryan Hooi, Fei Wang, Yujun Cai, Yuxuan Liang, Wenxuan Zhou, Jing Tang, Manjuan Duan, Muhao Chen
cs.CL, cs.AI
Relation extraction (RE) aims to extract the relations between entity names from the textual context. In principle, textual context determines the ground-truth relation and the RE models should be able to correctly identify the relations reflected by the textual context. However, existing work has found that the RE mod...
http://arxiv.org/pdf/2305.13551v2
2023-05-30T02:25:53
2023-05-22T23:53:32
{ "title": [ 0.010978699, -0.0017728806, -0.043701172, -0.062164307, 0.009864807, -0.014198303, -0.019958496, -0.002023697, -0.01600647, 0.013183594, -0.016174316, -0.0024757385, 0.018737793, -0.01953125, -0.013214111, -0.0007119179, -0.025497437, ...
http://arxiv.org/abs/2305.13547v2
Self-Evolution Learning for Mixup: Enhance Data Augmentation on Few-Shot Text Classification Tasks
Haoqi Zheng, Qihuang Zhong, Liang Ding, Zhiliang Tian, Xin Niu, Dongsheng Li, Dacheng Tao
cs.CL, cs.NI
Text classification tasks often encounter few shot scenarios with limited labeled data, and addressing data scarcity is crucial. Data augmentation with mixup has shown to be effective on various text classification tasks. However, most of the mixup methods do not consider the varying degree of learning difficulty in di...
http://arxiv.org/pdf/2305.13547v2
2023-10-22T16:38:51
2023-05-22T23:43:23
{ "title": [ -0.0040130615, 0.018295288, -0.05508423, -0.015266418, 0.016326904, -0.029525757, 0.010772705, -0.0063591003, -0.0042533875, 0.03161621, -0.02218628, 0.030258179, -0.036621094, -0.009002686, -0.05380249, -0.017440796, -0.0058021545, ...
http://arxiv.org/abs/2305.13535v1
Improving Classifier Robustness through Active Generation of Pairwise Counterfactuals
Ananth Balashankar, Xuezhi Wang, Yao Qin, Ben Packer, Nithum Thain, Jilin Chen, Ed H. Chi, Alex Beutel
cs.CL, cs.LG
Counterfactual Data Augmentation (CDA) is a commonly used technique for improving robustness in natural language classifiers. However, one fundamental challenge is how to discover meaningful counterfactuals and efficiently label them, with minimal human labeling cost. Most existing methods either completely rely on hum...
http://arxiv.org/pdf/2305.13535v1
2023-05-22T23:19:01
2023-05-22T23:19:01
{ "title": [ 0.020767212, 0.015945435, -0.007698059, -0.021606445, 0.026901245, -0.024719238, -0.0019044876, -0.05593872, 0.042236328, 0.02432251, -0.025146484, 0.018035889, -0.03137207, -0.01878357, -0.05130005, -0.0034160614, -0.015670776, 0.04...
http://arxiv.org/abs/2305.13534v1
How Language Model Hallucinations Can Snowball
Muru Zhang, Ofir Press, William Merrill, Alisa Liu, Noah A. Smith
cs.CL
A major risk of using language models in practical applications is their tendency to hallucinate incorrect statements. Hallucinations are often attributed to knowledge gaps in LMs, but we hypothesize that in some cases, when justifying previously generated hallucinations, LMs output false claims that they can separatel...
http://arxiv.org/pdf/2305.13534v1
2023-05-22T23:14:28
2023-05-22T23:14:28
{ "title": [ 0.04373169, 0.008354187, -0.016937256, -0.05307007, -0.04333496, -0.023406982, -0.04940796, -0.0076026917, -0.042999268, 0.05014038, -0.035308838, 0.0104904175, 0.027282715, -0.026916504, -0.00037240982, -0.0031871796, 0.027572632, 0...
http://arxiv.org/abs/2305.13533v2
Open-world Semi-supervised Generalized Relation Discovery Aligned in a Real-world Setting
William Hogan, Jiacheng Li, Jingbo Shang
cs.CL, I.2.7
Open-world Relation Extraction (OpenRE) has recently garnered significant attention. However, existing approaches tend to oversimplify the problem by assuming that all unlabeled texts belong to novel classes, thereby limiting the practicality of these methods. We argue that the OpenRE setting should be more aligned wit...
http://arxiv.org/pdf/2305.13533v2
2023-11-01T18:36:28
2023-05-22T23:12:57
{ "title": [ -0.00724411, -0.011192322, -0.064208984, -0.079589844, 0.0013923645, -0.011054993, -0.041229248, 0.019866943, -0.0013818741, 0.012565613, -0.029678345, -0.00006604195, -0.003326416, -0.014030457, -0.017425537, -0.015655518, 0.0051498413,...
http://arxiv.org/abs/2305.13530v1
The Grammar and Syntax Based Corpus Analysis Tool For The Ukrainian Language
Daria Stetsenko, Inez Okulska
cs.CL, cs.AI
This paper provides an overview of a text mining tool the StyloMetrix developed initially for the Polish language and further extended for English and recently for Ukrainian. The StyloMetrix is built upon various metrics crafted manually by computational linguists and researchers from literary studies to analyze gramma...
http://arxiv.org/pdf/2305.13530v1
2023-05-22T22:52:47
2023-05-22T22:52:47
{ "title": [ 0.013038635, -0.0061798096, -0.024597168, -0.051208496, -0.013145447, -0.0038642883, -0.018295288, 0.008682251, -0.020965576, 0.052612305, -0.062561035, -0.0019054413, 0.03753662, -0.007255554, -0.028076172, -0.019332886, -0.0047836304, ...
http://arxiv.org/abs/2305.13528v2
Transfer-Free Data-Efficient Multilingual Slot Labeling
Evgeniia Razumovskaia, Ivan Vulić, Anna Korhonen
cs.CL
Slot labeling (SL) is a core component of task-oriented dialogue (ToD) systems, where slots and corresponding values are usually language-, task- and domain-specific. Therefore, extending the system to any new language-domain-task configuration requires (re)running an expensive and resource-intensive data annotation pr...
http://arxiv.org/pdf/2305.13528v2
2023-11-12T10:17:53
2023-05-22T22:47:32
{ "title": [ -0.017959595, -0.01651001, -0.026062012, -0.080078125, -0.009109497, 0.015899658, -0.051696777, 0.010848999, -0.006855011, 0.05239868, -0.007095337, 0.007587433, -0.016220093, -0.016494751, -0.041137695, 0.0054779053, -0.00074386597, ...
http://arxiv.org/abs/2305.13527v2
Aligning the Norwegian UD Treebank with Entity and Coreference Information
Tollef Emil Jørgensen, Andre Kåsen
cs.CL, I.2.7
This paper presents a merged collection of entity and coreference annotated data grounded in the Universal Dependencies (UD) treebanks for the two written forms of Norwegian: Bokm{\aa}l and Nynorsk. The aligned and converted corpora are the Norwegian Named Entities (NorNE) and Norwegian Anaphora Resolution Corpus (NARC...
http://arxiv.org/pdf/2305.13527v2
2023-05-25T22:36:36
2023-05-22T22:44:53
{ "title": [ -0.002872467, -0.0119018555, -0.04147339, -0.09326172, 0.0031814575, 0.0011463165, -0.054656982, 0.012313843, 0.009155273, 0.059417725, -0.027664185, 0.011924744, 0.040283203, -0.010246277, -0.036590576, 0.03164673, -0.008369446, 0.0...
http://arxiv.org/abs/2305.18322v1
REFinD: Relation Extraction Financial Dataset
Simerjot Kaur, Charese Smiley, Akshat Gupta, Joy Sain, Dongsheng Wang, Suchetha Siddagangappa, Toyin Aguda, Sameena Shah
cs.CL, cs.AI
A number of datasets for Relation Extraction (RE) have been created to aide downstream tasks such as information retrieval, semantic search, question answering and textual entailment. However, these datasets fail to capture financial-domain specific challenges since most of these datasets are compiled using general kno...
http://arxiv.org/pdf/2305.18322v1
2023-05-22T22:40:11
2023-05-22T22:40:11
{ "title": [ -0.014320374, 0.003774643, -0.07171631, -0.04031372, -0.0032367706, -0.01777649, -0.04876709, -0.0021362305, 0.012229919, 0.031677246, -0.049316406, -0.0011043549, 0.037384033, -0.024551392, -0.02519226, -0.014266968, -0.0054893494, ...
http://arxiv.org/abs/2305.13523v1
A Study of Generative Large Language Model for Medical Research and Healthcare
Cheng Peng, Xi Yang, Aokun Chen, Kaleb E Smith, Nima PourNejatian, Anthony B Costa, Cheryl Martin, Mona G Flores, Ying Zhang, Tanja Magoc, Gloria Lipori, Duane A Mitchell, Naykky S Ospina, Mustafa M Ahmed, William R Hogan, Elizabeth A Shenkman, Yi Guo, Jiang Bian, Yonghui Wu
cs.CL
There is enormous enthusiasm and concerns in using large language models (LLMs) in healthcare, yet current assumptions are all based on general-purpose LLMs such as ChatGPT. This study develops a clinical generative LLM, GatorTronGPT, using 277 billion words of mixed clinical and English text with a GPT-3 architecture ...
http://arxiv.org/pdf/2305.13523v1
2023-05-22T22:37:24
2023-05-22T22:37:24
{ "title": [ -0.032409668, -0.008346558, -0.056488037, -0.06781006, 0.009918213, -0.00022351742, -0.06939697, 0.0017414093, -0.013084412, 0.044830322, -0.0052871704, 0.015960693, 0.004055023, -0.022155762, -0.033172607, 0.011413574, -0.01058197, ...
http://arxiv.org/abs/2305.13521v1
CEO: Corpus-based Open-Domain Event Ontology Induction
Nan Xu, Hongming Zhang, Jianshu Chen
cs.CL
Existing event-centric NLP models often only apply to the pre-defined ontology, which significantly restricts their generalization capabilities. This paper presents CEO, a novel Corpus-based Event Ontology induction model to relax the restriction imposed by pre-defined event ontologies. Without direct supervision, CEO ...
http://arxiv.org/pdf/2305.13521v1
2023-05-22T22:26:47
2023-05-22T22:26:47
{ "title": [ 0.0049705505, 0.0059776306, -0.03656006, -0.09680176, 0.035003662, 0.009613037, -0.05041504, -0.0028629303, 0.027038574, 0.010673523, -0.013412476, 0.0063209534, 0.0073776245, -0.030227661, -0.021942139, 0.017089844, -0.023406982, 0....
http://arxiv.org/abs/2305.13516v1
Scaling Speech Technology to 1,000+ Languages
Vineel Pratap, Andros Tjandra, Bowen Shi, Paden Tomasello, Arun Babu, Sayani Kundu, Ali Elkahky, Zhaoheng Ni, Apoorv Vyas, Maryam Fazel-Zarandi, Alexei Baevski, Yossi Adi, Xiaohui Zhang, Wei-Ning Hsu, Alexis Conneau, Michael Auli
cs.CL, cs.SD, eess.AS
Expanding the language coverage of speech technology has the potential to improve access to information for many more people. However, current speech technology is restricted to about one hundred languages which is a small fraction of the over 7,000 languages spoken around the world. The Massively Multilingual Speech (...
http://arxiv.org/pdf/2305.13516v1
2023-05-22T22:09:41
2023-05-22T22:09:41
{ "title": [ 0.03805542, -0.003479004, -0.05987549, -0.07550049, -0.014137268, -0.0262146, -0.043945312, 0.03555298, -0.013793945, 0.034118652, -0.009643555, -0.0058784485, 0.022964478, -0.047729492, 0.020263672, -0.0070114136, 0.013381958, 0.030...
http://arxiv.org/abs/2305.13514v1
Small Language Models Improve Giants by Rewriting Their Outputs
Giorgos Vernikos, Arthur Bražinskas, Jakub Adamek, Jonathan Mallinson, Aliaksei Severyn, Eric Malmi
cs.CL, cs.LG
Large language models (LLMs) have demonstrated impressive few-shot learning capabilities, but they often underperform compared to fine-tuned models on challenging tasks. Furthermore, their large size and restricted access only through APIs make task-specific fine-tuning impractical. Moreover, LLMs are sensitive to diff...
http://arxiv.org/pdf/2305.13514v1
2023-05-22T22:07:50
2023-05-22T22:07:50
{ "title": [ -0.030227661, -0.034240723, -0.04434204, -0.053894043, -0.06756592, 0.0035209656, -0.09411621, 0.017425537, -0.04196167, 0.050476074, -0.017105103, -0.009880066, 0.015823364, -0.035217285, -0.008056641, -0.010986328, 0.013259888, -0....
http://arxiv.org/abs/2305.13512v2
Can ChatGPT Detect Intent? Evaluating Large Language Models for Spoken Language Understanding
Mutian He, Philip N. Garner
cs.CL, cs.AI, cs.SD, eess.AS
Recently, large pretrained language models have demonstrated strong language understanding capabilities. This is particularly reflected in their zero-shot and in-context learning abilities on downstream tasks through prompting. To assess their impact on spoken language understanding (SLU), we evaluate several such mode...
http://arxiv.org/pdf/2305.13512v2
2023-08-17T19:23:20
2023-05-22T21:59:26
{ "title": [ -0.014457703, 0.0008907318, -0.026901245, -0.0340271, -0.0034713745, -0.0037403107, -0.04248047, -0.00881958, 0.00042533875, 0.018676758, -0.01524353, -0.006702423, 0.0082092285, -0.016983032, -0.032348633, 0.0028095245, -0.024307251, ...
http://arxiv.org/abs/2305.13507v3
Multimodal Automated Fact-Checking: A Survey
Mubashara Akhtar, Michael Schlichtkrull, Zhijiang Guo, Oana Cocarascu, Elena Simperl, Andreas Vlachos
cs.CL, cs.AI, cs.CV
Misinformation is often conveyed in multiple modalities, e.g. a miscaptioned image. Multimodal misinformation is perceived as more credible by humans, and spreads faster than its text-only counterparts. While an increasing body of research investigates automated fact-checking (AFC), previous surveys mostly focus on tex...
http://arxiv.org/pdf/2305.13507v3
2023-10-25T18:43:02
2023-05-22T21:52:24
{ "title": [ -0.015205383, 0.009124756, -0.03817749, -0.06640625, 0.010261536, 0.0007119179, -0.035736084, -0.013748169, -0.023529053, 0.04824829, -0.0124435425, 0.015808105, 0.014923096, -0.009841919, -0.041778564, -0.005176544, -0.03567505, 0.0...
http://arxiv.org/abs/2305.13504v1
Neural Machine Translation for Code Generation
Dharma KC, Clayton T. Morrison
cs.CL, cs.LG, A.1
Neural machine translation (NMT) methods developed for natural language processing have been shown to be highly successful in automating translation from one natural language to another. Recently, these NMT methods have been adapted to the generation of program code. In NMT for code generation, the task is to generate ...
http://arxiv.org/pdf/2305.13504v1
2023-05-22T21:43:12
2023-05-22T21:43:12
{ "title": [ -0.047698975, -0.023544312, -0.015571594, -0.049072266, -0.014564514, 0.0234375, -0.05392456, 0.010879517, -0.00046348572, 0.045074463, -0.0076904297, -0.0062789917, 0.014007568, -0.016845703, -0.04107666, 0.014060974, -0.009048462, ...
http://arxiv.org/abs/2305.13499v2
Learning Easily Updated General Purpose Text Representations with Adaptable Task-Specific Prefixes
Kuan-Hao Huang, Liang Tan, Rui Hou, Sinong Wang, Amjad Almahairi, Ruty Rinott
cs.CL
Many real-world applications require making multiple predictions from the same text. Fine-tuning a large pre-trained language model for each downstream task causes computational burdens in the inference time due to several times of forward passes. To amortize the computational cost, freezing the language model and buil...
http://arxiv.org/pdf/2305.13499v2
2023-10-14T15:35:07
2023-05-22T21:31:03
{ "title": [ -0.0184021, -0.006160736, -0.028015137, -0.05722046, -0.004470825, 0.006717682, -0.055023193, 0.0044059753, 0.013496399, 0.03643799, 0.0013713837, 0.0082473755, -0.0038280487, -0.005050659, -0.043273926, 0.015960693, -0.0028705597, 0...
http://arxiv.org/abs/2305.13484v3
Flover: A Temporal Fusion Framework for Efficient Autoregressive Model Parallel Inference
Jinghan Yao, Nawras Alnaasan, Tian Chen, Aamir Shafi, Hari Subramoni, Dhabaleswar K., Panda
cs.DC, cs.AI, cs.CL, cs.CV, cs.LG
Autoregressive models, despite their commendable performance in a myriad of generative tasks, face challenges stemming from their inherently sequential structure. Inference on these models, by design, harnesses a temporal dependency, where the current token's probability distribution is conditioned on preceding tokens....
http://arxiv.org/pdf/2305.13484v3
2023-11-03T03:37:20
2023-05-22T20:58:09
{ "title": [ -0.015541077, 0.0067367554, -0.06298828, -0.099487305, 0.020858765, 0.002532959, -0.06088257, -0.0070114136, -0.0014181137, 0.0413208, -0.0061454773, 0.02128601, -0.0051727295, -0.011161804, -0.04067993, 0.00092983246, 0.00642395, 0....
http://arxiv.org/abs/2305.13478v2
Automatic Readability Assessment for Closely Related Languages
Joseph Marvin Imperial, Ekaterina Kochmar
cs.CL
In recent years, the main focus of research on automatic readability assessment (ARA) has shifted towards using expensive deep learning-based methods with the primary goal of increasing models' accuracy. This, however, is rarely applicable for low-resource languages where traditional handcrafted features are still wide...
http://arxiv.org/pdf/2305.13478v2
2023-05-25T10:41:50
2023-05-22T20:42:53
{ "title": [ -0.038970947, 0.014884949, -0.021347046, -0.032287598, -0.007843018, 0.010475159, -0.05432129, -0.027389526, 0.006855011, 0.0357666, -0.037384033, -0.00415802, 0.010467529, -0.0042304993, -0.055419922, -0.0049743652, -0.008781433, 0....
http://arxiv.org/abs/2305.13477v2
Look-back Decoding for Open-Ended Text Generation
Nan Xu, Chunting Zhou, Asli Celikyilmaz, Xuezhe Ma
cs.CL
Given a prefix (context), open-ended generation aims to decode texts that are coherent, which do not abruptly drift from previous topics, and informative, which do not suffer from undesired repetitions. In this paper, we propose Look-back, an improved decoding algorithm that leverages the Kullback-Leibler divergence to...
http://arxiv.org/pdf/2305.13477v2
2023-10-23T00:58:02
2023-05-22T20:42:37
{ "title": [ -0.054229736, -0.020629883, -0.016235352, -0.044281006, -0.04284668, 0.021896362, -0.080322266, -0.013206482, 0.019378662, 0.04837036, 0.0015182495, -0.003282547, 0.01725769, -0.01676941, -0.055236816, 0.028182983, 0.002002716, 0.024...
http://arxiv.org/abs/2305.13469v2
MAILEX: Email Event and Argument Extraction
Saurabh Srivastava, Gaurav Singh, Shou Matsumoto, Ali Raz, Paulo Costa, Joshua Poore, Ziyu Yao
cs.CL, cs.AI
In this work, we present the first dataset, MailEx, for performing event extraction from conversational email threads. To this end, we first proposed a new taxonomy covering 10 event types and 76 arguments in the email domain. Our final dataset includes 1.5K email threads and ~4K emails, which are annotated with totall...
http://arxiv.org/pdf/2305.13469v2
2023-10-21T02:15:22
2023-05-22T20:28:23
{ "title": [ 0.02406311, -0.01348114, -0.05834961, -0.07775879, -0.0033874512, -0.014221191, -0.04550171, 0.006500244, -0.04373169, 0.03451538, 0.006198883, -0.0014877319, 0.050994873, -0.055267334, 0.013580322, -0.0011949539, -0.012054443, 0.047...
http://arxiv.org/abs/2305.13455v2
Clembench: Using Game Play to Evaluate Chat-Optimized Language Models as Conversational Agents
Kranti Chalamalasetti, Jana Götze, Sherzod Hakimov, Brielen Madureira, Philipp Sadler, David Schlangen
cs.CL
Recent work has proposed a methodology for the systematic evaluation of "Situated Language Understanding Agents"-agents that operate in rich linguistic and non-linguistic contexts-through testing them in carefully constructed interactive settings. Other recent work has argued that Large Language Models (LLMs), if suita...
http://arxiv.org/pdf/2305.13455v2
2023-10-27T10:27:37
2023-05-22T19:56:10
{ "title": [ -0.027923584, -0.007926941, -0.009742737, -0.018493652, -0.013084412, -0.0021514893, -0.04638672, -0.020629883, -0.0010442734, 0.03552246, -0.024734497, 0.002046585, 0.0054473877, -0.013122559, -0.049438477, -0.00018334389, -0.025970459,...
http://arxiv.org/abs/2305.13417v1
Interpreting Transformer's Attention Dynamic Memory and Visualizing the Semantic Information Flow of GPT
Shahar Katz, Yonatan Belinkov
cs.CL, 68T50, I.2.7
Recent advances in interpretability suggest we can project weights and hidden states of transformer-based language models (LMs) to their vocabulary, a transformation that makes them human interpretable and enables us to assign semantics to what was seen only as numerical vectors. In this paper, we interpret LM attentio...
http://arxiv.org/pdf/2305.13417v1
2023-05-22T19:04:56
2023-05-22T19:04:56
{ "title": [ 0.009979248, 0.024551392, -0.016571045, 0.0010843277, -0.007648468, -0.018615723, -0.009094238, -0.04446411, -0.015304565, 0.026748657, -0.030792236, 0.0021915436, 0.0129776, -0.020080566, -0.016616821, -0.012428284, 0.018341064, 0.0...
http://arxiv.org/abs/2305.13413v1
Syntactic Knowledge via Graph Attention with BERT in Machine Translation
Yuqian Dai, Serge Sharoff, Marc de Kamps
cs.CL
Although the Transformer model can effectively acquire context features via a self-attention mechanism, deeper syntactic knowledge is still not effectively modeled. To alleviate the above problem, we propose Syntactic knowledge via Graph attention with BERT (SGB) in Machine Translation (MT) scenarios. Graph Attention N...
http://arxiv.org/pdf/2305.13413v1
2023-05-22T18:56:14
2023-05-22T18:56:14
{ "title": [ -0.026626587, -0.019378662, 0.008628845, -0.064086914, -0.049560547, 0.026657104, -0.070251465, 0.016494751, 0.008956909, 0.04840088, -0.021011353, 0.016647339, 0.037353516, -0.010696411, -0.03970337, 0.014511108, -0.011688232, 0.027...
http://arxiv.org/abs/2305.13412v1
Element-aware Summarization with Large Language Models: Expert-aligned Evaluation and Chain-of-Thought Method
Yiming Wang, Zhuosheng Zhang, Rui Wang
cs.CL
Automatic summarization generates concise summaries that contain key ideas of source documents. As the most mainstream datasets for the news sub-domain, CNN/DailyMail and BBC XSum have been widely used for performance benchmarking. However, the reference summaries of those datasets turn out to be noisy, mainly in terms...
http://arxiv.org/pdf/2305.13412v1
2023-05-22T18:54:35
2023-05-22T18:54:35
{ "title": [ -0.012756348, 0.0051193237, -0.025146484, -0.06665039, -0.010910034, 0.017105103, -0.06854248, -0.014549255, 0.009056091, 0.040252686, -0.0043945312, 0.0015077591, 0.00920105, -0.010749817, -0.060913086, 0.0060653687, -0.012054443, 0...
http://arxiv.org/abs/2305.13408v1
Modular Domain Adaptation for Conformer-Based Streaming ASR
Qiujia Li, Bo Li, Dongseong Hwang, Tara N. Sainath, Pedro M. Mengibar
eess.AS, cs.CL, cs.LG, cs.SD
Speech data from different domains has distinct acoustic and linguistic characteristics. It is common to train a single multidomain model such as a Conformer transducer for speech recognition on a mixture of data from all domains. However, changing data in one domain or adding a new domain would require the multidomain...
http://arxiv.org/pdf/2305.13408v1
2023-05-22T18:49:35
2023-05-22T18:49:35
{ "title": [ -0.022384644, 0.0068588257, -0.09112549, -0.10925293, 0.022201538, -0.007980347, -0.060394287, 0.057495117, -0.021194458, 0.053375244, 0.0053596497, -0.0026836395, 0.01600647, -0.018127441, -0.038757324, 0.010238647, -0.005268097, 0....
http://arxiv.org/abs/2305.16333v1
Text Generation with Speech Synthesis for ASR Data Augmentation
Zhuangqun Huang, Gil Keren, Ziran Jiang, Shashank Jain, David Goss-Grubbs, Nelson Cheng, Farnaz Abtahi, Duc Le, David Zhang, Antony D'Avirro, Ethan Campbell-Taylor, Jessie Salas, Irina-Elena Veliche, Xi Chen
cs.CL, cs.AI, cs.LG, eess.AS
Aiming at reducing the reliance on expensive human annotations, data synthesis for Automatic Speech Recognition (ASR) has remained an active area of research. While prior work mainly focuses on synthetic speech generation for ASR data augmentation, its combination with text generation methods is considerably less explo...
http://arxiv.org/pdf/2305.16333v1
2023-05-22T18:45:20
2023-05-22T18:45:20
{ "title": [ -0.03564453, -0.010971069, -0.06390381, -0.06903076, -0.0115737915, 0.003452301, -0.059143066, 0.04232788, -0.02331543, 0.04864502, -0.033813477, 0.009628296, 0.01727295, -0.012420654, -0.03250122, 0.00026369095, -0.007297516, 0.0526...