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/2307.01595v1
Prompt Tuning Pushes Farther, Contrastive Learning Pulls Closer: A Two-Stage Approach to Mitigate Social Biases
Yingji Li, Mengnan Du, Xin Wang, Ying Wang
cs.CL, cs.AI
As the representation capability of Pre-trained Language Models (PLMs) improve, there is growing concern that they will inherit social biases from unprocessed corpora. Most previous debiasing techniques used Counterfactual Data Augmentation (CDA) to balance the training corpus. However, CDA slightly modifies the origin...
http://arxiv.org/pdf/2307.01595v1
2023-07-04T09:35:03
2023-07-04T09:35:03
{ "title": [ 0.036102295, -0.008583069, 0.022216797, -0.019866943, -0.012161255, -0.023269653, 0.00749588, -0.024169922, -0.060455322, 0.017669678, -0.01586914, 0.012023926, -0.013877869, -0.036376953, 0.010955811, -0.040130615, -0.0030326843, 0....
http://arxiv.org/abs/2307.01542v1
Mitigating the Learning Bias towards Repetition by Self-Contrastive Training for Open-Ended Generation
Jian Guan, Minlie Huang
cs.CL
Despite the huge progress in myriad generation tasks, pretrained language models (LMs) such as GPT2 still tend to generate repetitive texts with maximization-based decoding algorithms for open-ended generation. We attribute their overestimation of token-level repetition probabilities to the learning bias: LMs capture s...
http://arxiv.org/pdf/2307.01542v1
2023-07-04T07:53:55
2023-07-04T07:53:55
{ "title": [ 0.018249512, 0.014465332, -0.01977539, -0.050689697, 0.00018763542, -0.021896362, -0.022979736, -0.033721924, -0.007583618, 0.033172607, 0.007637024, 0.016159058, -0.014793396, -0.039642334, -0.021606445, 0.008094788, -0.017349243, 0...
http://arxiv.org/abs/2307.01540v2
Learning to Prompt in the Classroom to Understand AI Limits: A pilot study
Emily Theophilou, Cansu Koyuturk, Mona Yavari, Sathya Bursic, Gregor Donabauer, Alessia Telari, Alessia Testa, Raffaele Boiano, Davinia Hernandez-Leo, Martin Ruskov, Davide Taibi, Alessandro Gabbiadini, Dimitri Ognibene
cs.HC, cs.AI, cs.CL
Artificial intelligence's (AI) progress holds great promise in tackling pressing societal concerns such as health and climate. Large Language Models (LLM) and the derived chatbots, like ChatGPT, have highly improved the natural language processing capabilities of AI systems allowing them to process an unprecedented amo...
http://arxiv.org/pdf/2307.01540v2
2023-09-01T15:31:21
2023-07-04T07:51:37
{ "title": [ -0.010261536, -0.0070381165, -0.0062408447, -0.02079773, -0.00038027763, -0.0104522705, 0.0029582977, -0.03390503, -0.014701843, -0.015510559, -0.07385254, 0.008117676, -0.028320312, -0.004333496, -0.014137268, -0.039093018, -0.003231048...
http://arxiv.org/abs/2307.01503v1
On Evaluating and Mitigating Gender Biases in Multilingual Settings
Aniket Vashishtha, Kabir Ahuja, Sunayana Sitaram
cs.CL
While understanding and removing gender biases in language models has been a long-standing problem in Natural Language Processing, prior research work has primarily been limited to English. In this work, we investigate some of the challenges with evaluating and mitigating biases in multilingual settings which stem from...
http://arxiv.org/pdf/2307.01503v1
2023-07-04T06:23:04
2023-07-04T06:23:04
{ "title": [ -0.023513794, -0.00084257126, -0.047454834, -0.019058228, -0.01676941, -0.029312134, -0.026275635, -0.005405426, -0.017196655, 0.057037354, -0.013511658, -0.011985779, 0.04232788, -0.021697998, -0.02835083, -0.0026283264, -0.0357666, ...
http://arxiv.org/abs/2307.01488v1
SCAT: Robust Self-supervised Contrastive Learning via Adversarial Training for Text Classification
Junjie Wu, Dit-Yan Yeung
cs.CL
Despite their promising performance across various natural language processing (NLP) tasks, current NLP systems are vulnerable to textual adversarial attacks. To defend against these attacks, most existing methods apply adversarial training by incorporating adversarial examples. However, these methods have to rely on g...
http://arxiv.org/pdf/2307.01488v1
2023-07-04T05:41:31
2023-07-04T05:41:31
{ "title": [ -0.0034217834, 0.017196655, -0.027679443, -0.03729248, 0.020629883, -0.011474609, -0.005962372, -0.04071045, -0.001821518, 0.032073975, -0.0064697266, 0.031280518, -0.021942139, -0.01524353, -0.05078125, -0.0066070557, -0.015945435, ...
http://arxiv.org/abs/2307.01458v4
CARE-MI: Chinese Benchmark for Misinformation Evaluation in Maternity and Infant Care
Tong Xiang, Liangzhi Li, Wangyue Li, Mingbai Bai, Lu Wei, Bowen Wang, Noa Garcia
cs.CL
The recent advances in natural language processing (NLP), have led to a new trend of applying large language models (LLMs) to real-world scenarios. While the latest LLMs are astonishingly fluent when interacting with humans, they suffer from the misinformation problem by unintentionally generating factually false state...
http://arxiv.org/pdf/2307.01458v4
2023-10-26T07:19:58
2023-07-04T03:34:19
{ "title": [ -0.02571106, -0.014511108, -0.033477783, -0.011482239, -0.009651184, -0.009986877, -0.018463135, -0.004798889, -0.049621582, 0.026000977, -0.0017623901, 0.027526855, 0.005569458, -0.053619385, 0.003583908, -0.016098022, -0.027282715, ...
http://arxiv.org/abs/2307.01453v1
Diverse Retrieval-Augmented In-Context Learning for Dialogue State Tracking
Brendan King, Jeffrey Flanigan
cs.CL
There has been significant interest in zero and few-shot learning for dialogue state tracking (DST) due to the high cost of collecting and annotating task-oriented dialogues. Recent work has demonstrated that in-context learning requires very little data and zero parameter updates, and even outperforms trained methods ...
http://arxiv.org/pdf/2307.01453v1
2023-07-04T03:15:52
2023-07-04T03:15:52
{ "title": [ 0.023498535, -0.0025730133, -0.04083252, -0.076171875, 0.013397217, -0.023239136, -0.022857666, 0.008178711, 0.00982666, 0.018936157, -0.023376465, 0.015525818, -0.012184143, -0.022750854, -0.028121948, -0.013717651, -0.011314392, 0....
http://arxiv.org/abs/2307.01448v1
ReactIE: Enhancing Chemical Reaction Extraction with Weak Supervision
Ming Zhong, Siru Ouyang, Minhao Jiang, Vivian Hu, Yizhu Jiao, Xuan Wang, Jiawei Han
cs.CL
Structured chemical reaction information plays a vital role for chemists engaged in laboratory work and advanced endeavors such as computer-aided drug design. Despite the importance of extracting structured reactions from scientific literature, data annotation for this purpose is cost-prohibitive due to the significant...
http://arxiv.org/pdf/2307.01448v1
2023-07-04T02:52:30
2023-07-04T02:52:30
{ "title": [ -0.017486572, 0.0006656647, -0.06335449, -0.087768555, 0.022354126, -0.003540039, -0.052734375, 0.010375977, -0.0064964294, 0.014831543, -0.014602661, 0.028182983, -0.030380249, -0.022979736, -0.02508545, 0.008705139, -0.0029792786, ...
http://arxiv.org/abs/2307.01446v1
On Conditional and Compositional Language Model Differentiable Prompting
Jonathan Pilault, Can Liu, Mohit Bansal, Markus Dreyer
cs.CL, cs.LG
Prompts have been shown to be an effective method to adapt a frozen Pretrained Language Model (PLM) to perform well on downstream tasks. Prompts can be represented by a human-engineered word sequence or by a learned continuous embedding. In this work, we investigate conditional and compositional differentiable promptin...
http://arxiv.org/pdf/2307.01446v1
2023-07-04T02:47:42
2023-07-04T02:47:42
{ "title": [ -0.021987915, 0.01210022, -0.03894043, -0.072509766, -0.0056877136, 0.004047394, -0.091918945, -0.0041275024, -0.018310547, 0.024291992, -0.035064697, 0.001958847, -0.004722595, -0.02720642, -0.004753113, -0.011688232, -0.017959595, ...
http://arxiv.org/abs/2307.06844v1
Garbage in, garbage out: Zero-shot detection of crime using Large Language Models
Anj Simmons, Rajesh Vasa
cs.CL, cs.AI, cs.CV
This paper proposes exploiting the common sense knowledge learned by large language models to perform zero-shot reasoning about crimes given textual descriptions of surveillance videos. We show that when video is (manually) converted to high quality textual descriptions, large language models are capable of detecting a...
http://arxiv.org/pdf/2307.06844v1
2023-07-04T01:29:15
2023-07-04T01:29:15
{ "title": [ -0.01713562, -0.009407043, -0.032196045, -0.03189087, -0.0022525787, -0.011909485, -0.023605347, -0.0071983337, 0.0029335022, 0.030075073, -0.020233154, 0.0050201416, -0.00057172775, -0.0077209473, -0.03829956, -0.003572464, -0.007686615...
http://arxiv.org/abs/2307.01420v1
Modeling Tag Prediction based on Question Tagging Behavior Analysis of CommunityQA Platform Users
Kuntal Kumar Pal, Michael Gamon, Nirupama Chandrasekaran, Silviu Cucerzan
cs.CL
In community question-answering platforms, tags play essential roles in effective information organization and retrieval, better question routing, faster response to questions, and assessment of topic popularity. Hence, automatic assistance for predicting and suggesting tags for posts is of high utility to users of suc...
http://arxiv.org/pdf/2307.01420v1
2023-07-04T01:24:26
2023-07-04T01:24:26
{ "title": [ 0.023147583, 0.01576233, -0.00015211105, -0.035980225, 0.029556274, -0.0028629303, 0.035095215, -0.024658203, -0.00818634, 0.048553467, 0.0038642883, 0.015701294, 0.002002716, -0.018417358, -0.07159424, 0.007774353, -0.0050697327, 0....
http://arxiv.org/abs/2307.01401v1
Multi-Task Learning Improves Performance In Deep Argument Mining Models
Amirhossein Farzam, Shashank Shekhar, Isaac Mehlhaff, Marco Morucci
cs.CL
The successful analysis of argumentative techniques from user-generated text is central to many downstream tasks such as political and market analysis. Recent argument mining tools use state-of-the-art deep learning methods to extract and annotate argumentative techniques from various online text corpora, however each ...
http://arxiv.org/pdf/2307.01401v1
2023-07-03T23:42:29
2023-07-03T23:42:29
{ "title": [ 0.017913818, 0.0002887249, -0.02923584, -0.07891846, 0.003929138, -0.010559082, -0.04373169, 0.0005044937, -0.009803772, 0.023712158, -0.005241394, 0.0152282715, -0.0231781, -0.0063095093, -0.037384033, -0.0073242188, -0.015930176, 0...
http://arxiv.org/abs/2307.01387v1
ALBERTI, a Multilingual Domain Specific Language Model for Poetry Analysis
Javier de la Rosa, Álvaro Pérez Pozo, Salvador Ros, Elena González-Blanco
cs.CL
The computational analysis of poetry is limited by the scarcity of tools to automatically analyze and scan poems. In a multilingual settings, the problem is exacerbated as scansion and rhyme systems only exist for individual languages, making comparative studies very challenging and time consuming. In this work, we pre...
http://arxiv.org/pdf/2307.01387v1
2023-07-03T22:50:53
2023-07-03T22:50:53
{ "title": [ -0.035949707, -0.037384033, -0.050048828, -0.022079468, -0.06323242, -0.006095886, -0.07635498, 0.018051147, -0.0013008118, 0.066223145, 0.027816772, -0.016983032, 0.064453125, -0.058685303, -0.002986908, 0.01939392, -0.0069465637, 0...
http://arxiv.org/abs/2307.01381v1
Implicit Memory Transformer for Computationally Efficient Simultaneous Speech Translation
Matthew Raffel, Lizhong Chen
cs.CL, cs.LG
Simultaneous speech translation is an essential communication task difficult for humans whereby a translation is generated concurrently with oncoming speech inputs. For such a streaming task, transformers using block processing to break an input sequence into segments have achieved state-of-the-art performance at a red...
http://arxiv.org/pdf/2307.01381v1
2023-07-03T22:20:21
2023-07-03T22:20:21
{ "title": [ -0.008544922, -0.0024986267, -0.015457153, -0.08721924, -0.00030636787, 0.00983429, -0.03866577, 0.018737793, 0.0037651062, 0.064086914, -0.011360168, -0.0048179626, 0.04425049, -0.011772156, -0.045715332, 0.0211792, 0.00074768066, 0...
http://arxiv.org/abs/2307.01379v2
Shifting Attention to Relevance: Towards the Uncertainty Estimation of Large Language Models
Jinhao Duan, Hao Cheng, Shiqi Wang, Alex Zavalny, Chenan Wang, Renjing Xu, Bhavya Kailkhura, Kaidi Xu
cs.CL, cs.AI, cs.LG
While Large Language Models (LLMs) have demonstrated remarkable potential in natural language generation and instruction following, a persistent challenge lies in their susceptibility to "hallucinations", which erodes trust in their outputs. Although Uncertainty Quantification (UQ) presents a promising solution, its ac...
http://arxiv.org/pdf/2307.01379v2
2023-10-09T14:26:59
2023-07-03T22:17:16
{ "title": [ -0.05267334, -0.005542755, -0.050994873, -0.050598145, -0.002904892, -0.0043907166, -0.072753906, -0.0028877258, 0.010124207, 0.02230835, -0.013404846, -0.008300781, 0.009185791, -0.022903442, -0.0496521, 0.011421204, -0.007106781, 0...
http://arxiv.org/abs/2307.01377v1
Shiftable Context: Addressing Training-Inference Context Mismatch in Simultaneous Speech Translation
Matthew Raffel, Drew Penney, Lizhong Chen
cs.CL, cs.LG
Transformer models using segment-based processing have been an effective architecture for simultaneous speech translation. However, such models create a context mismatch between training and inference environments, hindering potential translation accuracy. We solve this issue by proposing Shiftable Context, a simple ye...
http://arxiv.org/pdf/2307.01377v1
2023-07-03T22:11:51
2023-07-03T22:11:51
{ "title": [ -0.011909485, -0.0041656494, -0.022827148, -0.0826416, 0.0017280579, 0.0029067993, -0.04046631, 0.010795593, -0.02267456, 0.052520752, -0.01146698, 0.017929077, 0.019119263, -0.023895264, -0.017196655, 0.00096416473, -0.014015198, 0....
http://arxiv.org/abs/2307.01370v2
Multilingual Language Models are not Multicultural: A Case Study in Emotion
Shreya Havaldar, Sunny Rai, Bhumika Singhal, Langchen Liu, Sharath Chandra Guntuku, Lyle Ungar
cs.CL
Emotions are experienced and expressed differently across the world. In order to use Large Language Models (LMs) for multilingual tasks that require emotional sensitivity, LMs must reflect this cultural variation in emotion. In this study, we investigate whether the widely-used multilingual LMs in 2023 reflect differen...
http://arxiv.org/pdf/2307.01370v2
2023-07-09T15:21:22
2023-07-03T21:54:28
{ "title": [ -0.033233643, -0.003622055, -0.030914307, -0.03567505, -0.0146865845, -0.010124207, -0.039398193, -0.006263733, -0.0006518364, 0.042633057, 0.00079393387, -0.019607544, 0.014961243, -0.035583496, -0.059906006, 0.0065994263, -0.011489868,...
http://arxiv.org/abs/2307.01323v1
Semantic enrichment towards efficient speech representations
Gaëlle Laperrière, Ha Nguyen, Sahar Ghannay, Bassam Jabaian, Yannick Estève
cs.CL, cs.SD, eess.AS
Over the past few years, self-supervised learned speech representations have emerged as fruitful replacements for conventional surface representations when solving Spoken Language Understanding (SLU) tasks. Simultaneously, multilingual models trained on massive textual data were introduced to encode language agnostic s...
http://arxiv.org/pdf/2307.01323v1
2023-07-03T19:52:56
2023-07-03T19:52:56
{ "title": [ 0.008804321, 0.00027394295, -0.05078125, -0.099853516, 0.019134521, -0.006298065, -0.057556152, 0.021087646, 0.01158905, 0.056030273, -0.005645752, 0.004524231, 0.024520874, -0.014793396, -0.03982544, 0.02949524, -0.007797241, 0.0546...
http://arxiv.org/abs/2307.01310v1
Exploring Spoken Named Entity Recognition: A Cross-Lingual Perspective
Moncef Benaicha, David Thulke, M. A. Tuğtekin Turan
cs.CL
Recent advancements in Named Entity Recognition (NER) have significantly improved the identification of entities in textual data. However, spoken NER, a specialized field of spoken document retrieval, lags behind due to its limited research and scarce datasets. Moreover, cross-lingual transfer learning in spoken NER ha...
http://arxiv.org/pdf/2307.01310v1
2023-07-03T19:30:24
2023-07-03T19:30:24
{ "title": [ -0.013557434, 0.0062942505, -0.04171753, -0.06414795, 0.019958496, -0.0042877197, -0.048187256, 0.019515991, 0.0033626556, 0.022659302, -0.049102783, 0.011116028, 0.0052337646, 0.0044174194, -0.04525757, -0.0059547424, -0.012786865, ...
http://arxiv.org/abs/2307.01299v1
The Evolution of Substance Use Coverage in the Philadelphia Inquirer
Layla Bouzoubaa, Ramtin Ehsani, Preetha Chatterjee, Rezvaneh Rezapour
cs.CL
The media's representation of illicit substance use can lead to harmful stereotypes and stigmatization for individuals struggling with addiction, ultimately influencing public perception, policy, and public health outcomes. To explore how the discourse and coverage of illicit drug use changed over time, this study anal...
http://arxiv.org/pdf/2307.01299v1
2023-07-03T19:09:04
2023-07-03T19:09:04
{ "title": [ 0.049865723, -0.024368286, -0.03753662, -0.050445557, -0.004207611, -0.028701782, 0.007888794, 0.025375366, -0.003929138, 0.017303467, -0.068725586, 0.011405945, -0.023162842, -0.00017046928, -0.0031642914, -0.04208374, 0.011100769, ...
http://arxiv.org/abs/2307.01189v1
Trainable Transformer in Transformer
Abhishek Panigrahi, Sadhika Malladi, Mengzhou Xia, Sanjeev Arora
cs.CL, cs.LG
Recent works attribute the capability of in-context learning (ICL) in large pre-trained language models to implicitly simulating and fine-tuning an internal model (e.g., linear or 2-layer MLP) during inference. However, such constructions require large memory overhead, which makes simulation of more sophisticated inter...
http://arxiv.org/pdf/2307.01189v1
2023-07-03T17:53:39
2023-07-03T17:53:39
{ "title": [ 0.051727295, 0.0129852295, -0.00038957596, 0.0040130615, -0.0335083, -0.024414062, 0.0058670044, -0.030395508, -0.045654297, 0.039215088, -0.04421997, 0.011009216, 0.020126343, -0.012084961, 0.03451538, -0.03363037, 0.023529053, 0.03...
http://arxiv.org/abs/2307.01163v2
Improving Language Plasticity via Pretraining with Active Forgetting
Yihong Chen, Kelly Marchisio, Roberta Raileanu, David Ifeoluwa Adelani, Pontus Stenetorp, Sebastian Riedel, Mikel Artetxe
cs.CL, cs.LG, cs.NE
Pretrained language models (PLMs) are today the primary model for natural language processing. Despite their impressive downstream performance, it can be difficult to apply PLMs to new languages, a barrier to making their capabilities universally accessible. While prior work has shown it possible to address this issue ...
http://arxiv.org/pdf/2307.01163v2
2023-07-04T14:56:43
2023-07-03T17:12:44
{ "title": [ -0.005432129, 0.0060577393, -0.042388916, -0.07763672, 0.010292053, -0.00066280365, -0.04019165, 0.011726379, -0.013885498, 0.039794922, 0.011741638, 0.027893066, -0.007484436, -0.031021118, -0.034210205, 0.019256592, -0.0010375977, ...
http://arxiv.org/abs/2307.07520v1
Translating Latin with Artificial Intelligence
Sylvio R. Bistafa
math.HO, cs.CL, physics.ed-ph
The major hindrance in the study of earlier scientific literature is the availability of Latin translations into modern languages. This is particular true for the works of Euler who authored about 850 manuscripts and wrote a thousand letters and received back almost two thousand more. The translation of many of these m...
http://arxiv.org/pdf/2307.07520v1
2023-07-03T16:27:32
2023-07-03T16:27:32
{ "title": [ -0.048858643, -0.013542175, 0.0022964478, 0.049194336, -0.07788086, -0.0003657341, 0.019424438, -0.030151367, 0.004623413, 0.07165527, -0.035308838, -0.024047852, 0.045074463, -0.036346436, -0.033203125, 0.0020141602, 0.027908325, 0....
http://arxiv.org/abs/2307.01139v1
SCITUNE: Aligning Large Language Models with Scientific Multimodal Instructions
Sameera Horawalavithana, Sai Munikoti, Ian Stewart, Henry Kvinge
cs.CV, cs.AI, cs.CL, cs.LG
Instruction finetuning is a popular paradigm to align large language models (LLM) with human intent. Despite its popularity, this idea is less explored in improving the LLMs to align existing foundation models with scientific disciplines, concepts and goals. In this work, we present SciTune as a tuning framework to imp...
http://arxiv.org/pdf/2307.01139v1
2023-07-03T16:25:49
2023-07-03T16:25:49
{ "title": [ -0.061920166, -0.011779785, -0.059814453, -0.044006348, -0.018829346, 0.0048942566, -0.08251953, -0.009063721, 0.009346008, 0.03656006, 0.018173218, -0.014678955, 0.018753052, -0.033935547, -0.031585693, 0.020080566, -0.0029010773, 0...
http://arxiv.org/abs/2307.01137v1
Exploring the In-context Learning Ability of Large Language Model for Biomedical Concept Linking
Qinyong Wang, Zhenxiang Gao, Rong Xu
cs.CL, cs.AI
The biomedical field relies heavily on concept linking in various areas such as literature mining, graph alignment, information retrieval, question-answering, data, and knowledge integration. Although large language models (LLMs) have made significant strides in many natural language processing tasks, their effectivene...
http://arxiv.org/pdf/2307.01137v1
2023-07-03T16:19:50
2023-07-03T16:19:50
{ "title": [ -0.010063171, -0.01525116, -0.030929565, -0.045013428, -0.012435913, -0.0003156662, -0.03466797, -0.011123657, 0.018005371, 0.024627686, -0.0038013458, -0.019119263, 0.00315094, -0.03555298, -0.030807495, 0.0026817322, -0.0121154785, ...
http://arxiv.org/abs/2307.01128v1
Iterative Zero-Shot LLM Prompting for Knowledge Graph Construction
Salvatore Carta, Alessandro Giuliani, Leonardo Piano, Alessandro Sebastian Podda, Livio Pompianu, Sandro Gabriele Tiddia
cs.CL, cs.AI
In the current digitalization era, capturing and effectively representing knowledge is crucial in most real-world scenarios. In this context, knowledge graphs represent a potent tool for retrieving and organizing a vast amount of information in a properly interconnected and interpretable structure. However, their gener...
http://arxiv.org/pdf/2307.01128v1
2023-07-03T16:01:45
2023-07-03T16:01:45
{ "title": [ -0.06274414, -0.033569336, -0.07116699, -0.08074951, -0.05441284, 0.015899658, -0.11669922, 0.040924072, -0.0029258728, 0.0064849854, -0.01737976, 0.030639648, -0.029449463, -0.027542114, -0.00052404404, -0.006286621, 0.011390686, -0...
http://arxiv.org/abs/2307.01076v1
Analyzing Multiple-Choice Reading and Listening Comprehension Tests
Vatsal Raina, Adian Liusie, Mark Gales
cs.CL
Multiple-choice reading and listening comprehension tests are an important part of language assessment. Content creators for standard educational tests need to carefully curate questions that assess the comprehension abilities of candidates taking the tests. However, recent work has shown that a large number of questio...
http://arxiv.org/pdf/2307.01076v1
2023-07-03T14:55:02
2023-07-03T14:55:02
{ "title": [ -0.03262329, -0.01486969, -0.020629883, -0.0044174194, -0.03186035, -0.018875122, -0.013336182, -0.009284973, -0.010353088, 0.025268555, -0.01928711, -0.00019049644, -0.030715942, -0.019317627, -0.02784729, -0.013221741, -0.012680054, ...
http://arxiv.org/abs/2307.01020v1
Estimating Post-OCR Denoising Complexity on Numerical Texts
Arthur Hemmer, Jérôme Brachat, Mickaël Coustaty, Jean-Marc Ogier
cs.CL
Post-OCR processing has significantly improved over the past few years. However, these have been primarily beneficial for texts consisting of natural, alphabetical words, as opposed to documents of numerical nature such as invoices, payslips, medical certificates, etc. To evaluate the OCR post-processing difficulty of ...
http://arxiv.org/pdf/2307.01020v1
2023-07-03T13:49:14
2023-07-03T13:49:14
{ "title": [ -0.018966675, 0.022872925, -0.011009216, -0.026733398, 0.010192871, -0.012046814, -0.0043678284, -0.029052734, 0.00097084045, 0.04510498, -0.02734375, 0.02293396, -0.0032100677, -0.009017944, -0.062042236, -0.016433716, 0.0012626648, ...
http://arxiv.org/abs/2307.01003v1
Visual Instruction Tuning with Polite Flamingo
Delong Chen, Jianfeng Liu, Wenliang Dai, Baoyuan Wang
cs.CV, cs.CL
Recent research has demonstrated that the multi-task fine-tuning of multi-modal Large Language Models (LLMs) using an assortment of annotated downstream vision-language datasets significantly enhances their performance. Yet, during this process, a side effect, which we termed as the "multi-modal alignment tax", surface...
http://arxiv.org/pdf/2307.01003v1
2023-07-03T13:37:00
2023-07-03T13:37:00
{ "title": [ -0.038269043, 0.011650085, -0.040527344, -0.0107040405, -0.050048828, -0.0069885254, -0.05734253, -0.02293396, -0.013259888, 0.02859497, -0.04940796, 0.0072631836, -0.0070762634, -0.012138367, -0.0024032593, -0.028747559, 0.008651733, ...
http://arxiv.org/abs/2307.00995v1
Towards Suicide Prevention from Bipolar Disorder with Temporal Symptom-Aware Multitask Learning
Daeun Lee, Sejung Son, Hyolim Jeon, Seungbae Kim, Jinyoung Han
cs.CL
Bipolar disorder (BD) is closely associated with an increased risk of suicide. However, while the prior work has revealed valuable insight into understanding the behavior of BD patients on social media, little attention has been paid to developing a model that can predict the future suicidality of a BD patient. Therefo...
http://arxiv.org/pdf/2307.00995v1
2023-07-03T13:18:55
2023-07-03T13:18:55
{ "title": [ 0.050994873, 0.012130737, -0.01838684, -0.074035645, 0.02293396, -0.022705078, -0.028823853, -0.016220093, -0.0107040405, 0.051086426, 0.0073242188, 0.010772705, 0.0184021, -0.051452637, -0.030349731, -0.004711151, 0.0016555786, 0.01...
http://arxiv.org/abs/2307.00963v1
Challenges in Domain-Specific Abstractive Summarization and How to Overcome them
Anum Afzal, Juraj Vladika, Daniel Braun, Florian Matthes
cs.CL, cs.AI
Large Language Models work quite well with general-purpose data and many tasks in Natural Language Processing. However, they show several limitations when used for a task such as domain-specific abstractive text summarization. This paper identifies three of those limitations as research problems in the context of abstr...
http://arxiv.org/pdf/2307.00963v1
2023-07-03T12:26:44
2023-07-03T12:26:44
{ "title": [ 0.011070251, 0.0049362183, -0.023498535, -0.08898926, -0.01473999, -0.0049209595, -0.056610107, -0.012413025, 0.0034503937, 0.05444336, -0.0102005005, 0.016311646, 0.0014591217, -0.007713318, -0.04916382, 0.021896362, -0.0051612854, ...
http://arxiv.org/abs/2307.00933v1
Data-Driven Information Extraction and Enrichment of Molecular Profiling Data for Cancer Cell Lines
Ellery Smith, Rahel Paloots, Dimitris Giagkos, Michael Baudis, Kurt Stockinger
cs.CL, cs.CE, cs.DB
With the proliferation of research means and computational methodologies, published biomedical literature is growing exponentially in numbers and volume. As a consequence, in the fields of biological, medical and clinical research, domain experts have to sift through massive amounts of scientific text to find relevant ...
http://arxiv.org/pdf/2307.00933v1
2023-07-03T11:15:42
2023-07-03T11:15:42
{ "title": [ -0.032989502, 0.009521484, -0.08001709, -0.07043457, 0.019714355, -0.013198853, -0.039154053, 0.013710022, -0.014831543, 0.015808105, -0.027069092, 0.01838684, -0.018875122, -0.0056152344, -0.023147583, 0.004043579, 0.0014877319, 0.0...
http://arxiv.org/abs/2307.00925v5
Automatic Design of Semantic Similarity Ensembles Using Grammatical Evolution
Jorge Martinez-Gil
cs.CL, cs.AI
Semantic similarity measures are widely used in natural language processing to catalyze various computer-related tasks. However, no single semantic similarity measure is the most appropriate for all tasks, and researchers often use ensemble strategies to ensure performance. This research work proposes a method for auto...
http://arxiv.org/pdf/2307.00925v5
2023-08-30T19:32:00
2023-07-03T10:53:05
{ "title": [ -0.036590576, -0.0071754456, -0.035186768, -0.019897461, 0.0036735535, 0.0063819885, -0.01574707, -0.00674057, 0.021881104, 0.042907715, -0.008552551, 0.019927979, -0.013900757, -0.007297516, -0.057525635, 0.01499176, -0.011528015, 0...
http://arxiv.org/abs/2307.00920v1
Node-weighted Graph Convolutional Network for Depression Detection in Transcribed Clinical Interviews
Sergio Burdisso, Esaú Villatoro-Tello, Srikanth Madikeri, Petr Motlicek
cs.CL, cs.AI
We propose a simple approach for weighting self-connecting edges in a Graph Convolutional Network (GCN) and show its impact on depression detection from transcribed clinical interviews. To this end, we use a GCN for modeling non-consecutive and long-distance semantics to classify the transcriptions into depressed or co...
http://arxiv.org/pdf/2307.00920v1
2023-07-03T10:44:07
2023-07-03T10:44:07
{ "title": [ 0.0075263977, -0.023376465, -0.02078247, -0.06982422, -0.018157959, -0.011680603, -0.060302734, 0.0057373047, 0.024642944, 0.021743774, -0.017044067, -0.014450073, 0.0035362244, -0.039855957, -0.04611206, -0.0066337585, -0.009216309, ...
http://arxiv.org/abs/2307.02377v2
Fraunhofer SIT at CheckThat! 2023: Tackling Classification Uncertainty Using Model Souping on the Example of Check-Worthiness Classification
Raphael Frick, Inna Vogel, Jeong-Eun Choi
cs.CL, cs.LG
This paper describes the second-placed approach developed by the Fraunhofer SIT team in the CLEF-2023 CheckThat! lab Task 1B for English. Given a text snippet from a political debate, the aim of this task is to determine whether it should be assessed for check-worthiness. Detecting check-worthy statements aims to facil...
http://arxiv.org/pdf/2307.02377v2
2023-07-27T14:43:56
2023-07-03T09:27:46
{ "title": [ 0.016143799, 0.006164551, -0.025787354, -0.014022827, -0.021057129, -0.027557373, -0.020019531, -0.02619934, 0.01902771, 0.017578125, -0.019073486, 0.011131287, 0.02720642, -0.044281006, 0.019210815, -0.00919342, -0.013748169, 0.0779...
http://arxiv.org/abs/2307.00866v2
Mining Clues from Incomplete Utterance: A Query-enhanced Network for Incomplete Utterance Rewriting
Shuzheng Si, Shuang Zeng, Baobao Chang
cs.CL, cs.AI
Incomplete utterance rewriting has recently raised wide attention. However, previous works do not consider the semantic structural information between incomplete utterance and rewritten utterance or model the semantic structure implicitly and insufficiently. To address this problem, we propose a QUEry-Enhanced Network ...
http://arxiv.org/pdf/2307.00866v2
2023-07-27T17:55:41
2023-07-03T09:08:06
{ "title": [ -0.0054359436, -0.023361206, -0.018981934, -0.10266113, -0.025619507, 0.027023315, -0.091674805, 0.02229309, -0.017745972, 0.012161255, -0.046203613, 0.02305603, 0.00034713745, -0.0022850037, -0.031463623, -0.015403748, -0.015808105, ...
http://arxiv.org/abs/2307.00862v1
UniFine: A Unified and Fine-grained Approach for Zero-shot Vision-Language Understanding
Rui Sun, Zhecan Wang, Haoxuan You, Noel Codella, Kai-Wei Chang, Shih-Fu Chang
cs.CV, cs.CL
Vision-language tasks, such as VQA, SNLI-VE, and VCR are challenging because they require the model's reasoning ability to understand the semantics of the visual world and natural language. Supervised methods working for vision-language tasks have been well-studied. However, solving these tasks in a zero-shot setting i...
http://arxiv.org/pdf/2307.00862v1
2023-07-03T09:03:12
2023-07-03T09:03:12
{ "title": [ -0.0209198, -0.01007843, -0.030426025, -0.06976318, 0.0038318634, 0.005542755, -0.048461914, 0.00712204, 0.026901245, 0.02104187, -0.018295288, -0.0025157928, -0.0038909912, -0.01638794, -0.047668457, 0.002960205, 0.012313843, 0.0181...
http://arxiv.org/abs/2307.00852v1
VOLTA: Diverse and Controllable Question-Answer Pair Generation with Variational Mutual Information Maximizing Autoencoder
Yueen Ma, Dafeng Chi, Jingjing Li, Yuzheng Zhuang, Jianye Hao, Irwin King
cs.CL
Previous question-answer pair generation methods aimed to produce fluent and meaningful question-answer pairs but tend to have poor diversity. Recent attempts addressing this issue suffer from either low model capacity or overcomplicated architecture. Furthermore, they overlooked the problem where the controllability o...
http://arxiv.org/pdf/2307.00852v1
2023-07-03T08:45:42
2023-07-03T08:45:42
{ "title": [ -0.020553589, -0.003189087, -0.040130615, -0.025390625, -0.016815186, -0.008430481, -0.035614014, -0.008605957, 0.018600464, 0.022109985, 0.0062942505, 0.021469116, -0.023620605, -0.017074585, -0.03994751, 0.013450623, -0.0032978058, ...
http://arxiv.org/abs/2307.01230v1
Large Language and Text-to-3D Models for Engineering Design Optimization
Thiago Rios, Stefan Menzel, Bernhard Sendhoff
cs.CL, cs.LG, cs.NE
The current advances in generative AI for learning large neural network models with the capability to produce essays, images, music and even 3D assets from text prompts create opportunities for a manifold of disciplines. In the present paper, we study the potential of deep text-to-3D models in the engineering domain, w...
http://arxiv.org/pdf/2307.01230v1
2023-07-03T07:54:09
2023-07-03T07:54:09
{ "title": [ -0.05505371, 0.0010528564, -0.034301758, -0.019378662, -0.004047394, -0.0020751953, -0.041503906, -0.022399902, -0.004623413, 0.02508545, -0.0013656616, 0.0013723373, -0.004966736, -0.021972656, -0.012893677, 0.0036010742, 0.006187439, ...
http://arxiv.org/abs/2307.00787v1
Evaluating Shutdown Avoidance of Language Models in Textual Scenarios
Teun van der Weij, Simon Lermen, Leon lang
cs.CL, cs.AI
Recently, there has been an increase in interest in evaluating large language models for emergent and dangerous capabilities. Importantly, agents could reason that in some scenarios their goal is better achieved if they are not turned off, which can lead to undesirable behaviors. In this paper, we investigate the poten...
http://arxiv.org/pdf/2307.00787v1
2023-07-03T07:05:59
2023-07-03T07:05:59
{ "title": [ -0.011940002, -0.014854431, -0.009048462, -0.06286621, -0.013511658, 0.014961243, -0.060028076, -0.016952515, 0.017547607, 0.04135132, -0.00198555, -0.015403748, 0.01953125, -0.01309967, -0.065979004, 0.01525116, -0.017318726, 0.0259...
http://arxiv.org/abs/2307.00782v2
ContextSpeech: Expressive and Efficient Text-to-Speech for Paragraph Reading
Yujia Xiao, Shaofei Zhang, Xi Wang, Xu Tan, Lei He, Sheng Zhao, Frank K. Soong, Tan Lee
cs.CL, cs.AI, eess.AS
While state-of-the-art Text-to-Speech systems can generate natural speech of very high quality at sentence level, they still meet great challenges in speech generation for paragraph / long-form reading. Such deficiencies are due to i) ignorance of cross-sentence contextual information, and ii) high computation and memo...
http://arxiv.org/pdf/2307.00782v2
2023-10-07T08:32:36
2023-07-03T06:55:03
{ "title": [ -0.009849548, -0.009086609, -0.03253174, -0.06866455, 0.016693115, 0.0039405823, -0.034851074, 0.0062408447, 0.006832123, 0.050048828, -0.021835327, 0.015617371, 0.026306152, -0.001291275, -0.05053711, 0.012001038, -0.012168884, 0.03...
http://arxiv.org/abs/2307.00769v1
CollabKG: A Learnable Human-Machine-Cooperative Information Extraction Toolkit for (Event) Knowledge Graph Construction
Xiang Wei, Yufeng Chen, Ning Cheng, Xingyu Cui, Jinan Xu, Wenjuan Han
cs.CL
In order to construct or extend entity-centric and event-centric knowledge graphs (KG and EKG), the information extraction (IE) annotation toolkit is essential. However, existing IE toolkits have several non-trivial problems, such as not supporting multi-tasks, not supporting automatic updates. In this work, we present...
http://arxiv.org/pdf/2307.00769v1
2023-07-03T06:18:13
2023-07-03T06:18:13
{ "title": [ -0.037963867, -0.018859863, -0.059051514, -0.07904053, -0.0010318756, 0.011329651, -0.07385254, 0.0065460205, 0.0042037964, 0.019638062, 0.011291504, 0.014076233, -0.019927979, -0.017333984, -0.04309082, 0.0011482239, -0.007949829, 0...
http://arxiv.org/abs/2307.00759v1
Multilingual Contextual Adapters To Improve Custom Word Recognition In Low-resource Languages
Devang Kulshreshtha, Saket Dingliwal, Brady Houston, Sravan Bodapati
cs.CL, cs.SD, eess.AS
Connectionist Temporal Classification (CTC) models are popular for their balance between speed and performance for Automatic Speech Recognition (ASR). However, these CTC models still struggle in other areas, such as personalization towards custom words. A recent approach explores Contextual Adapters, wherein an attenti...
http://arxiv.org/pdf/2307.00759v1
2023-07-03T05:29:38
2023-07-03T05:29:38
{ "title": [ -0.00945282, -0.0032367706, -0.024917603, -0.06414795, -0.016845703, 0.0032463074, -0.06311035, 0.0069084167, -0.0149383545, 0.06109619, -0.029373169, -0.0023536682, 0.008995056, -0.002588272, -0.0597229, -0.002046585, -0.0018844604, ...
http://arxiv.org/abs/2307.01226v2
vONTSS: vMF based semi-supervised neural topic modeling with optimal transport
Weijie Xu, Xiaoyu Jiang, Srinivasan H. Sengamedu, Francis Iannacci, Jinjin Zhao
cs.LG, cs.AI, cs.CL, cs.IT, math.IT
Recently, Neural Topic Models (NTM), inspired by variational autoencoders, have attracted a lot of research interest; however, these methods have limited applications in the real world due to the challenge of incorporating human knowledge. This work presents a semi-supervised neural topic modeling method, vONTSS, which...
http://arxiv.org/pdf/2307.01226v2
2023-09-16T16:51:17
2023-07-03T04:23:41
{ "title": [ -0.009597778, 0.002319336, -0.031677246, -0.056243896, 0.0029830933, 0.0017318726, -0.041503906, -0.014404297, 0.029586792, 0.010559082, 0.010498047, 0.027633667, -0.019973755, -0.030548096, -0.024673462, 0.0184021, -0.011886597, 0.0...
http://arxiv.org/abs/2307.00729v1
An End-to-End Multi-Module Audio Deepfake Generation System for ADD Challenge 2023
Sheng Zhao, Qilong Yuan, Yibo Duan, Zhuoyue Chen
cs.SD, cs.CL, eess.AS
The task of synthetic speech generation is to generate language content from a given text, then simulating fake human voice.The key factors that determine the effect of synthetic speech generation mainly include speed of generation, accuracy of word segmentation, naturalness of synthesized speech, etc. This paper build...
http://arxiv.org/pdf/2307.00729v1
2023-07-03T03:21:23
2023-07-03T03:21:23
{ "title": [ 0.02192688, 0.0005745888, -0.05456543, -0.094177246, 0.024856567, -0.020339966, -0.037231445, 0.023269653, -0.029724121, 0.04421997, 0.008529663, 0.014572144, 0.02015686, -0.0385437, -0.013336182, 0.021453857, -0.021331787, 0.0166168...
http://arxiv.org/abs/2307.01225v1
Interpretability and Transparency-Driven Detection and Transformation of Textual Adversarial Examples (IT-DT)
Bushra Sabir, M. Ali Babar, Sharif Abuadbba
cs.CL, cs.AI, cs.LG
Transformer-based text classifiers like BERT, Roberta, T5, and GPT-3 have shown impressive performance in NLP. However, their vulnerability to adversarial examples poses a security risk. Existing defense methods lack interpretability, making it hard to understand adversarial classifications and identify model vulnerabi...
http://arxiv.org/pdf/2307.01225v1
2023-07-03T03:17:20
2023-07-03T03:17:20
{ "title": [ 0.0062942505, 0.026870728, -0.0050849915, -0.036102295, 0.015625, 0.004261017, -0.03265381, -0.049713135, -0.030151367, 0.019729614, -0.008514404, 0.012573242, -0.0077552795, -0.02949524, -0.018814087, -0.03668213, -0.01197052, 0.039...
http://arxiv.org/abs/2307.00610v2
Fraunhofer SIT at CheckThat! 2023: Mixing Single-Modal Classifiers to Estimate the Check-Worthiness of Multi-Modal Tweets
Raphael Frick, Inna Vogel
cs.LG, cs.CL, cs.SI
The option of sharing images, videos and audio files on social media opens up new possibilities for distinguishing between false information and fake news on the Internet. Due to the vast amount of data shared every second on social media, not all data can be verified by a computer or a human expert. Here, a check-wort...
http://arxiv.org/pdf/2307.00610v2
2023-07-27T14:54:13
2023-07-02T16:35:54
{ "title": [ 0.016433716, -0.0014791489, 0.007633209, 0.00039243698, -0.029296875, 0.0014429092, -0.016998291, -0.026931763, 0.0032978058, 0.050231934, -0.05508423, 0.043823242, 0.052124023, -0.012367249, 0.030136108, -0.0058403015, -0.013084412, ...
http://arxiv.org/abs/2307.00589v2
MedCPT: Contrastive Pre-trained Transformers with Large-scale PubMed Search Logs for Zero-shot Biomedical Information Retrieval
Qiao Jin, Won Kim, Qingyu Chen, Donald C. Comeau, Lana Yeganova, W. John Wilbur, Zhiyong Lu
cs.IR, cs.AI, cs.CL, q-bio.QM
Information retrieval (IR) is essential in biomedical knowledge acquisition and clinical decision support. While recent progress has shown that language model encoders perform better semantic retrieval, training such models requires abundant query-article annotations that are difficult to obtain in biomedicine. As a re...
http://arxiv.org/pdf/2307.00589v2
2023-10-04T01:43:15
2023-07-02T15:11:59
{ "title": [ 0.04019165, -0.005470276, -0.012741089, -0.072387695, -0.0070114136, -0.0037345886, -0.023208618, -0.017959595, 0.015045166, 0.04336548, -0.018463135, -0.00017404556, 0.008811951, -0.020080566, -0.04309082, -0.0013332367, -0.0030441284, ...
http://arxiv.org/abs/2307.00569v1
SSP: Self-Supervised Post-training for Conversational Search
Quan Tu, Shen Gao, Xiaolong Wu, Zhao Cao, Ji-Rong Wen, Rui Yan
cs.CL
Conversational search has been regarded as the next-generation search paradigm. Constrained by data scarcity, most existing methods distill the well-trained ad-hoc retriever to the conversational retriever. However, these methods, which usually initialize parameters by query reformulation to discover contextualized dep...
http://arxiv.org/pdf/2307.00569v1
2023-07-02T13:36:36
2023-07-02T13:36:36
{ "title": [ 0.021514893, 0.0039863586, -0.007297516, -0.055389404, 0.004524231, -0.0056877136, -0.03353882, -0.019805908, 0.0115356445, 0.02519226, -0.0037403107, 0.018875122, -0.013626099, -0.01550293, -0.037963867, 0.01071167, -0.021636963, 0....
http://arxiv.org/abs/2307.00526v1
TensorGPT: Efficient Compression of the Embedding Layer in LLMs based on the Tensor-Train Decomposition
Mingxue Xu, Yao Lei Xu, Danilo P. Mandic
cs.CL, cs.LG, cs.NA, cs.NE, math.NA
High-dimensional token embeddings underpin Large Language Models (LLMs), as they can capture subtle semantic information and significantly enhance the modelling of complex language patterns. However, the associated high dimensionality also introduces considerable model parameters, and a prohibitively high model storage...
http://arxiv.org/pdf/2307.00526v1
2023-07-02T09:33:09
2023-07-02T09:33:09
{ "title": [ -0.0345459, -0.01007843, -0.04550171, -0.0055618286, -0.030899048, -0.007797241, -0.034057617, -0.00944519, 0.0061035156, 0.014289856, 0.0034656525, 0.015777588, 0.008415222, -0.058685303, 0.0021839142, 0.008598328, 0.012557983, -0.0...
http://arxiv.org/abs/2307.00524v1
Large Language Models Enable Few-Shot Clustering
Vijay Viswanathan, Kiril Gashteovski, Carolin Lawrence, Tongshuang Wu, Graham Neubig
cs.CL
Unlike traditional unsupervised clustering, semi-supervised clustering allows users to provide meaningful structure to the data, which helps the clustering algorithm to match the user's intent. Existing approaches to semi-supervised clustering require a significant amount of feedback from an expert to improve the clust...
http://arxiv.org/pdf/2307.00524v1
2023-07-02T09:17:11
2023-07-02T09:17:11
{ "title": [ 0.0048942566, -0.024978638, -0.025680542, -0.03881836, -0.052764893, -0.013389587, -0.04748535, 0.014518738, -0.009513855, 0.029464722, -0.024841309, 0.0043411255, 0.007080078, -0.030303955, -0.00090169907, 0.00030255318, 0.022460938, ...
http://arxiv.org/abs/2307.00509v1
HeGeL: A Novel Dataset for Geo-Location from Hebrew Text
Tzuf Paz-Argaman, Tal Bauman, Itai Mondshine, Itzhak Omer, Sagi Dalyot, Reut Tsarfaty
cs.CL, cs.AI, cs.IR, cs.LG
The task of textual geolocation - retrieving the coordinates of a place based on a free-form language description - calls for not only grounding but also natural language understanding and geospatial reasoning. Even though there are quite a few datasets in English used for geolocation, they are currently based on open-...
http://arxiv.org/pdf/2307.00509v1
2023-07-02T08:09:10
2023-07-02T08:09:10
{ "title": [ 0.0058898926, 0.0054130554, -0.028625488, -0.06896973, 0.02935791, 0.0049552917, -0.029632568, -0.024780273, 0.013076782, 0.036315918, -0.02015686, 0.016723633, 0.046539307, -0.018218994, -0.009796143, 0.035217285, 0.0033226013, 0.01...
http://arxiv.org/abs/2307.00470v4
PatternGPT :A Pattern-Driven Framework for Large Language Model Text Generation
Le Xiao, Xin Shan
cs.CL, cs.AI
Large language models(LLMS)have shown excellent text generation capabilities, capable of generating fluent human-like responses for many downstream tasks. However, applying large language models to real-world critical tasks remains challenging due to their susceptibility to hallucinations and inability to directly use ...
http://arxiv.org/pdf/2307.00470v4
2023-07-20T03:03:25
2023-07-02T04:32:41
{ "title": [ -0.02545166, -0.01953125, -0.039733887, -0.08001709, 0.00434494, 0.024124146, -0.077819824, 0.001250267, 0.003227234, 0.039001465, 0.004890442, 0.015655518, 0.015701294, -0.010696411, -0.045440674, 0.02659607, -0.018325806, 0.0395813...
http://arxiv.org/abs/2307.00461v1
Conformer LLMs -- Convolution Augmented Large Language Models
Prateek Verma
cs.CL, cs.AI, cs.LG, cs.MM, cs.SD
This work builds together two popular blocks of neural architecture, namely convolutional layers and Transformers, for large language models (LLMs). Non-causal conformers are used ubiquitously in automatic speech recognition. This work aims to adapt these architectures in a causal setup for training LLMs. Transformers ...
http://arxiv.org/pdf/2307.00461v1
2023-07-02T03:05:41
2023-07-02T03:05:41
{ "title": [ -0.029220581, -0.01537323, -0.07122803, -0.08648682, -0.016082764, -0.00806427, -0.09436035, 0.024368286, -0.011642456, 0.035186768, -0.030899048, -0.0027313232, 0.022750854, -0.031021118, -0.008934021, 0.00027489662, -0.0058135986, ...
http://arxiv.org/abs/2307.00457v2
GenRec: Large Language Model for Generative Recommendation
Jianchao Ji, Zelong Li, Shuyuan Xu, Wenyue Hua, Yingqiang Ge, Juntao Tan, Yongfeng Zhang
cs.IR, cs.AI, cs.CL, cs.LG
In recent years, large language models (LLM) have emerged as powerful tools for diverse natural language processing tasks. However, their potential for recommender systems under the generative recommendation paradigm remains relatively unexplored. This paper presents an innovative approach to recommendation systems usi...
http://arxiv.org/pdf/2307.00457v2
2023-07-04T20:04:58
2023-07-02T02:37:07
{ "title": [ -0.010032654, -0.014389038, -0.037506104, -0.05630493, -0.0034332275, 0.0015821457, -0.037017822, -0.00027561188, -0.003063202, 0.02999878, -0.0066452026, 0.022903442, -0.007255554, -0.016067505, -0.023391724, 0.008171082, -0.017745972, ...
http://arxiv.org/abs/2307.00456v1
Make Text Unlearnable: Exploiting Effective Patterns to Protect Personal Data
Xinzhe Li, Ming Liu, Shang Gao
cs.CL
This paper addresses the ethical concerns arising from the use of unauthorized public data in deep learning models and proposes a novel solution. Specifically, building on the work of Huang et al. (2021), we extend their bi-level optimization approach to generate unlearnable text using a gradient-based search technique...
http://arxiv.org/pdf/2307.00456v1
2023-07-02T02:34:57
2023-07-02T02:34:57
{ "title": [ 0.013473511, 0.0074691772, -0.051483154, -0.07409668, 0.033569336, -0.016143799, -0.017593384, 0.00060606, -0.036010742, 0.030761719, 0.012992859, 0.016159058, 0.00995636, -0.030212402, -0.016143799, 0.0011548996, -0.0033416748, 0.05...
http://arxiv.org/abs/2307.00453v1
Don't Stop Self-Supervision: Accent Adaptation of Speech Representations via Residual Adapters
Anshu Bhatia, Sanchit Sinha, Saket Dingliwal, Karthik Gopalakrishnan, Sravan Bodapati, Katrin Kirchhoff
cs.CL, cs.SD, eess.AS
Speech representations learned in a self-supervised fashion from massive unlabeled speech corpora have been adapted successfully toward several downstream tasks. However, such representations may be skewed toward canonical data characteristics of such corpora and perform poorly on atypical, non-native accented speaker ...
http://arxiv.org/pdf/2307.00453v1
2023-07-02T02:21:29
2023-07-02T02:21:29
{ "title": [ 0.021896362, -0.0008068085, -0.0059165955, -0.042144775, -0.016555786, -0.006942749, -0.027374268, -0.003932953, -0.0075569153, 0.049835205, 0.009719849, -0.008735657, 0.037353516, -0.034576416, -0.03668213, 0.022644043, 0.020858765, ...
http://arxiv.org/abs/2307.00449v1
A Dual-Stream Recurrence-Attention Network with Global-Local Awareness for Emotion Recognition in Textual Dialogue
Jiang Li, Xiaoping Wang, Zhigang Zeng
cs.CL
In real-world dialogue systems, the ability to understand the user's emotions and interact anthropomorphically is of great significance. Emotion Recognition in Conversation (ERC) is one of the key ways to accomplish this goal and has attracted growing attention. How to model the context in a conversation is a central a...
http://arxiv.org/pdf/2307.00449v1
2023-07-02T01:25:47
2023-07-02T01:25:47
{ "title": [ 0.0073890686, 0.009757996, -0.025527954, -0.06365967, -0.0042266846, -0.0024929047, -0.04800415, 0.012275696, 0.049682617, 0.026794434, -0.0062332153, -0.021316528, 0.025161743, -0.028152466, -0.060424805, 0.009307861, -0.01259613, 0...
http://arxiv.org/abs/2307.04761v1
Understanding Counterspeech for Online Harm Mitigation
Yi-Ling Chung, Gavin Abercrombie, Florence Enock, Jonathan Bright, Verena Rieser
cs.CL, cs.AI, cs.CY
Counterspeech offers direct rebuttals to hateful speech by challenging perpetrators of hate and showing support to targets of abuse. It provides a promising alternative to more contentious measures, such as content moderation and deplatforming, by contributing a greater amount of positive online speech rather than atte...
http://arxiv.org/pdf/2307.04761v1
2023-07-01T20:54:01
2023-07-01T20:54:01
{ "title": [ 0.034210205, 0.0065612793, 0.010871887, -0.050628662, -0.003616333, -0.0059051514, -0.017028809, -0.02835083, -0.020690918, 0.05593872, 0.010177612, 0.0015573502, 0.013557434, -0.026351929, -0.047698975, -0.0033512115, -0.009849548, ...
http://arxiv.org/abs/2307.00382v1
Low-Resource Cross-Lingual Adaptive Training for Nigerian Pidgin
Pin-Jie Lin, Muhammed Saeed, Ernie Chang, Merel Scholman
cs.CL
Developing effective spoken language processing systems for low-resource languages poses several challenges due to the lack of parallel data and limited resources for fine-tuning models. In this work, we target on improving upon both text classification and translation of Nigerian Pidgin (Naija) by collecting a large-s...
http://arxiv.org/pdf/2307.00382v1
2023-07-01T16:47:36
2023-07-01T16:47:36
{ "title": [ -0.0473938, -0.027786255, -0.001253128, -0.020111084, -0.04095459, 0.0124435425, -0.051727295, -0.0025539398, -0.010185242, 0.036834717, 0.0031604767, 0.00094795227, 0.009269714, -0.029510498, -0.02619934, 0.00018203259, -0.0073013306, ...
http://arxiv.org/abs/2307.00381v1
Effective Matching of Patients to Clinical Trials using Entity Extraction and Neural Re-ranking
Wojciech Kusa, Óscar E. Mendoza, Petr Knoth, Gabriella Pasi, Allan Hanbury
cs.IR, cs.CL
Clinical trials (CTs) often fail due to inadequate patient recruitment. This paper tackles the challenges of CT retrieval by presenting an approach that addresses the patient-to-trials paradigm. Our approach involves two key components in a pipeline-based model: (i) a data enrichment technique for enhancing both querie...
http://arxiv.org/pdf/2307.00381v1
2023-07-01T16:42:39
2023-07-01T16:42:39
{ "title": [ 0.00819397, 0.036315918, -0.0357666, -0.037261963, 0.036224365, -0.0052948, -0.015541077, -0.05303955, -0.012634277, 0.046051025, 0.0023269653, 0.014160156, -0.0020122528, -0.008430481, -0.03845215, 0.005355835, -0.027923584, 0.02706...
http://arxiv.org/abs/2307.00374v1
Revisiting Sample Size Determination in Natural Language Understanding
Ernie Chang, Muhammad Hassan Rashid, Pin-Jie Lin, Changsheng Zhao, Vera Demberg, Yangyang Shi, Vikas Chandra
cs.CL
Knowing exactly how many data points need to be labeled to achieve a certain model performance is a hugely beneficial step towards reducing the overall budgets for annotation. It pertains to both active learning and traditional data annotation, and is particularly beneficial for low resource scenarios. Nevertheless, it...
http://arxiv.org/pdf/2307.00374v1
2023-07-01T16:08:52
2023-07-01T16:08:52
{ "title": [ -0.046844482, -0.030136108, -0.039733887, -0.027511597, -0.02798462, -0.006126404, -0.043151855, -0.013412476, -0.028961182, 0.0014820099, -0.06665039, 0.010009766, -0.035888672, -0.0044784546, -0.03289795, -0.03768921, -0.00048446655, ...
http://arxiv.org/abs/2307.00370v2
Improving Text Matching in E-Commerce Search with A Rationalizable, Intervenable and Fast Entity-Based Relevance Model
Jiong Cai, Yong Jiang, Yue Zhang, Chengyue Jiang, Ke Yu, Jianhui Ji, Rong Xiao, Haihong Tang, Tao Wang, Zhongqiang Huang, Pengjun Xie, Fei Huang, Kewei Tu
cs.IR, cs.CL
Discovering the intended items of user queries from a massive repository of items is one of the main goals of an e-commerce search system. Relevance prediction is essential to the search system since it helps improve performance. When online serving a relevance model, the model is required to perform fast and accurate ...
http://arxiv.org/pdf/2307.00370v2
2023-07-19T06:55:04
2023-07-01T15:44:53
{ "title": [ -0.017288208, 0.0079574585, -0.04257202, -0.042633057, 0.019332886, -0.005214691, -0.023117065, -0.0231781, 0.0074806213, 0.04714966, -0.012283325, 0.008003235, 0.029754639, -0.0008773804, -0.048980713, 0.004699707, -0.012832642, 0.0...
http://arxiv.org/abs/2307.07518v1
CephGPT-4: An Interactive Multimodal Cephalometric Measurement and Diagnostic System with Visual Large Language Model
Lei Ma, Jincong Han, Zhaoxin Wang, Dian Zhang
cs.AI, cs.CL, cs.CV, eess.IV
Large-scale multimodal language models (LMMs) have achieved remarkable success in general domains. However, the exploration of diagnostic language models based on multimodal cephalometric medical data remains limited. In this paper, we propose a novel multimodal cephalometric analysis and diagnostic dialogue model. Fir...
http://arxiv.org/pdf/2307.07518v1
2023-07-01T15:41:12
2023-07-01T15:41:12
{ "title": [ -0.060424805, -0.012039185, -0.057037354, -0.036956787, -0.017364502, 0.000012874603, -0.05697632, -0.0032920837, -0.028808594, 0.01890564, 0.015022278, 0.011955261, -0.002828598, -0.056243896, -0.006942749, 0.008934021, -0.0028247833, ...
http://arxiv.org/abs/2307.00360v2
BatGPT: A Bidirectional Autoregessive Talker from Generative Pre-trained Transformer
Zuchao Li, Shitou Zhang, Hai Zhao, Yifei Yang, Dongjie Yang
cs.CL
BatGPT is a large-scale language model designed and trained jointly by Wuhan University and Shanghai Jiao Tong University. It is capable of generating highly natural and fluent text in response to various types of input, including text prompts, images, and audio. In the modeling level, we employ a bidirectional autoreg...
http://arxiv.org/pdf/2307.00360v2
2023-08-15T13:59:42
2023-07-01T15:10:01
{ "title": [ 0.008743286, 0.0070266724, -0.04156494, -0.07910156, 0.006038666, -0.014541626, -0.057159424, -0.0053253174, -0.01121521, 0.038238525, -0.009094238, 0.014587402, 0.025466919, -0.022903442, -0.016494751, 0.005420685, -0.0070343018, 0....
http://arxiv.org/abs/2307.00342v1
Improving Multitask Retrieval by Promoting Task Specialization
Wenzheng Zhang, Chenyan Xiong, Karl Stratos, Arnold Overwijk
cs.CL, cs.IR
In multitask retrieval, a single retriever is trained to retrieve relevant contexts for multiple tasks. Despite its practical appeal, naive multitask retrieval lags behind task-specific retrieval in which a separate retriever is trained for each task. We show that it is possible to train a multitask retriever that outp...
http://arxiv.org/pdf/2307.00342v1
2023-07-01T13:45:15
2023-07-01T13:45:15
{ "title": [ 0.0093688965, 0.005718231, -0.04171753, -0.04626465, 0.015991211, -0.028045654, -0.0043258667, -0.0023021698, -0.018478394, 0.057037354, 0.00969696, 0.027282715, -0.0035705566, -0.011749268, -0.046539307, 0.020111084, -0.008094788, 0...
http://arxiv.org/abs/2307.00335v1
Single Sequence Prediction over Reasoning Graphs for Multi-hop QA
Gowtham Ramesh, Makesh Sreedhar, Junjie Hu
cs.CL, cs.LG
Recent generative approaches for multi-hop question answering (QA) utilize the fusion-in-decoder method~\cite{izacard-grave-2021-leveraging} to generate a single sequence output which includes both a final answer and a reasoning path taken to arrive at that answer, such as passage titles and key facts from those passag...
http://arxiv.org/pdf/2307.00335v1
2023-07-01T13:15:09
2023-07-01T13:15:09
{ "title": [ -0.018753052, -0.030136108, -0.044647217, -0.08911133, -0.03643799, 0.0058403015, -0.066711426, 0.05734253, 0.004337311, 0.04107666, -0.011329651, -0.014427185, 0.025619507, -0.031677246, -0.030166626, 0.0026245117, -0.0072669983, 0....
http://arxiv.org/abs/2307.00279v1
Let Me Teach You: Pedagogical Foundations of Feedback for Language Models
Beatriz Borges, Niket Tandon, Tanja Käser, Antoine Bosselut
cs.CL
Natural Language Feedback (NLF) is an increasingly popular avenue to align Large Language Models (LLMs) to human preferences. Despite the richness and diversity of the information it can convey, NLF is often hand-designed and arbitrary. In a different world, research in pedagogy has long established several effective f...
http://arxiv.org/pdf/2307.00279v1
2023-07-01T09:18:24
2023-07-01T09:18:24
{ "title": [ -0.026916504, -0.017700195, -0.0076675415, -0.047668457, -0.01524353, 0.008598328, -0.0524292, -0.011375427, 0.012649536, 0.02444458, -0.029891968, 0.002374649, 0.002216339, -0.016067505, -0.04269409, -0.0029335022, -0.008644104, 0.0...
http://arxiv.org/abs/2307.00266v1
Hierarchical Pretraining for Biomedical Term Embeddings
Bryan Cai, Sihang Zeng, Yucong Lin, Zheng Yuan, Doudou Zhou, Lu Tian
cs.CL, cs.AI
Electronic health records (EHR) contain narrative notes that provide extensive details on the medical condition and management of patients. Natural language processing (NLP) of clinical notes can use observed frequencies of clinical terms as predictive features for downstream applications such as clinical decision maki...
http://arxiv.org/pdf/2307.00266v1
2023-07-01T08:16:00
2023-07-01T08:16:00
{ "title": [ 0.0052757263, 0.010513306, -0.045928955, -0.08123779, 0.014144897, 0.00064086914, -0.05114746, -0.0022659302, 0.013244629, 0.03878784, 0.0017328262, 0.021591187, 0.012641907, -0.026550293, -0.032440186, 0.027435303, -0.00028204918, 0...
http://arxiv.org/abs/2307.00259v2
InstructEval: Systematic Evaluation of Instruction Selection Methods
Anirudh Ajith, Chris Pan, Mengzhou Xia, Ameet Deshpande, Karthik Narasimhan
cs.CL, cs.AI
In-context learning (ICL) performs tasks by prompting a large language model (LLM) using an instruction and a small set of annotated examples called demonstrations. Recent work has shown that precise details of the inputs used in the ICL prompt significantly impact performance, which has incentivized instruction select...
http://arxiv.org/pdf/2307.00259v2
2023-07-16T10:14:44
2023-07-01T07:45:38
{ "title": [ -0.03466797, -0.001200676, -0.028015137, -0.012016296, -0.0077056885, -0.021011353, -0.008529663, -0.03378296, -0.02609253, 0.014640808, -0.025222778, 0.018753052, -0.026626587, -0.031555176, -0.015686035, -0.008392334, -0.014167786, ...
http://arxiv.org/abs/2307.01216v1
Discovering Patterns of Definitions and Methods from Scientific Documents
Yutian Sun, Hai Zhuge
cs.CL
The difficulties of automatic extraction of definitions and methods from scientific documents lie in two aspects: (1) the complexity and diversity of natural language texts, which requests an analysis method to support the discovery of pattern; and, (2) a complete definition or method represented by a scientific paper ...
http://arxiv.org/pdf/2307.01216v1
2023-07-01T05:08:44
2023-07-01T05:08:44
{ "title": [ -0.023025513, 0.01424408, -0.060668945, -0.034729004, 0.014801025, -0.021652222, -0.016998291, -0.017608643, -0.01687622, 0.034942627, -0.0066566467, 0.005710602, -0.008880615, -0.023956299, -0.03262329, 0.013023376, -0.0020370483, 0...
http://arxiv.org/abs/2307.00209v2
Image Matters: A New Dataset and Empirical Study for Multimodal Hyperbole Detection
Huixuan Zhang, Xiaojun Wan
cs.CV, cs.AI, cs.CL
Hyperbole, or exaggeration, is a common linguistic phenomenon. The detection of hyperbole is an important part of understanding human expression. There have been several studies on hyperbole detection, but most of which focus on text modality only. However, with the development of social media, people can create hyperb...
http://arxiv.org/pdf/2307.00209v2
2023-07-06T11:19:22
2023-07-01T03:23:56
{ "title": [ 0.020996094, 0.022201538, -0.037841797, -0.04437256, 0.021820068, -0.022628784, -0.023651123, -0.025177002, 0.009597778, 0.036987305, -0.0057144165, -0.009521484, 0.015167236, -0.018234253, -0.0619812, 0.007888794, -0.016921997, 0.05...
http://arxiv.org/abs/2307.01214v1
Automatic Counterfactual Augmentation for Robust Text Classification Based on Word-Group Search
Rui Song, Fausto Giunchiglia, Yingji Li, Hao Xu
cs.CL, cs.AI
Despite large-scale pre-trained language models have achieved striking results for text classificaion, recent work has raised concerns about the challenge of shortcut learning. In general, a keyword is regarded as a shortcut if it creates a superficial association with the label, resulting in a false prediction. Conver...
http://arxiv.org/pdf/2307.01214v1
2023-07-01T02:26:34
2023-07-01T02:26:34
{ "title": [ 0.021835327, 0.021118164, -0.034729004, -0.012901306, 0.017929077, -0.02508545, -0.007381439, -0.045837402, 0.04244995, 0.03652954, -0.058441162, 0.03881836, 0.004547119, 0.009346008, -0.07122803, -0.0061187744, -0.02973938, 0.045989...
http://arxiv.org/abs/2307.00186v1
How far is Language Model from 100% Few-shot Named Entity Recognition in Medical Domain
Mingchen Li, Rui Zhang
cs.CL
Recent advancements in language models (LMs) have led to the emergence of powerful models such as Small LMs (e.g., T5) and Large LMs (e.g., GPT-4). These models have demonstrated exceptional capabilities across a wide range of tasks, such as name entity recognition (NER) in the general domain. (We define SLMs as pre-tr...
http://arxiv.org/pdf/2307.00186v1
2023-07-01T01:18:09
2023-07-01T01:18:09
{ "title": [ 0.0018510818, -0.0021038055, -0.045196533, -0.049682617, 0.002943039, -0.019851685, -0.024810791, -0.0018138885, -0.028015137, 0.020950317, -0.026733398, 0.0011024475, 0.010772705, -0.0211792, -0.02897644, -0.009025574, 0.005756378, ...
http://arxiv.org/abs/2307.00184v3
Personality Traits in Large Language Models
Greg Serapio-García, Mustafa Safdari, Clément Crepy, Luning Sun, Stephen Fitz, Peter Romero, Marwa Abdulhai, Aleksandra Faust, Maja Matarić
cs.CL, cs.AI, cs.CY, cs.HC, 68T35, I.2.7
The advent of large language models (LLMs) has revolutionized natural language processing, enabling the generation of coherent and contextually relevant human-like text. As LLMs increasingly power conversational agents used by the general public world-wide, the synthetic personality embedded in these models, by virtue ...
http://arxiv.org/pdf/2307.00184v3
2023-09-21T21:10:56
2023-07-01T00:58:51
{ "title": [ -0.032836914, -0.007751465, -0.03955078, -0.05114746, 0.0002708435, -0.0035305023, -0.030410767, 0.010528564, 0.004173279, 0.034942627, -0.0055007935, -0.004940033, 0.020324707, -0.025466919, -0.033416748, 0.031463623, -0.016952515, ...
http://arxiv.org/abs/2307.00175v1
Still No Lie Detector for Language Models: Probing Empirical and Conceptual Roadblocks
B. A. Levinstein, Daniel A. Herrmann
cs.CL, cs.LG
We consider the questions of whether or not large language models (LLMs) have beliefs, and, if they do, how we might measure them. First, we evaluate two existing approaches, one due to Azaria and Mitchell (2023) and the other to Burns et al. (2022). We provide empirical results that show that these methods fail to gen...
http://arxiv.org/pdf/2307.00175v1
2023-06-30T23:44:51
2023-06-30T23:44:51
{ "title": [ -0.036621094, -0.015014648, -0.03161621, -0.034454346, -0.009780884, -0.0011796951, -0.037994385, -0.016448975, -0.0053977966, 0.024627686, -0.018203735, -0.0023441315, 0.015792847, -0.014678955, -0.04812622, 0.011184692, -0.024093628, ...
http://arxiv.org/abs/2307.00171v1
The Integer Linear Programming Inference Cookbook
Vivek Srikumar, Dan Roth
cs.AI, cs.CL, cs.LG
Over the years, integer linear programs have been employed to model inference in many natural language processing problems. This survey is meant to guide the reader through the process of framing a new inference problem as an instance of an integer linear program and is structured as a collection of recipes. At the end...
http://arxiv.org/pdf/2307.00171v1
2023-06-30T23:33:11
2023-06-30T23:33:11
{ "title": [ -0.048187256, -0.011131287, -0.040283203, -0.0637207, 0.0072784424, 0.010414124, -0.043304443, -0.0082473755, -0.01826477, 0.004638672, -0.023452759, 0.030456543, -0.038330078, -0.032043457, -0.010719299, -0.007205963, -0.00017738342, ...
http://arxiv.org/abs/2307.00165v1
Counterfactual Collaborative Reasoning
Jianchao Ji, Zelong Li, Shuyuan Xu, Max Xiong, Juntao Tan, Yingqiang Ge, Hao Wang, Yongfeng Zhang
cs.IR, cs.AI, cs.CL, cs.LG
Causal reasoning and logical reasoning are two important types of reasoning abilities for human intelligence. However, their relationship has not been extensively explored under machine intelligence context. In this paper, we explore how the two reasoning abilities can be jointly modeled to enhance both accuracy and ex...
http://arxiv.org/pdf/2307.00165v1
2023-06-30T23:01:10
2023-06-30T23:01:10
{ "title": [ -0.0013742447, -0.00030612946, -0.04510498, -0.03427124, -0.0049972534, -0.016525269, -0.021591187, -0.017654419, 0.01966858, 0.014038086, -0.030258179, -0.0054626465, -0.0385437, -0.025863647, -0.022415161, -0.008140564, -0.007385254, ...
http://arxiv.org/abs/2307.00162v1
What do self-supervised speech models know about words?
Ankita Pasad, Chung-Ming Chien, Shane Settle, Karen Livescu
cs.CL, cs.LG, eess.AS
Many self-supervised speech models (S3Ms) have been introduced over the last few years, producing performance and data efficiency improvements for a variety of speech tasks. Evidence is emerging that different S3Ms encode linguistic information in different layers, and also that some S3Ms appear to learn phone-like sub...
http://arxiv.org/pdf/2307.00162v1
2023-06-30T22:36:41
2023-06-30T22:36:41
{ "title": [ -0.02519226, 0.02230835, -0.037109375, -0.011329651, -0.04522705, -0.011993408, -0.05819702, -0.00818634, -0.02407837, 0.022705078, -0.04071045, -0.009437561, 0.0058517456, -0.028305054, 0.0064315796, -0.004055023, 0.012329102, 0.055...
http://arxiv.org/abs/2307.00135v1
SMILE: Evaluation and Domain Adaptation for Social Media Language Understanding
Vasilisa Bashlovkina, Riley Matthews, Zhaobin Kuang, Simon Baumgartner, Michael Bendersky
cs.CL
We study the ability of transformer-based language models (LMs) to understand social media language. Social media (SM) language is distinct from standard written language, yet existing benchmarks fall short of capturing LM performance in this socially, economically, and politically important domain. We quantify the deg...
http://arxiv.org/pdf/2307.00135v1
2023-06-30T21:04:59
2023-06-30T21:04:59
{ "title": [ -0.018341064, -0.004169464, -0.02357483, -0.04840088, 0.0014362335, 0.00057554245, -0.031951904, -0.0036849976, -0.005004883, 0.008384705, -0.020553589, 0.0126953125, -0.011146545, -0.0059051514, -0.031951904, 0.0025901794, -0.014694214,...
http://arxiv.org/abs/2307.00132v1
iMETRE: Incorporating Markers of Entity Types for Relation Extraction
N Harsha Vardhan, Manav Chaudhary
cs.CL
Sentence-level relation extraction (RE) aims to identify the relationship between 2 entities given a contextual sentence. While there have been many attempts to solve this problem, the current solutions have a lot of room to improve. In this paper, we approach the task of relationship extraction in the financial datase...
http://arxiv.org/pdf/2307.00132v1
2023-06-30T20:54:41
2023-06-30T20:54:41
{ "title": [ -0.0044021606, -0.0065841675, -0.054229736, -0.08319092, 0.028656006, 0.0061531067, -0.03100586, -0.0014953613, 0.020935059, 0.009063721, -0.007587433, 0.007904053, 0.012229919, -0.016708374, -0.029586792, 0.009803772, -0.027496338, ...
http://arxiv.org/abs/2307.00130v1
Information Extraction in Domain and Generic Documents: Findings from Heuristic-based and Data-driven Approaches
Shiyu Yuan, Carlo Lipizzi
cs.CL
Information extraction (IE) plays very important role in natural language processing (NLP) and is fundamental to many NLP applications that used to extract structured information from unstructured text data. Heuristic-based searching and data-driven learning are two main stream implementation approaches. However, no mu...
http://arxiv.org/pdf/2307.00130v1
2023-06-30T20:43:27
2023-06-30T20:43:27
{ "title": [ -0.03881836, -0.008255005, -0.0637207, -0.050323486, 0.007381439, -0.011199951, -0.036071777, -0.004432678, 0.0022735596, 0.024261475, -0.012580872, 0.0119018555, -0.0007147789, -0.011375427, -0.034362793, 0.0002872944, -0.015151978, ...
http://arxiv.org/abs/2307.00119v1
Meta-training with Demonstration Retrieval for Efficient Few-shot Learning
Aaron Mueller, Kanika Narang, Lambert Mathias, Qifan Wang, Hamed Firooz
cs.CL
Large language models show impressive results on few-shot NLP tasks. However, these models are memory and computation-intensive. Meta-training allows one to leverage smaller models for few-shot generalization in a domain-general and task-agnostic manner; however, these methods alone results in models that may not have ...
http://arxiv.org/pdf/2307.00119v1
2023-06-30T20:16:22
2023-06-30T20:16:22
{ "title": [ 0.014732361, -0.016204834, -0.018707275, -0.037017822, -0.0021076202, -0.0018577576, -0.014221191, -0.020431519, 0.015274048, 0.02406311, -0.023269653, 0.00945282, -0.017288208, -0.021911621, -0.039398193, -0.010696411, 0.0077323914, ...
http://arxiv.org/abs/2307.00108v1
Ticket-BERT: Labeling Incident Management Tickets with Language Models
Zhexiong Liu, Cris Benge, Siduo Jiang
cs.CL, cs.AI, cs.LG
An essential aspect of prioritizing incident tickets for resolution is efficiently labeling tickets with fine-grained categories. However, ticket data is often complex and poses several unique challenges for modern machine learning methods: (1) tickets are created and updated either by machines with pre-defined algorit...
http://arxiv.org/pdf/2307.00108v1
2023-06-30T19:48:25
2023-06-30T19:48:25
{ "title": [ -0.000056028366, -0.01486969, -0.040130615, -0.087768555, -0.0029792786, -0.0011024475, -0.0597229, 0.017715454, -0.033813477, 0.023468018, -0.024887085, 0.017547607, 0.015052795, -0.029907227, 0.010047913, -0.01473999, -0.014640808, ...
http://arxiv.org/abs/2307.00101v1
Queer People are People First: Deconstructing Sexual Identity Stereotypes in Large Language Models
Harnoor Dhingra, Preetiha Jayashanker, Sayali Moghe, Emma Strubell
cs.CL, cs.AI
Large Language Models (LLMs) are trained primarily on minimally processed web text, which exhibits the same wide range of social biases held by the humans who created that content. Consequently, text generated by LLMs can inadvertently perpetuate stereotypes towards marginalized groups, like the LGBTQIA+ community. In ...
http://arxiv.org/pdf/2307.00101v1
2023-06-30T19:39:01
2023-06-30T19:39:01
{ "title": [ -0.019485474, 0.00756073, -0.046081543, -0.046722412, 0.013626099, -0.019485474, -0.032684326, -0.0018453598, 0.0012722015, 0.047180176, -0.016967773, 0.0024776459, 0.026443481, -0.013710022, -0.04260254, 0.01109314, -0.026947021, 0....
http://arxiv.org/abs/2306.17842v3
SPAE: Semantic Pyramid AutoEncoder for Multimodal Generation with Frozen LLMs
Lijun Yu, Yong Cheng, Zhiruo Wang, Vivek Kumar, Wolfgang Macherey, Yanping Huang, David A. Ross, Irfan Essa, Yonatan Bisk, Ming-Hsuan Yang, Kevin Murphy, Alexander G. Hauptmann, Lu Jiang
cs.CV, cs.CL, cs.MM
In this work, we introduce Semantic Pyramid AutoEncoder (SPAE) for enabling frozen LLMs to perform both understanding and generation tasks involving non-linguistic modalities such as images or videos. SPAE converts between raw pixels and interpretable lexical tokens (or words) extracted from the LLM's vocabulary. The r...
http://arxiv.org/pdf/2306.17842v3
2023-10-28T18:09:46
2023-06-30T17:59:07
{ "title": [ -0.032196045, -0.0049552917, -0.055786133, -0.05581665, -0.026184082, -0.001121521, -0.068725586, 0.0057029724, -0.0041389465, 0.04107666, -0.015625, 0.02859497, 0.016662598, -0.0256958, -0.0023899078, 0.011375427, -0.0030937195, 0.0...
http://arxiv.org/abs/2306.17840v2
Statler: State-Maintaining Language Models for Embodied Reasoning
Takuma Yoneda, Jiading Fang, Peng Li, Huanyu Zhang, Tianchong Jiang, Shengjie Lin, Ben Picker, David Yunis, Hongyuan Mei, Matthew R. Walter
cs.RO, cs.CL
Large language models (LLMs) provide a promising tool that enable robots to perform complex robot reasoning tasks. However, the limited context window of contemporary LLMs makes reasoning over long time horizons difficult. Embodied tasks such as those that one might expect a household robot to perform typically require...
http://arxiv.org/pdf/2306.17840v2
2023-07-03T16:32:03
2023-06-30T17:58:02
{ "title": [ -0.044189453, -0.005710602, -0.02583313, -0.05117798, -0.0076904297, 0.022750854, -0.05142212, -0.010055542, 0.0019874573, 0.01927185, -0.0048217773, -0.006378174, 0.0045776367, -0.02482605, -0.035217285, 0.0027484894, -0.006160736, ...
http://arxiv.org/abs/2306.17820v2
Meta-Reasoning: Semantics-Symbol Deconstruction For Large Language Models
Yiming Wang, Zhuosheng Zhang, Rui Wang
cs.CL
Neural-symbolic methods have shown their effectiveness in enhancing the reasoning abilities of large language models (LLMs). However, existing methods primarily rely on mapping natural languages to more syntactically complete formal languages (e.g., Python and SQL). Those approaches necessitate that reasoning tasks be ...
http://arxiv.org/pdf/2306.17820v2
2023-11-02T16:38:29
2023-06-30T17:38:10
{ "title": [ -0.03186035, -0.03112793, -0.019866943, -0.07659912, -0.008178711, 0.03338623, -0.056518555, -0.00082731247, 0.0154418945, 0.030136108, 0.00315094, -0.009147644, 0.01625061, -0.025222778, -0.042175293, 0.0052375793, -0.004673004, 0.0...
http://arxiv.org/abs/2306.17810v2
A Massive Scale Semantic Similarity Dataset of Historical English
Emily Silcock, Melissa Dell
cs.CL, econ.GN, q-fin.EC
A diversity of tasks use language models trained on semantic similarity data. While there are a variety of datasets that capture semantic similarity, they are either constructed from modern web data or are relatively small datasets created in the past decade by human annotators. This study utilizes a novel source, newl...
http://arxiv.org/pdf/2306.17810v2
2023-08-24T01:22:36
2023-06-30T17:16:04
{ "title": [ -0.0052490234, 0.020523071, -0.04425049, -0.05130005, 0.043060303, -0.011009216, -0.01725769, 0.0024642944, 0.035369873, 0.045562744, -0.02015686, 0.02973938, 0.028045654, 0.00076293945, -0.039245605, 0.04876709, -0.0039367676, 0.033...
http://arxiv.org/abs/2306.17806v1
Stay on topic with Classifier-Free Guidance
Guillaume Sanchez, Honglu Fan, Alexander Spangher, Elad Levi, Pawan Sasanka Ammanamanchi, Stella Biderman
cs.CL, cs.CV, cs.LG
Classifier-Free Guidance (CFG) has recently emerged in text-to-image generation as a lightweight technique to encourage prompt-adherence in generations. In this work, we demonstrate that CFG can be used broadly as an inference-time technique in pure language modeling. We show that CFG (1) improves the performance of Py...
http://arxiv.org/pdf/2306.17806v1
2023-06-30T17:07:02
2023-06-30T17:07:02
{ "title": [ 0.053497314, -0.020141602, 0.0023269653, -0.093811035, -0.029785156, 0.0065307617, -0.031066895, -0.0011262894, -0.013183594, 0.03366089, 0.043304443, -0.008201599, -0.024856567, -0.0552063, 0.004360199, 0.008674622, 0.01525116, 0.00...
http://arxiv.org/abs/2307.07516v1
Voting-based Multimodal Automatic Deception Detection
Lana Touma, Mohammad Al Horani, Manar Tailouni, Anas Dahabiah, Khloud Al Jallad
cs.LG, cs.CL, cs.CV, cs.HC
Automatic Deception Detection has been a hot research topic for a long time, using machine learning and deep learning to automatically detect deception, brings new light to this old field. In this paper, we proposed a voting-based method for automatic deception detection from videos using audio, visual and lexical feat...
http://arxiv.org/pdf/2307.07516v1
2023-06-30T17:05:11
2023-06-30T17:05:11
{ "title": [ -0.015716553, 0.023498535, -0.0440979, -0.0000123381615, 0.022705078, -0.021896362, 0.0020942688, -0.04534912, -0.003610611, 0.024658203, -0.0006747246, -0.0012845993, 0.010604858, -0.025222778, -0.03845215, -0.00090551376, -0.032928467,...
http://arxiv.org/abs/2306.17792v1
Towards Improving the Performance of Pre-Trained Speech Models for Low-Resource Languages Through Lateral Inhibition
Andrei-Marius Avram, Răzvan-Alexandru Smădu, Vasile Păiş, Dumitru-Clementin Cercel, Radu Ion, Dan Tufiş
cs.CL, cs.SD, eess.AS
With the rise of bidirectional encoder representations from Transformer models in natural language processing, the speech community has adopted some of their development methodologies. Therefore, the Wav2Vec models were introduced to reduce the data required to obtain state-of-the-art results. This work leverages this ...
http://arxiv.org/pdf/2306.17792v1
2023-06-30T16:48:22
2023-06-30T16:48:22
{ "title": [ -0.0053977966, -0.003435135, -0.03173828, -0.07086182, -0.013679504, 0.0026168823, -0.056121826, 0.02015686, -0.026275635, 0.04916382, 0.0018625259, 0.007457733, 0.019958496, -0.027999878, -0.039794922, 0.008049011, -0.0075569153, 0....
http://arxiv.org/abs/2306.17757v1
Should you marginalize over possible tokenizations?
Nadezhda Chirkova, Germán Kruszewski, Jos Rozen, Marc Dymetman
cs.CL
Autoregressive language models (LMs) map token sequences to probabilities. The usual practice for computing the probability of any character string (e.g. English sentences) is to first transform it into a sequence of tokens that is scored by the model. However, there are exponentially many token sequences that represen...
http://arxiv.org/pdf/2306.17757v1
2023-06-30T16:09:01
2023-06-30T16:09:01
{ "title": [ 0.030960083, -0.004814148, 0.02609253, 0.0018510818, -0.021621704, 0.010757446, 0.02658081, -0.02243042, -0.06512451, -0.0066375732, -0.021530151, -0.01914978, -0.06536865, -0.036468506, 0.031234741, -0.08428955, -0.023468018, 0.0340...
http://arxiv.org/abs/2306.17733v1
Token-Event-Role Structure-based Multi-Channel Document-Level Event Extraction
Qizhi Wan, Changxuan Wan, Keli Xiao, Hui Xiong, Dexi Liu, Xiping Liu
cs.CL, cs.AI
Document-level event extraction is a long-standing challenging information retrieval problem involving a sequence of sub-tasks: entity extraction, event type judgment, and event type-specific multi-event extraction. However, addressing the problem as multiple learning tasks leads to increased model complexity. Also, ex...
http://arxiv.org/pdf/2306.17733v1
2023-06-30T15:22:57
2023-06-30T15:22:57
{ "title": [ -0.009719849, 0.0016651154, -0.058654785, -0.10406494, 0.013687134, 0.0016145706, -0.059661865, 0.01576233, 0.005645752, 0.02218628, 0.0134887695, -0.004173279, 0.024932861, -0.04446411, -0.025039673, 0.010604858, -0.020446777, 0.001...
http://arxiv.org/abs/2306.17727v3
Improved NL2SQL based on Multi-layer Expert Network
Chenduo Hao, Xu Zhang
cs.CL
The Natural Language to SQL (NL2SQL) technique is used to convert natural language queries into executable SQL statements. Typically, slot-filling is employed as a classification method for multi-task cases to achieve this goal. However, slot-filling can result in inaccurate SQL statement generation due to negative mig...
http://arxiv.org/pdf/2306.17727v3
2023-09-18T03:39:23
2023-06-30T15:16:52
{ "title": [ -0.019515991, 0.00054979324, -0.064697266, -0.0836792, 0.035583496, -0.0009379387, -0.017196655, 0.036895752, 0.010795593, 0.045043945, 0.006916046, -0.010002136, -0.00066518784, -0.028457642, -0.0524292, 0.012001038, 0.006668091, -0...
http://arxiv.org/abs/2306.17700v1
Beyond Neural-on-Neural Approaches to Speaker Gender Protection
Loes van Bemmel, Zhuoran Liu, Nik Vaessen, Martha Larson
eess.AS, cs.CL, cs.CR, cs.CY, cs.LG, cs.SD
Recent research has proposed approaches that modify speech to defend against gender inference attacks. The goal of these protection algorithms is to control the availability of information about a speaker's gender, a privacy-sensitive attribute. Currently, the common practice for developing and testing gender protectio...
http://arxiv.org/pdf/2306.17700v1
2023-06-30T14:26:49
2023-06-30T14:26:49
{ "title": [ 0.010398865, 0.011146545, -0.040893555, -0.06573486, 0.0063972473, -0.012710571, -0.024887085, 0.017593384, -0.025924683, 0.050445557, 0.004322052, -0.019485474, 0.042022705, -0.03173828, -0.029266357, 0.0099487305, -0.027542114, 0.0...
http://arxiv.org/abs/2306.17695v1
A New Task and Dataset on Detecting Attacks on Human Rights Defenders
Shihao Ran, Di Lu, Joel Tetreault, Aoife Cahill, Alejandro Jaimes
cs.CL
The ability to conduct retrospective analyses of attacks on human rights defenders over time and by location is important for humanitarian organizations to better understand historical or ongoing human rights violations and thus better manage the global impact of such events. We hypothesize that NLP can support such ef...
http://arxiv.org/pdf/2306.17695v1
2023-06-30T14:20:06
2023-06-30T14:20:06
{ "title": [ 0.03640747, 0.0059318542, -0.0574646, -0.027755737, 0.01512146, -0.058929443, 0.01663208, 0.016036987, -0.041992188, 0.015701294, -0.020462036, 0.038330078, -0.0049438477, -0.03277588, 0.023864746, -0.018463135, 0.00025582314, 0.0312...
http://arxiv.org/abs/2306.17674v1
X-RiSAWOZ: High-Quality End-to-End Multilingual Dialogue Datasets and Few-shot Agents
Mehrad Moradshahi, Tianhao Shen, Kalika Bali, Monojit Choudhury, Gaël de Chalendar, Anmol Goel, Sungkyun Kim, Prashant Kodali, Ponnurangam Kumaraguru, Nasredine Semmar, Sina J. Semnani, Jiwon Seo, Vivek Seshadri, Manish Shrivastava, Michael Sun, Aditya Yadavalli, Chaobin You, Deyi Xiong, Monica S. Lam
cs.CL
Task-oriented dialogue research has mainly focused on a few popular languages like English and Chinese, due to the high dataset creation cost for a new language. To reduce the cost, we apply manual editing to automatically translated data. We create a new multilingual benchmark, X-RiSAWOZ, by translating the Chinese Ri...
http://arxiv.org/pdf/2306.17674v1
2023-06-30T14:03:30
2023-06-30T14:03:30
{ "title": [ -0.011665344, -0.026275635, -0.03189087, -0.068237305, 0.00053167343, -0.00030565262, -0.049743652, 0.016281128, 0.008102417, 0.040252686, -0.0028591156, -0.01247406, 0.0030078888, -0.016616821, -0.05340576, 0.0069999695, -0.014343262, ...
http://arxiv.org/abs/2306.17649v3
Biomedical Language Models are Robust to Sub-optimal Tokenization
Bernal Jiménez Gutiérrez, Huan Sun, Yu Su
cs.CL
As opposed to general English, many concepts in biomedical terminology have been designed in recent history by biomedical professionals with the goal of being precise and concise. This is often achieved by concatenating meaningful biomedical morphemes to create new semantic units. Nevertheless, most modern biomedical l...
http://arxiv.org/pdf/2306.17649v3
2023-07-10T16:22:41
2023-06-30T13:35:24
{ "title": [ -0.038513184, 0.0022659302, -0.034454346, -0.04434204, -0.013313293, 0.0047454834, -0.07824707, -0.01576233, -0.0021152496, 0.030227661, -0.01322937, -0.011329651, 0.013221741, -0.031585693, -0.047210693, 0.010238647, -0.004966736, 0...
http://arxiv.org/abs/2306.17646v1
Feature Representation Learning for NL2SQL Generation Based on Coupling and Decoupling
Chenduo Hao, Xu Zhang, Chuanbao Gao, Deyu Zhou
cs.CL
The NL2SQL task involves parsing natural language statements into SQL queries. While most state-of-the-art methods treat NL2SQL as a slot-filling task and use feature representation learning techniques, they overlook explicit correlation features between the SELECT and WHERE clauses and implicit correlation features be...
http://arxiv.org/pdf/2306.17646v1
2023-06-30T13:34:31
2023-06-30T13:34:31
{ "title": [ -0.014839172, 0.0013427734, -0.04425049, -0.0690918, 0.01374054, 0.011009216, -0.057556152, 0.006816864, 0.028259277, 0.047027588, -0.011245728, -0.027832031, 0.0025997162, -0.025985718, -0.06173706, -0.019454956, -0.0027446747, 0.03...