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/2311.00906v1
Re-weighting Tokens: A Simple and Effective Active Learning Strategy for Named Entity Recognition
Haocheng Luo, Wei Tan, Ngoc Dang Nguyen, Lan Du
cs.CL, cs.LG
Active learning, a widely adopted technique for enhancing machine learning models in text and image classification tasks with limited annotation resources, has received relatively little attention in the domain of Named Entity Recognition (NER). The challenge of data imbalance in NER has hindered the effectiveness of a...
http://arxiv.org/pdf/2311.00906v1
2023-11-02T00:19:02
2023-11-02T00:19:02
{ "title": [ -0.018371582, -0.002960205, -0.045898438, -0.06945801, 0.02659607, -0.001540184, -0.025177002, -0.0036563873, 0.016921997, 0.026794434, -0.020828247, 0.023269653, -0.0074272156, -0.009109497, -0.03527832, 0.0084991455, -0.0036239624, ...
http://arxiv.org/abs/2311.00897v1
On The Open Prompt Challenge In Conditional Audio Generation
Ernie Chang, Sidd Srinivasan, Mahi Luthra, Pin-Jie Lin, Varun Nagaraja, Forrest Iandola, Zechun Liu, Zhaoheng Ni, Changsheng Zhao, Yangyang Shi, Vikas Chandra
cs.SD, cs.CL, eess.AS
Text-to-audio generation (TTA) produces audio from a text description, learning from pairs of audio samples and hand-annotated text. However, commercializing audio generation is challenging as user-input prompts are often under-specified when compared to text descriptions used to train TTA models. In this work, we trea...
http://arxiv.org/pdf/2311.00897v1
2023-11-01T23:33:25
2023-11-01T23:33:25
{ "title": [ -0.013771057, 0.001996994, -0.06427002, -0.09539795, 0.020523071, 0.0062294006, -0.07098389, 0.013244629, 0.0017223358, 0.03451538, -0.019577026, -0.002796173, 0.0037403107, -0.02255249, -0.034240723, -0.00074338913, -0.013008118, 0....
http://arxiv.org/abs/2311.00895v1
In-Context Prompt Editing For Conditional Audio Generation
Ernie Chang, Pin-Jie Lin, Yang Li, Sidd Srinivasan, Gael Le Lan, David Kant, Yangyang Shi, Forrest Iandola, Vikas Chandra
cs.SD, cs.CL, eess.AS
Distributional shift is a central challenge in the deployment of machine learning models as they can be ill-equipped for real-world data. This is particularly evident in text-to-audio generation where the encoded representations are easily undermined by unseen prompts, which leads to the degradation of generated audio ...
http://arxiv.org/pdf/2311.00895v1
2023-11-01T23:31:51
2023-11-01T23:31:51
{ "title": [ -0.011314392, -0.00793457, -0.06414795, -0.08642578, 0.00819397, -0.004463196, -0.055908203, 0.016159058, 0.0019445419, 0.0418396, -0.02468872, -0.0074310303, 0.016098022, -0.026153564, -0.038024902, -0.0010099411, -0.012451172, 0.01...
http://arxiv.org/abs/2311.00871v1
Pretraining Data Mixtures Enable Narrow Model Selection Capabilities in Transformer Models
Steve Yadlowsky, Lyric Doshi, Nilesh Tripuraneni
cs.LG, cs.CL, stat.ML
Transformer models, notably large language models (LLMs), have the remarkable ability to perform in-context learning (ICL) -- to perform new tasks when prompted with unseen input-output examples without any explicit model training. In this work, we study how effectively transformers can bridge between their pretraining...
http://arxiv.org/pdf/2311.00871v1
2023-11-01T21:41:08
2023-11-01T21:41:08
{ "title": [ 0.044677734, 0.008560181, -0.02281189, -0.022323608, -0.0135650635, -0.0418396, -0.01914978, -0.0385437, -0.049957275, 0.01940918, -0.025405884, 0.03250122, 0.002916336, -0.048553467, 0.045196533, -0.007255554, 0.009765625, 0.0338745...
http://arxiv.org/abs/2311.00867v1
Automatic Disfluency Detection from Untranscribed Speech
Amrit Romana, Kazuhito Koishida, Emily Mower Provost
eess.AS, cs.CL
Speech disfluencies, such as filled pauses or repetitions, are disruptions in the typical flow of speech. Stuttering is a speech disorder characterized by a high rate of disfluencies, but all individuals speak with some disfluencies and the rates of disfluencies may by increased by factors such as cognitive load. Clini...
http://arxiv.org/pdf/2311.00867v1
2023-11-01T21:36:39
2023-11-01T21:36:39
{ "title": [ -0.011550903, 0.020370483, -0.060424805, -0.061767578, 0.01449585, -0.0055618286, -0.048461914, -0.016601562, 0.046539307, 0.037841797, -0.024459839, -0.03439331, 0.03930664, -0.023086548, -0.066589355, 0.0026760101, -0.013687134, 0....
http://arxiv.org/abs/2311.00863v1
Training Dynamics of Contextual N-Grams in Language Models
Lucia Quirke, Lovis Heindrich, Wes Gurnee, Neel Nanda
cs.LG, cs.AI, cs.CL
Prior work has shown the existence of contextual neurons in language models, including a neuron that activates on German text. We show that this neuron exists within a broader contextual n-gram circuit: we find late layer neurons which recognize and continue n-grams common in German text, but which only activate if the...
http://arxiv.org/pdf/2311.00863v1
2023-11-01T21:32:51
2023-11-01T21:32:51
{ "title": [ -0.0098724365, 0.0035324097, -0.021499634, -0.04660034, -0.0063819885, 0.0036144257, -0.047698975, -0.011749268, 0.00092983246, 0.05508423, -0.027236938, 0.014312744, 0.033172607, 0.0053977966, -0.052856445, 0.0105896, -0.019958496, ...
http://arxiv.org/abs/2311.00835v1
Calibrated Seq2seq Models for Efficient and Generalizable Ultra-fine Entity Typing
Yanlin Feng, Adithya Pratapa, David R Mortensen
cs.CL
Ultra-fine entity typing plays a crucial role in information extraction by predicting fine-grained semantic types for entity mentions in text. However, this task poses significant challenges due to the massive number of entity types in the output space. The current state-of-the-art approaches, based on standard multi-l...
http://arxiv.org/pdf/2311.00835v1
2023-11-01T20:39:12
2023-11-01T20:39:12
{ "title": [ -0.015792847, -0.005207062, -0.05609131, -0.08898926, 0.024414062, -0.0048942566, -0.064453125, 0.008117676, 0.008972168, 0.04626465, -0.0018377304, 0.015975952, 0.03555298, -0.013618469, -0.034851074, 0.025985718, -0.026916504, 0.03...
http://arxiv.org/abs/2311.00790v1
Construction Artifacts in Metaphor Identification Datasets
Joanne Boisson, Luis Espinosa-Anke, Jose Camacho-Collados
cs.CL
Metaphor identification aims at understanding whether a given expression is used figuratively in context. However, in this paper we show how existing metaphor identification datasets can be gamed by fully ignoring the potential metaphorical expression or the context in which it occurs. We test this hypothesis in a vari...
http://arxiv.org/pdf/2311.00790v1
2023-11-01T19:21:55
2023-11-01T19:21:55
{ "title": [ -0.01361084, 0.0025615692, -0.02003479, -0.0044059753, -0.0031051636, -0.01411438, 0.004917145, -0.0362854, -0.014350891, 0.048797607, -0.015274048, 0.009284973, 0.020202637, -0.010513306, -0.032043457, 0.012321472, -0.015792847, 0.0...
http://arxiv.org/abs/2311.04913v2
An Improved Transformer-based Model for Detecting Phishing, Spam, and Ham: A Large Language Model Approach
Suhaima Jamal, Hayden Wimmer
cs.CL, cs.AI
Phishing and spam detection is long standing challenge that has been the subject of much academic research. Large Language Models (LLM) have vast potential to transform society and provide new and innovative approaches to solve well-established challenges. Phishing and spam have caused financial hardships and lost time...
http://arxiv.org/pdf/2311.04913v2
2023-11-12T16:32:16
2023-11-01T18:41:50
{ "title": [ 0.014709473, 0.039367676, -0.036071777, -0.026626587, 0.03945923, -0.022018433, -0.021209717, -0.040527344, -0.03842163, 0.044891357, -0.0069847107, 0.001996994, 0.04284668, -0.023223877, -0.022232056, -0.0010061264, -0.042114258, 0....
http://arxiv.org/abs/2311.04911v1
From Text to Structure: Using Large Language Models to Support the Development of Legal Expert Systems
Samyar Janatian, Hannes Westermann, Jinzhe Tan, Jaromir Savelka, Karim Benyekhlef
cs.CL, cs.AI, cs.HC
Encoding legislative text in a formal representation is an important prerequisite to different tasks in the field of AI & Law. For example, rule-based expert systems focused on legislation can support laypeople in understanding how legislation applies to them and provide them with helpful context and information. Howev...
http://arxiv.org/pdf/2311.04911v1
2023-11-01T18:31:02
2023-11-01T18:31:02
{ "title": [ -0.05227661, -0.020507812, -0.044952393, -0.06274414, -0.0027236938, 0.026107788, -0.06707764, 0.008964539, 0.018737793, 0.028182983, -0.034851074, -0.0013952255, 0.02015686, -0.010574341, -0.04562378, 0.008560181, -0.015296936, 0.03...
http://arxiv.org/abs/2311.00768v1
Language Model Training Paradigms for Clinical Feature Embeddings
Yurong Hu, Manuel Burger, Gunnar Rätsch, Rita Kuznetsova
cs.LG, cs.CL
In research areas with scarce data, representation learning plays a significant role. This work aims to enhance representation learning for clinical time series by deriving universal embeddings for clinical features, such as heart rate and blood pressure. We use self-supervised training paradigms for language models to...
http://arxiv.org/pdf/2311.00768v1
2023-11-01T18:23:12
2023-11-01T18:23:12
{ "title": [ -0.024642944, -0.010871887, -0.03933716, -0.056152344, -0.011070251, 0.0012931824, -0.06008911, 0.0074386597, 0.0077781677, 0.02532959, 0.004146576, -0.0071868896, 0.016464233, -0.043792725, -0.022521973, 0.018493652, -0.0070114136, ...
http://arxiv.org/abs/2311.00762v1
Challenges for Linguistically-Driven Computer-Based Sign Recognition from Continuous Signing for American Sign Language
Carol Neidle
cs.CV, cs.CL
There have been recent advances in computer-based recognition of isolated, citation-form signs from video. There are many challenges for such a task, not least the naturally occurring inter- and intra- signer synchronic variation in sign production, including sociolinguistic variation in the realization of certain sign...
http://arxiv.org/pdf/2311.00762v1
2023-11-01T18:08:44
2023-11-01T18:08:44
{ "title": [ -0.0519104, -0.0058021545, -0.032684326, -0.015594482, -0.023986816, 0.0003170967, -0.02368164, 0.0001270771, 0.007858276, 0.0151901245, -0.060180664, 0.0010156631, -0.010971069, -0.015670776, -0.039855957, -0.027145386, -0.0066566467, ...
http://arxiv.org/abs/2311.00697v1
End-to-End Single-Channel Speaker-Turn Aware Conversational Speech Translation
Juan Zuluaga-Gomez, Zhaocheng Huang, Xing Niu, Rohit Paturi, Sundararajan Srinivasan, Prashant Mathur, Brian Thompson, Marcello Federico
cs.CL, eess.AS
Conventional speech-to-text translation (ST) systems are trained on single-speaker utterances, and they may not generalize to real-life scenarios where the audio contains conversations by multiple speakers. In this paper, we tackle single-channel multi-speaker conversational ST with an end-to-end and multi-task trainin...
http://arxiv.org/pdf/2311.00697v1
2023-11-01T17:55:09
2023-11-01T17:55:09
{ "title": [ -0.0074691772, 0.00074100494, -0.01576233, -0.07055664, -0.012527466, 0.0118255615, -0.072631836, 0.01210022, -0.00022816658, 0.058013916, 0.0020542145, 0.00038433075, 0.04144287, -0.034851074, -0.03729248, 0.022994995, -0.0007867813, ...
http://arxiv.org/abs/2311.00694v1
Unleashing the Creative Mind: Language Model As Hierarchical Policy For Improved Exploration on Challenging Problem Solving
Zhan Ling, Yunhao Fang, Xuanlin Li, Tongzhou Mu, Mingu Lee, Reza Pourreza, Roland Memisevic, Hao Su
cs.AI, cs.CL
Large Language Models (LLMs) have achieved tremendous progress, yet they still often struggle with challenging reasoning problems. Current approaches address this challenge by sampling or searching detailed and low-level reasoning chains. However, these methods are still limited in their exploration capabilities, makin...
http://arxiv.org/pdf/2311.00694v1
2023-11-01T17:52:15
2023-11-01T17:52:15
{ "title": [ -0.018005371, -0.014480591, -0.025924683, -0.028839111, -0.009147644, -0.011779785, -0.02960205, -0.020202637, -0.020111084, 0.029403687, 0.003522873, 0.0047302246, -0.011016846, -0.027359009, -0.04586792, -0.00058174133, -0.027297974, ...
http://arxiv.org/abs/2311.00687v2
Improving Interpersonal Communication by Simulating Audiences with Language Models
Ryan Liu, Howard Yen, Raja Marjieh, Thomas L. Griffiths, Ranjay Krishna
cs.AI, cs.CL, cs.HC, cs.LG
How do we communicate with others to achieve our goals? We use our prior experience or advice from others, or construct a candidate utterance by predicting how it will be received. However, our experiences are limited and biased, and reasoning about potential outcomes can be difficult and cognitively challenging. In th...
http://arxiv.org/pdf/2311.00687v2
2023-11-03T13:17:55
2023-11-01T17:44:50
{ "title": [ -0.016693115, 0.011161804, -0.024856567, -0.045135498, 0.004333496, 0.006401062, -0.058532715, -0.017288208, 0.00073862076, 0.036865234, 0.0030593872, -0.009994507, 0.0035419464, -0.024414062, -0.04815674, 0.011665344, -0.0101623535, ...
http://arxiv.org/abs/2311.00686v1
Little Giants: Exploring the Potential of Small LLMs as Evaluation Metrics in Summarization in the Eval4NLP 2023 Shared Task
Neema Kotonya, Saran Krishnasamy, Joel Tetreault, Alejandro Jaimes
cs.CL
This paper describes and analyzes our participation in the 2023 Eval4NLP shared task, which focuses on assessing the effectiveness of prompt-based techniques to empower Large Language Models to handle the task of quality estimation, particularly in the context of evaluating machine translations and summaries. We conduc...
http://arxiv.org/pdf/2311.00686v1
2023-11-01T17:44:35
2023-11-01T17:44:35
{ "title": [ -0.05178833, -0.024627686, -0.038085938, -0.027694702, -0.060058594, 0.0051841736, -0.06958008, -0.0009059906, -0.016586304, 0.021209717, 0.0061035156, 0.015464783, -0.040985107, -0.03765869, -0.0209198, -0.008682251, -0.0023975372, ...
http://arxiv.org/abs/2311.00684v2
Attention Alignment and Flexible Positional Embeddings Improve Transformer Length Extrapolation
Ta-Chung Chi, Ting-Han Fan, Alexander I. Rudnicky
cs.CL, cs.LG
An ideal length-extrapolatable Transformer language model can handle sequences longer than the training length without any fine-tuning. Such long-context utilization capability relies heavily on a flexible positional embedding design. Upon investigating the flexibility of existing large pre-trained Transformer language...
http://arxiv.org/pdf/2311.00684v2
2023-11-15T15:55:02
2023-11-01T17:43:35
{ "title": [ 0.0074424744, 0.012687683, -0.017196655, -0.009208679, -0.035369873, -0.024749756, -0.03866577, -0.050750732, -0.008682251, 0.04663086, -0.02607727, 0.0256958, 0.005569458, -0.021987915, -0.0053596497, 0.00079774857, 0.0003247261, 0....
http://arxiv.org/abs/2311.00681v1
Are Large Language Models Reliable Judges? A Study on the Factuality Evaluation Capabilities of LLMs
Xue-Yong Fu, Md Tahmid Rahman Laskar, Cheng Chen, Shashi Bhushan TN
cs.CL
In recent years, Large Language Models (LLMs) have gained immense attention due to their notable emergent capabilities, surpassing those seen in earlier language models. A particularly intriguing application of LLMs is their role as evaluators for texts produced by various generative models. In this study, we delve i...
http://arxiv.org/pdf/2311.00681v1
2023-11-01T17:42:45
2023-11-01T17:42:45
{ "title": [ -0.04058838, -0.0069465637, -0.047973633, -0.038024902, -0.032684326, -0.013648987, -0.06542969, -0.016784668, 0.0038585663, 0.026397705, -0.029159546, 0.00472641, 0.020645142, -0.0184021, -0.03842163, 0.002073288, -0.040283203, 0.04...
http://arxiv.org/abs/2311.00671v1
Emotion Detection for Misinformation: A Review
Zhiwei Liu, Tianlin Zhang, Kailai Yang, Paul Thompson, Zeping Yu, Sophia Ananiadou
cs.CL
With the advent of social media, an increasing number of netizens are sharing and reading posts and news online. However, the huge volumes of misinformation (e.g., fake news and rumors) that flood the internet can adversely affect people's lives, and have resulted in the emergence of rumor and fake news detection as a ...
http://arxiv.org/pdf/2311.00671v1
2023-11-01T17:21:09
2023-11-01T17:21:09
{ "title": [ 0.0065956116, 0.012611389, -0.026275635, -0.014976501, 0.010932922, -0.021560669, 0.0154418945, -0.025939941, -0.0044441223, 0.019424438, -0.01725769, -0.0035896301, 0.006629944, -0.035369873, -0.021713257, -0.0022468567, -0.01260376, ...
http://arxiv.org/abs/2311.00658v1
Explicit Morphological Knowledge Improves Pre-training of Language Models for Hebrew
Eylon Gueta, Omer Goldman, Reut Tsarfaty
cs.CL
Pre-trained language models (PLMs) have shown remarkable successes in acquiring a wide range of linguistic knowledge, relying solely on self-supervised training on text streams. Nevertheless, the effectiveness of this language-agnostic approach has been frequently questioned for its sub-optimal performance when applied...
http://arxiv.org/pdf/2311.00658v1
2023-11-01T17:02:49
2023-11-01T17:02:49
{ "title": [ 0.004337311, -0.0066833496, -0.01864624, -0.08453369, -0.021514893, 0.005672455, -0.07019043, 0.011680603, -0.040222168, 0.031143188, -0.0013837814, 0.033203125, 0.03213501, -0.051696777, 0.012031555, 0.025115967, -0.011741638, 0.030...
http://arxiv.org/abs/2311.00629v1
Formal Translation from Reversing Petri Nets to Coloured Petri Nets
Kamila Barylska, Anna Gogolinska, Lukasz Mikulski, Anna Philippou, Marcin Piatkowski, Kyriaki Psara
cs.LO, cs.CL, 03, F.2; G.0
Reversible computation is an emerging computing paradigm that allows any sequence of operations to be executed in reverse order at any point during computation. Its appeal lies in its potential for lowpower computation and its relevance to a wide array of applications such as chemical reactions, quantum computation, ro...
http://arxiv.org/pdf/2311.00629v1
2023-11-01T16:28:38
2023-11-01T16:28:38
{ "title": [ -0.06744385, 0.01663208, -0.057891846, -0.06359863, 0.002275467, 0.029205322, -0.068603516, -0.010139465, -0.016052246, 0.024795532, -0.01448822, 0.003036499, -0.02243042, -0.00907135, -0.041381836, -0.0061836243, -0.00920105, -0.016...
http://arxiv.org/abs/2311.00588v1
Boosting Summarization with Normalizing Flows and Aggressive Training
Yu Yang, Xiaotong Shen
cs.CL, cs.AI
This paper presents FlowSUM, a normalizing flows-based variational encoder-decoder framework for Transformer-based summarization. Our approach tackles two primary challenges in variational summarization: insufficient semantic information in latent representations and posterior collapse during training. To address these...
http://arxiv.org/pdf/2311.00588v1
2023-11-01T15:33:38
2023-11-01T15:33:38
{ "title": [ -0.00043010712, 0.012306213, -0.028762817, -0.08691406, 0.008895874, 0.014694214, -0.0647583, -0.011253357, -0.0034065247, 0.048706055, -0.021942139, 0.04852295, -0.014984131, 0.00818634, -0.04373169, 0.013076782, 0.0020370483, 0.022...
http://arxiv.org/abs/2311.00587v1
Crosslingual Retrieval Augmented In-context Learning for Bangla
Xiaoqian Li, Ercong Nie, Sheng Liang
cs.CL
The promise of Large Language Models (LLMs) in Natural Language Processing has often been overshadowed by their limited performance in low-resource languages such as Bangla. To address this, our paper presents a pioneering approach that utilizes cross-lingual retrieval augmented in-context learning. By strategically so...
http://arxiv.org/pdf/2311.00587v1
2023-11-01T15:32:50
2023-11-01T15:32:50
{ "title": [ 0.015602112, 0.0008511543, -0.027877808, -0.03729248, 0.00907135, -0.018005371, 0.004878998, -0.004108429, 0.0000410676, 0.05142212, -0.037628174, 0.008872986, 0.018554688, -0.012207031, -0.03479004, -0.011360168, -0.0154800415, 0.05...
http://arxiv.org/abs/2311.00739v1
Can Large Language Models Design Accurate Label Functions?
Naiqing Guan, Kaiwen Chen, Nick Koudas
cs.CL, cs.DB, cs.LG, H.2.8; I.5.4
Programmatic weak supervision methodologies facilitate the expedited labeling of extensive datasets through the use of label functions (LFs) that encapsulate heuristic data sources. Nonetheless, the creation of precise LFs necessitates domain expertise and substantial endeavors. Recent advances in pre-trained language ...
http://arxiv.org/pdf/2311.00739v1
2023-11-01T15:14:46
2023-11-01T15:14:46
{ "title": [ -0.028503418, -0.0038814545, -0.023422241, -0.031143188, -0.045562744, -0.00504303, -0.054473877, -0.0070495605, -0.02796936, 0.009902954, -0.02571106, -0.017456055, -0.001294136, -0.036895752, 0.012336731, -0.0184021, 0.001584053, 0...
http://arxiv.org/abs/2311.00571v1
LLaVA-Interactive: An All-in-One Demo for Image Chat, Segmentation, Generation and Editing
Wei-Ge Chen, Irina Spiridonova, Jianwei Yang, Jianfeng Gao, Chunyuan Li
cs.CV, cs.AI, cs.CL, cs.HC, cs.MM
LLaVA-Interactive is a research prototype for multimodal human-AI interaction. The system can have multi-turn dialogues with human users by taking multimodal user inputs and generating multimodal responses. Importantly, LLaVA-Interactive goes beyond language prompt, where visual prompt is enabled to align human intents...
http://arxiv.org/pdf/2311.00571v1
2023-11-01T15:13:43
2023-11-01T15:13:43
{ "title": [ 0.021743774, -0.014160156, -0.011238098, -0.052764893, -0.015266418, -0.011955261, -0.040985107, -0.003206253, -0.014633179, 0.042419434, -0.03805542, 0.041503906, 0.020324707, -0.013793945, 0.008132935, -0.0059318542, 0.023803711, -...
http://arxiv.org/abs/2311.00541v1
An Embedded Diachronic Sense Change Model with a Case Study from Ancient Greek
Schyan Zafar, Geoff K. Nicholls
cs.CL, stat.ME
Word meanings change over time, and word senses evolve, emerge or die out in the process. For ancient languages, where the corpora are often small, sparse and noisy, modelling such changes accurately proves challenging, and quantifying uncertainty in sense-change estimates consequently becomes important. GASC and DiSC ...
http://arxiv.org/pdf/2311.00541v1
2023-11-01T14:20:18
2023-11-01T14:20:18
{ "title": [ 0.013465881, 0.017044067, -0.027252197, -0.035858154, 0.0085372925, -0.018981934, -0.008781433, -0.0055618286, 0.04324341, 0.038635254, -0.06750488, -0.007335663, 0.036254883, -0.011428833, -0.052856445, 0.004421234, 0.0041923523, -0...
http://arxiv.org/abs/2311.00522v1
Text Rendering Strategies for Pixel Language Models
Jonas F. Lotz, Elizabeth Salesky, Phillip Rust, Desmond Elliott
cs.CL
Pixel-based language models process text rendered as images, which allows them to handle any script, making them a promising approach to open vocabulary language modelling. However, recent approaches use text renderers that produce a large set of almost-equivalent input patches, which may prove sub-optimal for downstre...
http://arxiv.org/pdf/2311.00522v1
2023-11-01T13:49:31
2023-11-01T13:49:31
{ "title": [ -0.04724121, -0.015533447, -0.0039253235, -0.029327393, -0.04562378, 0.013763428, -0.054260254, -0.005054474, 0.023025513, 0.04949951, -0.029724121, -0.0019397736, 0.020645142, -0.0035591125, -0.059265137, 0.011329651, -0.00044465065, ...
http://arxiv.org/abs/2311.00513v1
Rule-Based Error Classification for Analyzing Differences in Frequent Errors
Atsushi Shirafuji, Taku Matsumoto, Md Faizul Ibne Amin, Yutaka Watanobe
cs.SE, cs.CL, cs.PL
Finding and fixing errors is a time-consuming task not only for novice programmers but also for expert programmers. Prior work has identified frequent error patterns among various levels of programmers. However, the differences in the tendencies between novices and experts have yet to be revealed. From the knowledge of...
http://arxiv.org/pdf/2311.00513v1
2023-11-01T13:36:20
2023-11-01T13:36:20
{ "title": [ -0.021240234, -0.01537323, -0.06365967, -0.06414795, 0.017623901, -0.009841919, -0.025787354, 0.014793396, -0.0074386597, 0.027435303, -0.025314331, 0.015571594, -0.029266357, -0.02331543, -0.031158447, -0.011009216, -0.02168274, 0.0...
http://arxiv.org/abs/2311.00508v1
Robustness Tests for Automatic Machine Translation Metrics with Adversarial Attacks
Yichen Huang, Timothy Baldwin
cs.CL
We investigate MT evaluation metric performance on adversarially-synthesized texts, to shed light on metric robustness. We experiment with word- and character-level attacks on three popular machine translation metrics: BERTScore, BLEURT, and COMET. Our human experiments validate that automatic metrics tend to overpenal...
http://arxiv.org/pdf/2311.00508v1
2023-11-01T13:14:23
2023-11-01T13:14:23
{ "title": [ -0.041778564, 0.0070724487, -0.010749817, -0.019866943, 0.0044555664, 0.010726929, -0.01991272, -0.02519226, -0.034698486, 0.04272461, 0.016403198, 0.023330688, 0.018493652, -0.018463135, -0.038879395, 0.002407074, 0.0011882782, -0.0...
http://arxiv.org/abs/2311.00502v1
Efficient LLM Inference on CPUs
Haihao Shen, Hanwen Chang, Bo Dong, Yu Luo, Hengyu Meng
cs.LG, cs.AI, cs.CL
Large language models (LLMs) have demonstrated remarkable performance and tremendous potential across a wide range of tasks. However, deploying these models has been challenging due to the astronomical amount of model parameters, which requires a demand for large memory capacity and high memory bandwidth. In this paper...
http://arxiv.org/pdf/2311.00502v1
2023-11-01T13:08:50
2023-11-01T13:08:50
{ "title": [ -0.06555176, -0.012176514, -0.045013428, -0.07128906, -0.018463135, 0.021942139, -0.06536865, 0.022338867, 0.0070114136, 0.00868988, -0.060150146, 0.024749756, -0.00077295303, -0.023101807, 0.013069153, 0.009254456, -0.004070282, 0.0...
http://arxiv.org/abs/2311.00488v1
Comparing Optimization Targets for Contrast-Consistent Search
Hugo Fry, Seamus Fallows, Ian Fan, Jamie Wright, Nandi Schoots
cs.LG, cs.CL
We investigate the optimization target of Contrast-Consistent Search (CCS), which aims to recover the internal representations of truth of a large language model. We present a new loss function that we call the Midpoint-Displacement (MD) loss function. We demonstrate that for a certain hyper-parameter value this MD los...
http://arxiv.org/pdf/2311.00488v1
2023-11-01T12:42:14
2023-11-01T12:42:14
{ "title": [ -0.0146484375, 0.0027446747, -0.0058135986, -0.015266418, 0.00932312, 0.0018501282, -0.0024967194, -0.048217773, 0.0034713745, 0.04006958, 0.0056533813, 0.012771606, -0.017852783, -0.012367249, -0.045013428, 0.009391785, -0.001698494, ...
http://arxiv.org/abs/2311.00475v1
Style Locality for Controllable Generation with kNN Language Models
Gilles Nawezi, Lucie Flek, Charles Welch
cs.CL
Recent language models have been improved by the addition of external memory. Nearest neighbor language models retrieve similar contexts to assist in word prediction. The addition of locality levels allows a model to learn how to weight neighbors based on their relative location to the current text in source documents,...
http://arxiv.org/pdf/2311.00475v1
2023-11-01T12:21:53
2023-11-01T12:21:53
{ "title": [ -0.060791016, -0.019882202, -0.05090332, -0.059661865, -0.023971558, 0.027618408, -0.08972168, 0.016418457, -0.0001963377, 0.054992676, 0.01928711, 0.027313232, 0.03982544, -0.017913818, -0.007003784, 0.046661377, -0.011253357, 0.010...
http://arxiv.org/abs/2311.00451v1
Discourse Relations Classification and Cross-Framework Discourse Relation Classification Through the Lens of Cognitive Dimensions: An Empirical Investigation
Yingxue Fu
cs.CL
Existing discourse formalisms use different taxonomies of discourse relations, which require expert knowledge to understand, posing a challenge for annotation and automatic classification. We show that discourse relations can be effectively captured by some simple cognitively inspired dimensions proposed by Sanders et ...
http://arxiv.org/pdf/2311.00451v1
2023-11-01T11:38:19
2023-11-01T11:38:19
{ "title": [ 0.00137043, -0.004890442, -0.07147217, -0.07885742, 0.0070533752, -0.01625061, -0.037353516, 0.009918213, 0.0012569427, 0.0074691772, 0.007129669, 0.008407593, -0.009597778, -0.040802002, -0.03656006, -0.012504578, -0.037231445, 0.07...
http://arxiv.org/abs/2311.00445v1
A Systematic Comparison of Syllogistic Reasoning in Humans and Language Models
Tiwalayo Eisape, MH Tessler, Ishita Dasgupta, Fei Sha, Sjoerd van Steenkiste, Tal Linzen
cs.CL, cs.AI, cs.LG
A central component of rational behavior is logical inference: the process of determining which conclusions follow from a set of premises. Psychologists have documented several ways in which humans' inferences deviate from the rules of logic. Do language models, which are trained on text generated by humans, replicate ...
http://arxiv.org/pdf/2311.00445v1
2023-11-01T11:13:06
2023-11-01T11:13:06
{ "title": [ -0.016677856, 0.00002014637, -0.020126343, -0.06976318, -0.010726929, 0.005065918, -0.075805664, -0.01008606, -0.012954712, 0.03466797, -0.004722595, -0.017593384, 0.00079774857, -0.024765015, -0.040985107, -0.002073288, -0.019363403, ...
http://arxiv.org/abs/2311.00430v1
Distil-Whisper: Robust Knowledge Distillation via Large-Scale Pseudo Labelling
Sanchit Gandhi, Patrick von Platen, Alexander M. Rush
cs.CL, cs.SD, eess.AS
As the size of pre-trained speech recognition models increases, running these large models in low-latency or resource-constrained environments becomes challenging. In this work, we leverage pseudo-labelling to assemble a large-scale open-source dataset which we use to distill the Whisper model into a smaller variant, c...
http://arxiv.org/pdf/2311.00430v1
2023-11-01T10:45:07
2023-11-01T10:45:07
{ "title": [ -0.014305115, 0.0047836304, -0.0814209, -0.101257324, 0.01864624, -0.007511139, -0.064575195, 0.0013017654, 0.01663208, 0.013999939, -0.014030457, -0.0104599, -0.012359619, -0.034942627, -0.025177002, -0.018814087, 0.015640259, 0.004...
http://arxiv.org/abs/2311.00416v1
Efficient Human-AI Coordination via Preparatory Language-based Convention
Cong Guan, Lichao Zhang, Chunpeng Fan, Yichen Li, Feng Chen, Lihe Li, Yunjia Tian, Lei Yuan, Yang Yu
cs.LG, cs.CL
Developing intelligent agents capable of seamless coordination with humans is a critical step towards achieving artificial general intelligence. Existing methods for human-AI coordination typically train an agent to coordinate with a diverse set of policies or with human models fitted from real human data. However, the...
http://arxiv.org/pdf/2311.00416v1
2023-11-01T10:18:23
2023-11-01T10:18:23
{ "title": [ -0.020141602, -0.0044898987, -0.041046143, -0.06842041, 0.0041885376, -0.005844116, -0.04071045, -0.006538391, -0.022720337, 0.041656494, -0.010826111, 0.00598526, 0.012466431, -0.0061187744, -0.02810669, -0.014877319, -0.0052108765, ...
http://arxiv.org/abs/2311.00408v1
AdaSent: Efficient Domain-Adapted Sentence Embeddings for Few-Shot Classification
Yongxin Huang, Kexin Wang, Sourav Dutta, Raj Nath Patel, Goran Glavaš, Iryna Gurevych
cs.CL
Recent work has found that few-shot sentence classification based on pre-trained Sentence Encoders (SEs) is efficient, robust, and effective. In this work, we investigate strategies for domain-specialization in the context of few-shot sentence classification with SEs. We first establish that unsupervised Domain-Adaptiv...
http://arxiv.org/pdf/2311.00408v1
2023-11-01T10:00:15
2023-11-01T10:00:15
{ "title": [ 0.012359619, -0.010864258, -0.0262146, -0.04244995, 0.005508423, -0.007358551, -0.02041626, -0.016540527, 0.017807007, 0.0073165894, -0.015449524, 0.018859863, -0.017303467, -0.012962341, -0.034362793, -0.0041656494, -0.013580322, 0....
http://arxiv.org/abs/2311.00399v1
Enhanced Knowledge Injection for Radiology Report Generation
Qingqiu Li, Jilan Xu, Runtian Yuan, Mohan Chen, Yuejie Zhang, Rui Feng, Xiaobo Zhang, Shang Gao
cs.CV, cs.CL
Automatic generation of radiology reports holds crucial clinical value, as it can alleviate substantial workload on radiologists and remind less experienced ones of potential anomalies. Despite the remarkable performance of various image captioning methods in the natural image field, generating accurate reports for med...
http://arxiv.org/pdf/2311.00399v1
2023-11-01T09:50:55
2023-11-01T09:50:55
{ "title": [ -0.003332138, -0.013160706, -0.04925537, -0.07208252, -0.0020370483, -0.0054626465, -0.027618408, 0.020889282, -0.01928711, 0.012237549, -0.048614502, 0.007835388, -0.03768921, -0.019699097, -0.039215088, -0.025802612, 0.01020813, -0...
http://arxiv.org/abs/2311.00367v1
Prompt-based Logical Semantics Enhancement for Implicit Discourse Relation Recognition
Chenxu Wang, Ping Jian, Mu Huang
cs.CL, cs.AI
Implicit Discourse Relation Recognition (IDRR), which infers discourse relations without the help of explicit connectives, is still a crucial and challenging task for discourse parsing. Recent works tend to exploit the hierarchical structure information from the annotated senses, which demonstrate enhanced discourse re...
http://arxiv.org/pdf/2311.00367v1
2023-11-01T08:38:08
2023-11-01T08:38:08
{ "title": [ 0.0037212372, 0.0018758774, -0.050079346, -0.10357666, 0.0046653748, 0.009880066, -0.08001709, 0.021072388, 0.040161133, 0.01637268, -0.01977539, -0.0067367554, -0.0050811768, -0.008773804, -0.05987549, -0.0047836304, -0.033050537, 0...
http://arxiv.org/abs/2311.04910v1
Ontology-Driven Processing of Transdisciplinary Domain Knowledge
Oleksandr Palagin, Mykola Petrenko, Sergii Kryvyi, Mykola Boyko, Kyrylo Malakhov
cs.DL, cs.AI, cs.CL
The monograph discusses certain aspects of modern real-world problems facing humanity, which are much more challenging than scientific ones. Modern science is unable to solve them in a fundamental way. Vernadsky's noosphere thesis, in fact, appeals to the scientific worldview that needs to be built in a way that overco...
http://arxiv.org/pdf/2311.04910v1
2023-11-01T07:42:34
2023-11-01T07:42:34
{ "title": [ -0.004295349, -0.0031776428, -0.064208984, -0.07977295, 0.015975952, -0.020492554, -0.038116455, -0.0026683807, 0.00069999695, 0.01525116, -0.006969452, -0.008911133, -0.011543274, -0.027404785, -0.035461426, -0.00090932846, -0.008361816...
http://arxiv.org/abs/2311.00732v1
tmn at #SMM4H 2023: Comparing Text Preprocessing Techniques for Detecting Tweets Self-reporting a COVID-19 Diagnosis
Anna Glazkova
cs.CL, cs.AI, cs.LG, cs.SI, 68T50, I.2.7; I.7.m; H.3.3; J.3
The paper describes a system developed for Task 1 at SMM4H 2023. The goal of the task is to automatically distinguish tweets that self-report a COVID-19 diagnosis (for example, a positive test, clinical diagnosis, or hospitalization) from those that do not. We investigate the use of different techniques for preprocessi...
http://arxiv.org/pdf/2311.00732v1
2023-11-01T07:41:23
2023-11-01T07:41:23
{ "title": [ 0.0040016174, 0.010810852, -0.013031006, -0.031555176, -0.012504578, 0.019012451, -0.04348755, -0.0033130646, -0.025756836, 0.044799805, -0.018554688, 0.055541992, -0.02557373, -0.013214111, 0.0112838745, -0.020690918, 0.007335663, 0...
http://arxiv.org/abs/2311.00321v1
HARE: Explainable Hate Speech Detection with Step-by-Step Reasoning
Yongjin Yang, Joonkee Kim, Yujin Kim, Namgyu Ho, James Thorne, Se-young Yun
cs.CL
With the proliferation of social media, accurate detection of hate speech has become critical to ensure safety online. To combat nuanced forms of hate speech, it is important to identify and thoroughly explain hate speech to help users understand its harmful effects. Recent benchmarks have attempted to tackle this issu...
http://arxiv.org/pdf/2311.00321v1
2023-11-01T06:09:54
2023-11-01T06:09:54
{ "title": [ 0.014663696, -0.000831604, -0.028518677, -0.06060791, 0.02267456, -0.009971619, -0.0041389465, 0.00070619583, -0.015838623, 0.0076446533, -0.019165039, 0.015403748, -0.005908966, -0.017547607, -0.02470398, -0.017440796, -0.01953125, ...
http://arxiv.org/abs/2311.00317v1
Data Augmentation for Code Translation with Comparable Corpora and Multiple References
Yiqing Xie, Atharva Naik, Daniel Fried, Carolyn Rose
cs.CL, cs.LG, cs.SE
One major challenge of translating code between programming languages is that parallel training data is often limited. To overcome this challenge, we present two data augmentation techniques, one that builds comparable corpora (i.e., code pairs with similar functionality), and another that augments existing parallel da...
http://arxiv.org/pdf/2311.00317v1
2023-11-01T06:01:22
2023-11-01T06:01:22
{ "title": [ -0.0206604, -0.011474609, -0.040039062, -0.06088257, 0.0019130707, -0.008300781, -0.03414917, 0.015838623, 0.009292603, 0.0501709, -0.037078857, 0.013267517, 0.03793335, -0.005886078, -0.035583496, 0.01512146, -0.0006556511, 0.023468...
http://arxiv.org/abs/2311.00310v1
Unsupervised Lexical Simplification with Context Augmentation
Takashi Wada, Timothy Baldwin, Jey Han Lau
cs.CL, cs.AI
We propose a new unsupervised lexical simplification method that uses only monolingual data and pre-trained language models. Given a target word and its context, our method generates substitutes based on the target context and also additional contexts sampled from monolingual data. We conduct experiments in English, Po...
http://arxiv.org/pdf/2311.00310v1
2023-11-01T05:48:05
2023-11-01T05:48:05
{ "title": [ -0.014022827, 0.0036392212, -0.049102783, -0.089660645, 0.02178955, 0.021575928, -0.06359863, 0.011657715, 0.01864624, 0.047546387, -0.017852783, 0.016189575, 0.020080566, -0.0051994324, -0.032714844, 0.01675415, -0.0033512115, 0.030...
http://arxiv.org/abs/2311.00306v1
Probing Explicit and Implicit Gender Bias through LLM Conditional Text Generation
Xiangjue Dong, Yibo Wang, Philip S. Yu, James Caverlee
cs.CL
Large Language Models (LLMs) can generate biased and toxic responses. Yet most prior work on LLM gender bias evaluation requires predefined gender-related phrases or gender stereotypes, which are challenging to be comprehensively collected and are limited to explicit bias evaluation. In addition, we believe that instan...
http://arxiv.org/pdf/2311.00306v1
2023-11-01T05:31:46
2023-11-01T05:31:46
{ "title": [ -0.06976318, -0.0017385483, -0.07183838, -0.022384644, -0.025390625, -0.0020256042, -0.06652832, 0.0041542053, 0.000934124, 0.021072388, -0.05407715, 0.018493652, -0.003396988, -0.002702713, -0.04006958, -0.021774292, -0.047821045, 0...
http://arxiv.org/abs/2311.00301v1
Detecting Syllable-Level Pronunciation Stress with A Self-Attention Model
Wang Weiying, Nakajima Akinori
cs.SD, cs.CL, eess.AS
One precondition of effective oral communication is that words should be pronounced clearly, especially for non-native speakers. Word stress is the key to clear and correct English, and misplacement of syllable stress may lead to misunderstandings. Thus, knowing the stress level is important for English speakers and le...
http://arxiv.org/pdf/2311.00301v1
2023-11-01T05:05:49
2023-11-01T05:05:49
{ "title": [ 0.0006623268, 0.012207031, -0.04345703, -0.05227661, -0.011016846, -0.0069389343, -0.04675293, 0.008079529, 0.010765076, 0.037902832, -0.0062179565, -0.015548706, 0.007987976, -0.028930664, -0.062408447, 0.0019950867, -0.0025539398, ...
http://arxiv.org/abs/2311.00300v1
Entity Alignment Method of Science and Technology Patent based on Graph Convolution Network and Information Fusion
Runze Fang, Yawen Li, Yingxia Shao, Zeli Guan, Zhe Xue
cs.CL
The entity alignment of science and technology patents aims to link the equivalent entities in the knowledge graph of different science and technology patent data sources. Most entity alignment methods only use graph neural network to obtain the embedding of graph structure or use attribute text description to obtain s...
http://arxiv.org/pdf/2311.00300v1
2023-11-01T05:04:55
2023-11-01T05:04:55
{ "title": [ -0.021194458, 0.010635376, -0.074645996, -0.013877869, -0.0075302124, -0.046783447, -0.0079422, -0.00010704994, -0.010894775, 0.048431396, -0.006767273, -0.0087509155, 0.042175293, -0.040893555, -0.013900757, 0.003967285, -0.0056381226, ...
http://arxiv.org/abs/2311.00296v1
Semantic Representation Learning of Scientific Literature based on Adaptive Feature and Graph Neural Network
Hongrui Gao, Yawen Li, Meiyu Liang, Zeli Guan, Zhe Xue
cs.CL
Because most of the scientific literature data is unmarked, it makes semantic representation learning based on unsupervised graph become crucial. At the same time, in order to enrich the features of scientific literature, a learning method of semantic representation of scientific literature based on adaptive features a...
http://arxiv.org/pdf/2311.00296v1
2023-11-01T05:00:44
2023-11-01T05:00:44
{ "title": [ -0.045715332, 0.00048708916, -0.06414795, -0.041992188, 0.0038967133, -0.0088272095, -0.035186768, 0.005168915, -0.012207031, 0.029510498, 0.026428223, -0.0015010834, -0.019165039, -0.027572632, -0.04360962, -0.0025501251, 0.001282692, ...
http://arxiv.org/abs/2311.00292v1
IBADR: an Iterative Bias-Aware Dataset Refinement Framework for Debiasing NLU models
Xiaoyue Wang, Xin Liu, Lijie Wang, Yaoxiang Wang, Jinsong Su, Hua Wu
cs.CL
As commonly-used methods for debiasing natural language understanding (NLU) models, dataset refinement approaches heavily rely on manual data analysis, and thus maybe unable to cover all the potential biased features. In this paper, we propose IBADR, an Iterative Bias-Aware Dataset Refinement framework, which debiases ...
http://arxiv.org/pdf/2311.00292v1
2023-11-01T04:50:38
2023-11-01T04:50:38
{ "title": [ -0.011024475, 0.018112183, -0.044555664, -0.08227539, -0.007675171, 0.010864258, -0.10083008, -0.028411865, -0.003124237, 0.035461426, -0.031280518, -0.0008945465, 0.0034656525, -0.02027893, -0.039367676, -0.016143799, -0.012817383, ...
http://arxiv.org/abs/2311.00288v1
Active Instruction Tuning: Improving Cross-Task Generalization by Training on Prompt Sensitive Tasks
Po-Nien Kung, Fan Yin, Di Wu, Kai-Wei Chang, Nanyun Peng
cs.CL, cs.AI
Instruction tuning (IT) achieves impressive zero-shot generalization results by training large language models (LLMs) on a massive amount of diverse tasks with instructions. However, how to select new tasks to improve the performance and generalizability of IT models remains an open question. Training on all existing t...
http://arxiv.org/pdf/2311.00288v1
2023-11-01T04:40:05
2023-11-01T04:40:05
{ "title": [ -0.03515625, -0.004589081, -0.009605408, -0.02458191, -0.012886047, -0.010848999, -0.021377563, -0.01499939, -0.0040664673, 0.0362854, -0.019546509, 0.025405884, -0.017837524, -0.006351471, -0.027893066, -0.00091934204, -0.0007519722, ...
http://arxiv.org/abs/2311.00287v1
Knowledge-Infused Prompting: Assessing and Advancing Clinical Text Data Generation with Large Language Models
Ran Xu, Hejie Cui, Yue Yu, Xuan Kan, Wenqi Shi, Yuchen Zhuang, Wei Jin, Joyce Ho, Carl Yang
cs.CL, cs.AI, cs.LG, q-bio.QM
Clinical natural language processing requires methods that can address domain-specific challenges, such as complex medical terminology and clinical contexts. Recently, large language models (LLMs) have shown promise in this domain. Yet, their direct deployment can lead to privacy issues and are constrained by resources...
http://arxiv.org/pdf/2311.00287v1
2023-11-01T04:37:28
2023-11-01T04:37:28
{ "title": [ -0.040618896, -0.020584106, -0.041992188, -0.082336426, -0.02470398, 0.018798828, -0.097595215, 0.010406494, 0.0025196075, 0.022399902, -0.027404785, 0.009979248, 0.0022201538, -0.01789856, -0.032226562, 0.004962921, 0.0019702911, 0....
http://arxiv.org/abs/2311.00286v2
JADE: A Linguistics-based Safety Evaluation Platform for LLM
Mi Zhang, Xudong Pan, Min Yang
cs.CL, cs.AI, cs.CR, cs.LG
In this paper, we present JADE, a targeted linguistic fuzzing platform which strengthens the linguistic complexity of seed questions to simultaneously and consistently break a wide range of widely-used LLMs categorized in three groups: eight open-sourced Chinese, six commercial Chinese and four commercial English LLMs....
http://arxiv.org/pdf/2311.00286v2
2023-11-02T02:36:47
2023-11-01T04:36:45
{ "title": [ -0.06390381, -0.01727295, -0.049835205, -0.07232666, -0.0070724487, 0.01965332, -0.08081055, 0.015274048, -0.025756836, 0.022369385, -0.016555786, 0.018951416, -0.0045051575, -0.028244019, -0.004398346, -0.0058021545, -0.033599854, 0...
http://arxiv.org/abs/2311.00273v1
SoulChat: Improving LLMs' Empathy, Listening, and Comfort Abilities through Fine-tuning with Multi-turn Empathy Conversations
Yirong Chen, Xiaofen Xing, Jingkai Lin, Huimin Zheng, Zhenyu Wang, Qi Liu, Xiangmin Xu
cs.CL
Large language models (LLMs) have been widely applied in various fields due to their excellent capability for memorizing knowledge and chain of thought (CoT). When these language models are applied in the field of psychological counseling, they often rush to provide universal advice. However, when users seek psychologi...
http://arxiv.org/pdf/2311.00273v1
2023-11-01T03:49:52
2023-11-01T03:49:52
{ "title": [ -0.009124756, -0.036346436, 0.016799927, -0.0027885437, -0.06088257, 0.003982544, -0.039978027, -0.02218628, -0.0006608963, 0.034301758, -0.027648926, 0.042053223, 0.02305603, -0.023101807, 0.0022697449, 0.0018119812, -0.019226074, -...
http://arxiv.org/abs/2311.00268v1
Syntactic Inductive Bias in Transformer Language Models: Especially Helpful for Low-Resource Languages?
Luke Gessler, Nathan Schneider
cs.CL
A line of work on Transformer-based language models such as BERT has attempted to use syntactic inductive bias to enhance the pretraining process, on the theory that building syntactic structure into the training process should reduce the amount of data needed for training. But such methods are often tested for high-re...
http://arxiv.org/pdf/2311.00268v1
2023-11-01T03:32:46
2023-11-01T03:32:46
{ "title": [ -0.022537231, -0.0021800995, -0.022949219, -0.045898438, -0.0140686035, -0.0013227463, -0.0770874, -0.015235901, -0.012466431, 0.017456055, -0.037902832, 0.0069236755, 0.029830933, -0.0256958, -0.0016698837, -0.008529663, -0.029205322, ...
http://arxiv.org/abs/2311.00262v1
Plug-and-Play Policy Planner for Large Language Model Powered Dialogue Agents
Yang Deng, Wenxuan Zhang, Wai Lam, See-Kiong Ng, Tat-Seng Chua
cs.CL, cs.AI
Proactive dialogues serve as a practical yet challenging dialogue problem in the era of large language models (LLMs), where the dialogue policy planning is the key to improving the proactivity of LLMs. Most existing studies enable the dialogue policy planning of LLMs using various prompting schemes or iteratively enhan...
http://arxiv.org/pdf/2311.00262v1
2023-11-01T03:20:16
2023-11-01T03:20:16
{ "title": [ -0.013160706, -0.020858765, -0.029769897, -0.07751465, -0.009239197, 0.009025574, -0.044433594, 0.031097412, -0.03793335, 0.008636475, -0.010719299, 0.008903503, -0.013397217, -0.025665283, 0.0013494492, -0.013160706, -0.013587952, 0...
http://arxiv.org/abs/2311.00258v1
Noisy Exemplars Make Large Language Models More Robust: A Domain-Agnostic Behavioral Analysis
Hongyi Zheng, Abulhair Saparov
cs.CL, cs.LG
Recent advances in prompt engineering enable large language models (LLMs) to solve multi-hop logical reasoning problems with impressive accuracy. However, there is little existing work investigating the robustness of LLMs with few-shot prompting techniques. Therefore, we introduce a systematic approach to test the robu...
http://arxiv.org/pdf/2311.00258v1
2023-11-01T03:15:05
2023-11-01T03:15:05
{ "title": [ -0.0020771027, 0.004005432, -0.01713562, -0.074523926, 0.004272461, -0.0038986206, -0.06500244, -0.01586914, 0.0030097961, 0.01927185, -0.033050537, 0.011451721, 0.0061569214, -0.01374054, -0.03894043, -0.00605011, -0.017562866, 0.05...
http://arxiv.org/abs/2311.00237v1
The Mystery and Fascination of LLMs: A Comprehensive Survey on the Interpretation and Analysis of Emergent Abilities
Yuxiang Zhou, Jiazheng Li, Yanzheng Xiang, Hanqi Yan, Lin Gui, Yulan He
cs.CL
Understanding emergent abilities, such as in-context learning (ICL) and chain-of-thought (CoT) prompting in large language models (LLMs), is of utmost importance. This importance stems not only from the better utilization of these capabilities across various tasks, but also from the proactive identification and mitigat...
http://arxiv.org/pdf/2311.00237v1
2023-11-01T02:40:42
2023-11-01T02:40:42
{ "title": [ -0.031188965, -0.004360199, -0.027633667, 0.006904602, -0.0259552, -0.021408081, -0.00035834312, -0.013053894, -0.005970001, -0.003932953, -0.035308838, 0.021820068, -0.0057296753, -0.030715942, 0.0047302246, -0.008262634, -0.012832642, ...
http://arxiv.org/abs/2311.00233v1
Distort, Distract, Decode: Instruction-Tuned Model Can Refine its Response from Noisy Instructions
Taehyeon Kim, Joonkee Kim, Gihun Lee, Se-Young Yun
cs.CL
While instruction-tuned language models have demonstrated impressive zero-shot generalization, these models often struggle to generate accurate responses when faced with instructions that fall outside their training set. This paper presents Instructive Decoding (ID), a simple yet effective approach that augments the ef...
http://arxiv.org/pdf/2311.00233v1
2023-11-01T02:31:35
2023-11-01T02:31:35
{ "title": [ -0.03366089, -0.025650024, 0.009765625, -0.009437561, -0.048217773, -0.0001989603, -0.036224365, -0.024276733, -0.007534027, 0.0335083, -0.06237793, 0.02279663, -0.004459381, 0.006755829, -0.03930664, -0.039215088, 0.00749588, 0.0207...
http://arxiv.org/abs/2311.00223v1
Is GPT Powerful Enough to Analyze the Emotions of Memes?
Jingjing Wang, Joshua Luo, Grace Yang, Allen Hong, Feng Luo
cs.CL, cs.MM
Large Language Models (LLMs), representing a significant achievement in artificial intelligence (AI) research, have demonstrated their ability in a multitude of tasks. This project aims to explore the capabilities of GPT-3.5, a leading example of LLMs, in processing the sentiment analysis of Internet memes. Memes, whic...
http://arxiv.org/pdf/2311.00223v1
2023-11-01T01:57:48
2023-11-01T01:57:48
{ "title": [ 0.051727295, 0.013244629, -0.0009303093, 0.028579712, -0.028198242, -0.034423828, 0.017837524, -0.064575195, 0.022460938, 0.02532959, 0.00030112267, -0.012420654, 0.017547607, -0.046051025, -0.02583313, 0.014640808, -0.0011262894, 0....
http://arxiv.org/abs/2311.00208v1
Transformers as Recognizers of Formal Languages: A Survey on Expressivity
Lena Strobl, William Merrill, Gail Weiss, David Chiang, Dana Angluin
cs.LG, cs.CL, cs.FL, cs.LO
As transformers have gained prominence in natural language processing, some researchers have investigated theoretically what problems they can and cannot solve, by treating problems as formal languages. Exploring questions such as this will help to compare transformers with other models, and transformer variants with o...
http://arxiv.org/pdf/2311.00208v1
2023-11-01T00:38:26
2023-11-01T00:38:26
{ "title": [ -0.01259613, 0.017578125, -0.013755798, -0.0859375, 0.029647827, 0.026824951, -0.04711914, -0.006668091, 0.0051116943, 0.005607605, -0.037902832, -0.0001604557, -0.0107803345, -0.012969971, -0.030044556, -0.006401062, -0.025985718, 0...
http://arxiv.org/abs/2311.00204v1
Continuous Training and Fine-tuning for Domain-Specific Language Models in Medical Question Answering
Zhen Guo, Yining Hua
cs.CL, cs.AI
Large language models exhibit promising general capabilities but often lack specialized knowledge for domain-specific tasks. Developing domain experts from a base model enables a range of applications without prohibitive training costs. This work demonstrates a method using continuous training and instruction fine-tuni...
http://arxiv.org/pdf/2311.00204v1
2023-11-01T00:18:00
2023-11-01T00:18:00
{ "title": [ -0.013641357, -0.0035133362, -0.03878784, -0.06896973, 0.0062713623, 0.002357483, -0.04156494, 0.016479492, -0.013648987, 0.03729248, 0.0031833649, 0.006790161, 0.009651184, -0.018798828, -0.052246094, 0.010894775, 0.0024585724, 0.02...
http://arxiv.org/abs/2311.00189v1
XAI-CLASS: Explanation-Enhanced Text Classification with Extremely Weak Supervision
Daniel Hajialigol, Hanwen Liu, Xuan Wang
cs.CL, cs.AI
Text classification aims to effectively categorize documents into pre-defined categories. Traditional methods for text classification often rely on large amounts of manually annotated training data, making the process time-consuming and labor-intensive. To address this issue, recent studies have focused on weakly-super...
http://arxiv.org/pdf/2311.00189v1
2023-10-31T23:24:22
2023-10-31T23:24:22
{ "title": [ -0.020706177, -0.015457153, -0.027160645, -0.061309814, -0.00365448, 0.0027065277, -0.033294678, -0.003944397, 0.0057029724, -0.012145996, -0.030395508, 0.022354126, -0.025299072, -0.0064735413, -0.019134521, -0.014015198, -0.032318115, ...
http://arxiv.org/abs/2311.01472v1
Relation Extraction from News Articles (RENA): A Tool for Epidemic Surveillance
Jaeff Hong, Duong Dung, Danielle Hutchinson, Zubair Akhtar, Rosalie Chen, Rebecca Dawson, Aditya Joshi, Samsung Lim, C Raina MacIntyre, Deepti Gurdasani
cs.CL, cs.AI
Relation Extraction from News Articles (RENA) is a browser-based tool designed to extract key entities and their semantic relationships in English language news articles related to infectious diseases. Constructed using the React framework, this system presents users with an elegant and user-friendly interface. It enab...
http://arxiv.org/pdf/2311.01472v1
2023-10-31T23:14:54
2023-10-31T23:14:54
{ "title": [ -0.016189575, -0.0014791489, -0.06213379, -0.08166504, 0.0143585205, 0.0004105568, -0.053619385, -0.014007568, 0.013008118, 0.014335632, -0.026901245, 0.010536194, -0.033355713, -0.009933472, -0.04559326, -0.012046814, -0.015686035, ...
http://arxiv.org/abs/2311.00176v2
ChipNeMo: Domain-Adapted LLMs for Chip Design
Mingjie Liu, Teodor-Dumitru Ene, Robert Kirby, Chris Cheng, Nathaniel Pinckney, Rongjian Liang, Jonah Alben, Himyanshu Anand, Sanmitra Banerjee, Ismet Bayraktaroglu, Bonita Bhaskaran, Bryan Catanzaro, Arjun Chaudhuri, Sharon Clay, Bill Dally, Laura Dang, Parikshit Deshpande, Siddhanth Dhodhi, Sameer Halepete, Eric Hill...
cs.CL
ChipNeMo aims to explore the applications of large language models (LLMs) for industrial chip design. Instead of directly deploying off-the-shelf commercial or open-source LLMs, we instead adopt the following domain adaptation techniques: custom tokenizers, domain-adaptive continued pretraining, supervised fine-tuning ...
http://arxiv.org/pdf/2311.00176v2
2023-11-13T23:07:55
2023-10-31T22:35:58
{ "title": [ -0.08306885, -0.017486572, -0.05126953, -0.016555786, -0.025146484, 0.00248909, -0.048553467, -0.0004119873, 0.0020217896, 0.021743774, -0.020935059, 0.010253906, 0.00006043911, -0.0209198, -0.019973755, 0.0053520203, -0.011169434, 0...
http://arxiv.org/abs/2311.00172v1
Robust Safety Classifier for Large Language Models: Adversarial Prompt Shield
Jinhwa Kim, Ali Derakhshan, Ian G. Harris
cs.CL, cs.AI
Large Language Models' safety remains a critical concern due to their vulnerability to adversarial attacks, which can prompt these systems to produce harmful responses. In the heart of these systems lies a safety classifier, a computational model trained to discern and mitigate potentially harmful, offensive, or unethi...
http://arxiv.org/pdf/2311.00172v1
2023-10-31T22:22:10
2023-10-31T22:22:10
{ "title": [ -0.016296387, 0.018478394, -0.053894043, -0.0703125, 0.024765015, -0.0062446594, -0.065979004, -0.0070381165, -0.023620605, 0.024597168, -0.009681702, 0.018661499, -0.002401352, -0.020889282, -0.008850098, -0.0041275024, -0.014144897, ...
http://arxiv.org/abs/2311.00161v1
Beyond Denouncing Hate: Strategies for Countering Implied Biases and Stereotypes in Language
Jimin Mun, Emily Allaway, Akhila Yerukola, Laura Vianna, Sarah-Jane Leslie, Maarten Sap
cs.CL, cs.AI
Counterspeech, i.e., responses to counteract potential harms of hateful speech, has become an increasingly popular solution to address online hate speech without censorship. However, properly countering hateful language requires countering and dispelling the underlying inaccurate stereotypes implied by such language. I...
http://arxiv.org/pdf/2311.00161v1
2023-10-31T21:33:46
2023-10-31T21:33:46
{ "title": [ 0.023544312, -0.0003209114, -0.026657104, -0.048553467, 0.015602112, -0.033050537, -0.00009781122, -0.013648987, -0.019073486, 0.029067993, -0.005531311, 0.012420654, 0.0065727234, -0.026443481, -0.035858154, -0.0070762634, -0.017745972,...
http://arxiv.org/abs/2311.00159v1
Longer Fixations, More Computation: Gaze-Guided Recurrent Neural Networks
Xinting Huang, Jiajing Wan, Ioannis Kritikos, Nora Hollenstein
cs.CL
Humans read texts at a varying pace, while machine learning models treat each token in the same way in terms of a computational process. Therefore, we ask, does it help to make models act more like humans? In this paper, we convert this intuition into a set of novel models with fixation-guided parallel RNNs or layers a...
http://arxiv.org/pdf/2311.00159v1
2023-10-31T21:32:11
2023-10-31T21:32:11
{ "title": [ 0.006702423, -0.0055999756, -0.0077934265, -0.03414917, 0.00598526, -0.02571106, 0.014602661, -0.013397217, 0.022659302, 0.0262146, -0.025939941, -0.008712769, 0.0016946793, -0.015106201, -0.052612305, 0.008918762, 0.005088806, 0.012...
http://arxiv.org/abs/2311.00143v1
Two-Stage Classifier for Campaign Negativity Detection using Axis Embeddings: A Case Study on Tweets of Political Users during 2021 Presidential Election in Iran
Fatemeh Rajabi, Ali Mohades
cs.LG, cs.AI, cs.CL
In elections around the world, the candidates may turn their campaigns toward negativity due to the prospect of failure and time pressure. In the digital age, social media platforms such as Twitter are rich sources of political discourse. Therefore, despite the large amount of data that is published on Twitter, the aut...
http://arxiv.org/pdf/2311.00143v1
2023-10-31T20:31:41
2023-10-31T20:31:41
{ "title": [ 0.06933594, -0.006175995, 0.039520264, -0.019851685, -0.004383087, -0.0039901733, 0.027145386, -0.021331787, 0.012268066, 0.0036888123, -0.05014038, 0.024093628, 0.0028133392, -0.030212402, -0.014572144, -0.028564453, -0.016464233, 0...
http://arxiv.org/abs/2311.00128v2
On the effect of curriculum learning with developmental data for grammar acquisition
Mattia Opper, J. Morrison, N. Siddharth
cs.CL
This work explores the degree to which grammar acquisition is driven by language `simplicity' and the source modality (speech vs. text) of data. Using BabyBERTa as a probe, we find that grammar acquisition is largely driven by exposure to speech data, and in particular through exposure to two of the BabyLM training cor...
http://arxiv.org/pdf/2311.00128v2
2023-11-03T16:42:33
2023-10-31T20:05:30
{ "title": [ -0.027908325, -0.024810791, -0.057922363, -0.036743164, -0.018814087, -0.030090332, -0.011489868, 0.033050537, -0.038391113, -0.009407043, -0.029647827, 0.043762207, -0.07086182, -0.01739502, 0.0042533875, -0.037719727, -0.029876709, ...
http://arxiv.org/abs/2311.00117v1
BadLlama: cheaply removing safety fine-tuning from Llama 2-Chat 13B
Pranav Gade, Simon Lermen, Charlie Rogers-Smith, Jeffrey Ladish
cs.CL
Llama 2-Chat is a collection of large language models that Meta developed and released to the public. While Meta fine-tuned Llama 2-Chat to refuse to output harmful content, we hypothesize that public access to model weights enables bad actors to cheaply circumvent Llama 2-Chat's safeguards and weaponize Llama 2's capa...
http://arxiv.org/pdf/2311.00117v1
2023-10-31T19:45:15
2023-10-31T19:45:15
{ "title": [ 0.010391235, -0.002817154, -0.0011205673, -0.02986145, -0.012626648, 0.023025513, -0.020355225, -0.020507812, -0.060394287, 0.03781128, 0.028137207, 0.03652954, -0.02003479, -0.05645752, 0.03466797, -0.01663208, 0.0063285828, -0.0010...
http://arxiv.org/abs/2311.00116v1
BERTwich: Extending BERT's Capabilities to Model Dialectal and Noisy Text
Aarohi Srivastava, David Chiang
cs.CL
Real-world NLP applications often deal with nonstandard text (e.g., dialectal, informal, or misspelled text). However, language models like BERT deteriorate in the face of dialect variation or noise. How do we push BERT's modeling capabilities to encompass nonstandard text? Fine-tuning helps, but it is designed for spe...
http://arxiv.org/pdf/2311.00116v1
2023-10-31T19:44:50
2023-10-31T19:44:50
{ "title": [ -0.021270752, -0.015319824, -0.0073623657, -0.062469482, -0.019134521, 0.020309448, -0.05947876, -0.0013790131, 0.0065841675, 0.04788208, -0.02470398, 0.0043640137, 0.043182373, -0.00031638145, -0.06149292, 0.021530151, -0.014251709, ...
http://arxiv.org/abs/2311.00059v1
The Generative AI Paradox: "What It Can Create, It May Not Understand"
Peter West, Ximing Lu, Nouha Dziri, Faeze Brahman, Linjie Li, Jena D. Hwang, Liwei Jiang, Jillian Fisher, Abhilasha Ravichander, Khyathi Chandu, Benjamin Newman, Pang Wei Koh, Allyson Ettinger, Yejin Choi
cs.AI, cs.CL, cs.CV, cs.LG
The recent wave of generative AI has sparked unprecedented global attention, with both excitement and concern over potentially superhuman levels of artificial intelligence: models now take only seconds to produce outputs that would challenge or exceed the capabilities even of expert humans. At the same time, models sti...
http://arxiv.org/pdf/2311.00059v1
2023-10-31T18:07:07
2023-10-31T18:07:07
{ "title": [ -0.037506104, 0.0070228577, -0.03881836, 0.009094238, -0.01083374, -0.021469116, 0.0008621216, -0.04623413, 0.007587433, -0.0054512024, -0.052581787, 0.007865906, -0.028045654, -0.03378296, -0.027893066, -0.035308838, 0.00856781, 0.0...
http://arxiv.org/abs/2311.00047v1
Grounding Visual Illusions in Language: Do Vision-Language Models Perceive Illusions Like Humans?
Yichi Zhang, Jiayi Pan, Yuchen Zhou, Rui Pan, Joyce Chai
cs.AI, cs.CL, cs.CV, cs.LG
Vision-Language Models (VLMs) are trained on vast amounts of data captured by humans emulating our understanding of the world. However, known as visual illusions, human's perception of reality isn't always faithful to the physical world. This raises a key question: do VLMs have the similar kind of illusions as humans d...
http://arxiv.org/pdf/2311.00047v1
2023-10-31T18:01:11
2023-10-31T18:01:11
{ "title": [ -0.04272461, 0.008255005, -0.029083252, -0.005214691, -0.011604309, -0.018325806, -0.015930176, -0.028945923, 0.03805542, 0.033203125, -0.051696777, -0.008430481, 0.03945923, -0.007926941, -0.0647583, 0.013648987, -0.013618469, 0.053...
http://arxiv.org/abs/2310.20707v1
What's In My Big Data?
Yanai Elazar, Akshita Bhagia, Ian Magnusson, Abhilasha Ravichander, Dustin Schwenk, Alane Suhr, Pete Walsh, Dirk Groeneveld, Luca Soldaini, Sameer Singh, Hanna Hajishirzi, Noah A. Smith, Jesse Dodge
cs.CL, cs.LG
Large text corpora are the backbone of language models. However, we have a limited understanding of the content of these corpora, including general statistics, quality, social factors, and inclusion of evaluation data (contamination). In this work, we propose What's In My Big Data? (WIMBD), a platform and a set of sixt...
http://arxiv.org/pdf/2310.20707v1
2023-10-31T17:59:38
2023-10-31T17:59:38
{ "title": [ 0.009407043, 0.016921997, -0.060577393, -0.045562744, -0.021514893, -0.021743774, -0.05593872, 0.008590698, -0.040618896, 0.0031604767, -0.03277588, 0.009651184, -0.07080078, -0.028182983, 0.016937256, -0.049224854, 0.048217773, -0.0...
http://arxiv.org/abs/2310.20703v1
Vanishing Gradients in Reinforcement Finetuning of Language Models
Noam Razin, Hattie Zhou, Omid Saremi, Vimal Thilak, Arwen Bradley, Preetum Nakkiran, Joshua Susskind, Etai Littwin
cs.LG, cs.AI, cs.CL, stat.ML
Pretrained language models are commonly aligned with human preferences and downstream tasks via reinforcement finetuning (RFT), which entails maximizing a (possibly learned) reward function using policy gradient algorithms. This work highlights a fundamental optimization obstacle in RFT: we prove that the expected grad...
http://arxiv.org/pdf/2310.20703v1
2023-10-31T17:59:05
2023-10-31T17:59:05
{ "title": [ 0.009490967, 0.0044441223, -0.009025574, -0.031082153, 0.001821518, -0.01524353, -0.006515503, -0.016708374, -0.0345459, 0.037902832, -0.040222168, 0.024932861, -0.0075035095, -0.0030765533, -0.033416748, -0.01727295, -0.023620605, 0...
http://arxiv.org/abs/2310.20697v1
Text-Transport: Toward Learning Causal Effects of Natural Language
Victoria Lin, Louis-Philippe Morency, Eli Ben-Michael
cs.CL, stat.ME
As language technologies gain prominence in real-world settings, it is important to understand how changes to language affect reader perceptions. This can be formalized as the causal effect of varying a linguistic attribute (e.g., sentiment) on a reader's response to the text. In this paper, we introduce Text-Transport...
http://arxiv.org/pdf/2310.20697v1
2023-10-31T17:56:51
2023-10-31T17:56:51
{ "title": [ -0.012664795, -0.019607544, -0.020492554, -0.061279297, -0.005302429, 0.003004074, -0.02243042, -0.0004274845, -0.0071525574, 0.008110046, -0.008071899, 0.018814087, -0.01638794, -0.023147583, -0.01687622, -0.00447464, -0.01625061, 0...
http://arxiv.org/abs/2310.20689v2
Learning From Mistakes Makes LLM Better Reasoner
Shengnan An, Zexiong Ma, Zeqi Lin, Nanning Zheng, Jian-Guang Lou, Weizhu Chen
cs.CL, cs.AI
Large language models (LLMs) recently exhibited remarkable reasoning capabilities on solving math problems. To further improve this capability, this work proposes Learning from Mistakes (LeMa), akin to human learning processes. Consider a human student who failed to solve a math problem, he will learn from what mistake...
http://arxiv.org/pdf/2310.20689v2
2023-11-14T02:37:55
2023-10-31T17:52:22
{ "title": [ -0.010475159, -0.019821167, -0.010482788, -0.025177002, -0.059020996, 0.009361267, -0.052978516, -0.005126953, -0.024658203, 0.014053345, -0.023498535, -0.0102005005, -0.025817871, -0.042633057, 0.024673462, -0.036499023, -0.009155273, ...
http://arxiv.org/abs/2310.20656v1
Non-Compositionality in Sentiment: New Data and Analyses
Verna Dankers, Christopher G. Lucas
cs.CL
When natural language phrases are combined, their meaning is often more than the sum of their parts. In the context of NLP tasks such as sentiment analysis, where the meaning of a phrase is its sentiment, that still applies. Many NLP studies on sentiment analysis, however, focus on the fact that sentiment computations ...
http://arxiv.org/pdf/2310.20656v1
2023-10-31T17:25:07
2023-10-31T17:25:07
{ "title": [ 0.008483887, -0.0023517609, -0.028289795, -0.050445557, 0.008964539, -0.014083862, -0.021057129, -0.018341064, -0.0020503998, 0.042419434, -0.024612427, 0.0061416626, -0.009986877, 0.0047187805, -0.0713501, -0.0041770935, -0.019119263, ...
http://arxiv.org/abs/2310.20633v1
Defining a New NLP Playground
Sha Li, Chi Han, Pengfei Yu, Carl Edwards, Manling Li, Xingyao Wang, Yi R. Fung, Charles Yu, Joel R. Tetreault, Eduard H. Hovy, Heng Ji
cs.CL
The recent explosion of performance of large language models (LLMs) has changed the field of Natural Language Processing (NLP) more abruptly and seismically than any other shift in the field's 80-year history. This has resulted in concerns that the field will become homogenized and resource-intensive. The new status qu...
http://arxiv.org/pdf/2310.20633v1
2023-10-31T17:02:33
2023-10-31T17:02:33
{ "title": [ 0.024414062, -0.018630981, 0.0025901794, -0.054718018, -0.054534912, 0.0072517395, -0.03503418, 0.011856079, -0.06500244, 0.0061187744, -0.057250977, 0.03010559, -0.040924072, -0.011474609, 0.043060303, -0.050750732, 0.003583908, 0.0...
http://arxiv.org/abs/2310.20620v1
The Unreasonable Effectiveness of Random Target Embeddings for Continuous-Output Neural Machine Translation
Evgeniia Tokarchuk, Vlad Niculae
cs.CL, cs.LG
Continuous-output neural machine translation (CoNMT) replaces the discrete next-word prediction problem with an embedding prediction. The semantic structure of the target embedding space (i.e., closeness of related words) is intuitively believed to be crucial. We challenge this assumption and show that completely rando...
http://arxiv.org/pdf/2310.20620v1
2023-10-31T16:53:10
2023-10-31T16:53:10
{ "title": [ -0.01651001, -0.0045547485, -0.013298035, -0.026687622, 0.0012016296, -0.00299263, -0.0017280579, 0.00013434887, -0.0143585205, 0.03652954, 0.0023059845, 0.014961243, 0.0023212433, -0.016967773, -0.036071777, 0.013885498, -0.002653122, ...
http://arxiv.org/abs/2310.20589v1
Increasing The Performance of Cognitively Inspired Data-Efficient Language Models via Implicit Structure Building
Omar Momen, David Arps, Laura Kallmeyer
cs.CL
In this paper, we describe our submission to the BabyLM Challenge 2023 shared task on data-efficient language model (LM) pretraining (Warstadt et al., 2023). We train transformer-based masked language models that incorporate unsupervised predictions about hierarchical sentence structure into the model architecture. Con...
http://arxiv.org/pdf/2310.20589v1
2023-10-31T16:26:36
2023-10-31T16:26:36
{ "title": [ -0.024383545, -0.009864807, -0.020385742, -0.08123779, -0.012077332, 0.01727295, -0.07836914, 0.013916016, 0.02041626, 0.021118164, -0.03050232, -0.002840042, 0.0039405823, -0.015930176, -0.03881836, 0.011001587, -0.01272583, 0.03649...
http://arxiv.org/abs/2310.20588v1
Zero-Shot Medical Information Retrieval via Knowledge Graph Embedding
Yuqi Wang, Zeqiang Wang, Wei Wang, Qi Chen, Kaizhu Huang, Anh Nguyen, Suparna De
cs.CL, cs.IR
In the era of the Internet of Things (IoT), the retrieval of relevant medical information has become essential for efficient clinical decision-making. This paper introduces MedFusionRank, a novel approach to zero-shot medical information retrieval (MIR) that combines the strengths of pre-trained language models and sta...
http://arxiv.org/pdf/2310.20588v1
2023-10-31T16:26:33
2023-10-31T16:26:33
{ "title": [ 0.005466461, -0.002538681, -0.04675293, -0.06201172, -0.007724762, -0.013717651, -0.033599854, 0.0061149597, 0.012420654, 0.042236328, -0.0065307617, 0.004837036, 0.00029277802, -0.019332886, -0.049743652, 0.0073890686, 0.018112183, ...
http://arxiv.org/abs/2310.20558v1
Breaking the Token Barrier: Chunking and Convolution for Efficient Long Text Classification with BERT
Aman Jaiswal, Evangelos Milios
cs.CL, cs.AI, I.2.7
Transformer-based models, specifically BERT, have propelled research in various NLP tasks. However, these models are limited to a maximum token limit of 512 tokens. Consequently, this makes it non-trivial to apply it in a practical setting with long input. Various complex methods have claimed to overcome this limit, bu...
http://arxiv.org/pdf/2310.20558v1
2023-10-31T15:41:08
2023-10-31T15:41:08
{ "title": [ -0.005054474, -0.01776123, -0.03086853, -0.08135986, 0.0044288635, 0.0037612915, -0.039916992, 0.010604858, 0.0016212463, 0.030838013, -0.030227661, 0.022277832, 0.006969452, -0.0052986145, -0.043762207, -0.008666992, -0.021560669, 0...
http://arxiv.org/abs/2310.20550v2
CapsFusion: Rethinking Image-Text Data at Scale
Qiying Yu, Quan Sun, Xiaosong Zhang, Yufeng Cui, Fan Zhang, Yue Cao, Xinlong Wang, Jingjing Liu
cs.CV, cs.AI, cs.CL, cs.LG
Large multimodal models demonstrate remarkable generalist ability to perform diverse multimodal tasks in a zero-shot manner. Large-scale web-based image-text pairs contribute fundamentally to this success, but suffer from excessive noise. Recent studies use alternative captions synthesized by captioning models and have...
http://arxiv.org/pdf/2310.20550v2
2023-11-02T11:25:20
2023-10-31T15:31:39
{ "title": [ -0.02192688, -0.0070762634, -0.04159546, -0.043182373, 0.004432678, -0.011253357, -0.021972656, -0.026275635, 0.02331543, 0.025527954, -0.039001465, 0.016815186, -0.028930664, 0.003704071, -0.059509277, -0.0032539368, 0.011222839, 0....
http://arxiv.org/abs/2310.20501v1
LLMs may Dominate Information Access: Neural Retrievers are Biased Towards LLM-Generated Texts
Sunhao Dai, Yuqi Zhou, Liang Pang, Weihao Liu, Xiaolin Hu, Yong Liu, Xiao Zhang, Jun Xu
cs.IR, cs.AI, cs.CL
Recently, the emergence of large language models (LLMs) has revolutionized the paradigm of information retrieval (IR) applications, especially in web search. With their remarkable capabilities in generating human-like texts, LLMs have created enormous texts on the Internet. As a result, IR systems in the LLMs era are f...
http://arxiv.org/pdf/2310.20501v1
2023-10-31T14:42:23
2023-10-31T14:42:23
{ "title": [ -0.014884949, -0.013404846, -0.019836426, 0.023101807, -0.05505371, -0.0115356445, -0.020614624, -0.019378662, -0.03616333, 0.034484863, -0.029388428, 0.031341553, 0.017349243, -0.02368164, 0.029281616, -0.018966675, -0.022994995, 0....
http://arxiv.org/abs/2310.20499v2
Leveraging Word Guessing Games to Assess the Intelligence of Large Language Models
Tian Liang, Zhiwei He, Jen-tse Huang, Wenxuan Wang, Wenxiang Jiao, Rui Wang, Yujiu Yang, Zhaopeng Tu, Shuming Shi, Xing Wang
cs.CL
The automatic evaluation of LLM-based agent intelligence is critical in developing advanced LLM-based agents. Although considerable effort has been devoted to developing human-annotated evaluation datasets, such as AlpacaEval, existing techniques are costly, time-consuming, and lack adaptability. In this paper, inspire...
http://arxiv.org/pdf/2310.20499v2
2023-11-06T02:27:34
2023-10-31T14:37:42
{ "title": [ -0.050994873, -0.004020691, -0.030075073, -0.007320404, -0.022384644, -0.013824463, -0.033935547, -0.022384644, 0.002319336, 0.021896362, -0.036315918, 0.0022087097, -0.007701874, -0.013244629, -0.045776367, -0.011787415, -0.012741089, ...
http://arxiv.org/abs/2310.20479v1
Multi-User MultiWOZ: Task-Oriented Dialogues among Multiple Users
Yohan Jo, Xinyan Zhao, Arijit Biswas, Nikoletta Basiou, Vincent Auvray, Nikolaos Malandrakis, Angeliki Metallinou, Alexandros Potamianos
cs.CL
While most task-oriented dialogues assume conversations between the agent and one user at a time, dialogue systems are increasingly expected to communicate with multiple users simultaneously who make decisions collaboratively. To facilitate development of such systems, we release the Multi-User MultiWOZ dataset: task-o...
http://arxiv.org/pdf/2310.20479v1
2023-10-31T14:12:07
2023-10-31T14:12:07
{ "title": [ -0.015205383, -0.025253296, -0.060424805, -0.04949951, 0.006324768, -0.024398804, -0.016601562, 0.016937256, -0.007007599, 0.06689453, 0.012870789, 0.010917664, 0.0093688965, -0.016830444, -0.050933838, 0.010925293, -0.019622803, 0.0...
http://arxiv.org/abs/2310.20470v1
Representativeness as a Forgotten Lesson for Multilingual and Code-switched Data Collection and Preparation
A. Seza Doğruöz, Sunayana Sitaram, Zheng-Xin Yong
cs.CL
Multilingualism is widespread around the world and code-switching (CSW) is a common practice among different language pairs/tuples across locations and regions. However, there is still not much progress in building successful CSW systems, despite the recent advances in Massive Multilingual Language Models (MMLMs). We i...
http://arxiv.org/pdf/2310.20470v1
2023-10-31T14:04:07
2023-10-31T14:04:07
{ "title": [ -0.011665344, -0.005329132, -0.030548096, -0.054138184, -0.0037765503, -0.006980896, -0.023971558, -0.00035500526, -0.0046691895, 0.02999878, -0.025482178, -0.0029735565, -0.000086188316, -0.026992798, -0.031204224, -0.02279663, -0.02052...
http://arxiv.org/abs/2310.20467v1
ACL Anthology Helper: A Tool to Retrieve and Manage Literature from ACL Anthology
Chen Tang, Frank Guerin, Chenghua Lin
cs.CL, cs.AI
The ACL Anthology is an online repository that serves as a comprehensive collection of publications in the field of natural language processing (NLP) and computational linguistics (CL). This paper presents a tool called ``ACL Anthology Helper''. It automates the process of parsing and downloading papers along with thei...
http://arxiv.org/pdf/2310.20467v1
2023-10-31T13:59:05
2023-10-31T13:59:05
{ "title": [ 0.00036287308, 0.017913818, -0.04071045, -0.00050640106, -0.015991211, -0.019454956, -0.04434204, -0.031341553, -0.00957489, 0.07196045, 0.056549072, -0.023773193, 0.081970215, -0.058685303, -0.014083862, 0.01322937, 0.023468018, -0....
http://arxiv.org/abs/2310.20456v1
Towards a Deep Understanding of Multilingual End-to-End Speech Translation
Haoran Sun, Xiaohu Zhao, Yikun Lei, Shaolin Zhu, Deyi Xiong
cs.CL
In this paper, we employ Singular Value Canonical Correlation Analysis (SVCCA) to analyze representations learnt in a multilingual end-to-end speech translation model trained over 22 languages. SVCCA enables us to estimate representational similarity across languages and layers, enhancing our understanding of the funct...
http://arxiv.org/pdf/2310.20456v1
2023-10-31T13:50:55
2023-10-31T13:50:55
{ "title": [ -0.0178833, -0.0061302185, -0.016174316, -0.07556152, 0.009284973, 0.007881165, -0.04849243, 0.020309448, -0.01727295, 0.05050659, 0.012466431, 0.008850098, 0.025680542, -0.030349731, -0.038391113, 0.0284729, -0.0075187683, 0.0298461...
http://arxiv.org/abs/2310.20440v1
The SourceData-NLP dataset: integrating curation into scientific publishing for training large language models
Jorge Abreu-Vicente, Hannah Sonntag, Thomas Eidens, Thomas Lemberger
cs.CL
Introduction: The scientific publishing landscape is expanding rapidly, creating challenges for researchers to stay up-to-date with the evolution of the literature. Natural Language Processing (NLP) has emerged as a potent approach to automating knowledge extraction from this vast amount of publications and preprints. ...
http://arxiv.org/pdf/2310.20440v1
2023-10-31T13:22:38
2023-10-31T13:22:38
{ "title": [ -0.04095459, -0.022277832, -0.04144287, -0.054260254, -0.022155762, 0.009597778, -0.06304932, -0.0041999817, 0.025985718, 0.012680054, -0.026794434, 0.0011692047, 0.007511139, -0.00932312, -0.035064697, 0.019973755, -0.015945435, 0.0...
http://arxiv.org/abs/2310.20410v2
FollowBench: A Multi-level Fine-grained Constraints Following Benchmark for Large Language Models
Yuxin Jiang, Yufei Wang, Xingshan Zeng, Wanjun Zhong, Liangyou Li, Fei Mi, Lifeng Shang, Xin Jiang, Qun Liu, Wei Wang
cs.CL
The ability to follow instructions is crucial for Large Language Models (LLMs) to handle various real-world applications. Existing benchmarks primarily focus on evaluating pure response quality, rather than assessing whether the response follows constraints stated in the instruction. To fill this research gap, in this ...
http://arxiv.org/pdf/2310.20410v2
2023-11-14T11:01:06
2023-10-31T12:32:38
{ "title": [ -0.04046631, -0.019317627, -0.038513184, -0.09777832, -0.003993988, 0.026641846, -0.095703125, 0.02760315, 0.010665894, 0.0231781, -0.015357971, 0.008895874, 0.013847351, -0.020355225, -0.0345459, 0.014434814, -0.009101868, 0.0270996...
http://arxiv.org/abs/2310.20384v1
Do large language models solve verbal analogies like children do?
Claire E. Stevenson, Mathilde ter Veen, Rochelle Choenni, Han L. J. van der Maas, Ekaterina Shutova
cs.CL, cs.AI
Analogy-making lies at the heart of human cognition. Adults solve analogies such as \textit{Horse belongs to stable like chicken belongs to ...?} by mapping relations (\textit{kept in}) and answering \textit{chicken coop}. In contrast, children often use association, e.g., answering \textit{egg}. This paper investigate...
http://arxiv.org/pdf/2310.20384v1
2023-10-31T11:49:11
2023-10-31T11:49:11
{ "title": [ -0.012840271, -0.002702713, 0.008049011, -0.023284912, -0.052124023, 0.0052452087, -0.04257202, -0.02557373, 0.005203247, 0.03475952, -0.027435303, -0.015533447, 0.02394104, -0.02305603, -0.03010559, -0.0146484375, -0.0039901733, 0.0...
http://arxiv.org/abs/2310.20352v1
AMERICANO: Argument Generation with Discourse-driven Decomposition and Agent Interaction
Zhe Hu, Hou Pong Chan, Yu Yin
cs.CL
Argument generation is a challenging task in natural language processing, which requires rigorous reasoning and proper content organization. Inspired by recent chain-of-thought prompting that breaks down a complex task into intermediate steps, we propose Americano, a novel framework with agent interaction for argument ...
http://arxiv.org/pdf/2310.20352v1
2023-10-31T10:47:33
2023-10-31T10:47:33
{ "title": [ -0.0043907166, -0.022033691, -0.012367249, -0.08312988, -0.011749268, 0.021102905, -0.060699463, 0.0034179688, -0.020492554, 0.020690918, 0.025466919, 0.013885498, -0.030822754, -0.029785156, -0.02848816, 0.0004544258, -0.029418945, ...
http://arxiv.org/abs/2310.20347v1
Automatic Generators for a Family of Matrix Multiplication Routines with Apache TVM
Guillermo Alaejos, Adrián Castelló, Pedro Alonso-Jordá, Francisco D. Igual, Héctor Martínez, Enrique S. Quintana-Ortí
cs.CL
We explore the utilization of the Apache TVM open source framework to automatically generate a family of algorithms that follow the approach taken by popular linear algebra libraries, such as GotoBLAS2, BLIS and OpenBLAS, in order to obtain high-performance blocked formulations of the general matrix multiplication (GEM...
http://arxiv.org/pdf/2310.20347v1
2023-10-31T10:36:26
2023-10-31T10:36:26
{ "title": [ -0.0013561249, -0.007736206, -0.040283203, -0.064208984, -0.023361206, -0.015777588, -0.037200928, 0.01965332, -0.015930176, 0.049468994, -0.011474609, 0.0077819824, 0.04598999, -0.049102783, 0.0019950867, 0.028427124, 0.047973633, -...
http://arxiv.org/abs/2310.20329v1
InstructCoder: Empowering Language Models for Code Editing
Qisheng Hu, Kaixin Li, Xu Zhao, Yuxi Xie, Tiedong Liu, Hui Chen, Qizhe Xie, Junxian He
cs.CL, cs.SE
Code editing encompasses a variety of pragmatic tasks that developers deal with daily. Despite its relevance and practical usefulness, automatic code editing remains an underexplored area in the evolution of deep learning models, partly due to data scarcity. In this work, we explore the use of large language models (LL...
http://arxiv.org/pdf/2310.20329v1
2023-10-31T10:15:35
2023-10-31T10:15:35
{ "title": [ -0.04373169, -0.03805542, -0.041137695, -0.060180664, -0.009429932, 0.010444641, -0.0496521, 0.020050049, 0.012512207, 0.030090332, -0.006122589, -0.011741638, 0.020996094, -0.026641846, -0.041259766, 0.009681702, -0.023666382, 0.029...
http://arxiv.org/abs/2310.20328v1
ChiSCor: A Corpus of Freely Told Fantasy Stories by Dutch Children for Computational Linguistics and Cognitive Science
Bram M. A. van Dijk, Max J. van Duijn, Suzan Verberne, Marco R. Spruit
cs.CL
In this resource paper we release ChiSCor, a new corpus containing 619 fantasy stories, told freely by 442 Dutch children aged 4-12. ChiSCor was compiled for studying how children render character perspectives, and unravelling language and cognition in development, with computational tools. Unlike existing resources, C...
http://arxiv.org/pdf/2310.20328v1
2023-10-31T10:15:20
2023-10-31T10:15:20
{ "title": [ -0.032409668, -0.02217102, -0.028533936, -0.0062561035, -0.041381836, 0.0069084167, -0.01676941, -0.010032654, 0.011978149, 0.02772522, -0.012321472, -0.017440796, 0.0423584, -0.03970337, -0.026046753, -0.016098022, -0.027236938, 0.0...
http://arxiv.org/abs/2310.20326v1
Erato: Automatizing Poetry Evaluation
Manex Agirrezabal, Hugo Gonçalo Oliveira, Aitor Ormazabal
cs.CL
We present Erato, a framework designed to facilitate the automated evaluation of poetry, including that generated by poetry generation systems. Our framework employs a diverse set of features, and we offer a brief overview of Erato's capabilities and its potential for expansion. Using Erato, we compare and contrast hum...
http://arxiv.org/pdf/2310.20326v1
2023-10-31T10:06:37
2023-10-31T10:06:37
{ "title": [ -0.009246826, -0.0002169609, -0.0262146, -0.045166016, 0.0024433136, -0.012374878, -0.007358551, -0.013946533, 0.033172607, 0.035095215, -0.021270752, 0.0031871796, -0.0023422241, -0.013648987, -0.068725586, -0.0027942657, -0.023849487, ...
http://arxiv.org/abs/2310.20322v1
FA Team at the NTCIR-17 UFO Task
Yuki Okumura, Masato Fujitake
cs.CL, cs.IR
The FA team participated in the Table Data Extraction (TDE) and Text-to-Table Relationship Extraction (TTRE) tasks of the NTCIR-17 Understanding of Non-Financial Objects in Financial Reports (UFO). This paper reports our approach to solving the problems and discusses the official results. We successfully utilized vario...
http://arxiv.org/pdf/2310.20322v1
2023-10-31T09:56:12
2023-10-31T09:56:12
{ "title": [ -0.0026435852, -0.011772156, -0.010284424, 0.016967773, -0.044799805, -0.008605957, 0.0076141357, -0.0035152435, -0.04067993, 0.06616211, 0.0087890625, 0.012748718, 0.062347412, -0.01473999, 0.019546509, -0.00806427, -0.015434265, 0....
http://arxiv.org/abs/2310.20320v1
Theory of Mind in Large Language Models: Examining Performance of 11 State-of-the-Art models vs. Children Aged 7-10 on Advanced Tests
Max J. van Duijn, Bram M. A. van Dijk, Tom Kouwenhoven, Werner de Valk, Marco R. Spruit, Peter van der Putten
cs.CL, cs.AI
To what degree should we ascribe cognitive capacities to Large Language Models (LLMs), such as the ability to reason about intentions and beliefs known as Theory of Mind (ToM)? Here we add to this emerging debate by (i) testing 11 base- and instruction-tuned LLMs on capabilities relevant to ToM beyond the dominant fals...
http://arxiv.org/pdf/2310.20320v1
2023-10-31T09:55:07
2023-10-31T09:55:07
{ "title": [ -0.002861023, -0.021972656, 0.0031108856, -0.008056641, -0.033050537, -0.019424438, -0.017105103, -0.019958496, -0.019897461, 0.013473511, -0.010269165, -0.008407593, -0.017700195, -0.030319214, -0.018722534, -0.036102295, -0.021728516, ...