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/2306.04743v1
ScienceBenchmark: A Complex Real-World Benchmark for Evaluating Natural Language to SQL Systems
Yi Zhang, Jan Deriu, George Katsogiannis-Meimarakis, Catherine Kosten, Georgia Koutrika, Kurt Stockinger
cs.DB, cs.AI, cs.CL, H.2.4; I.2.7
Natural Language to SQL systems (NL-to-SQL) have recently shown a significant increase in accuracy for natural language to SQL query translation. This improvement is due to the emergence of transformer-based language models, and the popularity of the Spider benchmark - the de-facto standard for evaluating NL-to-SQL sys...
http://arxiv.org/pdf/2306.04743v1
2023-06-07T19:37:55
2023-06-07T19:37:55
{ "title": [ -0.020355225, -0.015823364, -0.04348755, -0.05505371, 0.0026283264, -0.014175415, -0.031585693, 0.006340027, 0.0152282715, 0.038604736, -0.040985107, 0.021575928, -0.005607605, -0.010169983, -0.045166016, -0.011810303, -0.012145996, ...
http://arxiv.org/abs/2306.04735v1
Soft-prompt Tuning for Large Language Models to Evaluate Bias
Jacob-Junqi Tian, David Emerson, Sevil Zanjani Miyandoab, Deval Pandya, Laleh Seyyed-Kalantari, Faiza Khan Khattak
cs.CL, cs.AI, cs.LG
Prompting large language models has gained immense popularity in recent years due to the advantage of producing good results even without the need for labelled data. However, this requires prompt tuning to get optimal prompts that lead to better model performances. In this paper, we explore the use of soft-prompt tunin...
http://arxiv.org/pdf/2306.04735v1
2023-06-07T19:11:25
2023-06-07T19:11:25
{ "title": [ -0.027145386, -0.019470215, -0.035064697, -0.062683105, -0.015312195, 0.0014696121, -0.06628418, 0.0030117035, -0.0010061264, 0.023010254, -0.02748108, -0.004234314, 0.00806427, -0.025100708, -0.02633667, 0.00051403046, -0.0178833, 0...
http://arxiv.org/abs/2306.04724v1
Prompter: Zero-shot Adaptive Prefixes for Dialogue State Tracking Domain Adaptation
Taha Aksu, Min-Yen Kan, Nancy F. Chen
cs.CL
A challenge in the Dialogue State Tracking (DST) field is adapting models to new domains without using any supervised data, zero-shot domain adaptation. Parameter-Efficient Transfer Learning (PETL) has the potential to address this problem due to its robustness. However, it has yet to be applied to the zero-shot scenar...
http://arxiv.org/pdf/2306.04724v1
2023-06-07T18:39:57
2023-06-07T18:39:57
{ "title": [ -0.0030498505, -0.019989014, -0.05050659, -0.056854248, -0.013130188, -0.02633667, -0.029190063, 0.010765076, -0.0012130737, 0.01953125, 0.037750244, -0.0045928955, 0.005519867, -0.04989624, -0.0146865845, 0.007762909, -0.016983032, ...
http://arxiv.org/abs/2306.04723v2
Intrinsic Dimension Estimation for Robust Detection of AI-Generated Texts
Eduard Tulchinskii, Kristian Kuznetsov, Laida Kushnareva, Daniil Cherniavskii, Serguei Barannikov, Irina Piontkovskaya, Sergey Nikolenko, Evgeny Burnaev
cs.CL, cs.AI, cs.LG, math.AT, 68T50
Rapidly increasing quality of AI-generated content makes it difficult to distinguish between human and AI-generated texts, which may lead to undesirable consequences for society. Therefore, it becomes increasingly important to study the properties of human texts that are invariant over different text domains and varyin...
http://arxiv.org/pdf/2306.04723v2
2023-10-31T19:25:36
2023-06-07T18:38:04
{ "title": [ -0.044006348, 0.010124207, -0.04421997, 0.00025367737, 0.0054359436, -0.010231018, -0.0040130615, -0.028961182, -0.008270264, 0.02468872, -0.033233643, 0.013771057, 0.0068626404, -0.0129852295, -0.030593872, -0.014839172, -0.010925293, ...
http://arxiv.org/abs/2306.04707v1
Improving Open Language Models by Learning from Organic Interactions
Jing Xu, Da Ju, Joshua Lane, Mojtaba Komeili, Eric Michael Smith, Megan Ung, Morteza Behrooz, William Ngan, Rashel Moritz, Sainbayar Sukhbaatar, Y-Lan Boureau, Jason Weston, Kurt Shuster
cs.CL, cs.AI
We present BlenderBot 3x, an update on the conversational model BlenderBot 3, which is now trained using organic conversation and feedback data from participating users of the system in order to improve both its skills and safety. We are publicly releasing the participating de-identified interaction data for use by the...
http://arxiv.org/pdf/2306.04707v1
2023-06-07T18:19:46
2023-06-07T18:19:46
{ "title": [ -0.009963989, -0.005744934, -0.018676758, -0.09088135, 0.00042057037, 0.025619507, -0.091308594, -0.009780884, -0.017837524, 0.018096924, -0.018035889, 0.01826477, -0.018432617, -0.009719849, -0.02684021, -0.011909485, -0.026870728, ...
http://arxiv.org/abs/2306.04695v1
ConceptBed: Evaluating Concept Learning Abilities of Text-to-Image Diffusion Models
Maitreya Patel, Tejas Gokhale, Chitta Baral, Yezhou Yang
cs.CV, cs.CL, cs.LG
The ability to understand visual concepts and replicate and compose these concepts from images is a central goal for computer vision. Recent advances in text-to-image (T2I) models have lead to high definition and realistic image quality generation by learning from large databases of images and their descriptions. Howev...
http://arxiv.org/pdf/2306.04695v1
2023-06-07T18:00:38
2023-06-07T18:00:38
{ "title": [ 0.003665924, 0.0018463135, -0.009750366, 0.001452446, -0.0015220642, -0.013168335, 0.023284912, -0.03326416, 0.031829834, 0.029296875, -0.030349731, 0.03111267, -0.013809204, -0.017593384, -0.04397583, 0.0055389404, -0.015411377, 0.0...
http://arxiv.org/abs/2306.04640v2
ModuleFormer: Modularity Emerges from Mixture-of-Experts
Yikang Shen, Zheyu Zhang, Tianyou Cao, Shawn Tan, Zhenfang Chen, Chuang Gan
cs.CL, cs.AI, cs.LG
Large Language Models (LLMs) have achieved remarkable results. However, existing models are expensive to train and deploy, and it is also difficult to expand their knowledge beyond pre-training data without forgetting previous knowledge. This paper proposes a new neural network architecture, ModuleFormer, that leverage...
http://arxiv.org/pdf/2306.04640v2
2023-09-11T19:31:26
2023-06-07T17:59:57
{ "title": [ 0.017852783, 0.0074157715, -0.089538574, -0.09863281, -0.0059394836, -0.026733398, -0.06933594, 0.024551392, -0.053466797, 0.019744873, 0.024932861, -0.002483368, -0.020568848, -0.066101074, 0.0031967163, -0.038085938, 0.0037765503, ...
http://arxiv.org/abs/2306.04637v2
Transformers as Statisticians: Provable In-Context Learning with In-Context Algorithm Selection
Yu Bai, Fan Chen, Huan Wang, Caiming Xiong, Song Mei
cs.LG, cs.AI, cs.CL, math.ST, stat.ML, stat.TH
Neural sequence models based on the transformer architecture have demonstrated remarkable \emph{in-context learning} (ICL) abilities, where they can perform new tasks when prompted with training and test examples, without any parameter update to the model. This work first provides a comprehensive statistical theory for...
http://arxiv.org/pdf/2306.04637v2
2023-07-06T16:55:36
2023-06-07T17:59:31
{ "title": [ 0.0048179626, 0.023422241, -0.029037476, -0.04901123, 0.03164673, -0.018737793, 0.0021457672, -0.034118652, -0.00017237663, 0.007205963, -0.056152344, 0.03366089, -0.030044556, -0.013031006, -0.0104904175, -0.012710571, 0.0008587837, ...
http://arxiv.org/abs/2306.04634v3
On the Reliability of Watermarks for Large Language Models
John Kirchenbauer, Jonas Geiping, Yuxin Wen, Manli Shu, Khalid Saifullah, Kezhi Kong, Kasun Fernando, Aniruddha Saha, Micah Goldblum, Tom Goldstein
cs.LG, cs.CL, cs.CR
As LLMs become commonplace, machine-generated text has the potential to flood the internet with spam, social media bots, and valueless content. Watermarking is a simple and effective strategy for mitigating such harms by enabling the detection and documentation of LLM-generated text. Yet a crucial question remains: How...
http://arxiv.org/pdf/2306.04634v3
2023-06-30T18:18:12
2023-06-07T17:58:48
{ "title": [ -0.017532349, 0.0011692047, -0.028640747, -0.05734253, 0.014564514, -0.0015296936, -0.04840088, -0.013381958, -0.0058517456, 0.03515625, -0.009017944, 0.0018997192, 0.005077362, -0.018188477, -0.042907715, 0.009544373, -0.02015686, 0...
http://arxiv.org/abs/2306.04618v2
Revisiting Out-of-distribution Robustness in NLP: Benchmark, Analysis, and LLMs Evaluations
Lifan Yuan, Yangyi Chen, Ganqu Cui, Hongcheng Gao, Fangyuan Zou, Xingyi Cheng, Heng Ji, Zhiyuan Liu, Maosong Sun
cs.CL, cs.CR, cs.LG
This paper reexamines the research on out-of-distribution (OOD) robustness in the field of NLP. We find that the distribution shift settings in previous studies commonly lack adequate challenges, hindering the accurate evaluation of OOD robustness. To address these issues, we propose a benchmark construction protocol t...
http://arxiv.org/pdf/2306.04618v2
2023-10-26T06:59:50
2023-06-07T17:47:03
{ "title": [ 0.0030765533, -0.0021781921, -0.007144928, -0.0390625, 0.018157959, -0.0063667297, -0.015182495, -0.027389526, 0.009376526, 0.00869751, -0.044677734, 0.032104492, -0.016555786, 0.002702713, -0.043945312, -0.0132751465, -0.0362854, 0....
http://arxiv.org/abs/2306.04610v1
The Two Word Test: A Semantic Benchmark for Large Language Models
Nicholas Riccardi, Rutvik H. Desai
cs.CL, cs.AI
Large Language Models (LLMs) have shown remarkable abilities recently, including passing advanced professional exams and demanding benchmark tests. This performance has led many to suggest that they are close to achieving humanlike or 'true' understanding of language, and even Artificial General Intelligence (AGI). Her...
http://arxiv.org/pdf/2306.04610v1
2023-06-07T17:22:03
2023-06-07T17:22:03
{ "title": [ -0.04324341, -0.024002075, -0.039001465, -0.066223145, 0.0043296814, 0.010017395, -0.06304932, 0.0038547516, 0.012550354, 0.028411865, -0.014053345, 0.010765076, 0.0013723373, -0.010734558, -0.050323486, 0.018814087, -0.01058197, 0.0...
http://arxiv.org/abs/2306.04597v1
Language Models Get a Gender Makeover: Mitigating Gender Bias with Few-Shot Data Interventions
Himanshu Thakur, Atishay Jain, Praneetha Vaddamanu, Paul Pu Liang, Louis-Philippe Morency
cs.CL, cs.LG
Societal biases present in pre-trained large language models are a critical issue as these models have been shown to propagate biases in countless downstream applications, rendering them unfair towards specific groups of people. Since large-scale retraining of these models from scratch is both time and compute-expensiv...
http://arxiv.org/pdf/2306.04597v1
2023-06-07T16:50:03
2023-06-07T16:50:03
{ "title": [ -0.03475952, -0.017593384, -0.05758667, -0.03753662, -0.016342163, -0.026748657, -0.044036865, 0.009017944, -0.009140015, 0.034362793, -0.01991272, -0.000078856945, 0.016693115, -0.011138916, -0.041381836, -0.008071899, -0.022094727, ...
http://arxiv.org/abs/2306.04573v1
Gender, names and other mysteries: Towards the ambiguous for gender-inclusive translation
Danielle Saunders, Katrina Olsen
cs.CL
The vast majority of work on gender in MT focuses on 'unambiguous' inputs, where gender markers in the source language are expected to be resolved in the output. Conversely, this paper explores the widespread case where the source sentence lacks explicit gender markers, but the target sentence contains them due to rich...
http://arxiv.org/pdf/2306.04573v1
2023-06-07T16:21:59
2023-06-07T16:21:59
{ "title": [ -0.08288574, -0.002981186, -0.040802002, -0.026382446, -0.01247406, 0.012008667, -0.026184082, 0.0051345825, -0.0016155243, 0.050628662, 0.00031208992, 0.0027160645, 0.04675293, -0.0146484375, -0.051483154, 0.01159668, -0.019546509, ...
http://arxiv.org/abs/2306.04563v1
ChatGPT is fun, but it is not funny! Humor is still challenging Large Language Models
Sophie Jentzsch, Kristian Kersting
cs.AI, cs.CL, cs.HC, cs.LG
Humor is a central aspect of human communication that has not been solved for artificial agents so far. Large language models (LLMs) are increasingly able to capture implicit and contextual information. Especially, OpenAI's ChatGPT recently gained immense public attention. The GPT3-based model almost seems to communica...
http://arxiv.org/pdf/2306.04563v1
2023-06-07T16:10:21
2023-06-07T16:10:21
{ "title": [ -0.026184082, -0.020187378, -0.0003077984, 0.016616821, -0.04663086, 0.020263672, -0.014335632, -0.01586914, 0.001376152, 0.033569336, -0.025314331, -0.008621216, 0.0074806213, -0.022232056, -0.03265381, -0.00003129244, -0.028503418, ...
http://arxiv.org/abs/2306.04551v2
Multi-Task Training with In-Domain Language Models for Diagnostic Reasoning
Brihat Sharma, Yanjun Gao, Timothy Miller, Matthew M. Churpek, Majid Afshar, Dmitriy Dligach
cs.CL, cs.LG
Generative artificial intelligence (AI) is a promising direction for augmenting clinical diagnostic decision support and reducing diagnostic errors, a leading contributor to medical errors. To further the development of clinical AI systems, the Diagnostic Reasoning Benchmark (DR.BENCH) was introduced as a comprehensive...
http://arxiv.org/pdf/2306.04551v2
2023-06-13T17:28:34
2023-06-07T15:55:34
{ "title": [ -0.009346008, -0.021606445, -0.051574707, -0.10064697, -0.010688782, 0.0029792786, -0.08093262, 0.015930176, -0.022018433, 0.015396118, 0.004524231, -0.004852295, -0.008331299, -0.0362854, 0.0038757324, -0.0051078796, -0.014480591, 0...
http://arxiv.org/abs/2306.04544v1
Contrastive Bootstrapping for Label Refinement
Shudi Hou, Yu Xia, Muhao Chen, Sujian Li
cs.CL, cs.AI
Traditional text classification typically categorizes texts into pre-defined coarse-grained classes, from which the produced models cannot handle the real-world scenario where finer categories emerge periodically for accurate services. In this work, we investigate the setting where fine-grained classification is done o...
http://arxiv.org/pdf/2306.04544v1
2023-06-07T15:49:04
2023-06-07T15:49:04
{ "title": [ -0.017929077, 0.008857727, -0.026977539, -0.06945801, -0.0026454926, 0.008079529, -0.058807373, -0.033966064, 0.022781372, 0.014434814, -0.022109985, 0.010246277, -0.008216858, -0.036987305, -0.03149414, 0.010894775, -0.016784668, 0....
http://arxiv.org/abs/2306.04539v1
Multimodal Learning Without Labeled Multimodal Data: Guarantees and Applications
Paul Pu Liang, Chun Kai Ling, Yun Cheng, Alex Obolenskiy, Yudong Liu, Rohan Pandey, Alex Wilf, Louis-Philippe Morency, Ruslan Salakhutdinov
cs.LG, cs.CL, cs.CV, cs.IT, math.IT, stat.ML
In many machine learning systems that jointly learn from multiple modalities, a core research question is to understand the nature of multimodal interactions: the emergence of new task-relevant information during learning from both modalities that was not present in either alone. We study this challenge of interaction ...
http://arxiv.org/pdf/2306.04539v1
2023-06-07T15:44:53
2023-06-07T15:44:53
{ "title": [ -0.024414062, 0.008323669, -0.06713867, -0.032440186, 0.016296387, -0.02355957, -0.020492554, -0.01940918, 0.008201599, 0.007785797, -0.026245117, 0.011306763, -0.033691406, -0.023864746, -0.019241333, -0.0039482117, -0.008728027, 0....
http://arxiv.org/abs/2306.04537v1
Long-form analogies generated by chatGPT lack human-like psycholinguistic properties
S. M. Seals, Valerie L. Shalin
cs.CL
Psycholinguistic analyses provide a means of evaluating large language model (LLM) output and making systematic comparisons to human-generated text. These methods can be used to characterize the psycholinguistic properties of LLM output and illustrate areas where LLMs fall short in comparison to human-generated text. I...
http://arxiv.org/pdf/2306.04537v1
2023-06-07T15:42:31
2023-06-07T15:42:31
{ "title": [ 0.00010675192, 0.0029067993, 0.020828247, 0.026153564, -0.05429077, -0.009483337, 0.012611389, -0.043273926, -0.01838684, 0.052947998, -0.014144897, 0.01033783, 0.016494751, -0.020492554, -0.0154953, 0.0040359497, -0.009735107, 0.058...
http://arxiv.org/abs/2306.04535v1
PromptAttack: Probing Dialogue State Trackers with Adversarial Prompts
Xiangjue Dong, Yun He, Ziwei Zhu, James Caverlee
cs.CL
A key component of modern conversational systems is the Dialogue State Tracker (or DST), which models a user's goals and needs. Toward building more robust and reliable DSTs, we introduce a prompt-based learning approach to automatically generate effective adversarial examples to probe DST models. Two key characteristi...
http://arxiv.org/pdf/2306.04535v1
2023-06-07T15:41:40
2023-06-07T15:41:40
{ "title": [ -0.0010309219, -0.017242432, -0.012557983, -0.054626465, -0.009666443, 0.0055618286, -0.025939941, -0.006565094, -0.026275635, -0.0009889603, 0.008598328, 0.0181427, -0.036071777, -0.041107178, -0.014823914, -0.016540527, -0.014862061, ...
http://arxiv.org/abs/2306.04530v1
Lenient Evaluation of Japanese Speech Recognition: Modeling Naturally Occurring Spelling Inconsistency
Shigeki Karita, Richard Sproat, Haruko Ishikawa
cs.CL
Word error rate (WER) and character error rate (CER) are standard metrics in Speech Recognition (ASR), but one problem has always been alternative spellings: If one's system transcribes adviser whereas the ground truth has advisor, this will count as an error even though the two spellings really represent the same word...
http://arxiv.org/pdf/2306.04530v1
2023-06-07T15:39:02
2023-06-07T15:39:02
{ "title": [ -0.033935547, 0.018966675, -0.0446167, -0.018173218, -0.003458023, -0.007865906, -0.036132812, -0.013755798, 0.0039787292, 0.031341553, -0.053771973, -0.016357422, 0.0066337585, -0.0075416565, -0.07458496, -0.035980225, -0.042510986, ...
http://arxiv.org/abs/2306.04528v4
PromptBench: Towards Evaluating the Robustness of Large Language Models on Adversarial Prompts
Kaijie Zhu, Jindong Wang, Jiaheng Zhou, Zichen Wang, Hao Chen, Yidong Wang, Linyi Yang, Wei Ye, Yue Zhang, Neil Zhenqiang Gong, Xing Xie
cs.CL, cs.CR, cs.LG
The increasing reliance on Large Language Models (LLMs) across academia and industry necessitates a comprehensive understanding of their robustness to prompts. In response to this vital need, we introduce PromptBench, a robustness benchmark designed to measure LLMs' resilience to adversarial prompts. This study uses a ...
http://arxiv.org/pdf/2306.04528v4
2023-10-18T14:16:01
2023-06-07T15:37:00
{ "title": [ -0.017181396, -0.0024280548, -0.031021118, -0.04888916, 0.009918213, 0.0030212402, -0.059143066, -0.011001587, -0.00806427, 0.020568848, -0.009780884, 0.01600647, -0.01020813, -0.014198303, -0.03768921, -0.0020961761, -0.0055122375, ...
http://arxiv.org/abs/2306.04523v1
Can current NLI systems handle German word order? Investigating language model performance on a new German challenge set of minimal pairs
Ines Reinig, Katja Markert
cs.CL
Compared to English, German word order is freer and therefore poses additional challenges for natural language inference (NLI). We create WOGLI (Word Order in German Language Inference), the first adversarial NLI dataset for German word order that has the following properties: (i) each premise has an entailed and a non...
http://arxiv.org/pdf/2306.04523v1
2023-06-07T15:33:07
2023-06-07T15:33:07
{ "title": [ -0.05596924, -0.017715454, -0.01928711, 0.009666443, -0.046966553, 0.008453369, -0.047576904, -0.01210022, -0.007293701, 0.050628662, -0.0027160645, -0.001250267, 0.053100586, -0.029464722, -0.035095215, 0.012191772, -0.024917603, 0....
http://arxiv.org/abs/2306.04508v1
Enhancing In-Context Learning with Answer Feedback for Multi-Span Question Answering
Zixian Huang, Jiaying Zhou, Gengyang Xiao, Gong Cheng
cs.CL, cs.AI
Whereas the recent emergence of large language models (LLMs) like ChatGPT has exhibited impressive general performance, it still has a large gap with fully-supervised models on specific tasks such as multi-span question answering. Previous researches found that in-context learning is an effective approach to exploiting...
http://arxiv.org/pdf/2306.04508v1
2023-06-07T15:20:24
2023-06-07T15:20:24
{ "title": [ -0.00013744831, 0.003080368, -0.032470703, -0.04925537, -0.00038981438, -0.014282227, -0.007762909, 0.005176544, 0.011894226, 0.026870728, -0.0446167, 0.017807007, -0.0013160706, 0.004940033, -0.04736328, -0.012023926, 0.0044898987, ...
http://arxiv.org/abs/2306.04504v3
Evaluation of ChatGPT on Biomedical Tasks: A Zero-Shot Comparison with Fine-Tuned Generative Transformers
Israt Jahan, Md Tahmid Rahman Laskar, Chun Peng, Jimmy Huang
cs.CL, cs.LG
ChatGPT is a large language model developed by OpenAI. Despite its impressive performance across various tasks, no prior work has investigated its capability in the biomedical domain yet. To this end, this paper aims to evaluate the performance of ChatGPT on various benchmark biomedical tasks, such as relation extracti...
http://arxiv.org/pdf/2306.04504v3
2023-08-24T13:39:17
2023-06-07T15:11:26
{ "title": [ 0.007068634, -0.01727295, 0.013069153, -0.018859863, -0.02494812, 0.011184692, -0.03111267, -0.024978638, 0.0015869141, 0.028656006, -0.0053367615, 0.025527954, -0.006652832, -0.035827637, -0.03567505, 0.015609741, -0.012435913, 0.02...
http://arxiv.org/abs/2306.04441v1
STEPS: A Benchmark for Order Reasoning in Sequential Tasks
Weizhi Wang, Hong Wang, Xifeng Yan
cs.CL
Various human activities can be abstracted into a sequence of actions in natural text, i.e. cooking, repairing, manufacturing, etc. Such action sequences heavily depend on the executing order, while disorder in action sequences leads to failure of further task execution by robots or AI agents. Therefore, to verify the ...
http://arxiv.org/pdf/2306.04441v1
2023-06-07T13:58:55
2023-06-07T13:58:55
{ "title": [ -0.019866943, -0.0040512085, -0.040039062, -0.09069824, 0.018463135, -0.00025963783, -0.04559326, -0.0035476685, 0.00554657, 0.02407837, 0.0053749084, 0.010391235, -0.0090408325, -0.033966064, -0.03488159, 0.011421204, 0.010604858, 0...
http://arxiv.org/abs/2306.04428v2
Zambezi Voice: A Multilingual Speech Corpus for Zambian Languages
Claytone Sikasote, Kalinda Siaminwe, Stanly Mwape, Bangiwe Zulu, Mofya Phiri, Martin Phiri, David Zulu, Mayumbo Nyirenda, Antonios Anastasopoulos
cs.CL, cs.SD, eess.AS
This work introduces Zambezi Voice, an open-source multilingual speech resource for Zambian languages. It contains two collections of datasets: unlabelled audio recordings of radio news and talk shows programs (160 hours) and labelled data (over 80 hours) consisting of read speech recorded from text sourced from public...
http://arxiv.org/pdf/2306.04428v2
2023-06-13T20:48:02
2023-06-07T13:36:37
{ "title": [ -0.04953003, -0.010734558, -0.054901123, -0.039855957, -0.014831543, 0.0011377335, -0.06793213, 0.0011577606, 0.023147583, 0.04446411, -0.0104599, -0.012718201, 0.049957275, -0.03753662, -0.034332275, 0.015472412, -0.014274597, 0.050...
http://arxiv.org/abs/2306.04424v1
Examining Bias in Opinion Summarisation Through the Perspective of Opinion Diversity
Nannan Huang, Lin Tian, Haytham Fayek, Xiuzhen Zhang
cs.CL
Opinion summarisation is a task that aims to condense the information presented in the source documents while retaining the core message and opinions. A summary that only represents the majority opinions will leave the minority opinions unrepresented in the summary. In this paper, we use the stance towards a certain ta...
http://arxiv.org/pdf/2306.04424v1
2023-06-07T13:31:02
2023-06-07T13:31:02
{ "title": [ 0.004043579, 0.008644104, -0.05480957, -0.053833008, 0.0129776, -0.029418945, -0.02267456, -0.017181396, -0.018005371, 0.04220581, -0.028503418, 0.0038909912, -0.014228821, -0.0033302307, -0.06994629, -0.015731812, -0.02330017, 0.062...
http://arxiv.org/abs/2306.04399v1
Transfer Learning of Transformer-based Speech Recognition Models from Czech to Slovak
Jan Lehečka, Josef V. Psutka, Josef Psutka
cs.CL
In this paper, we are comparing several methods of training the Slovak speech recognition models based on the Transformers architecture. Specifically, we are exploring the approach of transfer learning from the existing Czech pre-trained Wav2Vec 2.0 model into Slovak. We are demonstrating the benefits of the proposed a...
http://arxiv.org/pdf/2306.04399v1
2023-06-07T12:58:46
2023-06-07T12:58:46
{ "title": [ -0.014945984, 0.015403748, -0.046051025, -0.036621094, 0.011245728, 0.0010557175, -0.03262329, 0.0029563904, -0.041900635, 0.05529785, -0.039489746, 0.0064430237, 0.031921387, -0.01424408, -0.028915405, -0.010139465, -0.003320694, 0....
http://arxiv.org/abs/2306.04387v2
M$^3$IT: A Large-Scale Dataset towards Multi-Modal Multilingual Instruction Tuning
Lei Li, Yuwei Yin, Shicheng Li, Liang Chen, Peiyi Wang, Shuhuai Ren, Mukai Li, Yazheng Yang, Jingjing Xu, Xu Sun, Lingpeng Kong, Qi Liu
cs.CV, cs.CL
Instruction tuning has significantly advanced large language models (LLMs) such as ChatGPT, enabling them to align with human instructions across diverse tasks. However, progress in open vision-language models (VLMs) has been limited due to the scarcity of high-quality instruction datasets. To tackle this challenge and...
http://arxiv.org/pdf/2306.04387v2
2023-06-08T13:44:24
2023-06-07T12:35:37
{ "title": [ -0.05606079, -0.021911621, -0.02923584, -0.049743652, -0.022659302, 0.003358841, -0.08728027, 0.014518738, -0.00029206276, 0.045440674, -0.027923584, 0.020111084, 0.009170532, -0.011291504, -0.018661499, 0.0038719177, -0.006958008, 0...
http://arxiv.org/abs/2306.05524v1
Check Me If You Can: Detecting ChatGPT-Generated Academic Writing using CheckGPT
Zeyan Liu, Zijun Yao, Fengjun Li, Bo Luo
cs.CL, cs.CR, cs.LG
With ChatGPT under the spotlight, utilizing large language models (LLMs) for academic writing has drawn a significant amount of discussions and concerns in the community. While substantial research efforts have been stimulated for detecting LLM-Generated Content (LLM-content), most of the attempts are still in the earl...
http://arxiv.org/pdf/2306.05524v1
2023-06-07T12:33:24
2023-06-07T12:33:24
{ "title": [ 0.0036697388, 0.014656067, -0.011123657, 0.021194458, -0.01751709, 0.0025463104, -0.025115967, -0.058563232, -0.060516357, 0.03201294, -0.010093689, 0.026245117, -0.010696411, -0.021499634, -0.012359619, -0.030593872, -0.031951904, 0...
http://arxiv.org/abs/2306.07936v1
FOOCTTS: Generating Arabic Speech with Acoustic Environment for Football Commentator
Massa Baali, Ahmed Ali
eess.AS, cs.CL, cs.SD
This paper presents FOOCTTS, an automatic pipeline for a football commentator that generates speech with background crowd noise. The application gets the text from the user, applies text pre-processing such as vowelization, followed by the commentator's speech synthesizer. Our pipeline included Arabic automatic speech ...
http://arxiv.org/pdf/2306.07936v1
2023-06-07T12:33:02
2023-06-07T12:33:02
{ "title": [ -0.0059013367, -0.024093628, -0.013916016, -0.060791016, -0.019577026, 0.005214691, -0.04171753, 0.007194519, -0.017837524, 0.04788208, 0.0022735596, -0.008163452, 0.00491333, -0.035095215, -0.05105591, -0.010437012, -0.0017032623, 0...
http://arxiv.org/abs/2306.04384v1
Multilingual Clinical NER: Translation or Cross-lingual Transfer?
Xavier Fontaine, Félix Gaschi, Parisa Rastin, Yannick Toussaint
cs.CL, cs.AI, cs.LG
Natural language tasks like Named Entity Recognition (NER) in the clinical domain on non-English texts can be very time-consuming and expensive due to the lack of annotated data. Cross-lingual transfer (CLT) is a way to circumvent this issue thanks to the ability of multilingual large language models to be fine-tuned o...
http://arxiv.org/pdf/2306.04384v1
2023-06-07T12:31:07
2023-06-07T12:31:07
{ "title": [ -0.025283813, -0.022384644, -0.00015854836, -0.050231934, -0.022018433, 0.0262146, -0.053710938, -0.013137817, -0.03994751, 0.03378296, 0.030532837, -0.0068855286, 0.0099487305, -0.05987549, 0.0016908646, 0.019226074, -0.010719299, 0...
http://arxiv.org/abs/2306.04374v1
Label Aware Speech Representation Learning For Language Identification
Shikhar Vashishth, Shikhar Bharadwaj, Sriram Ganapathy, Ankur Bapna, Min Ma, Wei Han, Vera Axelrod, Partha Talukdar
cs.CL, cs.LG, cs.SD, eess.AS
Speech representation learning approaches for non-semantic tasks such as language recognition have either explored supervised embedding extraction methods using a classifier model or self-supervised representation learning approaches using raw data. In this paper, we propose a novel framework of combining self-supervis...
http://arxiv.org/pdf/2306.04374v1
2023-06-07T12:14:16
2023-06-07T12:14:16
{ "title": [ 0.0028266907, 0.021133423, -0.04660034, -0.05532837, 0.016403198, -0.013412476, -0.039245605, 0.0070724487, -0.037628174, 0.026275635, -0.0053863525, 0.0024738312, -0.0007004738, -0.041137695, -0.018035889, -0.010292053, -0.008308411, ...
http://arxiv.org/abs/2306.04368v1
Arabic Dysarthric Speech Recognition Using Adversarial and Signal-Based Augmentation
Massa Baali, Ibrahim Almakky, Shady Shehata, Fakhri Karray
cs.SD, cs.CL, eess.AS
Despite major advancements in Automatic Speech Recognition (ASR), the state-of-the-art ASR systems struggle to deal with impaired speech even with high-resource languages. In Arabic, this challenge gets amplified, with added complexities in collecting data from dysarthric speakers. In this paper, we aim to improve the ...
http://arxiv.org/pdf/2306.04368v1
2023-06-07T12:01:46
2023-06-07T12:01:46
{ "title": [ 0.0345459, 0.008926392, -0.020767212, -0.070617676, 0.012237549, 0.0029354095, -0.021499634, 0.0012559891, -0.035308838, 0.042236328, -0.0064964294, -0.0022964478, -0.028518677, -0.04019165, -0.04815674, -0.029052734, 0.005405426, 0....
http://arxiv.org/abs/2306.04362v1
Youku-mPLUG: A 10 Million Large-scale Chinese Video-Language Dataset for Pre-training and Benchmarks
Haiyang Xu, Qinghao Ye, Xuan Wu, Ming Yan, Yuan Miao, Jiabo Ye, Guohai Xu, Anwen Hu, Yaya Shi, Guangwei Xu, Chenliang Li, Qi Qian, Maofei Que, Ji Zhang, Xiao Zeng, Fei Huang
cs.CV, cs.CL
To promote the development of Vision-Language Pre-training (VLP) and multimodal Large Language Model (LLM) in the Chinese community, we firstly release the largest public Chinese high-quality video-language dataset named Youku-mPLUG, which is collected from Youku, a well-known Chinese video-sharing website, with strict...
http://arxiv.org/pdf/2306.04362v1
2023-06-07T11:52:36
2023-06-07T11:52:36
{ "title": [ -0.017303467, -0.0027561188, -0.046905518, -0.060302734, 0.011390686, -0.010917664, -0.03652954, -0.0013256073, -0.006866455, 0.039489746, -0.007396698, 0.006996155, 0.00055122375, -0.023635864, -0.038970947, -0.00077819824, -0.000699043...
http://arxiv.org/abs/2306.04357v3
ConTextual Masked Auto-Encoder for Retrieval-based Dialogue Systems
Zhenpeng Su, Xing Wu, Wei Zhou, Guangyuan Ma, Songlin Hu
cs.CL, cs.AI
Dialogue response selection aims to select an appropriate response from several candidates based on a given user and system utterance history. Recent studies have been improving the accuracy of dialogue response selection through post-training, mostly relying on naive masked language modeling methods. However, the rece...
http://arxiv.org/pdf/2306.04357v3
2023-06-14T04:18:17
2023-06-07T11:40:07
{ "title": [ -0.006526947, 0.012367249, -0.034942627, -0.04321289, 0.009002686, -0.0019550323, -0.044158936, -0.0069618225, 0.022079468, 0.05038452, -0.0060157776, -0.0041618347, 0.032806396, -0.025299072, -0.04034424, 0.020111084, -0.014083862, ...
http://arxiv.org/abs/2306.04349v2
GPT Self-Supervision for a Better Data Annotator
Xiaohuan Pei, Yanxi Li, Chang Xu
cs.CL, cs.DB
The task of annotating data into concise summaries poses a significant challenge across various domains, frequently requiring the allocation of significant time and specialized knowledge by human experts. Despite existing efforts to use large language models for annotation tasks, significant problems such as limited ap...
http://arxiv.org/pdf/2306.04349v2
2023-06-08T05:45:45
2023-06-07T11:33:14
{ "title": [ 0.027542114, 0.0146865845, -0.039123535, -0.07525635, -0.010498047, -0.026351929, -0.04559326, -0.012573242, -0.020721436, 0.026550293, -0.03086853, -0.0038223267, -0.014961243, -0.039123535, -0.0075416565, 0.0033226013, 0.011138916, ...
http://arxiv.org/abs/2306.04347v1
World Models for Math Story Problems
Andreas Opedal, Niklas Stoehr, Abulhair Saparov, Mrinmaya Sachan
cs.CL
Solving math story problems is a complex task for students and NLP models alike, requiring them to understand the world as described in the story and reason over it to compute an answer. Recent years have seen impressive performance on automatically solving these problems with large pre-trained language models and inno...
http://arxiv.org/pdf/2306.04347v1
2023-06-07T11:25:20
2023-06-07T11:25:20
{ "title": [ 0.0034542084, -0.018630981, 0.004310608, 0.014122009, -0.060668945, -0.014678955, -0.003206253, -0.0044784546, -0.054229736, 0.027282715, -0.026473999, -0.009284973, -0.0059051514, -0.03353882, 0.024307251, -0.032196045, 0.0075950623, ...
http://arxiv.org/abs/2306.04340v1
Co-evolving Graph Reasoning Network for Emotion-Cause Pair Extraction
Bowen Xing, Ivor W. Tsang
cs.CL, cs.AI
Emotion-Cause Pair Extraction (ECPE) aims to extract all emotion clauses and their corresponding cause clauses from a document. Existing approaches tackle this task through multi-task learning (MTL) framework in which the two subtasks provide indicative clues for ECPE. However, the previous MTL framework considers only...
http://arxiv.org/pdf/2306.04340v1
2023-06-07T11:11:12
2023-06-07T11:11:12
{ "title": [ -0.010772705, -0.012763977, -0.03451538, -0.04486084, -0.019165039, -0.013557434, -0.03643799, 0.0104599, 0.031829834, 0.0127334595, 0.023361206, -0.012161255, -0.005622864, -0.050231934, -0.030548096, 0.015701294, -0.001742363, 0.00...
http://arxiv.org/abs/2306.04337v1
A Study on the Reliability of Automatic Dysarthric Speech Assessments
Xavier F. Cadet, Ranya Aloufi, Sara Ahmadi-Abhari, Hamed Haddadi
cs.CL
Automating dysarthria assessments offers the opportunity to develop effective, low-cost tools that address the current limitations of manual and subjective assessments. Nonetheless, it is unclear whether current approaches rely on dysarthria-related speech patterns or external factors. We aim toward obtaining a clearer...
http://arxiv.org/pdf/2306.04337v1
2023-06-07T11:04:02
2023-06-07T11:04:02
{ "title": [ 0.016403198, 0.00819397, -0.026168823, -0.0602417, 0.0116119385, -0.001326561, -0.001493454, -0.005809784, -0.016479492, 0.044403076, -0.04019165, 0.0209198, -0.002029419, -0.015144348, -0.053131104, -0.016662598, -0.033294678, 0.055...
http://arxiv.org/abs/2306.04334v1
Echoes from Alexandria: A Large Resource for Multilingual Book Summarization
Alessandro Scirè, Simone Conia, Simone Ciciliano, Roberto Navigli
cs.CL
In recent years, research in text summarization has mainly focused on the news domain, where texts are typically short and have strong layout features. The task of full-book summarization presents additional challenges which are hard to tackle with current resources, due to their limited size and availability in Englis...
http://arxiv.org/pdf/2306.04334v1
2023-06-07T11:01:39
2023-06-07T11:01:39
{ "title": [ -0.01651001, -0.0061912537, -0.021270752, -0.0635376, -0.0039405823, 0.008224487, -0.046844482, -0.01838684, -0.003145218, 0.04598999, 0.006549835, 0.03439331, 0.02708435, -0.012954712, -0.020935059, 0.023529053, 0.0064315796, 0.0253...
http://arxiv.org/abs/2306.04328v1
IUTEAM1 at MEDIQA-Chat 2023: Is simple fine tuning effective for multilayer summarization of clinical conversations?
Dhananjay Srivastava
cs.CL
Clinical conversation summarization has become an important application of Natural language Processing. In this work, we intend to analyze summarization model ensembling approaches, that can be utilized to improve the overall accuracy of the generated medical report called chart note. The work starts with a single summ...
http://arxiv.org/pdf/2306.04328v1
2023-06-07T10:47:33
2023-06-07T10:47:33
{ "title": [ -0.0070877075, -0.017044067, 0.0044517517, -0.051879883, -0.029953003, 0.019485474, -0.059539795, 0.0010309219, -0.009979248, 0.05441284, 0.047729492, 0.0043640137, 0.028884888, -0.047912598, -0.014343262, 0.00491333, 0.005882263, 0....
http://arxiv.org/abs/2306.04314v1
Cross-Genre Argument Mining: Can Language Models Automatically Fill in Missing Discourse Markers?
Gil Rocha, Henrique Lopes Cardoso, Jonas Belouadi, Steffen Eger
cs.CL
Available corpora for Argument Mining differ along several axes, and one of the key differences is the presence (or absence) of discourse markers to signal argumentative content. Exploring effective ways to use discourse markers has received wide attention in various discourse parsing tasks, from which it is well-known...
http://arxiv.org/pdf/2306.04314v1
2023-06-07T10:19:50
2023-06-07T10:19:50
{ "title": [ -0.0021266937, -0.0032653809, -0.008049011, -0.05114746, -0.0065345764, 0.011550903, -0.0395813, -0.015182495, -0.034454346, 0.013183594, 0.014480591, 0.008255005, -0.019241333, -0.030563354, -0.028717041, -0.020599365, -0.031921387, ...
http://arxiv.org/abs/2306.04308v1
Personality testing of GPT-3: Limited temporal reliability, but highlighted social desirability of GPT-3's personality instruments results
Bojana Bodroza, Bojana M. Dinic, Ljubisa Bojic
cs.AI, cs.CL, cs.HC
To assess the potential applications and limitations of chatbot GPT-3 Davinci-003, this study explored the temporal reliability of personality questionnaires applied to the chatbot and its personality profile. Psychological questionnaires were administered to the chatbot on two separate occasions, followed by a compari...
http://arxiv.org/pdf/2306.04308v1
2023-06-07T10:14:17
2023-06-07T10:14:17
{ "title": [ 0.02041626, 0.034484863, -0.019805908, -0.0036830902, 0.012809753, -0.025283813, -0.010383606, -0.052337646, 0.030426025, 0.015289307, -0.05670166, -0.0018939972, -0.0024852753, -0.017929077, -0.015266418, -0.011138916, -0.013168335, ...
http://arxiv.org/abs/2306.04306v2
Allophant: Cross-lingual Phoneme Recognition with Articulatory Attributes
Kevin Glocker, Aaricia Herygers, Munir Georges
cs.CL, cs.SD, eess.AS, I.2.7
This paper proposes Allophant, a multilingual phoneme recognizer. It requires only a phoneme inventory for cross-lingual transfer to a target language, allowing for low-resource recognition. The architecture combines a compositional phone embedding approach with individually supervised phonetic attribute classifiers in...
http://arxiv.org/pdf/2306.04306v2
2023-08-16T17:44:59
2023-06-07T10:11:09
{ "title": [ -0.015777588, -0.010383606, -0.030380249, -0.055633545, -0.018295288, -0.0021514893, -0.055603027, 0.010856628, -0.011657715, 0.041229248, -0.01008606, 0.023010254, 0.009208679, -0.021392822, -0.017364502, 0.011146545, -0.0066070557, ...
http://arxiv.org/abs/2306.04293v1
Phrase Retrieval for Open-Domain Conversational Question Answering with Conversational Dependency Modeling via Contrastive Learning
Soyeong Jeong, Jinheon Baek, Sung Ju Hwang, Jong C. Park
cs.CL, cs.IR, cs.LG
Open-Domain Conversational Question Answering (ODConvQA) aims at answering questions through a multi-turn conversation based on a retriever-reader pipeline, which retrieves passages and then predicts answers with them. However, such a pipeline approach not only makes the reader vulnerable to the errors propagated from ...
http://arxiv.org/pdf/2306.04293v1
2023-06-07T09:46:38
2023-06-07T09:46:38
{ "title": [ 0.009529114, 0.00737381, -0.017211914, -0.07104492, 0.0023880005, 0.0036792755, -0.048828125, -0.005393982, 0.009635925, 0.036712646, -0.012527466, 0.014129639, 0.012588501, -0.008575439, -0.052886963, -0.000084638596, -0.008056641, ...
http://arxiv.org/abs/2306.04277v1
Analysis of the Fed's communication by using textual entailment model of Zero-Shot classification
Yasuhiro Nakayama, Tomochika Sawaki
cs.CL
In this study, we analyze documents published by central banks using text mining techniques and propose a method to evaluate the policy tone of central banks. Since the monetary policies of major central banks have a broad impact on financial market trends, the pricing of risky assets, and the real economy, market part...
http://arxiv.org/pdf/2306.04277v1
2023-06-07T09:23:26
2023-06-07T09:23:26
{ "title": [ 0.034484863, -0.0067329407, -0.04220581, -0.036834717, -0.0069999695, -0.036590576, 0.005836487, 0.026229858, -0.014305115, 0.04937744, 0.0017538071, 0.0035438538, 0.043395996, -0.034118652, -0.008895874, -0.009338379, -0.0103302, 0....
http://arxiv.org/abs/2306.04268v1
Multi-microphone Automatic Speech Segmentation in Meetings Based on Circular Harmonics Features
Théo Mariotte, Anthony Larcher, Silvio Montrésor, Jean-Hugh Thomas
cs.SD, cs.CL, eess.AS
Speaker diarization is the task of answering Who spoke and when? in an audio stream. Pipeline systems rely on speech segmentation to extract speakers' segments and achieve robust speaker diarization. This paper proposes a common framework to solve three segmentation tasks in the distant speech scenario: Voice Activity ...
http://arxiv.org/pdf/2306.04268v1
2023-06-07T09:09:00
2023-06-07T09:09:00
{ "title": [ -0.008354187, 0.017837524, -0.06512451, -0.07330322, 0.0049819946, 0.0028438568, -0.06732178, -0.0038166046, -0.0070152283, 0.061523438, 0.000012636185, 0.019546509, 0.015151978, -0.018432617, -0.055358887, 0.0018548965, -0.007461548, ...
http://arxiv.org/abs/2306.04233v1
Transfer Learning from Pre-trained Language Models Improves End-to-End Speech Summarization
Kohei Matsuura, Takanori Ashihara, Takafumi Moriya, Tomohiro Tanaka, Takatomo Kano, Atsunori Ogawa, Marc Delcroix
cs.CL, cs.SD, eess.AS
End-to-end speech summarization (E2E SSum) directly summarizes input speech into easy-to-read short sentences with a single model. This approach is promising because it, in contrast to the conventional cascade approach, can utilize full acoustical information and mitigate to the propagation of transcription errors. How...
http://arxiv.org/pdf/2306.04233v1
2023-06-07T08:23:58
2023-06-07T08:23:58
{ "title": [ 0.016921997, 0.0062942505, -0.03111267, -0.08178711, -0.0071792603, -0.010231018, -0.06628418, -0.00034070015, -0.026885986, 0.05038452, 0.019851685, 0.02418518, 0.00907135, -0.019699097, -0.024002075, 0.018463135, 0.0021972656, 0.03...
http://arxiv.org/abs/2306.04217v1
Effective Neural Topic Modeling with Embedding Clustering Regularization
Xiaobao Wu, Xinshuai Dong, Thong Nguyen, Anh Tuan Luu
cs.CL
Topic models have been prevalent for decades with various applications. However, existing topic models commonly suffer from the notorious topic collapsing: discovered topics semantically collapse towards each other, leading to highly repetitive topics, insufficient topic discovery, and damaged model interpretability. I...
http://arxiv.org/pdf/2306.04217v1
2023-06-07T07:45:38
2023-06-07T07:45:38
{ "title": [ -0.006843567, 0.016784668, -0.034973145, -0.030471802, 0.017562866, -0.009391785, -0.021713257, -0.03427124, 0.021636963, 0.017318726, 0.00983429, 0.021316528, -0.013999939, -0.032806396, -0.018493652, 0.027023315, -0.006965637, 0.02...
http://arxiv.org/abs/2306.04203v1
Leveraging Knowledge Graph Embeddings to Enhance Contextual Representations for Relation Extraction
Fréjus A. A. Laleye, Loïc Rakotoson, Sylvain Massip
cs.CL, cs.LG
Relation extraction task is a crucial and challenging aspect of Natural Language Processing. Several methods have surfaced as of late, exhibiting notable performance in addressing the task; however, most of these approaches rely on vast amounts of data from large-scale knowledge graphs or language models pretrained on ...
http://arxiv.org/pdf/2306.04203v1
2023-06-07T07:15:20
2023-06-07T07:15:20
{ "title": [ 0.0077285767, -0.0013942719, -0.060516357, -0.08868408, 0.014472961, -0.006752014, -0.04751587, 0.016036987, 0.028640747, 0.025131226, -0.027633667, 0.016693115, 0.0064048767, -0.0007534027, -0.04046631, 0.011276245, -0.010688782, 0....
http://arxiv.org/abs/2306.04190v1
An ASR-Based Tutor for Learning to Read: How to Optimize Feedback to First Graders
Yu Bai, Cristian Tejedor-Garcia, Ferdy Hubers, Catia Cucchiarini, Helmer Strik
cs.CL, cs.LG, cs.SD, eess.AS
The interest in employing automatic speech recognition (ASR) in applications for reading practice has been growing in recent years. In a previous study, we presented an ASR-based Dutch reading tutor application that was developed to provide instantaneous feedback to first-graders learning to read. We saw that ASR has p...
http://arxiv.org/pdf/2306.04190v1
2023-06-07T06:58:38
2023-06-07T06:58:38
{ "title": [ -0.04824829, -0.03060913, -0.012237549, -0.012023926, -0.06439209, 0.004840851, -0.05331421, 0.0031585693, -0.048309326, 0.057373047, -0.015792847, 0.0062561035, 0.023162842, -0.0066604614, -0.022003174, -0.026367188, 0.008094788, 0....
http://arxiv.org/abs/2306.04188v1
A New Dataset and Empirical Study for Sentence Simplification in Chinese
Shiping Yang, Renliang Sun, Xiaojun Wan
cs.CL
Sentence Simplification is a valuable technique that can benefit language learners and children a lot. However, current research focuses more on English sentence simplification. The development of Chinese sentence simplification is relatively slow due to the lack of data. To alleviate this limitation, this paper introd...
http://arxiv.org/pdf/2306.04188v1
2023-06-07T06:47:34
2023-06-07T06:47:34
{ "title": [ -0.02130127, -0.0065307617, -0.025421143, -0.0552063, -0.006095886, 0.00141716, -0.05114746, -0.0020484924, -0.0132751465, 0.046813965, -0.0053901672, 0.018371582, 0.017166138, -0.018508911, -0.043121338, 0.0064430237, -0.0073776245, ...
http://arxiv.org/abs/2306.04187v2
Knowing-how & Knowing-that: A New Task for Machine Comprehension of User Manuals
Hongru Liang, Jia Liu, Weihong Du, Dingnan Jin, Wenqiang Lei, Zujie Wen, Jiancheng Lv
cs.CL
The machine reading comprehension (MRC) of user manuals has huge potential in customer service. However, current methods have trouble answering complex questions. Therefore, we introduce the Knowing-how & Knowing-that task that requires the model to answer factoid-style, procedure-style, and inconsistent questions abou...
http://arxiv.org/pdf/2306.04187v2
2023-08-09T01:56:01
2023-06-07T06:46:56
{ "title": [ -0.04385376, -0.020812988, -0.026870728, -0.046295166, -0.0074539185, -0.0024929047, -0.048095703, -0.012107849, 0.006855011, 0.022369385, -0.026779175, 0.0057907104, 0.010971069, -0.012451172, -0.052581787, -0.006126404, -0.002521515, ...
http://arxiv.org/abs/2306.04181v2
Benchmarking Foundation Models with Language-Model-as-an-Examiner
Yushi Bai, Jiahao Ying, Yixin Cao, Xin Lv, Yuze He, Xiaozhi Wang, Jifan Yu, Kaisheng Zeng, Yijia Xiao, Haozhe Lyu, Jiayin Zhang, Juanzi Li, Lei Hou
cs.CL, cs.LG
Numerous benchmarks have been established to assess the performance of foundation models on open-ended question answering, which serves as a comprehensive test of a model's ability to understand and generate language in a manner similar to humans. Most of these works focus on proposing new datasets, however, we see two...
http://arxiv.org/pdf/2306.04181v2
2023-11-04T11:50:13
2023-06-07T06:29:58
{ "title": [ -0.0569458, -0.015335083, -0.04168701, -0.06341553, -0.024765015, 0.0181427, -0.078186035, -0.0037250519, -0.00075244904, 0.031097412, -0.0063438416, 0.01096344, -0.011428833, -0.022079468, -0.04525757, 0.0012674332, -0.026123047, 0....
http://arxiv.org/abs/2306.04176v1
When to Read Documents or QA History: On Unified and Selective Open-domain QA
Kyungjae Lee, Sang-eun Han, Seung-won Hwang, Moontae Lee
cs.CL, cs.AI, I.2.7
This paper studies the problem of open-domain question answering, with the aim of answering a diverse range of questions leveraging knowledge resources. Two types of sources, QA-pair and document corpora, have been actively leveraged with the following complementary strength. The former is highly precise when the parap...
http://arxiv.org/pdf/2306.04176v1
2023-06-07T06:03:39
2023-06-07T06:03:39
{ "title": [ -0.015167236, 0.01247406, -0.038482666, -0.050964355, -0.0048713684, -0.014907837, -0.04055786, 0.0038375854, -0.04336548, 0.027648926, -0.049621582, 0.025863647, -0.0031280518, -0.032592773, 0.002796173, -0.04626465, 0.003791809, 0....
http://arxiv.org/abs/2306.04170v1
From the One, Judge of the Whole: Typed Entailment Graph Construction with Predicate Generation
Zhibin Chen, Yansong Feng, Dongyan Zhao
cs.CL
Entailment Graphs (EGs) have been constructed based on extracted corpora as a strong and explainable form to indicate context-independent entailment relations in natural languages. However, EGs built by previous methods often suffer from the severe sparsity issues, due to limited corpora available and the long-tail phe...
http://arxiv.org/pdf/2306.04170v1
2023-06-07T05:46:19
2023-06-07T05:46:19
{ "title": [ -0.02017212, -0.029815674, -0.036590576, -0.121154785, -0.005126953, 0.031707764, -0.064331055, 0.043304443, -0.010002136, 0.01272583, 0.0048713684, 0.008049011, -0.0128479, -0.03564453, -0.0099105835, 0.007129669, -0.025009155, 0.03...
http://arxiv.org/abs/2306.04140v1
Increasing Diversity While Maintaining Accuracy: Text Data Generation with Large Language Models and Human Interventions
John Joon Young Chung, Ece Kamar, Saleema Amershi
cs.CL
Large language models (LLMs) can be used to generate text data for training and evaluating other models. However, creating high-quality datasets with LLMs can be challenging. In this work, we explore human-AI partnerships to facilitate high diversity and accuracy in LLM-based text data generation. We first examine two ...
http://arxiv.org/pdf/2306.04140v1
2023-06-07T04:27:09
2023-06-07T04:27:09
{ "title": [ -0.03375244, 0.0014257431, -0.052581787, -0.06085205, -0.00092220306, -0.002735138, -0.07598877, -0.009613037, -0.008911133, 0.031051636, -0.013343811, 0.021484375, -0.0031280518, -0.0039405823, -0.039215088, 0.010047913, -0.01864624, ...
http://arxiv.org/abs/2306.04136v1
Knowledge-Augmented Language Model Prompting for Zero-Shot Knowledge Graph Question Answering
Jinheon Baek, Alham Fikri Aji, Amir Saffari
cs.CL
Large Language Models (LLMs) are capable of performing zero-shot closed-book question answering tasks, based on their internal knowledge stored in parameters during pre-training. However, such internalized knowledge might be insufficient and incorrect, which could lead LLMs to generate factually wrong answers. Furtherm...
http://arxiv.org/pdf/2306.04136v1
2023-06-07T04:15:21
2023-06-07T04:15:21
{ "title": [ -0.032073975, -0.02607727, -0.045715332, -0.07220459, -0.04071045, 0.0019702911, -0.07110596, 0.039154053, 0.013870239, 0.031402588, -0.014060974, 0.0050582886, 0.008720398, -0.012779236, -0.04876709, 0.0132751465, 0.019805908, 0.000...
http://arxiv.org/abs/2306.04125v2
Multimodal Fusion Interactions: A Study of Human and Automatic Quantification
Paul Pu Liang, Yun Cheng, Ruslan Salakhutdinov, Louis-Philippe Morency
cs.LG, cs.CL, cs.HC
In order to perform multimodal fusion of heterogeneous signals, we need to understand their interactions: how each modality individually provides information useful for a task and how this information changes in the presence of other modalities. In this paper, we perform a comparative study of how humans annotate two c...
http://arxiv.org/pdf/2306.04125v2
2023-10-30T18:06:46
2023-06-07T03:44:50
{ "title": [ -0.010749817, 0.026916504, -0.04449463, -0.03555298, 0.01576233, -0.025772095, -0.022140503, -0.03302002, -0.0015859604, 0.0362854, -0.015182495, -0.0045661926, 0.011398315, -0.023040771, -0.037322998, -0.0076828003, -0.016021729, 0....
http://arxiv.org/abs/2306.04116v1
Unbalanced Optimal Transport for Unbalanced Word Alignment
Yuki Arase, Han Bao, Sho Yokoi
cs.CL, cs.LG
Monolingual word alignment is crucial to model semantic interactions between sentences. In particular, null alignment, a phenomenon in which words have no corresponding counterparts, is pervasive and critical in handling semantically divergent sentences. Identification of null alignment is useful on its own to reason a...
http://arxiv.org/pdf/2306.04116v1
2023-06-07T03:03:41
2023-06-07T03:03:41
{ "title": [ -0.034423828, 0.0044517517, -0.010627747, -0.006427765, -0.03012085, 0.0026626587, -0.037475586, -0.04034424, 0.046966553, 0.059417725, -0.011940002, 0.0004811287, 0.04321289, -0.014434814, -0.035125732, 0.028060913, -0.008338928, 0....
http://arxiv.org/abs/2306.04101v1
Gotta: Generative Few-shot Question Answering by Prompt-based Cloze Data Augmentation
Xiusi Chen, Yu Zhang, Jinliang Deng, Jyun-Yu Jiang, Wei Wang
cs.CL, cs.AI
Few-shot question answering (QA) aims at precisely discovering answers to a set of questions from context passages while only a few training samples are available. Although existing studies have made some progress and can usually achieve proper results, they suffer from understanding deep semantics for reasoning out th...
http://arxiv.org/pdf/2306.04101v1
2023-06-07T01:44:43
2023-06-07T01:44:43
{ "title": [ -0.005458832, -0.020706177, -0.045776367, -0.057159424, -0.013618469, 0.0010442734, -0.02848816, 0.029266357, 0.025344849, 0.019851685, -0.024597168, 0.009765625, -0.0206604, -0.004638672, -0.06549072, -0.0056152344, 0.0039367676, 0....
http://arxiv.org/abs/2306.04085v1
XSemPLR: Cross-Lingual Semantic Parsing in Multiple Natural Languages and Meaning Representations
Yusen Zhang, Jun Wang, Zhiguo Wang, Rui Zhang
cs.CL
Cross-Lingual Semantic Parsing (CLSP) aims to translate queries in multiple natural languages (NLs) into meaning representations (MRs) such as SQL, lambda calculus, and logic forms. However, existing CLSP models are separately proposed and evaluated on datasets of limited tasks and applications, impeding a comprehensiv...
http://arxiv.org/pdf/2306.04085v1
2023-06-07T01:09:37
2023-06-07T01:09:37
{ "title": [ -0.03012085, -0.0113220215, -0.030792236, -0.06695557, 0.007381439, 0.01574707, -0.07495117, -0.01096344, -0.0040397644, 0.029251099, -0.016845703, 0.02178955, 0.0007376671, -0.0079422, -0.031097412, 0.01751709, -0.021072388, 0.07110...
http://arxiv.org/abs/2306.04076v1
Text-only Domain Adaptation using Unified Speech-Text Representation in Transducer
Lu Huang, Boyu Li, Jun Zhang, Lu Lu, Zejun Ma
cs.CL, cs.SD, eess.AS
Domain adaptation using text-only corpus is challenging in end-to-end(E2E) speech recognition. Adaptation by synthesizing audio from text through TTS is resource-consuming. We present a method to learn Unified Speech-Text Representation in Conformer Transducer(USTR-CT) to enable fast domain adaptation using the text-on...
http://arxiv.org/pdf/2306.04076v1
2023-06-07T00:33:02
2023-06-07T00:33:02
{ "title": [ -0.012168884, -0.0033493042, -0.04159546, -0.050079346, 0.0010147095, -0.022323608, -0.03366089, 0.0051879883, 0.00038218498, 0.062561035, 0.0061187744, 0.0039863586, 0.03845215, -0.026885986, -0.030792236, 0.03237915, -0.016021729, ...
http://arxiv.org/abs/2306.04067v1
An Empirical Analysis of Parameter-Efficient Methods for Debiasing Pre-Trained Language Models
Zhongbin Xie, Thomas Lukasiewicz
cs.CL
The increasingly large size of modern pretrained language models not only makes them inherit more human-like biases from the training corpora, but also makes it computationally expensive to mitigate such biases. In this paper, we investigate recent parameter-efficient methods in combination with counterfactual data aug...
http://arxiv.org/pdf/2306.04067v1
2023-06-06T23:56:18
2023-06-06T23:56:18
{ "title": [ -0.011390686, 0.0072669983, -0.021194458, -0.07635498, 0.008460999, 0.0058021545, -0.07513428, -0.028274536, -0.010643005, 0.02003479, -0.005329132, 0.00096178055, 0.013946533, -0.027755737, -0.028320312, 0.0058403015, -0.012382507, ...
http://arxiv.org/abs/2306.04059v1
Augmenting Reddit Posts to Determine Wellness Dimensions impacting Mental Health
Chandreen Liyanage, Muskan Garg, Vijay Mago, Sunghwan Sohn
cs.CL, cs.CY
Amid ongoing health crisis, there is a growing necessity to discern possible signs of Wellness Dimensions (WD) manifested in self-narrated text. As the distribution of WD on social media data is intrinsically imbalanced, we experiment the generative NLP models for data augmentation to enable further improvement in the ...
http://arxiv.org/pdf/2306.04059v1
2023-06-06T23:15:59
2023-06-06T23:15:59
{ "title": [ 0.0418396, 0.019927979, -0.016815186, -0.027160645, 0.003643036, -0.029052734, -0.004283905, -0.031341553, -0.049102783, 0.033050537, -0.00844574, 0.014793396, -0.0071411133, -0.043640137, 0.00085926056, 0.0018644333, 0.038116455, 0....
http://arxiv.org/abs/2306.04050v2
LLMZip: Lossless Text Compression using Large Language Models
Chandra Shekhara Kaushik Valmeekam, Krishna Narayanan, Dileep Kalathil, Jean-Francois Chamberland, Srinivas Shakkottai
cs.IT, cs.CL, cs.LG, math.IT
We provide new estimates of an asymptotic upper bound on the entropy of English using the large language model LLaMA-7B as a predictor for the next token given a window of past tokens. This estimate is significantly smaller than currently available estimates in \cite{cover1978convergent}, \cite{lutati2023focus}. A natu...
http://arxiv.org/pdf/2306.04050v2
2023-06-26T18:03:12
2023-06-06T22:42:00
{ "title": [ -0.07598877, -0.022003174, -0.054870605, -0.04107666, -0.012283325, 0.014831543, -0.05255127, 0.020080566, -0.0039520264, 0.03643799, -0.0012950897, 0.0040283203, 0.019882202, -0.025390625, -0.031173706, 0.01576233, 0.0061836243, 0.0...
http://arxiv.org/abs/2306.04047v1
Active Sparse Conversations for Improved Audio-Visual Embodied Navigation
Xiulong Liu, Sudipta Paul, Moitreya Chatterjee, Anoop Cherian
cs.CV, cs.AI, cs.CL, cs.LG
Efficient navigation towards an audio-goal necessitates an embodied agent to not only possess the ability to use audio-visual cues effectively, but also be equipped to actively (but occasionally) seek human/oracle assistance without sacrificing autonomy, e.g., when it is uncertain of where to navigate towards locating ...
http://arxiv.org/pdf/2306.04047v1
2023-06-06T22:32:49
2023-06-06T22:32:49
{ "title": [ -0.025146484, -0.017364502, -0.007293701, -0.032836914, -0.018447876, 0.00051259995, -0.0045928955, -0.00982666, 0.024215698, 0.038482666, -0.013877869, -0.002708435, 0.006576538, -0.020767212, -0.038909912, 0.0047721863, 0.02029419, ...
http://arxiv.org/abs/2306.04043v1
An Analysis of Reader Engagement in Literary Fiction through Eye Tracking and Linguistic Features
Rose Neis, Karin de Langis, Zae Myung Kim, Dongyeop Kang
cs.CL
Capturing readers' engagement in fiction is a challenging but important aspect of narrative understanding. In this study, we collected 23 readers' reactions to 2 short stories through eye tracking, sentence-level annotations, and an overall engagement scale survey. We analyzed the significance of various qualities of t...
http://arxiv.org/pdf/2306.04043v1
2023-06-06T22:14:59
2023-06-06T22:14:59
{ "title": [ -0.0017700195, -0.0075645447, -0.028396606, 0.026565552, -0.011634827, -0.040130615, 0.028152466, -0.024917603, -0.00032305717, 0.047698975, 0.004886627, -0.006904602, 0.029220581, -0.024154663, -0.052886963, -0.002817154, -0.0110321045,...
http://arxiv.org/abs/2306.06083v1
Improving Fairness and Robustness in End-to-End Speech Recognition through unsupervised clustering
Irina-Elena Veliche, Pascale Fung
cs.SD, cs.CL, cs.LG, eess.AS
The challenge of fairness arises when Automatic Speech Recognition (ASR) systems do not perform equally well for all sub-groups of the population. In the past few years there have been many improvements in overall speech recognition quality, but without any particular focus on advancing Equality and Equity for all user...
http://arxiv.org/pdf/2306.06083v1
2023-06-06T21:13:08
2023-06-06T21:13:08
{ "title": [ 0.004184723, 0.020126343, -0.0625, -0.06756592, 0.04638672, -0.016555786, -0.021484375, 0.007297516, -0.022094727, 0.03515625, -0.02609253, 0.02381897, 0.0047950745, -0.012916565, -0.023376465, 0.005252838, -0.014709473, 0.05441284, ...
http://arxiv.org/abs/2306.04009v1
Triggering Multi-Hop Reasoning for Question Answering in Language Models using Soft Prompts and Random Walks
Kanishka Misra, Cicero Nogueira dos Santos, Siamak Shakeri
cs.CL, cs.AI
Despite readily memorizing world knowledge about entities, pre-trained language models (LMs) struggle to compose together two or more facts to perform multi-hop reasoning in question-answering tasks. In this work, we propose techniques that improve upon this limitation by relying on random walks over structured knowled...
http://arxiv.org/pdf/2306.04009v1
2023-06-06T20:45:18
2023-06-06T20:45:18
{ "title": [ -0.018554688, -0.045318604, -0.02279663, -0.059417725, -0.03555298, 0.011955261, -0.04699707, 0.05053711, 0.021392822, 0.005542755, -0.0385437, 0.010818481, 0.0041236877, -0.017181396, -0.02243042, -0.006504059, 0.011634827, 0.004695...
http://arxiv.org/abs/2306.03997v1
Sentiment Analysis in Finance: From Transformers Back to eXplainable Lexicons (XLex)
Maryan Rizinski, Hristijan Peshov, Kostadin Mishev, Milos Jovanovik, Dimitar Trajanov
cs.CL
Lexicon-based sentiment analysis (SA) in finance leverages specialized, manually annotated lexicons created by human experts to extract sentiment from financial texts. Although lexicon-based methods are simple to implement and fast to operate on textual data, they require considerable manual annotation efforts to creat...
http://arxiv.org/pdf/2306.03997v1
2023-06-06T20:19:33
2023-06-06T20:19:33
{ "title": [ -0.025283813, -0.006465912, -0.049682617, -0.02017212, 0.004211426, -0.021713257, -0.012817383, -0.0031757355, -0.014930725, 0.039367676, -0.064086914, 0.009178162, 0.023956299, -0.005622864, -0.03387451, -0.025527954, -0.011711121, ...
http://arxiv.org/abs/2306.03984v2
Toward More Accurate and Generalizable Evaluation Metrics for Task-Oriented Dialogs
Abishek Komma, Nagesh Panyam Chandrasekarasastry, Timothy Leffel, Anuj Goyal, Angeliki Metallinou, Spyros Matsoukas, Aram Galstyan
cs.CL, cs.LG
Measurement of interaction quality is a critical task for the improvement of spoken dialog systems. Existing approaches to dialog quality estimation either focus on evaluating the quality of individual turns, or collect dialog-level quality measurements from end users immediately following an interaction. In contrast t...
http://arxiv.org/pdf/2306.03984v2
2023-06-09T01:17:39
2023-06-06T19:43:29
{ "title": [ -0.00623703, -0.007537842, -0.036193848, -0.048339844, 0.010299683, -0.01878357, -0.02709961, -0.004650116, 0.006980896, 0.037322998, 0.013122559, -0.009567261, -0.011550903, -0.01802063, -0.066589355, -0.0020694733, -0.01259613, 0.0...
http://arxiv.org/abs/2306.03978v1
Büyük dil modellerinin Türkçe verisetleri ile eğitilmesi ve ince ayarlanması
A. Taha Arslan
cs.CL, I.2.7
Large language models have advanced enormously, gained vast attraction and are having a phase of intensed research. Some of the developed models and training datasets have been made open-accessible. Hence these may be further fine-tuned with some techniques to obtain specialized models for specific tasks. When it comes...
http://arxiv.org/pdf/2306.03978v1
2023-06-06T19:31:08
2023-06-06T19:31:08
{ "title": [ 0.033569336, 0.00010281801, -0.041503906, -0.046142578, 0.029708862, -0.03289795, -0.007736206, -0.008071899, -0.015670776, 0.035858154, 0.020767212, -0.016494751, -0.012367249, -0.059692383, -0.031433105, -0.007358551, 0.02583313, -...
http://arxiv.org/abs/2306.03975v2
Revisiting Conversation Discourse for Dialogue Disentanglement
Bobo Li, Hao Fei, Fei Li, Shengqiong Wu, Lizi Liao, Yinwei Wei, Tat-Seng Chua, Donghong Ji
cs.CL
Dialogue disentanglement aims to detach the chronologically ordered utterances into several independent sessions. Conversation utterances are essentially organized and described by the underlying discourse, and thus dialogue disentanglement requires the full understanding and harnessing of the intrinsic discourse attri...
http://arxiv.org/pdf/2306.03975v2
2023-06-10T06:42:03
2023-06-06T19:17:47
{ "title": [ -0.012863159, -0.00548172, -0.05505371, -0.06945801, 0.011642456, -0.010284424, -0.04159546, 0.013870239, 0.009864807, 0.025115967, 0.00027012825, 0.0005636215, 0.00077342987, -0.031280518, -0.038757324, 0.010238647, -0.019058228, 0....
http://arxiv.org/abs/2306.03974v2
TKDP: Threefold Knowledge-enriched Deep Prompt Tuning for Few-shot Named Entity Recognition
Jiang Liu, Hao Fei, Fei Li, Jingye Li, Bobo Li, Liang Zhao, Chong Teng, Donghong Ji
cs.CL
Few-shot named entity recognition (NER) exploits limited annotated instances to identify named mentions. Effectively transferring the internal or external resources thus becomes the key to few-shot NER. While the existing prompt tuning methods have shown remarkable few-shot performances, they still fail to make full us...
http://arxiv.org/pdf/2306.03974v2
2023-06-10T06:50:35
2023-06-06T19:11:59
{ "title": [ -0.01058197, -0.01436615, -0.045410156, -0.079589844, -0.0074691772, -0.0029582977, -0.061462402, 0.026290894, 0.021911621, 0.016281128, -0.048675537, 0.026687622, 0.012413025, 0.0029087067, -0.033325195, 0.006023407, 0.012283325, 0....
http://arxiv.org/abs/2306.03969v2
ECQED: Emotion-Cause Quadruple Extraction in Dialogs
Li Zheng, Donghong Ji, Fei Li, Hao Fei, Shengqiong Wu, Jingye Li, Bobo Li, Chong Teng
cs.CL
The existing emotion-cause pair extraction (ECPE) task, unfortunately, ignores extracting the emotion type and cause type, while these fine-grained meta-information can be practically useful in real-world applications, i.e., chat robots and empathic dialog generation. Also the current ECPE is limited to the scenario of...
http://arxiv.org/pdf/2306.03969v2
2023-06-10T06:52:49
2023-06-06T19:04:30
{ "title": [ 0.014259338, -0.02557373, -0.041809082, -0.07281494, -0.003419876, -0.028442383, -0.021469116, 0.00818634, 0.027893066, 0.030288696, 0.015052795, -0.0058174133, 0.009712219, -0.043182373, -0.045806885, 0.024642944, -0.01285553, 0.016...
http://arxiv.org/abs/2306.03959v1
Leveraging Explicit Procedural Instructions for Data-Efficient Action Prediction
Julia White, Arushi Raghuvanshi, Yada Pruksachatkun
cs.CL, cs.IR
Task-oriented dialogues often require agents to enact complex, multi-step procedures in order to meet user requests. While large language models have found success automating these dialogues in constrained environments, their widespread deployment is limited by the substantial quantities of task-specific data required ...
http://arxiv.org/pdf/2306.03959v1
2023-06-06T18:42:08
2023-06-06T18:42:08
{ "title": [ -0.013763428, -0.012207031, -0.037963867, -0.08648682, -0.0072288513, -0.0020446777, -0.0501709, 0.01625061, 0.0104599, 0.02848816, -0.019226074, 0.007068634, 0.005935669, -0.01121521, -0.033721924, 0.0053253174, 0.01777649, 0.031433...
http://arxiv.org/abs/2306.03954v1
Recognition of Handwritten Japanese Characters Using Ensemble of Convolutional Neural Networks
Angel I. Solis, Justin Zarkovacki, John Ly, Adham Atyabi
cs.CV, cs.AI, cs.CL, cs.LG
The Japanese writing system is complex, with three character types of Hiragana, Katakana, and Kanji. Kanji consists of thousands of unique characters, further adding to the complexity of character identification and literature understanding. Being able to translate handwritten Japanese characters into digital text is u...
http://arxiv.org/pdf/2306.03954v1
2023-06-06T18:30:51
2023-06-06T18:30:51
{ "title": [ -0.018310547, 0.011367798, -0.037139893, -0.034179688, 0.009292603, -0.009712219, -0.023147583, -0.020401001, 0.01713562, 0.03967285, -0.046142578, -0.00044965744, 0.014427185, 0.00006878376, -0.06933594, -0.012779236, 0.009773254, 0...
http://arxiv.org/abs/2306.03950v2
MISGENDERED: Limits of Large Language Models in Understanding Pronouns
Tamanna Hossain, Sunipa Dev, Sameer Singh
cs.CL
Content Warning: This paper contains examples of misgendering and erasure that could be offensive and potentially triggering. Gender bias in language technologies has been widely studied, but research has mostly been restricted to a binary paradigm of gender. It is essential also to consider non-binary gender identit...
http://arxiv.org/pdf/2306.03950v2
2023-07-07T05:18:34
2023-06-06T18:27:52
{ "title": [ -0.023773193, -0.011650085, -0.032226562, -0.05999756, -0.0042533875, 0.00046038628, -0.0284729, 0.00040245056, 0.017349243, 0.030197144, -0.01626587, -0.01524353, 0.03164673, -0.015220642, -0.045440674, 0.0059432983, -0.027557373, 0...
http://arxiv.org/abs/2306.05434v1
How Good is the Model in Model-in-the-loop Event Coreference Resolution Annotation?
Shafiuddin Rehan Ahmed, Abhijnan Nath, Michael Regan, Adam Pollins, Nikhil Krishnaswamy, James H. Martin
cs.CL
Annotating cross-document event coreference links is a time-consuming and cognitively demanding task that can compromise annotation quality and efficiency. To address this, we propose a model-in-the-loop annotation approach for event coreference resolution, where a machine learning model suggests likely corefering even...
http://arxiv.org/pdf/2306.05434v1
2023-06-06T18:06:24
2023-06-06T18:06:24
{ "title": [ 0.048980713, 0.00042629242, -0.006767273, -0.061706543, -0.00932312, -0.025390625, -0.024856567, -0.015594482, 0.01878357, 0.0524292, 0.013534546, -0.023468018, 0.062805176, -0.038635254, -0.012298584, 0.028579712, 0.00038671494, 0.0...
http://arxiv.org/abs/2306.03917v1
Turning large language models into cognitive models
Marcel Binz, Eric Schulz
cs.CL, cs.AI, cs.LG
Large language models are powerful systems that excel at many tasks, ranging from translation to mathematical reasoning. Yet, at the same time, these models often show unhuman-like characteristics. In the present paper, we address this gap and ask whether large language models can be turned into cognitive models. We fi...
http://arxiv.org/pdf/2306.03917v1
2023-06-06T18:00:01
2023-06-06T18:00:01
{ "title": [ -0.025360107, -0.0146484375, -0.02355957, -0.05718994, -0.0019302368, 0.006790161, -0.050598145, -0.010108948, 0.016860962, 0.021911621, -0.018951416, -0.02079773, -0.0009202957, -0.01852417, -0.049743652, 0.009765625, -0.00957489, 0...
http://arxiv.org/abs/2306.03907v1
CL-UZH at SemEval-2023 Task 10: Sexism Detection through Incremental Fine-Tuning and Multi-Task Learning with Label Descriptions
Janis Goldzycher
cs.CL, cs.CY
The widespread popularity of social media has led to an increase in hateful, abusive, and sexist language, motivating methods for the automatic detection of such phenomena. The goal of the SemEval shared task \textit{Towards Explainable Detection of Online Sexism} (EDOS 2023) is to detect sexism in English social media...
http://arxiv.org/pdf/2306.03907v1
2023-06-06T17:59:49
2023-06-06T17:59:49
{ "title": [ 0.017715454, -0.012840271, -0.045959473, -0.037994385, 0.0036392212, -0.043060303, -0.013893127, -0.0020275116, -0.017944336, 0.05126953, -0.010734558, 0.01928711, -0.00390625, -0.006626129, -0.037750244, -0.019058228, -0.024139404, ...
http://arxiv.org/abs/2306.03902v1
Utterance Classification with Logical Neural Network: Explainable AI for Mental Disorder Diagnosis
Yeldar Toleubay, Don Joven Agravante, Daiki Kimura, Baihan Lin, Djallel Bouneffouf, Michiaki Tatsubori
cs.CL, cs.AI, cs.LO, q-bio.NC
In response to the global challenge of mental health problems, we proposes a Logical Neural Network (LNN) based Neuro-Symbolic AI method for the diagnosis of mental disorders. Due to the lack of effective therapy coverage for mental disorders, there is a need for an AI solution that can assist therapists with the diagn...
http://arxiv.org/pdf/2306.03902v1
2023-06-06T17:58:44
2023-06-06T17:58:44
{ "title": [ -0.013343811, 0.009162903, -0.04284668, -0.05810547, 0.020904541, -0.017654419, -0.019851685, -0.0023555756, -0.00039219856, 0.02619934, -0.006652832, 0.0013608932, 0.0011978149, -0.02748108, -0.047302246, -0.0009098053, -0.026687622, ...
http://arxiv.org/abs/2306.03901v2
ChatDB: Augmenting LLMs with Databases as Their Symbolic Memory
Chenxu Hu, Jie Fu, Chenzhuang Du, Simian Luo, Junbo Zhao, Hang Zhao
cs.AI, cs.CL, cs.DB, cs.LG
Large language models (LLMs) with memory are computationally universal. However, mainstream LLMs are not taking full advantage of memory, and the designs are heavily influenced by biological brains. Due to their approximate nature and proneness to the accumulation of errors, conventional neural memory mechanisms cannot...
http://arxiv.org/pdf/2306.03901v2
2023-06-07T17:22:22
2023-06-06T17:58:24
{ "title": [ -0.045654297, -0.04724121, -0.05328369, -0.07305908, -0.0385437, 0.011138916, -0.07141113, 0.03213501, 0.0008883476, 0.042999268, -0.047332764, 0.011054993, 0.004940033, -0.013549805, -0.053466797, -0.016525269, -0.0041999817, 0.0026...
http://arxiv.org/abs/2306.03882v2
Causal interventions expose implicit situation models for commonsense language understanding
Takateru Yamakoshi, James L. McClelland, Adele E. Goldberg, Robert D. Hawkins
cs.CL
Accounts of human language processing have long appealed to implicit ``situation models'' that enrich comprehension with relevant but unstated world knowledge. Here, we apply causal intervention techniques to recent transformer models to analyze performance on the Winograd Schema Challenge (WSC), where a single context...
http://arxiv.org/pdf/2306.03882v2
2023-06-07T13:17:04
2023-06-06T17:36:43
{ "title": [ 0.02609253, -0.016479492, 0.00016832352, -0.074401855, -0.020523071, -0.0065612793, -0.045013428, -0.019760132, 0.012039185, -0.0061912537, -0.028625488, -0.0051956177, -0.012565613, -0.030593872, -0.018692017, -0.017044067, -0.025344849...
http://arxiv.org/abs/2306.03872v3
Deductive Verification of Chain-of-Thought Reasoning
Zhan Ling, Yunhao Fang, Xuanlin Li, Zhiao Huang, Mingu Lee, Roland Memisevic, Hao Su
cs.CL, cs.AI, cs.LG
Large Language Models (LLMs) significantly benefit from Chain-of-Thought (CoT) prompting in performing various reasoning tasks. While CoT allows models to produce more comprehensive reasoning processes, its emphasis on intermediate reasoning steps can inadvertently introduce hallucinations and accumulated errors, there...
http://arxiv.org/pdf/2306.03872v3
2023-10-03T19:48:22
2023-06-06T17:18:56
{ "title": [ 0.00843811, 0.005554199, -0.028121948, -0.08050537, 0.017974854, -0.0015907288, -0.040771484, -0.016815186, -0.0048942566, 0.011177063, -0.019424438, -0.018951416, -0.0057907104, -0.03488159, -0.022003174, -0.0079956055, -0.0050468445, ...
http://arxiv.org/abs/2306.03866v1
Correction of Errors in Preference Ratings from Automated Metrics for Text Generation
Jan Deriu, Pius von Däniken, Don Tuggener, Mark Cieliebak
cs.CL, cs.AI
A major challenge in the field of Text Generation is evaluation: Human evaluations are cost-intensive, and automated metrics often display considerable disagreement with human judgments. In this paper, we propose a statistical model of Text Generation evaluation that accounts for the error-proneness of automated metric...
http://arxiv.org/pdf/2306.03866v1
2023-06-06T17:09:29
2023-06-06T17:09:29
{ "title": [ -0.045135498, -0.0028190613, -0.02758789, 0.0033569336, -0.021636963, 0.001203537, -0.016235352, -0.032684326, 0.0033893585, 0.024734497, -0.0259552, -0.005153656, 0.008377075, -0.018936157, -0.048858643, -0.008056641, -0.04473877, 0...
http://arxiv.org/abs/2306.03856v1
Iterative Translation Refinement with Large Language Models
Pinzhen Chen, Zhicheng Guo, Barry Haddow, Kenneth Heafield
cs.CL, cs.AI
Large language models have shown surprising performances in understanding instructions and performing natural language tasks. In this paper, we propose iterative translation refinement to leverage the power of large language models for more natural translation and post-editing. We show that by simply involving a large ...
http://arxiv.org/pdf/2306.03856v1
2023-06-06T16:51:03
2023-06-06T16:51:03
{ "title": [ -0.05871582, -0.007167816, -0.02470398, -0.077697754, -0.010093689, 0.031555176, -0.1048584, -0.0022678375, 0.017807007, 0.04446411, -0.009674072, -0.0011825562, 0.025650024, -0.01360321, -0.050933838, 0.021316528, -0.00982666, 0.032...
http://arxiv.org/abs/2306.03853v1
From Key Points to Key Point Hierarchy: Structured and Expressive Opinion Summarization
Arie Cattan, Lilach Eden, Yoav Kantor, Roy Bar-Haim
cs.CL
Key Point Analysis (KPA) has been recently proposed for deriving fine-grained insights from collections of textual comments. KPA extracts the main points in the data as a list of concise sentences or phrases, termed key points, and quantifies their prevalence. While key points are more expressive than word clouds and k...
http://arxiv.org/pdf/2306.03853v1
2023-06-06T16:45:44
2023-06-06T16:45:44
{ "title": [ -0.0044822693, -0.004737854, -0.020568848, -0.06854248, -0.0010404587, 0.010017395, -0.03387451, -0.016281128, 0.029052734, 0.04046631, -0.026565552, 0.010772705, -0.010871887, 0.0076675415, -0.080322266, 0.007598877, -0.008369446, 0...
http://arxiv.org/abs/2306.03819v3
LEACE: Perfect linear concept erasure in closed form
Nora Belrose, David Schneider-Joseph, Shauli Ravfogel, Ryan Cotterell, Edward Raff, Stella Biderman
cs.LG, cs.CL, cs.CY
Concept erasure aims to remove specified features from a representation. It can improve fairness (e.g. preventing a classifier from using gender or race) and interpretability (e.g. removing a concept to observe changes in model behavior). We introduce LEAst-squares Concept Erasure (LEACE), a closed-form method which pr...
http://arxiv.org/pdf/2306.03819v3
2023-10-29T21:41:46
2023-06-06T16:07:24
{ "title": [ 0.018920898, -0.011909485, -0.01927185, -0.015808105, -0.049346924, -0.023803711, -0.010559082, -0.03463745, 0.0053253174, 0.04067993, -0.010635376, -0.0050735474, 0.034851074, -0.045806885, -0.018707275, 0.012451172, -0.013175964, 0...
http://arxiv.org/abs/2306.03799v1
Prompt Space Optimizing Few-shot Reasoning Success with Large Language Models
Fobo Shi, Peijun Qing, Dong Yang, Nan Wang, Youbo Lei, Haonan Lu, Xiaodong Lin
cs.CL
Prompt engineering is an essential technique for enhancing the abilities of large language models (LLMs) by providing explicit and specific instructions. It enables LLMs to excel in various tasks, such as arithmetic reasoning, question answering, summarization, relation extraction, machine translation, and sentiment an...
http://arxiv.org/pdf/2306.03799v1
2023-06-06T15:43:16
2023-06-06T15:43:16
{ "title": [ -0.020263672, -0.043762207, -0.041046143, -0.05126953, -0.03466797, -0.0004041195, -0.06329346, 0.018341064, 0.00687027, 0.015808105, -0.031585693, -0.00415802, 0.00024557114, -0.02470398, -0.025527954, -0.006832123, 0.0060195923, 0....
http://arxiv.org/abs/2306.03774v3
Exploring Hybrid Linguistic Features for Turkish Text Readability
Ahmet Yavuz Uluslu, Gerold Schneider
cs.CL
This paper presents the first comprehensive study on automatic readability assessment of Turkish texts. We combine state-of-the-art neural network models with linguistic features at lexical, morphosyntactic, syntactic and discourse levels to develop an advanced readability tool. We evaluate the effectiveness of traditi...
http://arxiv.org/pdf/2306.03774v3
2023-11-04T13:03:35
2023-06-06T15:32:22
{ "title": [ -0.023620605, 0.0096588135, -0.018249512, -0.010612488, 0.0028495789, 0.005924225, -0.0048675537, -0.01586914, 0.0024604797, 0.03375244, -0.04107666, -0.022033691, 0.037353516, -0.00166893, -0.05078125, -0.016952515, -0.021865845, 0....
http://arxiv.org/abs/2306.05432v1
Towards End-to-end Speech-to-text Summarization
Raul Monteiro, Diogo Pernes
cs.CL, cs.AI, cs.LG, eess.AS
Speech-to-text (S2T) summarization is a time-saving technique for filtering and keeping up with the broadcast news uploaded online on a daily basis. The rise of large language models from deep learning with impressive text generation capabilities has placed the research focus on summarization systems that produce parap...
http://arxiv.org/pdf/2306.05432v1
2023-06-06T15:22:16
2023-06-06T15:22:16
{ "title": [ 0.010093689, 0.0059890747, -0.016540527, -0.09613037, 0.014198303, 0.01436615, -0.05368042, 0.008766174, 0.0015649796, 0.06866455, 0.0015439987, 0.016036987, 0.017593384, 0.0013132095, -0.06451416, 0.027008057, -0.010192871, 0.034759...
http://arxiv.org/abs/2306.03736v1
FinRED: A Dataset for Relation Extraction in Financial Domain
Soumya Sharma, Tapas Nayak, Arusarka Bose, Ajay Kumar Meena, Koustuv Dasgupta, Niloy Ganguly, Pawan Goyal
cs.CL
Relation extraction models trained on a source domain cannot be applied on a different target domain due to the mismatch between relation sets. In the current literature, there is no extensive open-source relation extraction dataset specific to the finance domain. In this paper, we release FinRED, a relation extraction...
http://arxiv.org/pdf/2306.03736v1
2023-06-06T14:52:47
2023-06-06T14:52:47
{ "title": [ -0.0053253174, 0.004638672, -0.07775879, -0.068481445, 0.03640747, -0.01058197, -0.033569336, -0.0007033348, 0.023239136, 0.015449524, -0.030471802, -0.009010315, -0.0010147095, -0.0037822723, -0.061584473, -0.007282257, -0.022537231, ...
http://arxiv.org/abs/2306.03734v2
A Cross-Linguistic Pressure for Uniform Information Density in Word Order
Thomas Hikaru Clark, Clara Meister, Tiago Pimentel, Michael Hahn, Ryan Cotterell, Richard Futrell, Roger Levy
cs.CL
While natural languages differ widely in both canonical word order and word order flexibility, their word orders still follow shared cross-linguistic statistical patterns, often attributed to functional pressures. In the effort to identify these pressures, prior work has compared real and counterfactual word orders. Ye...
http://arxiv.org/pdf/2306.03734v2
2023-07-09T17:17:39
2023-06-06T14:52:15
{ "title": [ -0.02243042, -0.007003784, -0.027542114, -0.066345215, -0.0029315948, 0.006877899, -0.05355835, -0.0014314651, -0.013801575, 0.048950195, -0.0035057068, 0.0052757263, 0.03213501, -0.01789856, -0.03164673, 0.02180481, -0.00349617, 0.0...
http://arxiv.org/abs/2306.03733v1
A Novel Approach To User Agent String Parsing For Vulnerability Analysis Using Mutli-Headed Attention
Dhruv Nandakumar, Sathvik Murli, Ankur Khosla, Kevin Choi, Abdul Rahman, Drew Walsh, Scott Riede, Eric Dull, Edward Bowen
cs.CR, cs.CL
The increasing reliance on the internet has led to the proliferation of a diverse set of web-browsers and operating systems (OSs) capable of browsing the web. User agent strings (UASs) are a component of web browsing that are transmitted with every Hypertext Transfer Protocol (HTTP) request. They contain information ab...
http://arxiv.org/pdf/2306.03733v1
2023-06-06T14:49:25
2023-06-06T14:49:25
{ "title": [ 0.02027893, -0.00793457, -0.016479492, -0.060394287, -0.0007004738, -0.010734558, -0.024215698, 0.0066566467, -0.013572693, 0.045288086, -0.017471313, -0.00793457, 0.029037476, -0.016098022, -0.040130615, -0.01197052, -0.029174805, 0...
http://arxiv.org/abs/2306.03723v1
Financial Numeric Extreme Labelling: A Dataset and Benchmarking for XBRL Tagging
Soumya Sharma, Subhendu Khatuya, Manjunath Hegde, Afreen Shaikh. Koustuv Dasgupta, Pawan Goyal, Niloy Ganguly
cs.CL, cs.AI, cs.CE
The U.S. Securities and Exchange Commission (SEC) mandates all public companies to file periodic financial statements that should contain numerals annotated with a particular label from a taxonomy. In this paper, we formulate the task of automating the assignment of a label to a particular numeral span in a sentence fr...
http://arxiv.org/pdf/2306.03723v1
2023-06-06T14:41:30
2023-06-06T14:41:30
{ "title": [ 0.00016462803, 0.026687622, -0.056121826, -0.07220459, 0.027008057, -0.015594482, -0.038330078, -0.011405945, -0.0021762848, 0.04043579, -0.031311035, 0.014472961, 0.014633179, -0.027420044, -0.021652222, 0.00091457367, -0.016189575, ...
http://arxiv.org/abs/2306.03722v2
Evaluating the Effectiveness of Natural Language Inference for Hate Speech Detection in Languages with Limited Labeled Data
Janis Goldzycher, Moritz Preisig, Chantal Amrhein, Gerold Schneider
cs.CL, cs.CY
Most research on hate speech detection has focused on English where a sizeable amount of labeled training data is available. However, to expand hate speech detection into more languages, approaches that require minimal training data are needed. In this paper, we test whether natural language inference (NLI) models whic...
http://arxiv.org/pdf/2306.03722v2
2023-06-10T09:20:59
2023-06-06T14:40:41
{ "title": [ 0.0013380051, -0.014541626, -0.028930664, -0.045928955, -0.0069236755, -0.01852417, -0.022537231, -0.0058135986, -0.010887146, 0.0099487305, -0.032714844, 0.0070762634, 0.0009851456, -0.011947632, -0.03375244, -0.016204834, -0.022827148,...
http://arxiv.org/abs/2306.03678v1
On the Difference of BERT-style and CLIP-style Text Encoders
Zhihong Chen, Guiming Hardy Chen, Shizhe Diao, Xiang Wan, Benyou Wang
cs.CL
Masked language modeling (MLM) has been one of the most popular pretraining recipes in natural language processing, e.g., BERT, one of the representative models. Recently, contrastive language-image pretraining (CLIP) has also attracted attention, especially its vision models that achieve excellent performance on a bro...
http://arxiv.org/pdf/2306.03678v1
2023-06-06T13:41:09
2023-06-06T13:41:09
{ "title": [ -0.018981934, -0.015419006, -0.012924194, -0.011672974, -0.06088257, -0.0073623657, -0.041748047, -0.0029621124, -0.022949219, 0.060302734, -0.009429932, -0.012481689, 0.03866577, -0.026107788, -0.007736206, 0.019989014, -0.0026550293, ...