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/2306.02514v1
Jambu: A historical linguistic database for South Asian languages
Aryaman Arora, Adam Farris, Samopriya Basu, Suresh Kolichala
cs.CL
We introduce Jambu, a cognate database of South Asian languages which unifies dozens of previous sources in a structured and accessible format. The database includes 287k lemmata from 602 lects, grouped together in 23k sets of cognates. We outline the data wrangling necessary to compile the dataset and train neural mod...
http://arxiv.org/pdf/2306.02514v1
2023-06-05T00:32:57
2023-06-05T00:32:57
{ "title": [ -0.026123047, -0.002292633, -0.048034668, -0.07409668, 0.016860962, 0.00566864, -0.03933716, 0.02432251, 0.014022827, 0.021453857, -0.00844574, -0.0013914108, -0.00010204315, -0.029373169, -0.045043945, 0.008392334, -0.0023517609, 0....
http://arxiv.org/abs/2306.02492v1
RadLing: Towards Efficient Radiology Report Understanding
Rikhiya Ghosh, Sanjeev Kumar Karn, Manuela Daniela Danu, Larisa Micu, Ramya Vunikili, Oladimeji Farri
cs.CL
Most natural language tasks in the radiology domain use language models pre-trained on biomedical corpus. There are few pretrained language models trained specifically for radiology, and fewer still that have been trained in a low data setting and gone on to produce comparable results in fine-tuning tasks. We present R...
http://arxiv.org/pdf/2306.02492v1
2023-06-04T21:53:04
2023-06-04T21:53:04
{ "title": [ -0.016235352, -0.010032654, -0.026657104, -0.075805664, 0.008331299, 0.009567261, -0.03604126, 0.008758545, 0.00806427, 0.015686035, -0.014823914, -0.013473511, -0.020553589, -0.020523071, -0.04321289, 0.00036644936, 0.00026631355, 0...
http://arxiv.org/abs/2306.02475v1
Modeling Cross-Cultural Pragmatic Inference with Codenames Duet
Omar Shaikh, Caleb Ziems, William Held, Aryan J. Pariani, Fred Morstatter, Diyi Yang
cs.CL
Pragmatic reference enables efficient interpersonal communication. Prior work uses simple reference games to test models of pragmatic reasoning, often with unidentified speakers and listeners. In practice, however, speakers' sociocultural background shapes their pragmatic assumptions. For example, readers of this paper...
http://arxiv.org/pdf/2306.02475v1
2023-06-04T20:47:07
2023-06-04T20:47:07
{ "title": [ -0.04159546, -0.023406982, -0.017868042, -0.031021118, -0.028442383, -0.0019168854, -0.049926758, 0.009094238, -0.0024642944, 0.018936157, -0.033050537, 0.0039863586, -0.0116119385, -0.018249512, -0.027282715, -0.0050621033, -0.030654907...
http://arxiv.org/abs/2306.02457v1
Adaptive and Personalized Exercise Generation for Online Language Learning
Peng Cui, Mrinmaya Sachan
cs.CL, cs.AI
Adaptive learning aims to provide customized educational activities (e.g., exercises) to address individual learning needs. However, manual construction and delivery of such activities is a laborious process. Thus, in this paper, we study a novel task of adaptive and personalized exercise generation for online language...
http://arxiv.org/pdf/2306.02457v1
2023-06-04T20:18:40
2023-06-04T20:18:40
{ "title": [ -0.017791748, -0.008522034, -0.009864807, -0.026275635, -0.008384705, 0.0037441254, -0.019378662, -0.010147095, -0.038269043, 0.03717041, 0.024230957, 0.010185242, -0.016601562, -0.03137207, -0.01828003, -0.0043907166, -0.0025043488, ...
http://arxiv.org/abs/2306.02428v1
Taught by the Internet, Exploring Bias in OpenAIs GPT3
Ali Ayaz, Aditya Nawalgaria, Ruilian Yin
cs.CL, cs.AI
This research delves into the current literature on bias in Natural Language Processing Models and the techniques proposed to mitigate the problem of bias, including why it is important to tackle bias in the first place. Additionally, these techniques are further analysed in the light of newly developed models that tow...
http://arxiv.org/pdf/2306.02428v1
2023-06-04T18:21:44
2023-06-04T18:21:44
{ "title": [ 0.017501831, 0.012458801, -0.029388428, 0.019195557, -0.017654419, -0.037994385, 0.027664185, -0.0340271, -0.02116394, 0.0030612946, -0.058135986, 0.010871887, -0.004673004, -0.033813477, 0.028427124, -0.051696777, -0.01499176, 0.058...
http://arxiv.org/abs/2306.02408v1
Evaluating and Improving Tool-Augmented Computation-Intensive Math Reasoning
Beichen Zhang, Kun Zhou, Xilin Wei, Wayne Xin Zhao, Jing Sha, Shijin Wang, Ji-Rong Wen
cs.CL
Chain-of-thought prompting~(CoT) and tool augmentation have been validated in recent work as effective practices for improving large language models~(LLMs) to perform step-by-step reasoning on complex math-related tasks. However, most existing math reasoning datasets may be not able to fully evaluate and analyze the ab...
http://arxiv.org/pdf/2306.02408v1
2023-06-04T17:02:59
2023-06-04T17:02:59
{ "title": [ -0.016708374, -0.008201599, -0.024398804, -0.046020508, -0.00088739395, -0.0033912659, 0.0066184998, 0.015434265, 3.5762787e-7, 0.028411865, -0.014472961, -0.0053520203, 0.002714157, -0.012619019, -0.040649414, -0.0022144318, -0.0052948,...
http://arxiv.org/abs/2306.02405v1
An Information-Theoretic Analysis of Self-supervised Discrete Representations of Speech
Badr M. Abdullah, Mohammed Maqsood Shaik, Bernd Möbius, Dietrich Klakow
cs.CL
Self-supervised representation learning for speech often involves a quantization step that transforms the acoustic input into discrete units. However, it remains unclear how to characterize the relationship between these discrete units and abstract phonetic categories such as phonemes. In this paper, we develop an info...
http://arxiv.org/pdf/2306.02405v1
2023-06-04T16:52:11
2023-06-04T16:52:11
{ "title": [ -0.004611969, 0.013290405, -0.054473877, -0.06707764, 0.020233154, -0.012298584, -0.030975342, 0.022628784, -0.035949707, 0.054382324, -0.0035266876, 0.022735596, -0.018203735, -0.010253906, -0.047973633, -0.0007648468, 0.009483337, ...
http://arxiv.org/abs/2306.02388v1
Commonsense Knowledge Transfer for Pre-trained Language Models
Wangchunshu Zhou, Ronan Le Bras, Yejin Choi
cs.CL, cs.AI, cs.LG
Despite serving as the foundation models for a wide range of NLP benchmarks, pre-trained language models have shown limited capabilities of acquiring implicit commonsense knowledge from self-supervision alone, compared to learning linguistic and factual knowledge that appear more explicitly in the surface patterns in t...
http://arxiv.org/pdf/2306.02388v1
2023-06-04T15:44:51
2023-06-04T15:44:51
{ "title": [ -0.013420105, -0.034606934, -0.026168823, -0.066101074, -0.02130127, -0.004550934, -0.04699707, 0.009666443, 0.010520935, 0.027999878, -0.0069999695, 0.00023937225, -0.0009698868, -0.022994995, -0.036193848, 0.011184692, 0.0035533905, ...
http://arxiv.org/abs/2306.02383v2
Evolution of Efficient Symbolic Communication Codes
Anton Kolonin
cs.CL, cs.IT, cs.LG, math.IT
The paper explores how the human natural language structure can be seen as a product of evolution of inter-personal communication code, targeting maximisation of such culture-agnostic and cross-lingual metrics such as anti-entropy, compression factor and cross-split F1 score. The exploration is done as part of a larger...
http://arxiv.org/pdf/2306.02383v2
2023-06-11T06:38:03
2023-06-04T15:33:16
{ "title": [ -0.06640625, -0.0090789795, -0.03213501, -0.022354126, -0.00756073, 0.007091522, -0.0101623535, -0.0028095245, 0.008644104, 0.06304932, 0.018569946, 0.011520386, 0.018218994, -0.009338379, -0.05206299, 0.03756714, 0.009590149, 0.0041...
http://arxiv.org/abs/2306.02379v1
Modular Transformers: Compressing Transformers into Modularized Layers for Flexible Efficient Inference
Wangchunshu Zhou, Ronan Le Bras, Yejin Choi
cs.CL, cs.LG
Pre-trained Transformer models like T5 and BART have advanced the state of the art on a wide range of text generation tasks. Compressing these models into smaller ones has become critically important for practical use. Common neural network compression techniques such as knowledge distillation or quantization are limit...
http://arxiv.org/pdf/2306.02379v1
2023-06-04T15:26:28
2023-06-04T15:26:28
{ "title": [ -0.029281616, -0.005317688, -0.06286621, -0.090026855, 0.024291992, 0.0070495605, -0.05847168, 0.010406494, -0.0042915344, 0.02267456, -0.01625061, 0.016281128, -0.0027503967, -0.023117065, -0.034820557, 0.004776001, 0.013130188, 0.0...
http://arxiv.org/abs/2306.02377v2
"Are you telling me to put glasses on the dog?'' Content-Grounded Annotation of Instruction Clarification Requests in the CoDraw Dataset
Brielen Madureira, David Schlangen
cs.CL
Instruction Clarification Requests are a mechanism to solve communication problems, which is very functional in instruction-following interactions. Recent work has argued that the CoDraw dataset is a valuable source of naturally occurring iCRs. Beyond identifying when iCRs should be made, dialogue models should also be...
http://arxiv.org/pdf/2306.02377v2
2023-07-26T08:54:03
2023-06-04T15:23:16
{ "title": [ -0.011886597, -0.021469116, -0.024551392, 0.016036987, -0.037139893, -0.026535034, -0.01966858, -0.0021953583, 0.03451538, 0.024169922, -0.021133423, 0.018249512, 0.0035743713, -0.01625061, -0.02293396, -0.024520874, 0.011672974, 0.0...
http://arxiv.org/abs/2306.02349v2
bgGLUE: A Bulgarian General Language Understanding Evaluation Benchmark
Momchil Hardalov, Pepa Atanasova, Todor Mihaylov, Galia Angelova, Kiril Simov, Petya Osenova, Ves Stoyanov, Ivan Koychev, Preslav Nakov, Dragomir Radev
cs.CL, cs.IR, cs.LG, 68T50, F.2.2; I.2.7
We present bgGLUE(Bulgarian General Language Understanding Evaluation), a benchmark for evaluating language models on Natural Language Understanding (NLU) tasks in Bulgarian. Our benchmark includes NLU tasks targeting a variety of NLP problems (e.g., natural language inference, fact-checking, named entity recognition, ...
http://arxiv.org/pdf/2306.02349v2
2023-06-07T03:57:51
2023-06-04T12:54:00
{ "title": [ 0.0051727295, -0.013191223, -0.015174866, -0.064331055, 0.016418457, 0.008758545, -0.017166138, 0.0006914139, 0.00020885468, 0.022506714, -0.0050354004, -0.00415802, 0.0129776, -0.029769897, -0.03186035, 0.0032405853, -0.019805908, 0...
http://arxiv.org/abs/2306.02348v1
Leverage Points in Modality Shifts: Comparing Language-only and Multimodal Word Representations
Aleksey Tikhonov, Lisa Bylinina, Denis Paperno
cs.CL
Multimodal embeddings aim to enrich the semantic information in neural representations of language compared to text-only models. While different embeddings exhibit different applicability and performance on downstream tasks, little is known about the systematic representation differences attributed to the visual modali...
http://arxiv.org/pdf/2306.02348v1
2023-06-04T12:53:12
2023-06-04T12:53:12
{ "title": [ -0.01838684, -0.009147644, -0.036468506, -0.026885986, -0.027770996, -0.024627686, -0.033569336, -0.014884949, 0.010948181, 0.04864502, -0.014808655, 0.017288208, 0.012771606, -0.011619568, -0.032104492, 0.0178833, 0.0022640228, 0.01...
http://arxiv.org/abs/2306.02334v1
Long Text Generation Challenge
Nikolay Mikhaylovskiy
cs.CL, I.2.7
We propose a shared task of human-like long text generation, LTG Challenge, that asks models to output a consistent human-like long text (a Harry Potter generic audience fanfic in English), given a prompt of about 1000 tokens. We suggest a novel statistical metric of the text structuredness, GloVe Autocorrelations Powe...
http://arxiv.org/pdf/2306.02334v1
2023-06-04T11:52:36
2023-06-04T11:52:36
{ "title": [ 0.0025749207, -0.0036640167, -0.02104187, -0.03564453, -0.051879883, -0.019683838, -0.028961182, -0.0008993149, -0.029083252, 0.06561279, -0.020690918, -0.004486084, 0.026748657, -0.029037476, 0.0072517395, -0.014831543, 0.008140564, ...
http://arxiv.org/abs/2306.02320v1
Arbitrary Few Parameters are Good Enough for Adapting Large-scale Pre-trained Language Models
Yusheng Su, Chi-Min Chan, Jiali Cheng, Yujia Qin, Yankai Lin, Shengding Hu, Zonghan Yang, Ning Ding, Zhiyuan Liu, Maosong Sun
cs.CL, cs.AI
Parameter-efficient tuning (PET) methods can effectively drive extremely large pre-trained language models (PLMs) by only training minimal parameters. Different PET methods utilize different manually designed modules. In a small PLM, there are usually noticeable performance differences among PET methods. Nevertheless, ...
http://arxiv.org/pdf/2306.02320v1
2023-06-04T10:10:54
2023-06-04T10:10:54
{ "title": [ 0.017837524, -0.024383545, -0.01663208, -0.073791504, -0.045166016, -0.013046265, -0.0690918, 0.015525818, -0.023223877, 0.015548706, -0.005836487, -0.022033691, -0.0035095215, -0.050842285, 0.00089740753, -0.028442383, 0.0023174286, ...
http://arxiv.org/abs/2306.02317v1
SpellMapper: A non-autoregressive neural spellchecker for ASR customization with candidate retrieval based on n-gram mappings
Alexandra Antonova, Evelina Bakhturina, Boris Ginsburg
cs.CL, cs.AI, cs.SD, eess.AS
Contextual spelling correction models are an alternative to shallow fusion to improve automatic speech recognition (ASR) quality given user vocabulary. To deal with large user vocabularies, most of these models include candidate retrieval mechanisms, usually based on minimum edit distance between fragments of ASR hypot...
http://arxiv.org/pdf/2306.02317v1
2023-06-04T10:00:12
2023-06-04T10:00:12
{ "title": [ -0.010726929, 0.0050354004, 0.0036697388, -0.01423645, -0.026046753, 0.02104187, -0.034210205, -0.019866943, -0.022857666, 0.052581787, -0.011650085, 0.00032114983, 0.033935547, -0.023986816, -0.029846191, -0.012481689, -0.0060920715, ...
http://arxiv.org/abs/2306.02307v1
Finding the SWEET Spot: Analysis and Improvement of Adaptive Inference in Low Resource Settings
Daniel Rotem, Michael Hassid, Jonathan Mamou, Roy Schwartz
cs.CL, cs.AI, cs.LG
Adaptive inference is a simple method for reducing inference costs. The method works by maintaining multiple classifiers of different capacities, and allocating resources to each test instance according to its difficulty. In this work, we compare the two main approaches for adaptive inference, Early-Exit and Multi-Mode...
http://arxiv.org/pdf/2306.02307v1
2023-06-04T09:16:39
2023-06-04T09:16:39
{ "title": [ -0.046813965, -0.0063972473, -0.0340271, -0.03137207, 0.007827759, 0.0034389496, -0.040740967, -0.028244019, 0.008071899, 0.0044288635, -0.020141602, 0.021499634, -0.029251099, -0.019805908, -0.025680542, 0.0036258698, -0.003479004, ...
http://arxiv.org/abs/2306.02302v1
Does Character-level Information Always Improve DRS-based Semantic Parsing?
Tomoya Kurosawa, Hitomi Yanaka
cs.CL
Even in the era of massive language models, it has been suggested that character-level representations improve the performance of neural models. The state-of-the-art neural semantic parser for Discourse Representation Structures uses character-level representations, improving performance in the four languages (i.e., En...
http://arxiv.org/pdf/2306.02302v1
2023-06-04T08:54:32
2023-06-04T08:54:32
{ "title": [ -0.008666992, 0.009429932, -0.034820557, -0.0725708, -0.0231781, 0.002506256, -0.07159424, 0.00012457371, -0.01725769, 0.014457703, -0.030944824, 0.005455017, -0.0055007935, -0.019943237, -0.018371582, -0.017730713, 0.009552002, 0.02...
http://arxiv.org/abs/2306.02295v1
A Mathematical Abstraction for Balancing the Trade-off Between Creativity and Reality in Large Language Models
Ritwik Sinha, Zhao Song, Tianyi Zhou
cs.CL, cs.LG
Large Language Models have become popular for their remarkable capabilities in human-oriented tasks and traditional natural language processing tasks. Its efficient functioning is attributed to the attention mechanism in the Transformer architecture, enabling it to concentrate on particular aspects of the input. LLMs...
http://arxiv.org/pdf/2306.02295v1
2023-06-04T08:12:34
2023-06-04T08:12:34
{ "title": [ 0.0046577454, -0.0060424805, -0.04321289, -0.05038452, -0.00415802, -0.048614502, -0.034332275, 0.01663208, -0.0071487427, 0.053497314, -0.016921997, 0.00630188, 0.036071777, -0.011650085, -0.028625488, -0.00046133995, -0.014259338, ...
http://arxiv.org/abs/2306.02294v1
Exposing Bias in Online Communities through Large-Scale Language Models
Celine Wald, Lukas Pfahler
cs.CL, cs.CY, cs.LG
Progress in natural language generation research has been shaped by the ever-growing size of language models. While large language models pre-trained on web data can generate human-sounding text, they also reproduce social biases and contribute to the propagation of harmful stereotypes. This work utilises the flaw of b...
http://arxiv.org/pdf/2306.02294v1
2023-06-04T08:09:26
2023-06-04T08:09:26
{ "title": [ 0.007091522, 0.0036411285, -0.0385437, -0.04928589, 0.020507812, -0.017349243, -0.013198853, -0.008865356, -0.0090789795, 0.04208374, -0.016555786, 0.0021877289, 0.022659302, -0.0049591064, -0.037353516, 0.010528564, -0.021209717, 0....
http://arxiv.org/abs/2306.02282v1
Exploring and Verbalizing Academic Ideas by Concept Co-occurrence
Yi Xu, Shuqian Sheng, Bo Xue, Luoyi Fu, Xinbing Wang, Chenghu Zhou
cs.CL, cs.AI
Researchers usually come up with new ideas only after thoroughly comprehending vast quantities of literature. The difficulty of this procedure is exacerbated by the fact that the number of academic publications is growing exponentially. In this study, we devise a framework based on concept co-occurrence for academic id...
http://arxiv.org/pdf/2306.02282v1
2023-06-04T07:01:30
2023-06-04T07:01:30
{ "title": [ 0.021499634, 0.021072388, -0.0158844, -0.021530151, -0.00617218, -0.02178955, -0.0061836243, -0.05871582, -0.02116394, 0.04788208, -0.018539429, 0.01739502, 0.010124207, -0.017959595, -0.017578125, -0.011695862, -0.007949829, 0.02400...
http://arxiv.org/abs/2306.02273v1
End-to-End Joint Target and Non-Target Speakers ASR
Ryo Masumura, Naoki Makishima, Taiga Yamane, Yoshihiko Yamazaki, Saki Mizuno, Mana Ihori, Mihiro Uchida, Keita Suzuki, Hiroshi Sato, Tomohiro Tanaka, Akihiko Takashima, Satoshi Suzuki, Takafumi Moriya, Nobukatsu Hojo, Atsushi Ando
cs.CL, cs.SD, eess.AS
This paper proposes a novel automatic speech recognition (ASR) system that can transcribe individual speaker's speech while identifying whether they are target or non-target speakers from multi-talker overlapped speech. Target-speaker ASR systems are a promising way to only transcribe a target speaker's speech by enrol...
http://arxiv.org/pdf/2306.02273v1
2023-06-04T06:38:15
2023-06-04T06:38:15
{ "title": [ 0.01096344, 0.017196655, -0.061431885, -0.062164307, 0.0046463013, -0.023391724, -0.025390625, 0.030181885, -0.031829834, 0.040985107, 0.0021305084, -0.010391235, 0.04019165, -0.06402588, 0.025604248, 0.015457153, -0.001414299, 0.031...
http://arxiv.org/abs/2306.02272v2
OWQ: Lessons learned from activation outliers for weight quantization in large language models
Changhun Lee, Jungyu Jin, Taesu Kim, Hyungjun Kim, Eunhyeok Park
cs.CL
Large language models (LLMs) with hundreds of billions of parameters show impressive results across various language tasks using simple prompt tuning and few-shot examples, without the need for task-specific fine-tuning. However, their enormous size requires multiple server-grade GPUs even for inference, creating a sig...
http://arxiv.org/pdf/2306.02272v2
2023-06-13T07:31:38
2023-06-04T06:33:13
{ "title": [ -0.020889282, 0.009498596, -0.04916382, -0.02949524, 0.0053482056, -0.017547607, -0.033233643, -0.014213562, -0.007873535, 0.051574707, 0.0030384064, 0.012168884, 0.0068855286, -0.026275635, -0.047454834, 0.004306793, -0.020858765, 0...
http://arxiv.org/abs/2306.02258v1
Probing Physical Reasoning with Counter-Commonsense Context
Kazushi Kondo, Saku Sugawara, Akiko Aizawa
cs.CL
In this study, we create a CConS (Counter-commonsense Contextual Size comparison) dataset to investigate how physical commonsense affects the contextualized size comparison task; the proposed dataset consists of both contexts that fit physical commonsense and those that do not. This dataset tests the ability of languag...
http://arxiv.org/pdf/2306.02258v1
2023-06-04T04:24:43
2023-06-04T04:24:43
{ "title": [ 0.008544922, -0.011917114, -0.012825012, -0.030471802, -0.0069618225, -0.02633667, 0.0099105835, -0.03225708, -0.0005221367, 0.01940918, -0.016159058, 0.0035266876, 0.0027332306, -0.030456543, -0.007648468, -0.00034332275, 0.0051002502, ...
http://arxiv.org/abs/2306.02254v2
A Technical Report for Polyglot-Ko: Open-Source Large-Scale Korean Language Models
Hyunwoong Ko, Kichang Yang, Minho Ryu, Taekyoon Choi, Seungmu Yang, Jiwung Hyun, Sungho Park, Kyubyong Park
cs.CL
Polyglot is a pioneering project aimed at enhancing the non-English language performance of multilingual language models. Despite the availability of various multilingual models such as mBERT (Devlin et al., 2019), XGLM (Lin et al., 2022), and BLOOM (Scao et al., 2022), researchers and developers often resort to buildi...
http://arxiv.org/pdf/2306.02254v2
2023-06-06T03:27:33
2023-06-04T04:04:04
{ "title": [ -0.03253174, -0.0098724365, -0.007850647, -0.05041504, -0.031982422, 0.016433716, -0.081726074, -0.02418518, -0.03289795, 0.041625977, -0.011199951, 0.011657715, 0.04776001, -0.03540039, 0.0060195923, 0.0048179626, -0.008666992, 0.04...
http://arxiv.org/abs/2306.02250v2
Large Language Model Augmented Narrative Driven Recommendations
Sheshera Mysore, Andrew McCallum, Hamed Zamani
cs.IR, cs.CL
Narrative-driven recommendation (NDR) presents an information access problem where users solicit recommendations with verbose descriptions of their preferences and context, for example, travelers soliciting recommendations for points of interest while describing their likes/dislikes and travel circumstances. These requ...
http://arxiv.org/pdf/2306.02250v2
2023-07-21T07:46:03
2023-06-04T03:46:45
{ "title": [ -0.011268616, -0.013847351, -0.045043945, -0.050689697, -0.024429321, -0.008903503, -0.036987305, 0.014808655, -0.039154053, 0.046813965, -0.004737854, 0.00017726421, 0.011016846, -0.027770996, -0.0024757385, -0.0004916191, -0.012252808,...
http://arxiv.org/abs/2306.02247v1
Sen2Pro: A Probabilistic Perspective to Sentence Embedding from Pre-trained Language Model
Lingfeng Shen, Haiyun Jiang, Lemao Liu, Shuming Shi
cs.CL, cs.AI
Sentence embedding is one of the most fundamental tasks in Natural Language Processing and plays an important role in various tasks. The recent breakthrough in sentence embedding is achieved by pre-trained language models (PLMs). Despite its success, an embedded vector (Sen2Vec) representing a point estimate does not n...
http://arxiv.org/pdf/2306.02247v1
2023-06-04T03:26:43
2023-06-04T03:26:43
{ "title": [ -0.008377075, 0.0011434555, -0.027496338, -0.04977417, 0.004917145, -0.0031280518, -0.0440979, -0.013198853, 0.016357422, 0.013259888, -0.009963989, 0.010612488, 0.00086307526, -0.020401001, -0.031158447, 0.0071105957, -0.023147583, ...
http://arxiv.org/abs/2306.02242v1
Extract and Attend: Improving Entity Translation in Neural Machine Translation
Zixin Zeng, Rui Wang, Yichong Leng, Junliang Guo, Xu Tan, Tao Qin, Tie-yan Liu
cs.CL, cs.AI
While Neural Machine Translation(NMT) has achieved great progress in recent years, it still suffers from inaccurate translation of entities (e.g., person/organization name, location), due to the lack of entity training instances. When we humans encounter an unknown entity during translation, we usually first look up in...
http://arxiv.org/pdf/2306.02242v1
2023-06-04T03:05:25
2023-06-04T03:05:25
{ "title": [ -0.013877869, 0.0016298294, -0.017852783, -0.061065674, 0.013336182, 0.010475159, -0.037353516, -0.0023326874, 0.0015630722, 0.03933716, 0.0019464493, 0.005455017, 0.027420044, -0.012611389, -0.04046631, 0.03253174, -0.012580872, 0.0...
http://arxiv.org/abs/2306.02231v3
Fine-Tuning Language Models with Advantage-Induced Policy Alignment
Banghua Zhu, Hiteshi Sharma, Felipe Vieira Frujeri, Shi Dong, Chenguang Zhu, Michael I. Jordan, Jiantao Jiao
cs.CL, cs.AI, cs.LG, cs.SY, eess.SY
Reinforcement learning from human feedback (RLHF) has emerged as a reliable approach to aligning large language models (LLMs) to human preferences. Among the plethora of RLHF techniques, proximal policy optimization (PPO) is of the most widely used methods. Despite its popularity, however, PPO may suffer from mode coll...
http://arxiv.org/pdf/2306.02231v3
2023-11-02T22:47:14
2023-06-04T01:59:40
{ "title": [ -0.01171875, -0.014343262, -0.03503418, -0.07800293, -0.0053367615, -0.005176544, -0.045959473, 0.014778137, -0.041870117, 0.020812988, -0.007507324, 0.033935547, -0.025421143, -0.018341064, -0.01121521, -0.00041985512, -0.022537231, ...
http://arxiv.org/abs/2306.02213v3
Evaluating Emotion Arcs Across Languages: Bridging the Global Divide in Sentiment Analysis
Daniela Teodorescu, Saif M. Mohammad
cs.CL
Emotion arcs capture how an individual (or a population) feels over time. They are widely used in industry and research; however, there is little work on evaluating the automatically generated arcs. This is because of the difficulty of establishing the true (gold) emotion arc. Our work, for the first time, systematical...
http://arxiv.org/pdf/2306.02213v3
2023-11-04T18:00:32
2023-06-03T23:34:55
{ "title": [ -0.010025024, -0.016403198, -0.019729614, -0.03933716, 0.005973816, -0.00623703, -0.013298035, -0.006286621, 0.017211914, 0.031234741, -0.033996582, 0.009902954, 0.0076293945, -0.009674072, -0.060760498, 0.0036773682, -0.0022068024, ...
http://arxiv.org/abs/2306.02207v3
SpeechGen: Unlocking the Generative Power of Speech Language Models with Prompts
Haibin Wu, Kai-Wei Chang, Yuan-Kuei Wu, Hung-yi Lee
eess.AS, cs.AI, cs.CL, cs.LG
Large language models (LLMs) have gained considerable attention for Artificial Intelligence Generated Content (AIGC), particularly with the emergence of ChatGPT. However, the direct adaptation of continuous speech to LLMs that process discrete tokens remains an unsolved challenge, hindering the application of LLMs for ...
http://arxiv.org/pdf/2306.02207v3
2023-08-25T16:10:18
2023-06-03T22:35:27
{ "title": [ -0.01979065, -0.022827148, -0.04977417, -0.07720947, 0.010826111, 0.0068740845, -0.06732178, 0.014846802, 0.0002810955, 0.02293396, -0.018814087, 0.007789612, 0.024978638, -0.026306152, -0.027114868, 0.015617371, -0.007911682, 0.0358...
http://arxiv.org/abs/2306.05500v1
Word-Level Explanations for Analyzing Bias in Text-to-Image Models
Alexander Lin, Lucas Monteiro Paes, Sree Harsha Tanneru, Suraj Srinivas, Himabindu Lakkaraju
cs.CL, cs.AI, cs.CV, cs.LG
Text-to-image models take a sentence (i.e., prompt) and generate images associated with this input prompt. These models have created award wining-art, videos, and even synthetic datasets. However, text-to-image (T2I) models can generate images that underrepresent minorities based on race and sex. This paper investigate...
http://arxiv.org/pdf/2306.05500v1
2023-06-03T21:39:07
2023-06-03T21:39:07
{ "title": [ -0.0041236877, 0.005466461, -0.0033130646, -0.027328491, -0.01676941, -0.005207062, -0.026321411, -0.043762207, 0.014984131, 0.033569336, -0.03552246, 0.0061950684, 0.010795593, -0.017150879, -0.016647339, -0.00042891502, -0.0078125, ...
http://arxiv.org/abs/2306.02196v1
Question-Context Alignment and Answer-Context Dependencies for Effective Answer Sentence Selection
Minh Van Nguyen, Kishan KC, Toan Nguyen, Thien Huu Nguyen, Ankit Chadha, Thuy Vu
cs.CL
Answer sentence selection (AS2) in open-domain question answering finds answer for a question by ranking candidate sentences extracted from web documents. Recent work exploits answer context, i.e., sentences around a candidate, by incorporating them as additional input string to the Transformer models to improve the co...
http://arxiv.org/pdf/2306.02196v1
2023-06-03T20:59:19
2023-06-03T20:59:19
{ "title": [ -0.0069885254, 0.0073280334, -0.034851074, -0.044555664, 0.00075149536, -0.0035247803, -0.016525269, -0.014305115, 0.018447876, 0.031158447, -0.005050659, 0.029067993, -0.0050201416, -0.0075531006, -0.046417236, 0.019058228, -0.003545761...
http://arxiv.org/abs/2306.02193v1
LDEB -- Label Digitization with Emotion Binarization and Machine Learning for Emotion Recognition in Conversational Dialogues
Amitabha Dey, Shan Suthaharan
cs.CL, cs.LG
Emotion recognition in conversations (ERC) is vital to the advancements of conversational AI and its applications. Therefore, the development of an automated ERC model using the concepts of machine learning (ML) would be beneficial. However, the conversational dialogues present a unique problem where each dialogue depi...
http://arxiv.org/pdf/2306.02193v1
2023-06-03T20:37:46
2023-06-03T20:37:46
{ "title": [ -0.012817383, 0.0020961761, -0.036376953, -0.034606934, -0.022521973, -0.014060974, -0.036010742, -0.005344391, 0.006717682, 0.0446167, -0.020828247, -0.008377075, 0.04724121, -0.039031982, -0.030075073, -0.0035820007, -0.0026626587, ...
http://arxiv.org/abs/2306.02190v1
Stubborn Lexical Bias in Data and Models
Sofia Serrano, Jesse Dodge, Noah A. Smith
cs.CL
In NLP, recent work has seen increased focus on spurious correlations between various features and labels in training data, and how these influence model behavior. However, the presence and effect of such correlations are typically examined feature by feature. We investigate the cumulative impact on a model of many suc...
http://arxiv.org/pdf/2306.02190v1
2023-06-03T20:12:27
2023-06-03T20:12:27
{ "title": [ -0.003168106, 0.0059661865, -0.032226562, -0.009300232, -0.052581787, -0.024108887, -0.020828247, -0.04458618, 0.0069847107, 0.034179688, -0.07147217, -0.027999878, -0.0036125183, -0.034606934, -0.019760132, -0.020324707, -0.007713318, ...
http://arxiv.org/abs/2306.02182v1
FlairNLP at SemEval-2023 Task 6b: Extraction of Legal Named Entities from Legal Texts using Contextual String Embeddings
Vinay N Ramesh, Rohan Eswara
cs.CL
Indian court legal texts and processes are essential towards the integrity of the judicial system and towards maintaining the social and political order of the nation. Due to the increase in number of pending court cases, there is an urgent need to develop tools to automate many of the legal processes with the knowledg...
http://arxiv.org/pdf/2306.02182v1
2023-06-03T19:38:04
2023-06-03T19:38:04
{ "title": [ -0.008674622, -0.0016145706, -0.055511475, -0.032226562, 0.0063972473, -0.017868042, -0.025924683, -0.0064697266, 0.048461914, 0.030014038, -0.04144287, 0.039245605, 0.052734375, -0.0049324036, -0.00945282, 0.03729248, -0.039886475, ...
http://arxiv.org/abs/2306.02175v1
TART: Improved Few-shot Text Classification Using Task-Adaptive Reference Transformation
Shuo Lei, Xuchao Zhang, Jianfeng He, Fanglan Chen, Chang-Tien Lu
cs.CL
Meta-learning has emerged as a trending technique to tackle few-shot text classification and achieve state-of-the-art performance. However, the performance of existing approaches heavily depends on the inter-class variance of the support set. As a result, it can perform well on tasks when the semantics of sampled class...
http://arxiv.org/pdf/2306.02175v1
2023-06-03T18:38:02
2023-06-03T18:38:02
{ "title": [ -0.011199951, -0.007774353, -0.034576416, -0.0178833, 0.001461029, -0.014877319, -0.0020370483, -0.005317688, 0.019134521, 0.029968262, -0.031463623, 0.008659363, 0.008003235, -0.0018949509, -0.0501709, -0.016952515, -0.0038070679, 0...
http://arxiv.org/abs/2306.02153v1
Acoustic Word Embeddings for Untranscribed Target Languages with Continued Pretraining and Learned Pooling
Ramon Sanabria, Ondrej Klejch, Hao Tang, Sharon Goldwater
cs.CL, cs.LG, cs.SD, eess.AS
Acoustic word embeddings are typically created by training a pooling function using pairs of word-like units. For unsupervised systems, these are mined using k-nearest neighbor (KNN) search, which is slow. Recently, mean-pooled representations from a pre-trained self-supervised English model were suggested as a promisi...
http://arxiv.org/pdf/2306.02153v1
2023-06-03T16:44:21
2023-06-03T16:44:21
{ "title": [ -0.013771057, 0.0012102127, -0.024520874, -0.052368164, -0.012641907, 0.0015554428, -0.05960083, 0.0068740845, -0.010574341, 0.028198242, -0.020477295, 0.03555298, 0.0065078735, -0.01940918, -0.00920105, 0.009765625, -0.0030097961, 0...
http://arxiv.org/abs/2306.02140v1
Unsupervised Human Activity Recognition through Two-stage Prompting with ChatGPT
Qingxin Xia, Takuya Maekawa, Takahiro Hara
cs.HC, cs.CL
Wearable sensor devices, which offer the advantage of recording daily objects used by a person while performing an activity, enable the feasibility of unsupervised Human Activity Recognition (HAR). Unfortunately, previous unsupervised approaches using the usage sequence of objects usually require a proper description o...
http://arxiv.org/pdf/2306.02140v1
2023-06-03T15:41:59
2023-06-03T15:41:59
{ "title": [ -0.0048294067, 0.02279663, -0.034729004, -0.042114258, 0.019592285, -0.0032348633, -0.039855957, -0.025756836, -0.02444458, 0.020385742, -0.036193848, 0.035858154, -0.021743774, -0.008148193, -0.023788452, -0.005016327, -0.011314392, ...
http://arxiv.org/abs/2306.02130v2
Extending an Event-type Ontology: Adding Verbs and Classes Using Fine-tuned LLMs Suggestions
Jana Straková, Eva Fučíková, Jan Hajič, Zdeňka Urešová
cs.CL
In this project, we have investigated the use of advanced machine learning methods, specifically fine-tuned large language models, for pre-annotating data for a lexical extension task, namely adding descriptive words (verbs) to an existing (but incomplete, as of yet) ontology of event types. Several research questions ...
http://arxiv.org/pdf/2306.02130v2
2023-08-10T11:13:02
2023-06-03T14:57:47
{ "title": [ 0.012413025, -0.013633728, -0.043640137, -0.09112549, 0.01826477, -0.018417358, -0.03955078, 0.0031471252, -0.031799316, 0.017974854, -0.008583069, 0.031829834, 0.0033817291, -0.02796936, 0.011413574, -0.0031547546, -0.004550934, 0.0...
http://arxiv.org/abs/2306.02115v2
Table and Image Generation for Investigating Knowledge of Entities in Pre-trained Vision and Language Models
Hidetaka Kamigaito, Katsuhiko Hayashi, Taro Watanabe
cs.CL, cs.CV, cs.LG
In this paper, we propose a table and image generation task to verify how the knowledge about entities acquired from natural language is retained in Vision & Language (V&L) models. This task consists of two parts: the first is to generate a table containing knowledge about an entity and its related image, and the secon...
http://arxiv.org/pdf/2306.02115v2
2023-07-26T02:20:30
2023-06-03T14:01:54
{ "title": [ 0.008903503, -0.02243042, -0.045288086, -0.027999878, -0.022491455, -0.04598999, -0.023849487, -0.007911682, 0.002035141, 0.03567505, -0.039367676, -0.0043182373, 0.032073975, -0.023071289, -0.008323669, -0.011756897, 0.030456543, 0....
http://arxiv.org/abs/2306.02105v2
Adapting Pretrained ASR Models to Low-resource Clinical Speech using Epistemic Uncertainty-based Data Selection
Bonaventure F. P. Dossou, Atnafu Lambebo Tonja, Chris Chinenye Emezue, Tobi Olatunji, Naome A Etori, Salomey Osei, Tosin Adewumi, Sahib Singh
cs.CL, cs.SD, eess.AS
While there has been significant progress in ASR, African-accented clinical ASR has been understudied due to a lack of training datasets. Building robust ASR systems in this domain requires large amounts of annotated or labeled data, for a wide variety of linguistically and morphologically rich accents, which are expen...
http://arxiv.org/pdf/2306.02105v2
2023-10-08T18:07:56
2023-06-03T13:11:37
{ "title": [ -0.009414673, -0.0054473877, -0.046691895, -0.08862305, -0.007144928, -0.011947632, -0.07556152, 0.022003174, -0.02078247, 0.031341553, -0.0013942719, -0.005405426, -0.003736496, -0.045684814, -0.031311035, -0.015838623, 0.005783081, ...
http://arxiv.org/abs/2306.02080v2
Benchmarking Robustness of Adaptation Methods on Pre-trained Vision-Language Models
Shuo Chen, Jindong Gu, Zhen Han, Yunpu Ma, Philip Torr, Volker Tresp
cs.CV, cs.CL, cs.LG
Various adaptation methods, such as LoRA, prompts, and adapters, have been proposed to enhance the performance of pre-trained vision-language models in specific domains. The robustness of these adaptation methods against distribution shifts have not been studied. In this study, we assess the robustness of 11 widely-use...
http://arxiv.org/pdf/2306.02080v2
2023-06-11T13:00:15
2023-06-03T11:05:04
{ "title": [ -0.023468018, 0.00982666, -0.023712158, -0.036254883, -0.0014820099, -0.012039185, -0.049316406, -0.030136108, 0.013954163, 0.04309082, -0.012123108, 0.0060920715, 0.00466156, -0.013702393, -0.059051514, 0.0011796951, 0.026824951, 0....
http://arxiv.org/abs/2306.02078v1
Incorporating Deep Syntactic and Semantic Knowledge for Chinese Sequence Labeling with GCN
Xuemei Tang, Jun Wang, Qi Su
cs.CL, cs.AI
Recently, it is quite common to integrate Chinese sequence labeling results to enhance syntactic and semantic parsing. However, little attention has been paid to the utility of hierarchy and structure information encoded in syntactic and semantic features for Chinese sequence labeling tasks. In this paper, we propose a...
http://arxiv.org/pdf/2306.02078v1
2023-06-03T10:56:44
2023-06-03T10:56:44
{ "title": [ -0.035858154, -0.0055656433, -0.042114258, -0.07421875, -0.014144897, 0.002960205, -0.06738281, 0.022460938, 0.0024852753, 0.03262329, -0.031402588, 0.00026750565, 0.00712204, -0.0209198, -0.04019165, -0.0028915405, 0.0034770966, 0.0...
http://arxiv.org/abs/2306.02077v1
Utilizing ChatGPT to Enhance Clinical Trial Enrollment
Georgios Peikos, Symeon Symeonidis, Pranav Kasela, Gabriella Pasi
cs.IR, cs.CL
Clinical trials are a critical component of evaluating the effectiveness of new medical interventions and driving advancements in medical research. Therefore, timely enrollment of patients is crucial to prevent delays or premature termination of trials. In this context, Electronic Health Records (EHRs) have emerged as ...
http://arxiv.org/pdf/2306.02077v1
2023-06-03T10:54:23
2023-06-03T10:54:23
{ "title": [ 0.006752014, -0.004142761, -0.025634766, -0.028793335, -0.0026283264, 0.0026245117, -0.02470398, -0.022827148, -0.02355957, 0.030853271, -0.022750854, 0.0184021, -0.028030396, -0.019897461, -0.027633667, 0.0011873245, -0.025161743, 0...
http://arxiv.org/abs/2306.02074v2
A Conditional Generative Chatbot using Transformer Model
Nura Esfandiari, Kourosh Kiani, Razieh Rastgoo
cs.CL, cs.CV
A Chatbot serves as a communication tool between a human user and a machine to achieve an appropriate answer based on the human input. In more recent approaches, a combination of Natural Language Processing and sequential models are used to build a generative Chatbot. The main challenge of these models is their sequent...
http://arxiv.org/pdf/2306.02074v2
2023-09-08T17:15:37
2023-06-03T10:35:04
{ "title": [ 0.00015223026, 0.0014305115, -0.027191162, -0.05819702, 0.0041885376, 0.01473999, -0.056365967, -0.0071792603, 0.00762558, 0.037200928, -0.040527344, -0.0037460327, 0.020584106, -0.010643005, -0.053955078, -0.012565613, -0.013824463, ...
http://arxiv.org/abs/2306.02069v2
MultiLegalPile: A 689GB Multilingual Legal Corpus
Joel Niklaus, Veton Matoshi, Matthias Stürmer, Ilias Chalkidis, Daniel E. Ho
cs.CL, cs.AI, cs.LG, 68T50, I.2
Large, high-quality datasets are crucial for training Large Language Models (LLMs). However, so far, there are few datasets available for specialized critical domains such as law and the available ones are often only for the English language. We curate and release MultiLegalPile, a 689GB corpus in 24 languages from 17 ...
http://arxiv.org/pdf/2306.02069v2
2023-06-06T06:43:07
2023-06-03T10:10:38
{ "title": [ -0.034606934, -0.024734497, -0.07318115, -0.064575195, -0.03112793, 0.001745224, -0.091552734, 0.025970459, 0.014137268, 0.055847168, -0.043945312, 0.0045814514, 0.05328369, -0.025772095, -0.0064201355, 0.0032958984, -0.004486084, 0....
http://arxiv.org/abs/2306.02052v1
Conflicts, Villains, Resolutions: Towards models of Narrative Media Framing
Lea Frermann, Jiatong Li, Shima Khanehzar, Gosia Mikolajczak
cs.CL
Despite increasing interest in the automatic detection of media frames in NLP, the problem is typically simplified as single-label classification and adopts a topic-like view on frames, evading modelling the broader document-level narrative. In this work, we revisit a widely used conceptualization of framing from the c...
http://arxiv.org/pdf/2306.02052v1
2023-06-03T08:50:13
2023-06-03T08:50:13
{ "title": [ -0.004753113, -0.003540039, -0.032104492, -0.03866577, 0.015464783, -0.015411377, 0.0030117035, -0.015304565, -0.047912598, 0.027023315, 0.000749588, 0.018066406, -0.01965332, -0.025360107, -0.027404785, -0.009857178, -0.018341064, 0...
http://arxiv.org/abs/2306.02051v2
A Comprehensive Survey on Deep Learning for Relation Extraction: Recent Advances and New Frontiers
Xiaoyan Zhao, Yang Deng, Min Yang, Lingzhi Wang, Rui Zhang, Hong Cheng, Wai Lam, Ying Shen, Ruifeng Xu
cs.CL, cs.AI
Relation extraction (RE) involves identifying the relations between entities from unstructured texts. RE serves as the foundation for many natural language processing (NLP) applications, such as knowledge graph completion, question answering, and information retrieval. In recent years, deep neural networks have dominat...
http://arxiv.org/pdf/2306.02051v2
2023-06-06T12:39:21
2023-06-03T08:39:25
{ "title": [ -0.0055160522, -0.00034546852, -0.057159424, -0.08312988, 0.028747559, -0.0014591217, -0.037597656, 0.003124237, 0.030822754, 0.013900757, -0.0134887695, 0.009315491, -0.0004286766, -0.014427185, -0.042541504, 0.0078125, -0.015594482, ...
http://arxiv.org/abs/2306.02042v1
Impact of translation on biomedical information extraction from real-life clinical notes
Christel Gérardin, Yuhan Xiong, Perceval Wajsbürt, Fabrice Carrat, Xavier Tannier
cs.CL
The objective of our study is to determine whether using English tools to extract and normalize French medical concepts on translations provides comparable performance to French models trained on a set of annotated French clinical notes. We compare two methods: a method involving French language models and a method inv...
http://arxiv.org/pdf/2306.02042v1
2023-06-03T07:48:00
2023-06-03T07:48:00
{ "title": [ -0.048706055, -0.014175415, -0.022415161, -0.040893555, -0.015380859, 0.014671326, -0.051696777, -0.0022296906, -0.022979736, 0.03591919, -0.011917114, 0.0074653625, 0.0036334991, -0.01008606, -0.037200928, -0.0005555153, 0.0113220215, ...
http://arxiv.org/abs/2306.02038v1
Span Identification of Epistemic Stance-Taking in Academic Written English
Masaki Eguchi, Kristopher Kyle
cs.CL, cs.AI, cs.LG
Responding to the increasing need for automated writing evaluation (AWE) systems to assess language use beyond lexis and grammar (Burstein et al., 2016), we introduce a new approach to identify rhetorical features of stance in academic English writing. Drawing on the discourse-analytic framework of engagement in the Ap...
http://arxiv.org/pdf/2306.02038v1
2023-06-03T07:32:25
2023-06-03T07:32:25
{ "title": [ -0.016067505, 0.0032138824, -0.036315918, -0.0135650635, -0.005886078, -0.0262146, 0.005466461, -0.014457703, -0.025238037, 0.012069702, -0.019180298, -0.0024299622, -0.017120361, -0.030334473, -0.033935547, -0.035461426, -0.02079773, ...
http://arxiv.org/abs/2306.02022v1
ACI-BENCH: a Novel Ambient Clinical Intelligence Dataset for Benchmarking Automatic Visit Note Generation
Wen-wai Yim, Yujuan Fu, Asma Ben Abacha, Neal Snider, Thomas Lin, Meliha Yetisgen
cs.CL
Recent immense breakthroughs in generative models such as in GPT4 have precipitated re-imagined ubiquitous usage of these models in all applications. One area that can benefit by improvements in artificial intelligence (AI) is healthcare. The note generation task from doctor-patient encounters, and its associated elect...
http://arxiv.org/pdf/2306.02022v1
2023-06-03T06:42:17
2023-06-03T06:42:17
{ "title": [ -0.0029087067, -0.03161621, -0.018203735, -0.064208984, -0.017364502, 0.008399963, -0.05807495, -0.00079393387, -0.004398346, 0.04196167, -0.055541992, 0.032470703, -0.015899658, -0.006263733, -0.039031982, -0.021224976, 0.015930176, ...
http://arxiv.org/abs/2306.01985v2
COBRA Frames: Contextual Reasoning about Effects and Harms of Offensive Statements
Xuhui Zhou, Hao Zhu, Akhila Yerukola, Thomas Davidson, Jena D. Hwang, Swabha Swayamdipta, Maarten Sap
cs.CL
Warning: This paper contains content that may be offensive or upsetting. Understanding the harms and offensiveness of statements requires reasoning about the social and situational context in which statements are made. For example, the utterance "your English is very good" may implicitly signal an insult when uttered b...
http://arxiv.org/pdf/2306.01985v2
2023-06-09T01:49:06
2023-06-03T02:47:24
{ "title": [ 0.023742676, 0.0025997162, -0.03692627, -0.059295654, 0.008987427, -0.01398468, -0.0020656586, -0.018005371, -0.017730713, 0.031234741, -0.018859863, 0.027236938, 0.0115356445, -0.027832031, -0.02357483, -0.014328003, -0.01008606, 0....
http://arxiv.org/abs/2306.01966v2
GENTLE: A Genre-Diverse Multilayer Challenge Set for English NLP and Linguistic Evaluation
Tatsuya Aoyama, Shabnam Behzad, Luke Gessler, Lauren Levine, Jessica Lin, Yang Janet Liu, Siyao Peng, Yilun Zhu, Amir Zeldes
cs.CL
We present GENTLE, a new mixed-genre English challenge corpus totaling 17K tokens and consisting of 8 unusual text types for out-of domain evaluation: dictionary entries, esports commentaries, legal documents, medical notes, poetry, mathematical proofs, syllabuses, and threat letters. GENTLE is manually annotated for a...
http://arxiv.org/pdf/2306.01966v2
2023-09-22T03:31:17
2023-06-03T00:20:15
{ "title": [ -0.02734375, -0.0073509216, -0.048339844, -0.03164673, 0.021530151, -0.011222839, -0.016067505, 0.003955841, -0.0041046143, 0.0036125183, 0.012672424, 0.020202637, -0.012916565, -0.031799316, -0.020187378, 0.014968872, -0.03918457, 0...
http://arxiv.org/abs/2306.01945v1
Efficient Spoken Language Recognition via Multilabel Classification
Oriol Nieto, Zeyu Jin, Franck Dernoncourt, Justin Salamon
cs.CL, cs.LG
Spoken language recognition (SLR) is the task of automatically identifying the language present in a speech signal. Existing SLR models are either too computationally expensive or too large to run effectively on devices with limited resources. For real-world deployment, a model should also gracefully handle unseen lang...
http://arxiv.org/pdf/2306.01945v1
2023-06-02T23:04:19
2023-06-02T23:04:19
{ "title": [ -0.005016327, 0.0047950745, -0.042114258, -0.07086182, 0.008300781, -0.013244629, -0.05569458, 0.005443573, -0.0064430237, 0.021392822, -0.025558472, 0.007369995, -0.023361206, -0.010795593, -0.04034424, -0.020462036, -0.019897461, 0...
http://arxiv.org/abs/2306.01944v1
EdGCon: Auto-assigner of Iconicity Ratings Grounded by Lexical Properties to Aid in Generation of Technical Gestures
Sameena Hossain, Payal Kamboj, Aranyak Maity, Tamiko Azuma, Ayan Banerjee, Sandeep K. S. Gupta
cs.HC, cs.AI, cs.CL
Gestures that share similarities in their forms and are related in their meanings, should be easier for learners to recognize and incorporate into their existing lexicon. In that regard, to be more readily accepted as standard by the Deaf and Hard of Hearing community, technical gestures in American Sign Language (ASL)...
http://arxiv.org/pdf/2306.01944v1
2023-06-02T23:04:01
2023-06-02T23:04:01
{ "title": [ 0.0022277832, 0.01991272, -0.032989502, 0.0016269684, -0.021972656, -0.008804321, -0.022003174, -0.037231445, 0.01687622, 0.038146973, -0.021224976, 0.020645142, 0.066833496, -0.025253296, 0.012832642, 0.0051193237, -0.014778137, 0.0...
http://arxiv.org/abs/2306.01943v1
NLPositionality: Characterizing Design Biases of Datasets and Models
Sebastin Santy, Jenny T. Liang, Ronan Le Bras, Katharina Reinecke, Maarten Sap
cs.CL, cs.CY, cs.HC
Design biases in NLP systems, such as performance differences for different populations, often stem from their creator's positionality, i.e., views and lived experiences shaped by identity and background. Despite the prevalence and risks of design biases, they are hard to quantify because researcher, system, and datase...
http://arxiv.org/pdf/2306.01943v1
2023-06-02T23:02:09
2023-06-02T23:02:09
{ "title": [ -0.02368164, 0.0015325546, -0.0619812, -0.009559631, 0.0101623535, -0.03314209, -0.019714355, -0.051452637, 0.016204834, 0.0082473755, -0.042419434, 0.007320404, -0.019607544, -0.019180298, -0.038085938, -0.014862061, -0.023254395, 0...
http://arxiv.org/abs/2306.01942v1
Can Contextual Biasing Remain Effective with Whisper and GPT-2?
Guangzhi Sun, Xianrui Zheng, Chao Zhang, Philip C. Woodland
cs.CL, cs.SD, eess.AS
End-to-end automatic speech recognition (ASR) and large language models, such as Whisper and GPT-2, have recently been scaled to use vast amounts of training data. Despite the large amount of training data, infrequent content words that occur in a particular task may still exhibit poor ASR performance, with contextual ...
http://arxiv.org/pdf/2306.01942v1
2023-06-02T22:56:01
2023-06-02T22:56:01
{ "title": [ 0.046295166, 0.018615723, -0.013519287, -0.014793396, -0.015548706, -0.025344849, -0.013511658, -0.053009033, -0.025482178, 0.036376953, -0.023712158, -0.0006766319, 0.019134521, -0.028305054, -0.005794525, 0.00008028746, -0.025543213, ...
http://arxiv.org/abs/2306.01937v1
LIC-GAN: Language Information Conditioned Graph Generative GAN Model
Robert Lo, Arnhav Datar, Abishek Sridhar
cs.LG, cs.AI, cs.CL
Deep generative models for Natural Language data offer a new angle on the problem of graph synthesis: by optimizing differentiable models that directly generate graphs, it is possible to side-step expensive search procedures in the discrete and vast space of possible graphs. We introduce LIC-GAN, an implicit, likelihoo...
http://arxiv.org/pdf/2306.01937v1
2023-06-02T22:39:14
2023-06-02T22:39:14
{ "title": [ -0.04244995, -0.016311646, -0.06439209, -0.034851074, -0.019317627, -0.00969696, -0.04876709, 0.014877319, -0.011009216, 0.008895874, -0.006866455, 0.0064811707, 0.0042495728, -0.050964355, 0.015701294, 0.0024375916, -0.0087890625, 0...
http://arxiv.org/abs/2306.01931v1
Exploring semantic information in disease: Simple Data Augmentation Techniques for Chinese Disease Normalization
Wenqian Cui, Shaohui Liu, Xiangling Fu, Xien Liu, Ji Wu
cs.CL, cs.AI
The disease is a core concept in the medical field, and the task of normalizing disease names is the basis of all disease-related tasks. However, due to the multi-axis and multi-grain nature of disease names, incorrect information is often injected and harms the performance when using general text data augmentation tec...
http://arxiv.org/pdf/2306.01931v1
2023-06-02T22:12:05
2023-06-02T22:12:05
{ "title": [ 0.0004477501, 0.008590698, -0.051727295, -0.0826416, 0.028518677, 0.0017871857, -0.034362793, -0.0010166168, -0.0020370483, 0.016708374, -0.02670288, 0.033081055, -0.02658081, -0.026168823, -0.018814087, 0.0026664734, 0.0022201538, 0...
http://arxiv.org/abs/2306.01930v2
Structural Similarities Between Language Models and Neural Response Measurements
Jiaang Li, Antonia Karamolegkou, Yova Kementchedjhieva, Mostafa Abdou, Sune Lehmann, Anders Søgaard
cs.CL, cs.AI
Large language models (LLMs) have complicated internal dynamics, but induce representations of words and phrases whose geometry we can study. Human language processing is also opaque, but neural response measurements can provide (noisy) recordings of activation during listening or reading, from which we can extract sim...
http://arxiv.org/pdf/2306.01930v2
2023-10-31T15:02:05
2023-06-02T22:09:46
{ "title": [ -0.03375244, 0.015823364, -0.025283813, -0.0006828308, -0.0211792, -0.017410278, -0.031188965, -0.035705566, -0.010597229, 0.048797607, 0.002784729, -0.004096985, -0.0006518364, -0.037322998, -0.05593872, 0.020324707, -0.0018758774, ...
http://arxiv.org/abs/2306.01907v1
A Simple yet Effective Self-Debiasing Framework for Transformer Models
Xiaoyue Wang, Lijie Wang, Xin Liu, Suhang Wu, Jinsong Su, Hua Wu
cs.CL
Current Transformer-based natural language understanding (NLU) models heavily rely on dataset biases, while failing to handle real-world out-of-distribution (OOD) instances. Many methods have been proposed to deal with this issue, but they ignore the fact that the features learned in different layers of Transformer-bas...
http://arxiv.org/pdf/2306.01907v1
2023-06-02T20:31:58
2023-06-02T20:31:58
{ "title": [ 0.007549286, 0.031143188, -0.018615723, -0.054779053, 0.02960205, 0.005695343, -0.042175293, -0.061431885, -0.016723633, 0.04159546, 0.017333984, 0.0018663406, 0.016433716, -0.035858154, -0.025253296, 0.023788452, 0.010940552, 0.0213...
http://arxiv.org/abs/2306.03778v1
Streaming Speech-to-Confusion Network Speech Recognition
Denis Filimonov, Prabhat Pandey, Ariya Rastrow, Ankur Gandhe, Andreas Stolcke
eess.AS, cs.CL
In interactive automatic speech recognition (ASR) systems, low-latency requirements limit the amount of search space that can be explored during decoding, particularly in end-to-end neural ASR. In this paper, we present a novel streaming ASR architecture that outputs a confusion network while maintaining limited latenc...
http://arxiv.org/pdf/2306.03778v1
2023-06-02T20:28:14
2023-06-02T20:28:14
{ "title": [ 0.0056648254, 0.019882202, -0.0597229, -0.04776001, 0.00041747093, -0.023040771, -0.03994751, 0.02494812, 0.006187439, 0.053741455, -0.049194336, 0.0072135925, 0.023376465, -0.019317627, -0.0024166107, 0.008270264, 0.00031399727, 0.0...
http://arxiv.org/abs/2306.01879v2
Revisiting the Role of Language Priors in Vision-Language Models
Zhiqiu Lin, Xinyue Chen, Deepak Pathak, Pengchuan Zhang, Deva Ramanan
cs.CV, cs.AI, cs.CL
Vision-language models (VLMs) are impactful in part because they can be applied to a variety of visual understanding tasks in a zero-shot fashion, without any fine-tuning. We study $\textit{generative VLMs}$ that are trained for next-word generation given an image. We explore their zero-shot performance on the illustra...
http://arxiv.org/pdf/2306.01879v2
2023-10-05T04:12:28
2023-06-02T19:19:43
{ "title": [ -0.033966064, -0.01473999, -0.029205322, -0.048431396, -0.022521973, -0.0047721863, -0.04727173, -0.0060768127, 0.013092041, 0.011932373, -0.023498535, -0.014266968, -0.002248764, -0.024307251, -0.042053223, 0.002210617, 0.0011386871, ...
http://arxiv.org/abs/2306.01857v1
Knowledge of cultural moral norms in large language models
Aida Ramezani, Yang Xu
cs.CL
Moral norms vary across cultures. A recent line of work suggests that English large language models contain human-like moral biases, but these studies typically do not examine moral variation in a diverse cultural setting. We investigate the extent to which monolingual English language models contain knowledge about mo...
http://arxiv.org/pdf/2306.01857v1
2023-06-02T18:23:35
2023-06-02T18:23:35
{ "title": [ -0.01638794, -0.0024871826, -0.026565552, -0.049468994, -0.005405426, -0.022537231, -0.033569336, 0.0006570816, 0.004798889, 0.020217896, -0.042297363, -0.0005917549, 0.004951477, -0.010299683, -0.02571106, 0.0002477169, -0.0048294067, ...
http://arxiv.org/abs/2306.01855v1
5IDER: Unified Query Rewriting for Steering, Intent Carryover, Disfluencies, Entity Carryover and Repair
Jiarui Lu, Bo-Hsiang Tseng, Joel Ruben Antony Moniz, Site Li, Xueyun Zhu, Hong Yu, Murat Akbacak
cs.CL, cs.AI, cs.LG
Providing voice assistants the ability to navigate multi-turn conversations is a challenging problem. Handling multi-turn interactions requires the system to understand various conversational use-cases, such as steering, intent carryover, disfluencies, entity carryover, and repair. The complexity of this problem is com...
http://arxiv.org/pdf/2306.01855v1
2023-06-02T18:17:52
2023-06-02T18:17:52
{ "title": [ 0.020004272, -0.011138916, -0.060577393, -0.10522461, 0.0057525635, -0.002325058, -0.0546875, 0.042755127, -0.0059127808, 0.041168213, -0.0026950836, 0.015487671, 0.05114746, -0.045318604, 0.0119018555, 0.0018539429, -0.0044403076, -...
http://arxiv.org/abs/2306.01841v1
Binary and Ternary Natural Language Generation
Zechun Liu, Barlas Oguz, Aasish Pappu, Yangyang Shi, Raghuraman Krishnamoorthi
cs.CL
Ternary and binary neural networks enable multiplication-free computation and promise multiple orders of magnitude efficiency gains over full-precision networks if implemented on specialized hardware. However, since both the parameter and the output space are highly discretized, such networks have proven very difficult...
http://arxiv.org/pdf/2306.01841v1
2023-06-02T18:01:02
2023-06-02T18:01:02
{ "title": [ -0.044036865, -0.02420044, -0.03314209, -0.06695557, -0.026687622, 0.009170532, -0.06951904, 0.017822266, 0.000521183, 0.033843994, -0.024093628, -0.017089844, 0.005092621, -0.023773193, -0.035949707, -0.00019574165, -0.025177002, 0....
http://arxiv.org/abs/2306.01735v1
Multilingual Conceptual Coverage in Text-to-Image Models
Michael Saxon, William Yang Wang
cs.CL, cs.AI, cs.CV, eess.IV
We propose "Conceptual Coverage Across Languages" (CoCo-CroLa), a technique for benchmarking the degree to which any generative text-to-image system provides multilingual parity to its training language in terms of tangible nouns. For each model we can assess "conceptual coverage" of a given target language relative to...
http://arxiv.org/pdf/2306.01735v1
2023-06-02T17:59:09
2023-06-02T17:59:09
{ "title": [ -0.02659607, -0.00031876564, 0.0062179565, -0.06488037, -0.012313843, 0.03265381, -0.077819824, -0.022705078, 0.030548096, 0.05130005, -0.026153564, 0.007167816, 0.005264282, -0.008125305, -0.049987793, 0.01360321, 0.0023918152, 0.01...
http://arxiv.org/abs/2306.01733v1
DocFormerv2: Local Features for Document Understanding
Srikar Appalaraju, Peng Tang, Qi Dong, Nishant Sankaran, Yichu Zhou, R. Manmatha
cs.CV, cs.CL, cs.LG
We propose DocFormerv2, a multi-modal transformer for Visual Document Understanding (VDU). The VDU domain entails understanding documents (beyond mere OCR predictions) e.g., extracting information from a form, VQA for documents and other tasks. VDU is challenging as it needs a model to make sense of multiple modalities...
http://arxiv.org/pdf/2306.01733v1
2023-06-02T17:58:03
2023-06-02T17:58:03
{ "title": [ -0.0423584, -0.014793396, -0.043701172, -0.06347656, 0.010070801, 0.007827759, -0.049560547, -0.01751709, 0.023956299, 0.010757446, -0.013046265, -0.015327454, 0.008972168, -0.036010742, -0.04269409, 0.0050315857, 0.016082764, 0.0339...
http://arxiv.org/abs/2306.01729v1
Improving Generalization in Task-oriented Dialogues with Workflows and Action Plans
Stefania Raimondo, Christopher Pal, Xiaotian Liu, David Vazquez, Hector Palacios
cs.CL, cs.AI
Task-oriented dialogue is difficult in part because it involves understanding user intent, collecting information from the user, executing API calls, and generating helpful and fluent responses. However, for complex tasks one must also correctly do all of these things over multiple steps, and in a specific order. While...
http://arxiv.org/pdf/2306.01729v1
2023-06-02T17:54:36
2023-06-02T17:54:36
{ "title": [ -0.023101807, 0.0031147003, -0.059783936, -0.089904785, 0.02520752, -0.0014514923, -0.046295166, 0.038330078, 0.0022621155, 0.050323486, -0.004852295, 0.012687683, 0.0035152435, -0.014411926, -0.048736572, -0.011909485, -0.0003979206, ...
http://arxiv.org/abs/2306.01709v1
Distilling Efficient Language-Specific Models for Cross-Lingual Transfer
Alan Ansell, Edoardo Maria Ponti, Anna Korhonen, Ivan Vulić
cs.CL
Massively multilingual Transformers (MMTs), such as mBERT and XLM-R, are widely used for cross-lingual transfer learning. While these are pretrained to represent hundreds of languages, end users of NLP systems are often interested only in individual languages. For such purposes, the MMTs' language coverage makes them u...
http://arxiv.org/pdf/2306.01709v1
2023-06-02T17:31:52
2023-06-02T17:31:52
{ "title": [ -0.012466431, -0.013389587, -0.017929077, -0.07824707, 0.004714966, 0.013542175, -0.062805176, -0.0043754578, 0.017349243, 0.03869629, -0.0040664673, 0.017791748, 0.028747559, -0.015533447, -0.038513184, 0.0335083, -0.01348114, 0.033...
http://arxiv.org/abs/2306.01708v2
TIES-Merging: Resolving Interference When Merging Models
Prateek Yadav, Derek Tam, Leshem Choshen, Colin Raffel, Mohit Bansal
cs.LG, cs.AI, cs.CL, cs.CV
Transfer learning - i.e., further fine-tuning a pre-trained model on a downstream task - can confer significant advantages, including improved downstream performance, faster convergence, and better sample efficiency. These advantages have led to a proliferation of task-specific fine-tuned models, which typically can on...
http://arxiv.org/pdf/2306.01708v2
2023-10-27T01:09:31
2023-06-02T17:31:32
{ "title": [ -0.0033035278, 0.013542175, -0.0357666, -0.09844971, 0.013465881, 0.02053833, -0.074401855, -0.03756714, -0.01977539, 0.025054932, -0.0115737915, -0.0031280518, 0.01586914, -0.037078857, -0.005455017, 0.010932922, -0.0071144104, 0.01...
http://arxiv.org/abs/2306.01707v3
Learning Multi-Step Reasoning by Solving Arithmetic Tasks
Tianduo Wang, Wei Lu
cs.CL
Mathematical reasoning is regarded as a necessary ability for Language Models (LMs). Recent works demonstrate large LMs' impressive performance in solving math problems. The success is attributed to their Chain-of-Thought (CoT) reasoning abilities, i.e., the ability to decompose complex questions into step-by-step reas...
http://arxiv.org/pdf/2306.01707v3
2023-06-07T03:45:15
2023-06-02T17:29:22
{ "title": [ -0.013755798, -0.0059547424, -0.056030273, -0.05279541, 0.0078125, -0.026046753, -0.017196655, 0.012718201, 0.002943039, 0.0020618439, -0.019973755, 0.003856659, -0.039489746, -0.025131226, -0.019546509, -0.0066986084, 0.008117676, 0...
http://arxiv.org/abs/2306.01693v2
Fine-Grained Human Feedback Gives Better Rewards for Language Model Training
Zeqiu Wu, Yushi Hu, Weijia Shi, Nouha Dziri, Alane Suhr, Prithviraj Ammanabrolu, Noah A. Smith, Mari Ostendorf, Hannaneh Hajishirzi
cs.CL
Language models (LMs) often exhibit undesirable text generation behaviors, including generating false, toxic, or irrelevant outputs. Reinforcement learning from human feedback (RLHF) - where human preference judgments on LM outputs are transformed into a learning signal - has recently shown promise in addressing these ...
http://arxiv.org/pdf/2306.01693v2
2023-10-30T06:34:35
2023-06-02T17:11:37
{ "title": [ 0.010772705, -0.0028572083, -0.019897461, -0.04269409, -0.03643799, -0.022583008, -0.051757812, -0.011001587, -0.029785156, 0.030914307, -0.029327393, 0.012252808, 0.0051231384, -0.024993896, -0.018127441, -0.016220093, -0.0032539368, ...
http://arxiv.org/abs/2306.01657v2
DiffusEmp: A Diffusion Model-Based Framework with Multi-Grained Control for Empathetic Response Generation
Guanqun Bi, Lei Shen, Yanan Cao, Meng Chen, Yuqiang Xie, Zheng Lin, Xiaodong He
cs.CL, cs.AI
Empathy is a crucial factor in open-domain conversations, which naturally shows one's caring and understanding to others. Though several methods have been proposed to generate empathetic responses, existing works often lead to monotonous empathy that refers to generic and safe expressions. In this paper, we propose to ...
http://arxiv.org/pdf/2306.01657v2
2023-07-09T08:09:49
2023-06-02T16:26:21
{ "title": [ 0.0028800964, -0.013092041, -0.040374756, -0.04043579, -0.0030136108, -0.019851685, -0.015838623, 0.0011348724, 0.010597229, 0.006286621, -0.005947113, -0.010643005, 0.0068969727, -0.043945312, -0.0256958, -0.007911682, -0.0115737915, ...
http://arxiv.org/abs/2306.01584v2
Learning from Partially Annotated Data: Example-aware Creation of Gap-filling Exercises for Language Learning
Semere Kiros Bitew, Johannes Deleu, A. Seza Doğruöz, Chris Develder, Thomas Demeester
cs.CL
Since performing exercises (including, e.g., practice tests) forms a crucial component of learning, and creating such exercises requires non-trivial effort from the teacher, there is a great value in automatic exercise generation in digital tools in education. In this paper, we particularly focus on automatic creation ...
http://arxiv.org/pdf/2306.01584v2
2023-06-15T10:23:53
2023-06-02T14:54:16
{ "title": [ 0.005016327, -0.0048980713, 0.0018262863, -0.043060303, -0.004814148, -0.0029830933, -0.018234253, -0.00573349, -0.0051994324, 0.0035095215, -0.02609253, 0.013511658, -0.035888672, -0.02885437, -0.038024902, -0.036834717, -0.0061416626, ...
http://arxiv.org/abs/2306.01579v1
EmoUS: Simulating User Emotions in Task-Oriented Dialogues
Hsien-Chin Lin, Shutong Feng, Christian Geishauser, Nurul Lubis, Carel van Niekerk, Michael Heck, Benjamin Ruppik, Renato Vukovic, Milica Gašić
cs.CL
Existing user simulators (USs) for task-oriented dialogue systems only model user behaviour on semantic and natural language levels without considering the user persona and emotions. Optimising dialogue systems with generic user policies, which cannot model diverse user behaviour driven by different emotional states, m...
http://arxiv.org/pdf/2306.01579v1
2023-06-02T14:48:19
2023-06-02T14:48:19
{ "title": [ -0.00605011, -0.018096924, -0.032104492, -0.04827881, 0.00009518862, -0.01586914, -0.027618408, -0.0033302307, 0.02998352, 0.02748108, -0.007698059, -0.023849487, 0.028869629, -0.036590576, -0.05621338, -0.0012340546, -0.016189575, 0...
http://arxiv.org/abs/2306.01551v1
Comparing a composite model versus chained models to locate a nearest visual object
Antoine Le Borgne, Xavier Marjou, Fanny Parzysz, Tayeb Lemlouma
cs.CL
Extracting information from geographic images and text is crucial for autonomous vehicles to determine in advance the best cell stations to connect to along their future path. Multiple artificial neural network models can address this challenge; however, there is no definitive guidance on the selection of an appropriat...
http://arxiv.org/pdf/2306.01551v1
2023-06-02T13:58:59
2023-06-02T13:58:59
{ "title": [ 0.0519104, 0.0016241074, 0.019683838, -0.011947632, -0.02116394, -0.022766113, 0.017181396, -0.05126953, -0.015731812, 0.028121948, -0.066833496, -0.006954193, 0.01423645, -0.019134521, -0.013473511, -0.045440674, 0.018600464, 0.0162...
http://arxiv.org/abs/2306.01549v1
Evaluating Machine Translation Quality with Conformal Predictive Distributions
Patrizio Giovannotti
cs.CL, stat.ML
This paper presents a new approach for assessing uncertainty in machine translation by simultaneously evaluating translation quality and providing a reliable confidence score. Our approach utilizes conformal predictive distributions to produce prediction intervals with guaranteed coverage, meaning that for any given si...
http://arxiv.org/pdf/2306.01549v1
2023-06-02T13:56:30
2023-06-02T13:56:30
{ "title": [ -0.04864502, -0.0011901855, -0.006965637, -0.02784729, -0.028762817, 0.020553589, -0.05114746, -0.01687622, -0.0006341934, 0.05529785, -0.0014715195, 0.0031414032, 0.013626099, -0.014411926, -0.074279785, 0.018493652, -0.012718201, 0...
http://arxiv.org/abs/2306.01545v2
PassGPT: Password Modeling and (Guided) Generation with Large Language Models
Javier Rando, Fernando Perez-Cruz, Briland Hitaj
cs.CL, cs.AI, cs.CR
Large language models (LLMs) successfully model natural language from vast amounts of text without the need for explicit supervision. In this paper, we investigate the efficacy of LLMs in modeling passwords. We present PassGPT, a LLM trained on password leaks for password generation. PassGPT outperforms existing method...
http://arxiv.org/pdf/2306.01545v2
2023-06-14T22:45:28
2023-06-02T13:49:53
{ "title": [ -0.04458618, -0.025604248, -0.04574585, -0.036590576, -0.009384155, 0.0035896301, -0.03186035, 0.0047836304, -0.014923096, 0.030151367, 0.0058288574, 0.0072746277, 0.015419006, -0.03173828, -0.022277832, 0.01399231, -0.023620605, 0.0...
http://arxiv.org/abs/2306.01506v2
BabySLM: language-acquisition-friendly benchmark of self-supervised spoken language models
Marvin Lavechin, Yaya Sy, Hadrien Titeux, María Andrea Cruz Blandón, Okko Räsänen, Hervé Bredin, Emmanuel Dupoux, Alejandrina Cristia
cs.CL, eess.AS, stat.ML
Self-supervised techniques for learning speech representations have been shown to develop linguistic competence from exposure to speech without the need for human labels. In order to fully realize the potential of these approaches and further our understanding of how infants learn language, simulations must closely emu...
http://arxiv.org/pdf/2306.01506v2
2023-06-08T12:22:30
2023-06-02T12:54:38
{ "title": [ -0.02961731, -0.0037708282, -0.036132812, -0.06561279, -0.011482239, 0.0018844604, -0.07733154, 0.016845703, -0.021575928, 0.026184082, -0.014533997, 0.021148682, -0.019119263, -0.014350891, -0.038269043, 0.0054779053, -0.010871887, ...
http://arxiv.org/abs/2306.01505v2
Supervised Adversarial Contrastive Learning for Emotion Recognition in Conversations
Dou Hu, Yinan Bao, Lingwei Wei, Wei Zhou, Songlin Hu
cs.CL, cs.AI, cs.LG
Extracting generalized and robust representations is a major challenge in emotion recognition in conversations (ERC). To address this, we propose a supervised adversarial contrastive learning (SACL) framework for learning class-spread structured representations in a supervised manner. SACL applies contrast-aware advers...
http://arxiv.org/pdf/2306.01505v2
2023-07-09T23:27:27
2023-06-02T12:52:38
{ "title": [ 0.0004968643, 0.01033783, -0.018234253, -0.042663574, 0.00983429, -0.0040893555, -0.026016235, -0.020812988, 0.025680542, 0.029022217, -0.01725769, -0.00051546097, 0.0019130707, -0.028167725, -0.052947998, -0.0013618469, -0.013305664, ...
http://arxiv.org/abs/2306.01499v1
Can LLMs like GPT-4 outperform traditional AI tools in dementia diagnosis? Maybe, but not today
Zhuo Wang, Rongzhen Li, Bowen Dong, Jie Wang, Xiuxing Li, Ning Liu, Chenhui Mao, Wei Zhang, Liling Dong, Jing Gao, Jianyong Wang
cs.CL, cs.LG
Recent investigations show that large language models (LLMs), specifically GPT-4, not only have remarkable capabilities in common Natural Language Processing (NLP) tasks but also exhibit human-level performance on various professional and academic benchmarks. However, whether GPT-4 can be directly used in practical app...
http://arxiv.org/pdf/2306.01499v1
2023-06-02T12:47:45
2023-06-02T12:47:45
{ "title": [ -0.038208008, 0.012619019, -0.039611816, -0.0047340393, -0.0054473877, -0.022415161, 0.0128479, 0.0018482208, -0.01171875, 0.006919861, -0.039245605, 0.021865845, -0.040496826, -0.028060913, -0.00970459, -0.054382324, 0.0019626617, 0...
http://arxiv.org/abs/2306.01497v1
Data-Efficient French Language Modeling with CamemBERTa
Wissam Antoun, Benoît Sagot, Djamé Seddah
cs.CL
Recent advances in NLP have significantly improved the performance of language models on a variety of tasks. While these advances are largely driven by the availability of large amounts of data and computational power, they also benefit from the development of better training methods and architectures. In this paper, w...
http://arxiv.org/pdf/2306.01497v1
2023-06-02T12:45:34
2023-06-02T12:45:34
{ "title": [ -0.047668457, -0.026947021, -0.027053833, -0.076293945, -0.009895325, 0.024139404, -0.061431885, 0.02381897, 0.027770996, 0.04055786, -0.007949829, -0.020492554, 0.021316528, -0.021469116, -0.05203247, 0.03137207, -0.013648987, 0.061...
http://arxiv.org/abs/2306.01481v1
GAIA Search: Hugging Face and Pyserini Interoperability for NLP Training Data Exploration
Aleksandra Piktus, Odunayo Ogundepo, Christopher Akiki, Akintunde Oladipo, Xinyu Zhang, Hailey Schoelkopf, Stella Biderman, Martin Potthast, Jimmy Lin
cs.CL
Noticing the urgent need to provide tools for fast and user-friendly qualitative analysis of large-scale textual corpora of the modern NLP, we propose to turn to the mature and well-tested methods from the domain of Information Retrieval (IR) - a research field with a long history of tackling TB-scale document collecti...
http://arxiv.org/pdf/2306.01481v1
2023-06-02T12:09:59
2023-06-02T12:09:59
{ "title": [ -0.03744507, -0.03201294, -0.05065918, -0.019973755, -0.029312134, -0.02519226, -0.030929565, -0.0031147003, 0.01927185, -0.0006184578, -0.041656494, -0.026535034, -0.012306213, -0.024780273, -0.03805542, -0.023803711, -0.006958008, ...
http://arxiv.org/abs/2306.01818v1
Beta Thalassemia Carriers detection empowered federated Learning
Muhammad Shoaib Farooq, Hafiz Ali Younas
cs.LG, cs.AI, cs.CL
Thalassemia is a group of inherited blood disorders that happen when hemoglobin, the protein in red blood cells that carries oxygen, is not made enough. It is found all over the body and is needed for survival. If both parents have thalassemia, a child's chance of getting it increases. Genetic counselling and early dia...
http://arxiv.org/pdf/2306.01818v1
2023-06-02T11:59:57
2023-06-02T11:59:57
{ "title": [ 0.029525757, 0.018737793, -0.05831909, -0.06817627, 0.0027599335, -0.023651123, -0.026351929, -0.0005631447, -0.02418518, 0.062561035, -0.019546509, -0.001367569, 0.0026931763, -0.02532959, -0.042633057, -0.020599365, 0.021438599, 0....
http://arxiv.org/abs/2306.01471v1
Guiding Text-to-Text Privatization by Syntax
Stefan Arnold, Dilara Yesilbas, Sven Weinzierl
cs.CL, cs.CR, cs.LG
Metric Differential Privacy is a generalization of differential privacy tailored to address the unique challenges of text-to-text privatization. By adding noise to the representation of words in the geometric space of embeddings, words are replaced with words located in the proximity of the noisy representation. Since ...
http://arxiv.org/pdf/2306.01471v1
2023-06-02T11:52:21
2023-06-02T11:52:21
{ "title": [ 0.011154175, -0.023040771, 0.010681152, -0.06323242, -0.059143066, 0.02154541, -0.06726074, 0.005596161, -0.046966553, 0.03793335, -0.034301758, 0.036499023, 0.02178955, -0.027664185, 0.02178955, -0.019546509, -0.014526367, 0.0264282...
http://arxiv.org/abs/2306.01465v1
Light Coreference Resolution for Russian with Hierarchical Discourse Features
Elena Chistova, Ivan Smirnov
cs.CL
Coreference resolution is the task of identifying and grouping mentions referring to the same real-world entity. Previous neural models have mainly focused on learning span representations and pairwise scores for coreference decisions. However, current methods do not explicitly capture the referential choice in the hie...
http://arxiv.org/pdf/2306.01465v1
2023-06-02T11:41:24
2023-06-02T11:41:24
{ "title": [ 0.011421204, 0.011695862, -0.028457642, -0.11029053, 0.029037476, 0.0063972473, -0.044067383, 0.022384644, -0.011421204, 0.040924072, -0.021697998, 0.018753052, 0.03829956, -0.022964478, -0.017974854, 0.011161804, -0.013549805, 0.031...
http://arxiv.org/abs/2306.01457v1
Driving Context into Text-to-Text Privatization
Stefan Arnold, Dilara Yesilbas, Sven Weinzierl
cs.CL, cs.LG
\textit{Metric Differential Privacy} enables text-to-text privatization by adding calibrated noise to the vector of a word derived from an embedding space and projecting this noisy vector back to a discrete vocabulary using a nearest neighbor search. Since words are substituted without context, this mechanism is expect...
http://arxiv.org/pdf/2306.01457v1
2023-06-02T11:33:06
2023-06-02T11:33:06
{ "title": [ 0.036071777, -0.007106781, 0.007598877, -0.06100464, -0.030059814, -0.0010538101, -0.026565552, 0.002538681, -0.0592041, 0.051971436, -0.04031372, 0.038757324, 0.022766113, -0.017974854, 0.038879395, -0.02331543, -0.011734009, 0.0238...
http://arxiv.org/abs/2306.01444v1
Unsupervised Extractive Summarization of Emotion Triggers
Tiberiu Sosea, Hongli Zhan, Junyi Jessy Li, Cornelia Caragea
cs.CL
Understanding what leads to emotions during large-scale crises is important as it can provide groundings for expressed emotions and subsequently improve the understanding of ongoing disasters. Recent approaches trained supervised models to both detect emotions and explain emotion triggers (events and appraisals) via ab...
http://arxiv.org/pdf/2306.01444v1
2023-06-02T11:07:13
2023-06-02T11:07:13
{ "title": [ 0.011123657, 0.009895325, -0.026260376, -0.07861328, 0.009597778, 0.0023975372, -0.053955078, -0.016021729, 0.035186768, 0.03555298, -0.014221191, -0.0009870529, 0.021514893, -0.014389038, -0.054107666, 0.025482178, 0.017425537, 0.00...
http://arxiv.org/abs/2306.01443v1
Unsupervised Paraphrasing of Multiword Expressions
Takashi Wada, Yuji Matsumoto, Timothy Baldwin, Jey Han Lau
cs.CL, cs.AI, cs.LG
We propose an unsupervised approach to paraphrasing multiword expressions (MWEs) in context. Our model employs only monolingual corpus data and pre-trained language models (without fine-tuning), and does not make use of any external resources such as dictionaries. We evaluate our method on the SemEval 2022 idiomatic se...
http://arxiv.org/pdf/2306.01443v1
2023-06-02T11:06:48
2023-06-02T11:06:48
{ "title": [ 0.008049011, -0.004234314, -0.0042419434, -0.07006836, -0.01210022, 0.009681702, -0.038146973, -0.012680054, -0.0054626465, 0.045684814, -0.02456665, 0.0048675537, 0.019744873, -0.008682251, -0.035125732, 0.0126571655, -0.011985779, ...
http://arxiv.org/abs/2306.01442v1
Towards Robust FastSpeech 2 by Modelling Residual Multimodality
Fabian Kögel, Bac Nguyen, Fabien Cardinaux
cs.SD, cs.CL, cs.LG, eess.AS
State-of-the-art non-autoregressive text-to-speech (TTS) models based on FastSpeech 2 can efficiently synthesise high-fidelity and natural speech. For expressive speech datasets however, we observe characteristic audio distortions. We demonstrate that such artefacts are introduced to the vocoder reconstruction by over-...
http://arxiv.org/pdf/2306.01442v1
2023-06-02T11:03:26
2023-06-02T11:03:26
{ "title": [ -0.025039673, 0.015235901, -0.060150146, -0.030593872, 0.00038337708, -0.020858765, -0.041259766, -0.018936157, 0.045898438, 0.044677734, 0.015449524, -0.0231781, 0.02079773, -0.036346436, -0.048736572, 0.010772705, -0.0017385483, 0....
http://arxiv.org/abs/2306.01439v2
Interpretable and Explainable Logical Policies via Neurally Guided Symbolic Abstraction
Quentin Delfosse, Hikaru Shindo, Devendra Dhami, Kristian Kersting
cs.LG, cs.AI, cs.CL, cs.LO, cs.SC
The limited priors required by neural networks make them the dominating choice to encode and learn policies using reinforcement learning (RL). However, they are also black-boxes, making it hard to understand the agent's behaviour, especially when working on the image level. Therefore, neuro-symbolic RL aims at creating...
http://arxiv.org/pdf/2306.01439v2
2023-10-25T16:40:27
2023-06-02T10:59:44
{ "title": [ 0.0013055801, 0.014328003, -0.016662598, -0.082336426, 0.013977051, 0.0061302185, -0.035461426, -0.0078125, -0.039398193, 0.011665344, -0.026351929, 0.0045700073, -0.027923584, -0.01651001, -0.020614624, -0.028167725, -0.015037537, 0...
http://arxiv.org/abs/2306.03103v1
Sampling and Ranking for Digital Ink Generation on a tight computational budget
Andrei Afonin, Andrii Maksai, Aleksandr Timofeev, Claudiu Musat
cs.HC, cs.CL
Digital ink (online handwriting) generation has a number of potential applications for creating user-visible content, such as handwriting autocompletion, spelling correction, and beautification. Writing is personal and usually the processing is done on-device. Ink generative models thus need to produce high quality con...
http://arxiv.org/pdf/2306.03103v1
2023-06-02T09:55:15
2023-06-02T09:55:15
{ "title": [ -0.03189087, -0.002002716, -0.0115737915, -0.026977539, -0.018920898, 0.0058631897, -0.045196533, -0.026290894, 0.0035381317, 0.029678345, -0.03704834, 0.017822266, -0.014907837, -0.003293991, -0.046447754, 0.004207611, 0.010856628, ...
http://arxiv.org/abs/2306.01399v1
Knowledge Graph Reasoning over Entities and Numerical Values
Jiaxin Bai, Chen Luo, Zheng Li, Qingyu Yin, Bing Yin, Yangqiu Song
cs.AI, cs.CL, cs.LO
A complex logic query in a knowledge graph refers to a query expressed in logic form that conveys a complex meaning, such as where did the Canadian Turing award winner graduate from? Knowledge graph reasoning-based applications, such as dialogue systems and interactive search engines, rely on the ability to answer comp...
http://arxiv.org/pdf/2306.01399v1
2023-06-02T09:46:29
2023-06-02T09:46:29
{ "title": [ -0.019699097, -0.0044021606, -0.06689453, -0.0736084, 0.0011863708, -0.01953125, -0.035980225, 0.01763916, 0.008354187, 0.014945984, -0.030822754, -0.0055160522, 0.015548706, -0.033691406, 0.0049743652, -0.0020999908, 0.002002716, 0....
http://arxiv.org/abs/2306.01393v2
Assessing the Importance of Frequency versus Compositionality for Subword-based Tokenization in NMT
Benoist Wolleb, Romain Silvestri, Giorgos Vernikos, Ljiljana Dolamic, Andrei Popescu-Belis
cs.CL
Subword tokenization is the de facto standard for tokenization in neural language models and machine translation systems. Three advantages are frequently cited in favor of subwords: shorter encoding of frequent tokens, compositionality of subwords, and ability to deal with unknown words. As their relative importance is...
http://arxiv.org/pdf/2306.01393v2
2023-06-05T09:04:32
2023-06-02T09:39:36
{ "title": [ -0.0007081032, 0.0021877289, -0.0077781677, -0.0625, 0.0057678223, 0.0069084167, -0.04940796, -0.026748657, -0.031188965, 0.040802002, -0.041229248, 0.02520752, 0.015136719, -0.0048065186, -0.016448975, -0.0075531006, -0.037475586, 0...
http://arxiv.org/abs/2306.01386v1
ChatGPT for Zero-shot Dialogue State Tracking: A Solution or an Opportunity?
Michael Heck, Nurul Lubis, Benjamin Ruppik, Renato Vukovic, Shutong Feng, Christian Geishauser, Hsien-Chin Lin, Carel van Niekerk, Milica Gašić
cs.CL, cs.AI
Recent research on dialogue state tracking (DST) focuses on methods that allow few- and zero-shot transfer to new domains or schemas. However, performance gains heavily depend on aggressive data augmentation and fine-tuning of ever larger language model based architectures. In contrast, general purpose language models,...
http://arxiv.org/pdf/2306.01386v1
2023-06-02T09:15:01
2023-06-02T09:15:01
{ "title": [ 0.0046958923, -0.030166626, -0.02104187, -0.047088623, -0.018234253, -0.013381958, -0.02015686, 0.0071029663, -0.008857727, 0.017303467, 0.0026416779, 0.013977051, -0.017318726, -0.03378296, -0.023117065, 0.0028533936, -0.02432251, 0...
http://arxiv.org/abs/2306.01385v2
Task-Agnostic Structured Pruning of Speech Representation Models
Haoyu Wang, Siyuan Wang, Wei-Qiang Zhang, Hongbin Suo, Yulong Wan
eess.AS, cs.CL, cs.SD
Self-supervised pre-trained models such as Wav2vec2, Hubert, and WavLM have been shown to significantly improve many speech tasks. However, their large memory and strong computational requirements hinder their industrial applicability. Structured pruning is a hardware-friendly model compression technique but usually re...
http://arxiv.org/pdf/2306.01385v2
2023-07-09T06:31:46
2023-06-02T09:11:06
{ "title": [ -0.01020813, 0.0029563904, -0.04220581, -0.095825195, 0.0012245178, 0.01638794, -0.08538818, 0.034942627, 0.009056091, 0.04321289, -0.002998352, -0.005256653, 0.022247314, -0.020828247, -0.042907715, 0.016906738, 0.0011730194, 0.0322...
http://arxiv.org/abs/2306.01382v2
Leveraging Auxiliary Domain Parallel Data in Intermediate Task Fine-tuning for Low-resource Translation
Shravan Nayak, Surangika Ranathunga, Sarubi Thillainathan, Rikki Hung, Anthony Rinaldi, Yining Wang, Jonah Mackey, Andrew Ho, En-Shiun Annie Lee
cs.CL
NMT systems trained on Pre-trained Multilingual Sequence-Sequence (PMSS) models flounder when sufficient amounts of parallel data is not available for fine-tuning. This specifically holds for languages missing/under-represented in these models. The problem gets aggravated when the data comes from different domains. In ...
http://arxiv.org/pdf/2306.01382v2
2023-09-24T01:33:13
2023-06-02T09:05:18
{ "title": [ -0.041137695, -0.01626587, -0.035003662, -0.07104492, -0.021697998, 0.003622055, -0.04949951, 0.017852783, -0.028030396, 0.082458496, 0.0041046143, 0.027160645, 0.027236938, -0.0070381165, -0.059448242, 0.019454956, -0.005088806, 0.0...
http://arxiv.org/abs/2306.01337v2
An Empirical Study on Challenging Math Problem Solving with GPT-4
Yiran Wu, Feiran Jia, Shaokun Zhang, Hangyu Li, Erkang Zhu, Yue Wang, Yin Tat Lee, Richard Peng, Qingyun Wu, Chi Wang
cs.CL, stat.ML
Employing Large Language Models (LLMs) to address mathematical problems is an intriguing research endeavor, considering the abundance of math problems expressed in natural language across numerous science and engineering fields. While several prior works have investigated solving elementary mathematics using LLMs, this...
http://arxiv.org/pdf/2306.01337v2
2023-06-08T02:34:35
2023-06-02T08:02:15
{ "title": [ 0.021743774, 0.0065841675, -0.023956299, -0.0023384094, 0.011741638, -0.048858643, 0.02897644, -0.029266357, -0.005607605, 0.009986877, -0.018066406, -0.0016059875, -0.0019130707, -0.030258179, -0.022827148, -0.007797241, -0.021911621, ...