id
stringlengths
33
34
title
stringlengths
6
218
authors
stringlengths
5
6.98k
categories
stringlengths
5
188
summary
stringlengths
34
3.26k
link_pdf
stringlengths
33
34
updated
timestamp[s]
published
timestamp[s]
embeddings
dict
http://arxiv.org/abs/2305.13406v2
DADA: Dialect Adaptation via Dynamic Aggregation of Linguistic Rules
Yanchen Liu, William Held, Diyi Yang
cs.CL
Existing large language models (LLMs) that mainly focus on Standard American English (SAE) often lead to significantly worse performance when being applied to other English dialects. While existing mitigations tackle discrepancies for individual target dialects, they assume access to high-accuracy dialect identificatio...
http://arxiv.org/pdf/2305.13406v2
2023-10-23T06:21:17
2023-05-22T18:43:31
{ "title": [ -0.01940918, -0.012763977, -0.05758667, -0.057281494, 0.00655365, -0.00008946657, -0.045135498, 0.014297485, -0.008361816, 0.059509277, 0.0013608932, 0.001707077, 0.039367676, -0.020370483, -0.041931152, 0.020812988, 0.011772156, 0.0...
http://arxiv.org/abs/2305.13403v1
GATology for Linguistics: What Syntactic Dependencies It Knows
Yuqian Dai, Serge Sharoff, Marc de Kamps
cs.CL
Graph Attention Network (GAT) is a graph neural network which is one of the strategies for modeling and representing explicit syntactic knowledge and can work with pre-trained models, such as BERT, in downstream tasks. Currently, there is still a lack of investigation into how GAT learns syntactic knowledge from the pe...
http://arxiv.org/pdf/2305.13403v1
2023-05-22T18:34:12
2023-05-22T18:34:12
{ "title": [ -0.014076233, -0.009597778, -0.030563354, -0.08538818, -0.013717651, 0.016723633, -0.070617676, 0.008071899, 0.00030636787, 0.020080566, -0.021713257, 0.00522995, -0.007774353, -0.026535034, -0.038085938, 0.0008716583, 0.005050659, 0...
http://arxiv.org/abs/2305.13401v1
A study of conceptual language similarity: comparison and evaluation
Haotian Ye, Yihong Liu, Hinrich Schütze
cs.CL
An interesting line of research in natural language processing (NLP) aims to incorporate linguistic typology to bridge linguistic diversity and assist the research of low-resource languages. While most works construct linguistic similarity measures based on lexical or typological features, such as word order and verbal...
http://arxiv.org/pdf/2305.13401v1
2023-05-22T18:28:02
2023-05-22T18:28:02
{ "title": [ -0.030822754, -0.009109497, -0.039855957, -0.058044434, 0.02218628, 0.0022697449, -0.03729248, -0.014884949, 0.003545761, 0.035217285, -0.018554688, 0.0158844, -0.0024852753, -0.0112838745, -0.051940918, 0.0064353943, -0.022003174, 0...
http://arxiv.org/abs/2305.13395v2
BioDEX: Large-Scale Biomedical Adverse Drug Event Extraction for Real-World Pharmacovigilance
Karel D'Oosterlinck, François Remy, Johannes Deleu, Thomas Demeester, Chris Develder, Klim Zaporojets, Aneiss Ghodsi, Simon Ellershaw, Jack Collins, Christopher Potts
cs.CL
Timely and accurate extraction of Adverse Drug Events (ADE) from biomedical literature is paramount for public safety, but involves slow and costly manual labor. We set out to improve drug safety monitoring (pharmacovigilance, PV) through the use of Natural Language Processing (NLP). We introduce BioDEX, a large-scale ...
http://arxiv.org/pdf/2305.13395v2
2023-10-20T15:51:45
2023-05-22T18:15:57
{ "title": [ -0.011634827, -0.004257202, -0.03744507, -0.093322754, 0.01687622, 0.001288414, -0.04272461, 0.00793457, -0.004749298, 0.02255249, -0.008705139, 0.036743164, -0.010948181, -0.018997192, -0.028030396, 0.0041046143, -0.010124207, 0.022...
http://arxiv.org/abs/2305.13388v1
The neural dynamics of auditory word recognition and integration
Jon Gauthier, Roger Levy
cs.CL, q-bio.NC
Listeners recognize and integrate words in rapid and noisy everyday speech by combining expectations about upcoming content with incremental sensory evidence. We present a computational model of word recognition which formalizes this perceptual process in Bayesian decision theory. We fit this model to explain scalp EEG...
http://arxiv.org/pdf/2305.13388v1
2023-05-22T18:06:32
2023-05-22T18:06:32
{ "title": [ -0.009498596, 0.019226074, -0.038208008, -0.020889282, 0.0022850037, -0.013160706, -0.0034503937, -0.007358551, -0.0075263977, 0.053710938, -0.022232056, -0.0012569427, 0.012428284, -0.011489868, -0.062469482, 0.0028038025, -0.010398865,...
http://arxiv.org/abs/2305.13386v2
Can LLMs facilitate interpretation of pre-trained language models?
Basel Mousi, Nadir Durrani, Fahim Dalvi
cs.CL
Work done to uncover the knowledge encoded within pre-trained language models rely on annotated corpora or human-in-the-loop methods. However, these approaches are limited in terms of scalability and the scope of interpretation. We propose using a large language model, ChatGPT, as an annotator to enable fine-grained in...
http://arxiv.org/pdf/2305.13386v2
2023-10-20T13:24:00
2023-05-22T18:03:13
{ "title": [ -0.02583313, -0.0070724487, -0.022598267, -0.04699707, -0.06567383, 0.010803223, -0.1126709, -0.01234436, -0.033294678, 0.014724731, -0.011444092, 0.0020542145, 0.001449585, -0.04574585, 0.03842163, -0.019989014, 0.01373291, 0.007926...
http://arxiv.org/abs/2305.13309v1
Evaluating Factual Consistency of Texts with Semantic Role Labeling
Jing Fan, Dennis Aumiller, Michael Gertz
cs.CL
Automated evaluation of text generation systems has recently seen increasing attention, particularly checking whether generated text stays truthful to input sources. Existing methods frequently rely on an evaluation using task-specific language models, which in turn allows for little interpretability of generated score...
http://arxiv.org/pdf/2305.13309v1
2023-05-22T17:59:42
2023-05-22T17:59:42
{ "title": [ -0.027557373, -0.0024490356, -0.06036377, -0.062561035, 0.0026760101, -0.0021438599, -0.04421997, 0.006214142, -0.035583496, 0.029830933, 0.005264282, 0.020858765, -0.016220093, -0.007911682, -0.045959473, -0.01626587, -0.036102295, ...
http://arxiv.org/abs/2305.13304v1
RecurrentGPT: Interactive Generation of (Arbitrarily) Long Text
Wangchunshu Zhou, Yuchen Eleanor Jiang, Peng Cui, Tiannan Wang, Zhenxin Xiao, Yifan Hou, Ryan Cotterell, Mrinmaya Sachan
cs.CL, cs.LG
The fixed-size context of Transformer makes GPT models incapable of generating arbitrarily long text. In this paper, we introduce RecurrentGPT, a language-based simulacrum of the recurrence mechanism in RNNs. RecurrentGPT is built upon a large language model (LLM) such as ChatGPT and uses natural language to simulate t...
http://arxiv.org/pdf/2305.13304v1
2023-05-22T17:58:10
2023-05-22T17:58:10
{ "title": [ -0.014770508, -0.02079773, -0.033172607, -0.05215454, -0.0074501038, -0.00062704086, -0.03842163, -0.0107040405, -0.007896423, 0.026657104, -0.0032043457, 0.020614624, -0.014076233, -0.020217896, -0.029830933, 0.009216309, -0.012763977, ...
http://arxiv.org/abs/2305.13303v3
Towards Unsupervised Recognition of Token-level Semantic Differences in Related Documents
Jannis Vamvas, Rico Sennrich
cs.CL
Automatically highlighting words that cause semantic differences between two documents could be useful for a wide range of applications. We formulate recognizing semantic differences (RSD) as a token-level regression task and study three unsupervised approaches that rely on a masked language model. To assess the approa...
http://arxiv.org/pdf/2305.13303v3
2023-10-20T12:27:41
2023-05-22T17:58:04
{ "title": [ -0.027313232, 0.0064735413, -0.06335449, -0.05343628, 0.013420105, -0.0129470825, -0.05255127, -0.010475159, 0.022598267, 0.015548706, -0.044647217, 0.009460449, -0.004085541, 0.009727478, -0.04144287, -0.01109314, -0.0036506653, 0.0...
http://arxiv.org/abs/2305.13302v1
Language-Agnostic Bias Detection in Language Models
Abdullatif Köksal, Omer Faruk Yalcin, Ahmet Akbiyik, M. Tahir Kilavuz, Anna Korhonen, Hinrich Schütze
cs.CL
Pretrained language models (PLMs) are key components in NLP, but they contain strong social biases. Quantifying these biases is challenging because current methods focusing on fill-the-mask objectives are sensitive to slight changes in input. To address this, we propose LABDet, a robust language-agnostic method for eva...
http://arxiv.org/pdf/2305.13302v1
2023-05-22T17:58:01
2023-05-22T17:58:01
{ "title": [ -0.028945923, 0.0044021606, -0.048736572, -0.042938232, 0.005924225, -0.0096588135, -0.066711426, -0.020431519, 0.013336182, 0.032043457, -0.021408081, -0.01260376, 0.02519226, -0.013694763, -0.04559326, 0.010063171, -0.032714844, 0....
http://arxiv.org/abs/2305.13300v3
Adaptive Chameleon or Stubborn Sloth: Revealing the Behavior of Large Language Models in Knowledge Conflicts
Jian Xie, Kai Zhang, Jiangjie Chen, Renze Lou, Yu Su
cs.CL, cs.AI
By providing external information to large language models (LLMs), tool augmentation (including retrieval augmentation) has emerged as a promising solution for addressing the limitations of LLMs' static parametric memory. However, how receptive are LLMs to such external evidence, especially when the evidence conflicts ...
http://arxiv.org/pdf/2305.13300v3
2023-10-03T02:47:41
2023-05-22T17:57:41
{ "title": [ -0.0385437, -0.009674072, -0.0637207, -0.044677734, -0.0019273758, -0.018661499, -0.04296875, 0.0015592575, -0.021072388, 0.022842407, -0.030944824, 0.0015659332, -0.009643555, -0.0068626404, -0.047912598, -0.03012085, 0.00013697147, ...
http://arxiv.org/abs/2305.13299v1
Measuring Inductive Biases of In-Context Learning with Underspecified Demonstrations
Chenglei Si, Dan Friedman, Nitish Joshi, Shi Feng, Danqi Chen, He He
cs.CL, cs.AI, cs.LG
In-context learning (ICL) is an important paradigm for adapting large language models (LLMs) to new tasks, but the generalization behavior of ICL remains poorly understood. We investigate the inductive biases of ICL from the perspective of feature bias: which feature ICL is more likely to use given a set of underspecif...
http://arxiv.org/pdf/2305.13299v1
2023-05-22T17:56:31
2023-05-22T17:56:31
{ "title": [ 0.018707275, 0.011245728, -0.04043579, -0.017807007, 0.023208618, -0.052337646, 0.025405884, -0.04324341, -0.0035324097, -0.004940033, -0.06781006, 0.02911377, -0.040405273, -0.01171875, -0.0079422, -0.043670654, -0.0007863045, 0.032...
http://arxiv.org/abs/2305.13298v1
DiffusionNER: Boundary Diffusion for Named Entity Recognition
Yongliang Shen, Kaitao Song, Xu Tan, Dongsheng Li, Weiming Lu, Yueting Zhuang
cs.CL
In this paper, we propose DiffusionNER, which formulates the named entity recognition task as a boundary-denoising diffusion process and thus generates named entities from noisy spans. During training, DiffusionNER gradually adds noises to the golden entity boundaries by a fixed forward diffusion process and learns a r...
http://arxiv.org/pdf/2305.13298v1
2023-05-22T17:56:12
2023-05-22T17:56:12
{ "title": [ -0.013916016, -0.0008816719, -0.05822754, -0.08062744, 0.034851074, 0.0056915283, -0.024963379, 0.016662598, 0.017059326, 0.015266418, -0.04196167, 0.03967285, 0.00081825256, 0.003332138, -0.02470398, 0.015777588, -0.019699097, 0.058...
http://arxiv.org/abs/2305.13297v2
Investigating the Role of Feed-Forward Networks in Transformers Using Parallel Attention and Feed-Forward Net Design
Shashank Sonkar, Richard G. Baraniuk
cs.CL
This paper investigates the key role of Feed-Forward Networks (FFNs) in transformer models by utilizing the Parallel Attention and Feed-Forward Net Design (PAF) architecture, and comparing it to their Series Attention and Feed-Forward Net Design (SAF) counterparts. Central to the effectiveness of PAF are two main assum...
http://arxiv.org/pdf/2305.13297v2
2023-05-25T17:01:13
2023-05-22T17:56:09
{ "title": [ 0.028182983, 0.022445679, -0.034484863, -0.03656006, 0.014411926, -0.032287598, 0.0022411346, -0.023712158, -0.008087158, 0.041290283, -0.028167725, 0.019073486, -0.009544373, -0.018325806, -0.028900146, 0.004798889, -0.008682251, 0....
http://arxiv.org/abs/2305.14387v2
AlpacaFarm: A Simulation Framework for Methods that Learn from Human Feedback
Yann Dubois, Xuechen Li, Rohan Taori, Tianyi Zhang, Ishaan Gulrajani, Jimmy Ba, Carlos Guestrin, Percy Liang, Tatsunori B. Hashimoto
cs.LG, cs.AI, cs.CL
Large language models (LLMs) such as ChatGPT have seen widespread adoption due to their ability to follow user instructions well. Developing these LLMs involves a complex yet poorly understood workflow requiring training with human feedback. Replicating and understanding this instruction-following process faces three m...
http://arxiv.org/pdf/2305.14387v2
2023-08-01T05:46:21
2023-05-22T17:55:50
{ "title": [ -0.03805542, -0.006801605, -0.032562256, -0.028366089, -0.0049552917, -0.0053367615, -0.04751587, -0.0345459, -0.016525269, 0.022705078, -0.01889038, 0.024353027, -0.019561768, -0.018661499, -0.01878357, 0.0023002625, 0.011619568, -0...
http://arxiv.org/abs/2305.18569v1
Fairness of ChatGPT
Yunqi Li, Yongfeng Zhang
cs.LG, cs.AI, cs.CL, cs.CY
Understanding and addressing unfairness in LLMs are crucial for responsible AI deployment. However, there is a limited availability of quantitative analyses and in-depth studies regarding fairness evaluations in LLMs, especially when applying LLMs to high-stakes fields. This work aims to fill this gap by providing a sy...
http://arxiv.org/pdf/2305.18569v1
2023-05-22T17:51:56
2023-05-22T17:51:56
{ "title": [ 0.030090332, -0.016555786, -0.01121521, -0.014076233, -0.03793335, -0.019729614, -0.029190063, -0.0121154785, 0.0054244995, 0.03289795, -0.04714966, 0.02104187, 0.014343262, -0.029342651, 0.011940002, 0.009033203, -0.018234253, 0.043...
http://arxiv.org/abs/2305.13286v1
How do languages influence each other? Studying cross-lingual data sharing during LLM fine-tuning
Rochelle Choenni, Dan Garrette, Ekaterina Shutova
cs.CL
Multilingual large language models (MLLMs) are jointly trained on data from many different languages such that representation of individual languages can benefit from other languages' data. Impressive performance on zero-shot cross-lingual transfer shows that these models are capable of exploiting data from other langu...
http://arxiv.org/pdf/2305.13286v1
2023-05-22T17:47:41
2023-05-22T17:47:41
{ "title": [ -0.03540039, -0.019195557, -0.064331055, -0.0446167, -0.026062012, -0.014274597, -0.051330566, 0.028930664, -0.04284668, 0.031677246, -0.029418945, 0.036865234, 0.0038661957, -0.0209198, 0.01335907, -0.023345947, 0.00038027763, 0.056...
http://arxiv.org/abs/2305.13282v1
Is Fine-tuning Needed? Pre-trained Language Models Are Near Perfect for Out-of-Domain Detection
Rheeya Uppaal, Junjie Hu, Yixuan Li
cs.CL, cs.LG
Out-of-distribution (OOD) detection is a critical task for reliable predictions over text. Fine-tuning with pre-trained language models has been a de facto procedure to derive OOD detectors with respect to in-distribution (ID) data. Despite its common use, the understanding of the role of fine-tuning and its necessity ...
http://arxiv.org/pdf/2305.13282v1
2023-05-22T17:42:44
2023-05-22T17:42:44
{ "title": [ 0.053588867, 0.006614685, 0.032440186, -0.0496521, -0.016418457, -0.018310547, -0.022445679, -0.030075073, -0.025482178, 0.022857666, -0.010505676, 0.0003349781, -0.009780884, -0.035339355, -0.0058555603, -0.04269409, -0.028045654, 0...
http://arxiv.org/abs/2305.13281v1
LM vs LM: Detecting Factual Errors via Cross Examination
Roi Cohen, May Hamri, Mor Geva, Amir Globerson
cs.CL
A prominent weakness of modern language models (LMs) is their tendency to generate factually incorrect text, which hinders their usability. A natural question is whether such factual errors can be detected automatically. Inspired by truth-seeking mechanisms in law, we propose a factuality evaluation framework for LMs t...
http://arxiv.org/pdf/2305.13281v1
2023-05-22T17:42:14
2023-05-22T17:42:14
{ "title": [ -0.010749817, -0.0020313263, -0.052764893, -0.026931763, -0.009086609, -0.022857666, -0.023132324, -0.02961731, -0.023086548, 0.023834229, -0.04144287, 0.01751709, 0.019439697, -0.01612854, 0.00554657, -0.0030784607, -0.028549194, 0....
http://arxiv.org/abs/2305.13276v2
Evaluating ChatGPT's Performance for Multilingual and Emoji-based Hate Speech Detection
Mithun Das, Saurabh Kumar Pandey, Animesh Mukherjee
cs.CL, cs.LG
Hate speech is a severe issue that affects many online platforms. So far, several studies have been performed to develop robust hate speech detection systems. Large language models like ChatGPT have recently shown a great promise in performing several tasks, including hate speech detection. However, it is crucial to co...
http://arxiv.org/pdf/2305.13276v2
2023-05-23T03:39:44
2023-05-22T17:36:58
{ "title": [ 0.0052833557, -0.0062713623, -0.011680603, -0.018707275, 0.000954628, -0.01084137, -0.0072746277, -0.009101868, -0.016433716, 0.040039062, -0.01234436, 0.01184082, 0.022079468, -0.013496399, -0.022232056, 0.0011301041, -0.029876709, ...
http://arxiv.org/abs/2305.14386v1
Let GPT be a Math Tutor: Teaching Math Word Problem Solvers with Customized Exercise Generation
Zhenwen Liang, Wenhao Yu, Tanmay Rajpurohit, Peter Clark, Xiangliang Zhang, Ashwin Kaylan
cs.LG, cs.AI, cs.CL
In this paper, we present a novel approach for distilling math word problem solving capabilities from large language models (LLMs) into smaller, more efficient student models. Our approach is designed to consider the student model's weaknesses and foster a tailored learning experience by generating targeted exercises a...
http://arxiv.org/pdf/2305.14386v1
2023-05-22T17:36:14
2023-05-22T17:36:14
{ "title": [ 0.00541687, -0.00504303, -0.0020694733, -0.0042381287, -0.021606445, -0.025787354, 0.006778717, -0.028335571, -0.034057617, 0.010978699, -0.029708862, 0.01663208, 0.0017490387, -0.027282715, 0.020141602, -0.021240234, 0.0012836456, 0...
http://arxiv.org/abs/2305.13272v2
CLASS: A Design Framework for building Intelligent Tutoring Systems based on Learning Science principles
Shashank Sonkar, Naiming Liu, Debshila Basu Mallick, Richard G. Baraniuk
cs.CL
We present a design framework called Conversational Learning with Analytical Step-by-Step Strategies (CLASS) for building advanced Intelligent Tutoring Systems (ITS) powered by high-performance Large Language Models (LLMs). The CLASS framework empowers ITS with two key capabilities. First, through a carefully curated s...
http://arxiv.org/pdf/2305.13272v2
2023-10-26T01:25:23
2023-05-22T17:35:05
{ "title": [ -0.009429932, -0.022064209, -0.040374756, -0.005088806, -0.00079774857, -0.04345703, 0.017486572, -0.016555786, -0.011619568, 0.0256958, -0.04043579, 0.019195557, -0.0052223206, -0.0073776245, -0.0317688, -0.026016235, 0.00037026405, ...
http://arxiv.org/abs/2305.13269v2
Chain-of-Knowledge: Grounding Large Language Models via Dynamic Knowledge Adapting over Heterogeneous Sources
Xingxuan Li, Ruochen Zhao, Yew Ken Chia, Bosheng Ding, Shafiq Joty, Soujanya Poria, Lidong Bing
cs.CL
We present chain-of-knowledge (CoK), a novel framework that augments large language models (LLMs) by dynamically incorporating grounding information from heterogeneous sources. It results in more factual rationales and reduced hallucination in generation. Specifically, CoK consists of three stages: reasoning preparatio...
http://arxiv.org/pdf/2305.13269v2
2023-10-03T12:30:20
2023-05-22T17:34:23
{ "title": [ -0.027694702, -0.010498047, -0.06021118, -0.08758545, 0.012245178, 0.00422287, -0.06085205, 0.022537231, 0.009460449, 0.018478394, -0.02947998, 0.0010499954, 0.0033836365, -0.010063171, -0.03805542, -0.009757996, 0.005355835, 0.03341...
http://arxiv.org/abs/2305.13267v1
Enhance Reasoning Ability of Visual-Language Models via Large Language Models
Yueting Yang, Xintong Zhang, Wenjuan Han
cs.CL, cs.AI
Pre-trained visual language models (VLM) have shown excellent performance in image caption tasks. However, it sometimes shows insufficient reasoning ability. In contrast, large language models (LLMs) emerge with powerful reasoning capabilities. Therefore, we propose a method called TReE, which transfers the reasoning a...
http://arxiv.org/pdf/2305.13267v1
2023-05-22T17:33:44
2023-05-22T17:33:44
{ "title": [ -0.020874023, -0.019088745, -0.03567505, -0.06915283, -0.013633728, 0.011604309, -0.06488037, 0.0000269413, 0.03253174, 0.013923645, -0.021530151, -0.022445679, 0.0019741058, -0.024887085, -0.031051636, 0.0046844482, 0.0002156496, 0....
http://arxiv.org/abs/2305.13264v2
Prompting is not a substitute for probability measurements in large language models
Jennifer Hu, Roger Levy
cs.CL, cs.AI
Prompting is now a dominant method for evaluating the linguistic knowledge of large language models (LLMs). While other methods directly read out models' probability distributions over strings, prompting requires models to access this internal information by processing linguistic input, thereby implicitly testing a new...
http://arxiv.org/pdf/2305.13264v2
2023-10-23T14:12:59
2023-05-22T17:33:17
{ "title": [ -0.0075187683, -0.0043640137, -0.010986328, -0.052825928, -0.03186035, -0.005882263, -0.047912598, -0.016357422, 0.00011014938, 0.0026741028, -0.04800415, -0.015716553, -0.0062026978, -0.041137695, -0.0027160645, -0.018051147, 0.00169754...
http://arxiv.org/abs/2305.13256v1
TaskWeb: Selecting Better Source Tasks for Multi-task NLP
Joongwon Kim, Akari Asai, Gabriel Ilharco, Hannaneh Hajishirzi
cs.CL, cs.AI
Recent work in NLP has shown promising results in training models on large amounts of tasks to achieve better generalization. However, it is not well-understood how tasks are related, and how helpful training tasks can be chosen for a new task. In this work, we investigate whether knowing task relationships via pairwis...
http://arxiv.org/pdf/2305.13256v1
2023-05-22T17:27:57
2023-05-22T17:27:57
{ "title": [ -0.016342163, -0.011817932, -0.046691895, -0.06536865, -0.0026664734, -0.0051498413, -0.04046631, 0.0041236877, -0.002336502, 0.025619507, -0.019744873, 0.022994995, -0.011749268, 0.0015707016, -0.035217285, 0.004901886, -0.014564514, ...
http://arxiv.org/abs/2305.13252v1
"According to ..." Prompting Language Models Improves Quoting from Pre-Training Data
Orion Weller, Marc Marone, Nathaniel Weir, Dawn Lawrie, Daniel Khashabi, Benjamin Van Durme
cs.CL, cs.AI
Large Language Models (LLMs) may hallucinate and generate fake information, despite pre-training on factual data. Inspired by the journalistic device of "according to sources", we propose according-to prompting: directing LLMs to ground responses against previously observed text. To quantify this grounding, we propose ...
http://arxiv.org/pdf/2305.13252v1
2023-05-22T17:25:24
2023-05-22T17:25:24
{ "title": [ 0.0015268326, -0.014953613, -0.040649414, -0.04763794, -0.043914795, -0.033996582, -0.068481445, -0.011367798, -0.031585693, 0.003320694, -0.024871826, 0.0099487305, 0.009735107, -0.046722412, 0.023498535, -0.0077209473, 0.0033245087, ...
http://arxiv.org/abs/2305.13246v1
Interactive Natural Language Processing
Zekun Wang, Ge Zhang, Kexin Yang, Ning Shi, Wangchunshu Zhou, Shaochun Hao, Guangzheng Xiong, Yizhi Li, Mong Yuan Sim, Xiuying Chen, Qingqing Zhu, Zhenzhu Yang, Adam Nik, Qi Liu, Chenghua Lin, Shi Wang, Ruibo Liu, Wenhu Chen, Ke Xu, Dayiheng Liu, Yike Guo, Jie Fu
cs.CL, cs.AI
Interactive Natural Language Processing (iNLP) has emerged as a novel paradigm within the field of NLP, aimed at addressing limitations in existing frameworks while aligning with the ultimate goals of artificial intelligence. This paradigm considers language models as agents capable of observing, acting, and receiving ...
http://arxiv.org/pdf/2305.13246v1
2023-05-22T17:18:29
2023-05-22T17:18:29
{ "title": [ -0.01777649, -0.0077590942, -0.029678345, -0.057250977, 0.0053749084, 0.008804321, -0.032928467, -0.011436462, 0.0005440712, 0.029647827, -0.039611816, -0.0018138885, -0.01802063, -0.009033203, -0.045532227, -0.00630188, -0.005558014, ...
http://arxiv.org/abs/2305.13245v2
GQA: Training Generalized Multi-Query Transformer Models from Multi-Head Checkpoints
Joshua Ainslie, James Lee-Thorp, Michiel de Jong, Yury Zemlyanskiy, Federico Lebrón, Sumit Sanghai
cs.CL, cs.LG
Multi-query attention (MQA), which only uses a single key-value head, drastically speeds up decoder inference. However, MQA can lead to quality degradation, and moreover it may not be desirable to train a separate model just for faster inference. We (1) propose a recipe for uptraining existing multi-head language model...
http://arxiv.org/pdf/2305.13245v2
2023-10-24T00:57:50
2023-05-22T17:16:38
{ "title": [ 0.010658264, 0.023925781, -0.040405273, -0.06964111, 0.037628174, 0.000025629997, -0.03475952, -0.01751709, 0.002948761, 0.033233643, -0.024032593, 0.0001642704, 0.0446167, -0.03527832, -0.009635925, 0.016571045, -0.0033454895, 0.015...
http://arxiv.org/abs/2305.13242v1
Deepfake Text Detection in the Wild
Yafu Li, Qintong Li, Leyang Cui, Wei Bi, Longyue Wang, Linyi Yang, Shuming Shi, Yue Zhang
cs.CL
Recent advances in large language models have enabled them to reach a level of text generation comparable to that of humans. These models show powerful capabilities across a wide range of content, including news article writing, story generation, and scientific writing. Such capability further narrows the gap between h...
http://arxiv.org/pdf/2305.13242v1
2023-05-22T17:13:29
2023-05-22T17:13:29
{ "title": [ -0.020111084, 0.020355225, -0.032073975, -0.027648926, 0.0134887695, -0.008850098, -0.010398865, -0.022033691, -0.027816772, 0.035827637, -0.021224976, 0.011703491, 0.0012750626, -0.006351471, -0.027297974, 0.0021343231, -0.016479492, ...
http://arxiv.org/abs/2305.13235v2
SPARSEFIT: Few-shot Prompting with Sparse Fine-tuning for Jointly Generating Predictions and Natural Language Explanations
Jesus Solano, Oana-Maria Camburu, Pasquale Minervini
cs.CL, cs.AI
Explaining the decisions of neural models is crucial for ensuring their trustworthiness at deployment time. Using Natural Language Explanations (NLEs) to justify a model's predictions has recently gained increasing interest. However, this approach usually demands large datasets of human-written NLEs for the ground-trut...
http://arxiv.org/pdf/2305.13235v2
2023-05-23T09:26:37
2023-05-22T17:06:41
{ "title": [ -0.005004883, -0.05831909, -0.007621765, -0.09442139, -0.045928955, 0.022521973, -0.058929443, 0.03567505, 0.0022468567, 0.002916336, -0.006401062, 0.014076233, -0.014205933, -0.032806396, -0.01737976, -0.0023155212, 0.019760132, -0....
http://arxiv.org/abs/2305.13230v2
To Repeat or Not To Repeat: Insights from Scaling LLM under Token-Crisis
Fuzhao Xue, Yao Fu, Wangchunshu Zhou, Zangwei Zheng, Yang You
cs.LG, cs.AI, cs.CL
Recent research has highlighted the importance of dataset size in scaling language models. However, large language models (LLMs) are notoriously token-hungry during pre-training, and high-quality text data on the web is approaching its scaling limit for LLMs. To further enhance LLMs, a straightforward approach is to re...
http://arxiv.org/pdf/2305.13230v2
2023-10-05T14:58:02
2023-05-22T17:02:15
{ "title": [ -0.049438477, -0.010147095, -0.07543945, -0.031158447, -0.035369873, -0.008583069, -0.054138184, 0.01411438, -0.01914978, 0.00008678436, -0.05053711, 0.019515991, -0.033569336, -0.021453857, -0.007785797, -0.03515625, -0.028152466, 0...
http://arxiv.org/abs/2306.01753v1
Preconditioned Visual Language Inference with Weak Supervision
Ehsan Qasemi, Amani R. Maina-Kilaas, Devadutta Dash, Khalid Alsaggaf, Muhao Chen
cs.CL, cs.AI, cs.CV
Humans can infer the affordance of objects by extracting related contextual preconditions for each scenario. For example, upon seeing an image of a broken cup, we can infer that this precondition prevents the cup from being used for drinking. Reasoning with preconditions of commonsense is studied in NLP where the model...
http://arxiv.org/pdf/2306.01753v1
2023-05-22T16:57:52
2023-05-22T16:57:52
{ "title": [ -0.0075531006, -0.016342163, -0.027297974, -0.08746338, -0.015853882, 0.010017395, -0.07788086, 0.0037002563, 0.011764526, 0.00566864, -0.03765869, 0.006843567, -0.010040283, -0.0073051453, -0.017608643, -0.013244629, 0.020431519, 0....
http://arxiv.org/abs/2305.13225v2
Multi-Task Instruction Tuning of LLaMa for Specific Scenarios: A Preliminary Study on Writing Assistance
Yue Zhang, Leyang Cui, Deng Cai, Xinting Huang, Tao Fang, Wei Bi
cs.CL
Proprietary Large Language Models (LLMs), such as ChatGPT, have garnered significant attention due to their exceptional capabilities in handling a diverse range of tasks. Recent studies demonstrate that open-sourced smaller foundational models, such as 7B-size LLaMA, can also display remarkable proficiency in tackling ...
http://arxiv.org/pdf/2305.13225v2
2023-10-09T07:55:26
2023-05-22T16:56:44
{ "title": [ -0.056243896, -0.015731812, -0.035247803, -0.0043411255, -0.051239014, -0.010643005, -0.057403564, -0.0038871765, -0.040924072, 0.046539307, -0.006954193, 0.028152466, 0.01247406, -0.0151901245, -0.0012779236, -0.011787415, -0.03012085, ...
http://arxiv.org/abs/2305.13214v1
Logical Reasoning for Natural Language Inference Using Generated Facts as Atoms
Joe Stacey, Pasquale Minervini, Haim Dubossarsky, Oana-Maria Camburu, Marek Rei
cs.CL, I.2.7
State-of-the-art neural models can now reach human performance levels across various natural language understanding tasks. However, despite this impressive performance, models are known to learn from annotation artefacts at the expense of the underlying task. While interpretability methods can identify influential feat...
http://arxiv.org/pdf/2305.13214v1
2023-05-22T16:45:50
2023-05-22T16:45:50
{ "title": [ -0.03567505, -0.033416748, -0.052001953, -0.08453369, -0.010398865, 0.01473999, -0.06201172, 0.028213501, -0.020751953, 0.027664185, -0.01737976, 0.010215759, -0.0042877197, -0.01574707, -0.024291992, 0.0037612915, -0.01927185, 0.046...
http://arxiv.org/abs/2305.13204v1
Improving Isochronous Machine Translation with Target Factors and Auxiliary Counters
Proyag Pal, Brian Thompson, Yogesh Virkar, Prashant Mathur, Alexandra Chronopoulou, Marcello Federico
cs.CL, cs.SD, eess.AS
To translate speech for automatic dubbing, machine translation needs to be isochronous, i.e. translated speech needs to be aligned with the source in terms of speech durations. We introduce target factors in a transformer model to predict durations jointly with target language phoneme sequences. We also introduce auxil...
http://arxiv.org/pdf/2305.13204v1
2023-05-22T16:36:04
2023-05-22T16:36:04
{ "title": [ -0.03567505, -0.007331848, -0.036010742, -0.051635742, -0.00869751, 0.0052375793, -0.059661865, 0.0020599365, -0.00061035156, 0.060150146, -0.018493652, 0.017929077, 0.030548096, -0.013916016, -0.04977417, 0.020553589, -0.015602112, ...
http://arxiv.org/abs/2305.13199v1
Knowledge-Retrieval Task-Oriented Dialog Systems with Semi-Supervision
Yucheng Cai, Hong Liu, Zhijian Ou, Yi Huang, Junlan Feng
cs.CL
Most existing task-oriented dialog (TOD) systems track dialog states in terms of slots and values and use them to query a database to get relevant knowledge to generate responses. In real-life applications, user utterances are noisier, and thus it is more difficult to accurately track dialog states and correctly secure...
http://arxiv.org/pdf/2305.13199v1
2023-05-22T16:29:20
2023-05-22T16:29:20
{ "title": [ -0.013648987, -0.0056762695, -0.05822754, -0.072631836, 0.013458252, -0.0131073, -0.045135498, 0.028518677, 0.017486572, 0.049072266, -0.021759033, -0.009315491, 0.012123108, -0.002506256, -0.072753906, -0.00793457, -0.0074768066, -0...
http://arxiv.org/abs/2305.13198v1
Multilingual Holistic Bias: Extending Descriptors and Patterns to Unveil Demographic Biases in Languages at Scale
Marta R. Costa-jussà, Pierre Andrews, Eric Smith, Prangthip Hansanti, Christophe Ropers, Elahe Kalbassi, Cynthia Gao, Daniel Licht, Carleigh Wood
cs.CL, I.2.7
We introduce a multilingual extension of the HOLISTICBIAS dataset, the largest English template-based taxonomy of textual people references: MULTILINGUALHOLISTICBIAS. This extension consists of 20,459 sentences in 50 languages distributed across all 13 demographic axes. Source sentences are built from combinations of 1...
http://arxiv.org/pdf/2305.13198v1
2023-05-22T16:29:04
2023-05-22T16:29:04
{ "title": [ 0.00014460087, -0.0029506683, -0.032440186, -0.036895752, 0.0020122528, -0.019454956, -0.019210815, -0.02619934, -0.0013570786, 0.016479492, -0.020965576, 0.008377075, 0.009262085, -0.024780273, -0.0032081604, -0.0017223358, -0.0362854, ...
http://arxiv.org/abs/2305.13197v1
Challenging Decoder helps in Masked Auto-Encoder Pre-training for Dense Passage Retrieval
Zehan Li, Yanzhao Zhang, Dingkun Long, Pengjun Xie
cs.IR, cs.CL
Recently, various studies have been directed towards exploring dense passage retrieval techniques employing pre-trained language models, among which the masked auto-encoder (MAE) pre-training architecture has emerged as the most promising. The conventional MAE framework relies on leveraging the passage reconstruction o...
http://arxiv.org/pdf/2305.13197v1
2023-05-22T16:27:10
2023-05-22T16:27:10
{ "title": [ 0.023956299, 0.008506775, -0.028945923, -0.018478394, -0.026809692, -0.04019165, -0.022109985, -0.024139404, -0.002336502, 0.054138184, -0.018081665, 0.019042969, 0.06878662, -0.036010742, 0.01727295, 0.023727417, 0.007785797, 0.0048...
http://arxiv.org/abs/2305.13194v2
SEAHORSE: A Multilingual, Multifaceted Dataset for Summarization Evaluation
Elizabeth Clark, Shruti Rijhwani, Sebastian Gehrmann, Joshua Maynez, Roee Aharoni, Vitaly Nikolaev, Thibault Sellam, Aditya Siddhant, Dipanjan Das, Ankur P. Parikh
cs.CL
Reliable automatic evaluation of summarization systems is challenging due to the multifaceted and subjective nature of the task. This is especially the case for languages other than English, where human evaluations are scarce. In this work, we introduce SEAHORSE, a dataset for multilingual, multifaceted summarization e...
http://arxiv.org/pdf/2305.13194v2
2023-11-01T22:29:53
2023-05-22T16:25:07
{ "title": [ -0.009140015, 0.0068359375, -0.03778076, -0.08093262, 0.008628845, 0.007949829, -0.070007324, -0.012046814, 0.020324707, 0.042938232, -0.017944336, 0.02696228, -0.0126571655, 0.008079529, -0.06323242, 0.0154800415, -0.007736206, 0.04...
http://arxiv.org/abs/2305.13192v2
SimCSE++: Improving Contrastive Learning for Sentence Embeddings from Two Perspectives
Jiahao Xu, Wei Shao, Lihui Chen, Lemao Liu
cs.CL
This paper improves contrastive learning for sentence embeddings from two perspectives: handling dropout noise and addressing feature corruption. Specifically, for the first perspective, we identify that the dropout noise from negative pairs affects the model's performance. Therefore, we propose a simple yet effective ...
http://arxiv.org/pdf/2305.13192v2
2023-10-20T13:07:23
2023-05-22T16:24:46
{ "title": [ 0.0064201355, -0.008171082, -0.01234436, -0.053588867, 0.0015058517, -0.004623413, -0.019454956, -0.014541626, 0.026016235, 0.010894775, -0.0211792, 0.024017334, -0.018203735, -0.011955261, -0.037628174, 0.001411438, -0.021713257, 0....
http://arxiv.org/abs/2305.13191v1
Taxonomy Expansion for Named Entity Recognition
Karthikeyan K, Yogarshi Vyas, Jie Ma, Giovanni Paolini, Neha Anna John, Shuai Wang, Yassine Benajiba, Vittorio Castelli, Dan Roth, Miguel Ballesteros
cs.CL, cs.AI, cs.LG
Training a Named Entity Recognition (NER) model often involves fixing a taxonomy of entity types. However, requirements evolve and we might need the NER model to recognize additional entity types. A simple approach is to re-annotate entire dataset with both existing and additional entity types and then train the model ...
http://arxiv.org/pdf/2305.13191v1
2023-05-22T16:23:46
2023-05-22T16:23:46
{ "title": [ -0.014076233, 0.0019245148, -0.059936523, -0.07312012, 0.035858154, -0.009895325, -0.02545166, 0.010566711, -0.00019001961, 0.012084961, -0.040924072, 0.02658081, -0.013069153, -0.00440979, -0.023132324, -0.0018091202, -0.008277893, ...
http://arxiv.org/abs/2305.13186v3
SCITAB: A Challenging Benchmark for Compositional Reasoning and Claim Verification on Scientific Tables
Xinyuan Lu, Liangming Pan, Qian Liu, Preslav Nakov, Min-Yen Kan
cs.CL, cs.AI
Current scientific fact-checking benchmarks exhibit several shortcomings, such as biases arising from crowd-sourced claims and an over-reliance on text-based evidence. We present SCITAB, a challenging evaluation dataset consisting of 1.2K expert-verified scientific claims that 1) originate from authentic scientific pub...
http://arxiv.org/pdf/2305.13186v3
2023-10-23T07:19:30
2023-05-22T16:13:50
{ "title": [ -0.009109497, 0.0055999756, -0.039367676, -0.10003662, 0.016448975, 0.017288208, -0.085632324, -0.009933472, 0.02508545, 0.028259277, -0.0013885498, 0.02154541, -0.0126953125, -0.024261475, -0.041870117, 0.018951416, -0.011070251, 0....
http://arxiv.org/abs/2305.13179v1
Teaching Probabilistic Logical Reasoning to Transformers
Aliakbar Nafar, Kristen Brent Venable, Parisa Kordjamshidi
cs.CL, cs.AI
Recent research on transformer-based language models investigates their reasoning ability over logical rules expressed in natural language text. However, their logic is not yet well-understood as we cannot explain the abstractions made by the models that help them in reasoning. These models are criticized for merely me...
http://arxiv.org/pdf/2305.13179v1
2023-05-22T16:08:20
2023-05-22T16:08:20
{ "title": [ -0.00062561035, 0.018447876, -0.037231445, -0.062438965, 0.025878906, -0.0053596497, -0.019104004, -0.007457733, -0.0146102905, -0.0026626587, -0.048065186, 0.02015686, -0.026977539, -0.026412964, 0.004802704, -0.017730713, -0.007789612,...
http://arxiv.org/abs/2305.13175v2
Discovering Universal Geometry in Embeddings with ICA
Hiroaki Yamagiwa, Momose Oyama, Hidetoshi Shimodaira
cs.CL
This study utilizes Independent Component Analysis (ICA) to unveil a consistent semantic structure within embeddings of words or images. Our approach extracts independent semantic components from the embeddings of a pre-trained model by leveraging anisotropic information that remains after the whitening process in Prin...
http://arxiv.org/pdf/2305.13175v2
2023-11-02T16:03:33
2023-05-22T16:04:44
{ "title": [ -0.03475952, -0.003534317, -0.047973633, -0.03982544, 0.0018625259, -0.00459671, -0.015640259, -0.005569458, 0.05038452, 0.021850586, -0.02609253, 0.0023002625, 0.0025920868, -0.018356323, -0.03274536, 0.03149414, -0.003490448, 0.058...
http://arxiv.org/abs/2305.13172v2
Editing Large Language Models: Problems, Methods, and Opportunities
Yunzhi Yao, Peng Wang, Bozhong Tian, Siyuan Cheng, Zhoubo Li, Shumin Deng, Huajun Chen, Ningyu Zhang
cs.CL, cs.AI, cs.CV, cs.IR, cs.LG
Despite the ability to train capable LLMs, the methodology for maintaining their relevancy and rectifying errors remains elusive. To this end, the past few years have witnessed a surge in techniques for editing LLMs, the objective of which is to efficiently alter the behavior of LLMs within a specific domain without ne...
http://arxiv.org/pdf/2305.13172v2
2023-10-11T16:51:50
2023-05-22T16:00:00
{ "title": [ -0.042144775, -0.018112183, -0.036895752, -0.070251465, 0.0024051666, 0.020996094, -0.06500244, 0.010429382, 0.00027418137, 0.03036499, -0.0024986267, -0.01701355, 0.014984131, -0.025421143, -0.045928955, 0.013473511, -0.020233154, 0...
http://arxiv.org/abs/2305.13169v2
A Pretrainer's Guide to Training Data: Measuring the Effects of Data Age, Domain Coverage, Quality, & Toxicity
Shayne Longpre, Gregory Yauney, Emily Reif, Katherine Lee, Adam Roberts, Barret Zoph, Denny Zhou, Jason Wei, Kevin Robinson, David Mimno, Daphne Ippolito
cs.CL, cs.LG
Pretraining is the preliminary and fundamental step in developing capable language models (LM). Despite this, pretraining data design is critically under-documented and often guided by empirically unsupported intuitions. To address this, we pretrain 28 1.5B parameter decoder-only models, training on data curated (1) at...
http://arxiv.org/pdf/2305.13169v2
2023-11-13T14:50:06
2023-05-22T15:57:53
{ "title": [ 0.041778564, 0.017745972, -0.060943604, -0.0793457, 0.032592773, -0.04055786, -0.033172607, 0.01725769, -0.030258179, 0.0030345917, -0.05239868, 0.029220581, -0.06121826, -0.0209198, 0.0022773743, -0.058776855, 0.02432251, -0.0016584...
http://arxiv.org/abs/2305.13168v1
LLMs for Knowledge Graph Construction and Reasoning: Recent Capabilities and Future Opportunities
Yuqi Zhu, Xiaohan Wang, Jing Chen, Shuofei Qiao, Yixin Ou, Yunzhi Yao, Shumin Deng, Huajun Chen, Ningyu Zhang
cs.CL, cs.AI, cs.DB, cs.IR, cs.LG
This paper presents an exhaustive quantitative and qualitative evaluation of Large Language Models (LLMs) for Knowledge Graph (KG) construction and reasoning. We employ eight distinct datasets that encompass aspects including entity, relation and event extraction, link prediction, and question answering. Empirically, o...
http://arxiv.org/pdf/2305.13168v1
2023-05-22T15:56:44
2023-05-22T15:56:44
{ "title": [ -0.06060791, -0.029388428, -0.07342529, -0.0769043, -0.016571045, 0.012275696, -0.07800293, 0.035369873, -0.016098022, 0.0032920837, -0.033813477, 0.018096924, -0.02420044, -0.026885986, -0.016784668, -0.017074585, -0.010574341, 0.03...
http://arxiv.org/abs/2305.13160v2
Can ChatGPT Defend its Belief in Truth? Evaluating LLM Reasoning via Debate
Boshi Wang, Xiang Yue, Huan Sun
cs.CL
Large language models (LLMs) such as ChatGPT and GPT-4 have shown impressive performance in complex reasoning tasks. However, it is difficult to know whether the models are reasoning based on deep understandings of truth and logic, or leveraging their memorized patterns in a relatively superficial way. In this work, we...
http://arxiv.org/pdf/2305.13160v2
2023-10-10T17:34:15
2023-05-22T15:47:31
{ "title": [ -0.03375244, -0.019958496, -0.03149414, -0.013084412, -0.041900635, -0.00092983246, -0.03414917, -0.0025577545, -0.040130615, 0.002500534, -0.019104004, 0.019317627, -0.03363037, -0.02708435, -0.004020691, -0.03010559, -0.037628174, ...
http://arxiv.org/abs/2305.13144v2
Response Length Perception and Sequence Scheduling: An LLM-Empowered LLM Inference Pipeline
Zangwei Zheng, Xiaozhe Ren, Fuzhao Xue, Yang Luo, Xin Jiang, Yang You
cs.CL
Large language models (LLMs) have revolutionized the field of AI, demonstrating unprecedented capacity across various tasks. However, the inference process for LLMs comes with significant computational costs. In this paper, we propose an efficient LLM inference pipeline that harnesses the power of LLMs. Our approach be...
http://arxiv.org/pdf/2305.13144v2
2023-05-28T08:22:19
2023-05-22T15:36:06
{ "title": [ -0.082092285, -0.027313232, -0.06286621, -0.046661377, -0.0569458, 0.006351471, -0.09020996, 0.028427124, -0.014213562, -0.002374649, -0.0501709, 0.030014038, -0.037872314, -0.01914978, -0.007686615, -0.021392822, -0.014839172, 0.037...
http://arxiv.org/abs/2305.13142v1
Better Sampling of Negatives for Distantly Supervised Named Entity Recognition
Lu Xu, Lidong Bing, Wei Lu
cs.CL
Distantly supervised named entity recognition (DS-NER) has been proposed to exploit the automatically labeled training data instead of human annotations. The distantly annotated datasets are often noisy and contain a considerable number of false negatives. The recent approach uses a weighted sampling approach to select...
http://arxiv.org/pdf/2305.13142v1
2023-05-22T15:35:39
2023-05-22T15:35:39
{ "title": [ -0.0039405823, 0.007587433, -0.030731201, -0.068847656, 0.02154541, -0.005531311, -0.032104492, -0.00818634, 0.013046265, 0.012176514, -0.068237305, 0.032348633, -0.005722046, 0.013450623, -0.028244019, -0.01134491, -0.009231567, 0.0...
http://arxiv.org/abs/2305.13140v1
Extrapolating Multilingual Understanding Models as Multilingual Generators
Bohong Wu, Fei Yuan, Hai Zhao, Lei Li, Jingjing Xu
cs.CL
Multilingual understanding models (or encoder-based), pre-trained via masked language modeling, have achieved promising results on many language understanding tasks (e.g., mBERT). However, these non-autoregressive (NAR) models still struggle to generate high-quality texts compared with autoregressive (AR) models. Consi...
http://arxiv.org/pdf/2305.13140v1
2023-05-22T15:33:21
2023-05-22T15:33:21
{ "title": [ -0.027694702, -0.018981934, -0.011924744, -0.07208252, -0.020462036, 0.020141602, -0.07720947, -0.009170532, -0.0043945312, 0.036834717, -0.009147644, -0.0045280457, -0.003610611, -0.023361206, -0.048553467, 0.0044784546, -0.011985779, ...
http://arxiv.org/abs/2305.13137v2
EMNS /Imz/ Corpus: An emotive single-speaker dataset for narrative storytelling in games, television and graphic novels
Kari Ali Noriy, Xiaosong Yang, Jian Jun Zhang
cs.CL, cs.LG, cs.MM
The increasing adoption of text-to-speech technologies has led to a growing demand for natural and emotive voices that adapt to a conversation's context and emotional tone. The Emotive Narrative Storytelling (EMNS) corpus is a unique speech dataset created to enhance conversations' expressiveness and emotive quality in...
http://arxiv.org/pdf/2305.13137v2
2023-05-25T16:17:24
2023-05-22T15:32:32
{ "title": [ -0.025772095, -0.045959473, -0.029800415, -0.005504608, -0.06707764, -0.018707275, -0.019622803, 0.019683838, 0.0044898987, 0.02053833, -0.0127334595, -0.029678345, 0.044708252, -0.04260254, -0.007663727, -0.012908936, -0.020065308, ...
http://arxiv.org/abs/2305.13120v1
Partial Annotation Learning for Biomedical Entity Recognition
Liangping Ding, Giovanni Colavizza, Zhixiong Zhang
cs.CL, cs.LG
Motivation: Named Entity Recognition (NER) is a key task to support biomedical research. In Biomedical Named Entity Recognition (BioNER), obtaining high-quality expert annotated data is laborious and expensive, leading to the development of automatic approaches such as distant supervision. However, manually and automat...
http://arxiv.org/pdf/2305.13120v1
2023-05-22T15:18:38
2023-05-22T15:18:38
{ "title": [ -0.024917603, -0.0009918213, -0.040740967, -0.081848145, 0.012741089, 0.0068740845, -0.04244995, 0.015197754, -0.0069503784, 0.034698486, -0.00027298927, 0.0036354065, -0.002752304, -0.024887085, -0.043182373, 0.009506226, 0.0076942444, ...
http://arxiv.org/abs/2305.13119v2
Ambiguity Meets Uncertainty: Investigating Uncertainty Estimation for Word Sense Disambiguation
Zhu Liu, Ying Liu
cs.CL
Word sense disambiguation (WSD), which aims to determine an appropriate sense for a target word given its context, is crucial for natural language understanding. Existing supervised methods treat WSD as a classification task and have achieved remarkable performance. However, they ignore uncertainty estimation (UE) in t...
http://arxiv.org/pdf/2305.13119v2
2023-06-10T03:27:31
2023-05-22T15:18:15
{ "title": [ -0.04562378, -0.0026817322, -0.04321289, -0.025634766, 0.00013172626, -0.008422852, -0.0317688, -0.018249512, 0.027542114, 0.021347046, -0.038024902, 0.000015437603, 0.012123108, -0.01008606, -0.053955078, -0.0011129379, -0.008056641, ...
http://arxiv.org/abs/2305.13117v3
AVeriTeC: A Dataset for Real-world Claim Verification with Evidence from the Web
Michael Schlichtkrull, Zhijiang Guo, Andreas Vlachos
cs.CL
Existing datasets for automated fact-checking have substantial limitations, such as relying on artificial claims, lacking annotations for evidence and intermediate reasoning, or including evidence published after the claim. In this paper we introduce AVeriTeC, a new dataset of 4,568 real-world claims covering fact-chec...
http://arxiv.org/pdf/2305.13117v3
2023-11-08T11:53:55
2023-05-22T15:17:18
{ "title": [ 0.0019607544, -0.009559631, -0.052368164, -0.07598877, 0.024536133, -0.017105103, -0.038360596, -0.0019931793, -0.034210205, 0.015464783, -0.020874023, 0.01626587, -0.0003721714, -0.015113831, -0.012786865, -0.008087158, -0.018722534, ...
http://arxiv.org/abs/2305.13112v2
Rethinking the Evaluation for Conversational Recommendation in the Era of Large Language Models
Xiaolei Wang, Xinyu Tang, Wayne Xin Zhao, Jingyuan Wang, Ji-Rong Wen
cs.CL, cs.IR
The recent success of large language models (LLMs) has shown great potential to develop more powerful conversational recommender systems (CRSs), which rely on natural language conversations to satisfy user needs. In this paper, we embark on an investigation into the utilization of ChatGPT for conversational recommendat...
http://arxiv.org/pdf/2305.13112v2
2023-11-03T02:49:46
2023-05-22T15:12:43
{ "title": [ -0.0047950745, -0.012077332, -0.03010559, -0.04397583, -0.003271103, -0.013725281, -0.035827637, -0.0047035217, 0.005748749, 0.024032593, -0.028320312, 0.0079422, -0.0051460266, -0.0121154785, -0.049682617, -0.0074310303, -0.023529053, ...
http://arxiv.org/abs/2305.13108v3
Debiased Automatic Speech Recognition for Dysarthric Speech via Sample Reweighting with Sample Affinity Test
Eungbeom Kim, Yunkee Chae, Jaeheon Sim, Kyogu Lee
eess.AS, cs.CL, cs.LG, cs.SD
Automatic speech recognition systems based on deep learning are mainly trained under empirical risk minimization (ERM). Since ERM utilizes the averaged performance on the data samples regardless of a group such as healthy or dysarthric speakers, ASR systems are unaware of the performance disparities across the groups. ...
http://arxiv.org/pdf/2305.13108v3
2023-06-27T13:19:19
2023-05-22T15:09:27
{ "title": [ 0.021255493, 0.021270752, -0.044799805, -0.0803833, 0.009849548, 0.0050697327, -0.05517578, -0.016235352, -0.015731812, 0.0569458, -0.055023193, 0.008300781, 0.0132369995, -0.006378174, -0.057281494, -0.03375244, -0.015327454, 0.0471...
http://arxiv.org/abs/2305.18320v1
Cognitive network science reveals bias in GPT-3, ChatGPT, and GPT-4 mirroring math anxiety in high-school students
Katherine Abramski, Salvatore Citraro, Luigi Lombardi, Giulio Rossetti, Massimo Stella
cs.CY, cs.AI, cs.CL
Large language models are becoming increasingly integrated into our lives. Hence, it is important to understand the biases present in their outputs in order to avoid perpetuating harmful stereotypes, which originate in our own flawed ways of thinking. This challenge requires developing new benchmarks and methods for qu...
http://arxiv.org/pdf/2305.18320v1
2023-05-22T15:06:51
2023-05-22T15:06:51
{ "title": [ 0.008720398, -0.0043144226, 0.004837036, 0.0635376, -0.035705566, -0.040252686, 0.02885437, -0.032073975, -0.0028858185, 0.02897644, -0.03186035, -0.00020885468, 0.03878784, -0.034484863, 0.015129089, -0.023590088, -0.019042969, 0.03...
http://arxiv.org/abs/2305.18319v1
Automated Feedback Generation for a Chemistry Database and Abstracting Exercise
Oscar Morris, Russell Morris
cs.CL, cs.AI, cs.LG
Timely feedback is an important part of teaching and learning. Here we describe how a readily available neural network transformer (machine-learning) model (BERT) can be used to give feedback on the structure of the response to an abstracting exercise where students are asked to summarise the contents of a published ar...
http://arxiv.org/pdf/2305.18319v1
2023-05-22T15:04:26
2023-05-22T15:04:26
{ "title": [ -0.006248474, -0.0027008057, -0.05102539, -0.04559326, -0.027740479, -0.016723633, -0.052459717, 0.011665344, -0.0055732727, 0.060699463, -0.0473938, 0.014892578, 0.025299072, -0.01763916, -0.015777588, -0.008979797, 0.024215698, 0.0...
http://arxiv.org/abs/2305.13102v1
Observations on LLMs for Telecom Domain: Capabilities and Limitations
Sumit Soman, Ranjani H G
cs.HC, cs.AI, cs.CL, cs.LG, 68T50
The landscape for building conversational interfaces (chatbots) has witnessed a paradigm shift with recent developments in generative Artificial Intelligence (AI) based Large Language Models (LLMs), such as ChatGPT by OpenAI (GPT3.5 and GPT4), Google's Bard, Large Language Model Meta AI (LLaMA), among others. In this p...
http://arxiv.org/pdf/2305.13102v1
2023-05-22T15:04:16
2023-05-22T15:04:16
{ "title": [ -0.030349731, -0.003982544, -0.04748535, -0.03366089, -0.027816772, -0.020111084, -0.043884277, 0.008895874, -0.025360107, 0.037017822, -0.04272461, 0.011405945, 0.004348755, -0.02268982, -0.0030117035, -0.00005275011, -0.011695862, ...
http://arxiv.org/abs/2305.13092v1
Improved Compositional Generalization by Generating Demonstrations for Meta-Learning
Sam Spilsbury, Alexander Ilin
cs.CL
Meta-learning and few-shot prompting are viable methods to induce certain types of compositional behaviour. However, these methods can be very sensitive to the choice of support examples used. Choosing good supports from the training data for a given test query is already a difficult problem, but in some cases solving ...
http://arxiv.org/pdf/2305.13092v1
2023-05-22T14:58:54
2023-05-22T14:58:54
{ "title": [ -0.00094366074, -0.0019273758, -0.01965332, -0.086364746, 0.0033016205, 0.015426636, -0.07788086, -0.012901306, 0.011817932, 0.030273438, -0.0070610046, 0.024414062, -0.0395813, -0.017166138, -0.051239014, 0.007297516, -0.01802063, 0...
http://arxiv.org/abs/2305.13091v2
Large Language Models are Not Yet Human-Level Evaluators for Abstractive Summarization
Chenhui Shen, Liying Cheng, Xuan-Phi Nguyen, Yang You, Lidong Bing
cs.CL
With the recent undeniable advancement in reasoning abilities in large language models (LLMs) like ChatGPT and GPT-4, there is a growing trend for using LLMs on various tasks. One area where LLMs can be employed is as an alternative evaluation metric for complex generative tasks, which generally demands expensive human...
http://arxiv.org/pdf/2305.13091v2
2023-10-20T03:47:27
2023-05-22T14:58:13
{ "title": [ -0.010032654, -0.010292053, -0.006038666, -0.037719727, -0.062347412, 0.006801605, -0.039611816, -0.016357422, 0.00920105, 0.048187256, -0.045654297, -0.021438599, -0.0104904175, -0.015029907, -0.06738281, -0.022354126, -0.000021219254, ...
http://arxiv.org/abs/2305.13088v1
Should We Attend More or Less? Modulating Attention for Fairness
Abdelrahman Zayed, Goncalo Mordido, Samira Shabanian, Sarath Chandar
cs.CL, cs.AI, cs.CY, cs.LG
The abundance of annotated data in natural language processing (NLP) poses both opportunities and challenges. While it enables the development of high-performing models for a variety of tasks, it also poses the risk of models learning harmful biases from the data, such as gender stereotypes. In this work, we investigat...
http://arxiv.org/pdf/2305.13088v1
2023-05-22T14:54:21
2023-05-22T14:54:21
{ "title": [ 0.046905518, 0.014282227, -0.039215088, -0.05456543, 0.04208374, -0.042236328, 0.011146545, -0.028396606, -0.022964478, 0.025268555, -0.02960205, 0.0071868896, -0.0063095093, -0.014640808, -0.027359009, -0.013061523, 0.002609253, 0.0...
http://arxiv.org/abs/2305.13086v1
LMGQS: A Large-scale Dataset for Query-focused Summarization
Ruochen Xu, Song Wang, Yang Liu, Shuohang Wang, Yichong Xu, Dan Iter, Chenguang Zhu, Michael Zeng
cs.CL
Query-focused summarization (QFS) aims to extract or generate a summary of an input document that directly answers or is relevant to a given query. The lack of large-scale datasets in the form of documents, queries, and summaries has hindered model development in this area. In contrast, multiple large-scale high-qualit...
http://arxiv.org/pdf/2305.13086v1
2023-05-22T14:53:45
2023-05-22T14:53:45
{ "title": [ 0.0041503906, 0.0059318542, -0.044433594, -0.09515381, 0.01826477, 0.009757996, -0.050354004, 0.0031166077, 0.0010499954, 0.049987793, -0.011367798, 0.021255493, 0.0017118454, 0.0024051666, -0.048095703, 0.011428833, 0.007434845, 0.0...
http://arxiv.org/abs/2305.13085v2
Decomposed Prompting for Machine Translation Between Related Languages using Large Language Models
Ratish Puduppully, Anoop Kunchukuttan, Raj Dabre, Ai Ti Aw, Nancy F. Chen
cs.CL
This study investigates machine translation between related languages i.e., languages within the same family that share linguistic characteristics such as word order and lexical similarity. Machine translation through few-shot prompting leverages a small set of translation pair examples to generate translations for tes...
http://arxiv.org/pdf/2305.13085v2
2023-10-23T02:30:20
2023-05-22T14:52:47
{ "title": [ -0.04156494, -0.025375366, -0.007080078, -0.060333252, -0.02822876, 0.029373169, -0.08074951, 0.0010585785, -0.003610611, 0.029708862, -0.0049209595, -0.005432129, 0.023757935, -0.02696228, -0.035125732, 0.016342163, 0.011413574, -0....
http://arxiv.org/abs/2305.13083v1
InheritSumm: A General, Versatile and Compact Summarizer by Distilling from GPT
Yichong Xu, Ruochen Xu, Dan Iter, Yang Liu, Shuohang Wang, Chenguang Zhu, Michael Zeng
cs.CL
While large models such as GPT-3 demonstrate exceptional performance in zeroshot and fewshot summarization tasks, their extensive serving and fine-tuning costs hinder their utilization in various applications. Conversely, previous studies have found that although automatic metrics tend to favor smaller fine-tuned model...
http://arxiv.org/pdf/2305.13083v1
2023-05-22T14:52:32
2023-05-22T14:52:32
{ "title": [ 0.024353027, -0.0028457642, -0.021102905, -0.10046387, -0.013336182, 0.014335632, -0.09082031, -0.027755737, 0.013923645, 0.036254883, -0.02607727, -0.002910614, 0.02305603, -0.023910522, -0.032196045, 0.012062073, -0.005584717, 0.05...
http://arxiv.org/abs/2305.13080v1
Mitigating Catastrophic Forgetting for Few-Shot Spoken Word Classification Through Meta-Learning
Ruan van der Merwe, Herman Kamper
cs.CL, cs.AI, eess.AS, I.2.7; I.2.6
We consider the problem of few-shot spoken word classification in a setting where a model is incrementally introduced to new word classes. This would occur in a user-defined keyword system where new words can be added as the system is used. In such a continual learning scenario, a model might start to misclassify earli...
http://arxiv.org/pdf/2305.13080v1
2023-05-22T14:51:15
2023-05-22T14:51:15
{ "title": [ 0.0026569366, -0.009346008, -0.029968262, -0.047576904, -0.019073486, -0.004085541, -0.03125, 0.0070610046, 0.0041160583, 0.05331421, -0.024902344, 0.014541626, -0.0052108765, -0.0017976761, -0.07287598, -0.017700195, 0.0059165955, 0...
http://arxiv.org/abs/2305.13076v1
An Abstract Specification of VoxML as an Annotation Language
Kiyong Lee, Nikhil Krishnaswamy, James Pustejovsky
cs.CL
VoxML is a modeling language used to map natural language expressions into real-time visualizations using commonsense semantic knowledge of objects and events. Its utility has been demonstrated in embodied simulation environments and in agent-object interactions in situated multimodal human-agent collaboration and comm...
http://arxiv.org/pdf/2305.13076v1
2023-05-22T14:47:59
2023-05-22T14:47:59
{ "title": [ -0.011222839, -0.01322937, -0.027069092, -0.08886719, 0.007904053, 0.016342163, -0.0574646, -0.0049552917, -0.005783081, 0.04525757, 0.009315491, -0.0010824203, -0.00422287, -0.023910522, -0.046051025, 0.0259552, -0.011352539, 0.0099...
http://arxiv.org/abs/2305.13073v2
Text-to-SQL Error Correction with Language Models of Code
Ziru Chen, Shijie Chen, Michael White, Raymond Mooney, Ali Payani, Jayanth Srinivasa, Yu Su, Huan Sun
cs.CL, cs.AI, cs.DB, cs.LG
Despite recent progress in text-to-SQL parsing, current semantic parsers are still not accurate enough for practical use. In this paper, we investigate how to build automatic text-to-SQL error correction models. Noticing that token-level edits are out of context and sometimes ambiguous, we propose building clause-level...
http://arxiv.org/pdf/2305.13073v2
2023-05-28T15:32:26
2023-05-22T14:42:39
{ "title": [ -0.010604858, -0.017868042, -0.03652954, -0.064941406, -0.0044898987, -0.0026168823, -0.0340271, 0.023254395, 0.029083252, 0.05593872, -0.03475952, -0.019729614, 0.03491211, -0.00920105, -0.062164307, -0.0057907104, -0.005870819, 0.0...
http://arxiv.org/abs/2305.13071v1
Machine-Created Universal Language for Cross-lingual Transfer
Yaobo Liang, Quanzhi Zhu, Junhe Zhao, Nan Duan
cs.CL
There are two types of approaches to solving cross-lingual transfer: multilingual pre-training implicitly aligns the hidden representations of different languages, while the translate-test explicitly translates different languages to an intermediate language, such as English. Translate-test has better interpretability ...
http://arxiv.org/pdf/2305.13071v1
2023-05-22T14:41:09
2023-05-22T14:41:09
{ "title": [ -0.01838684, -0.007369995, -0.041168213, -0.051818848, -0.015579224, -0.0034484863, -0.03744507, 0.007904053, -0.011695862, 0.05996704, 0.009979248, 0.007896423, 0.038726807, -0.0309906, -0.012748718, 0.031951904, 0.001200676, 0.0421...
http://arxiv.org/abs/2305.13067v1
Improving Robustness in Knowledge Distillation Using Domain-Targeted Data Augmentation
Joe Stacey, Marek Rei
cs.CL, cs.LG, I.2.7
Applying knowledge distillation encourages a student model to behave more like a teacher model, largely retaining the performance of the teacher model, even though the student model may have substantially fewer parameters. However, while distillation helps student models behave more like teacher models in-distribution,...
http://arxiv.org/pdf/2305.13067v1
2023-05-22T14:37:05
2023-05-22T14:37:05
{ "title": [ 0.0013866425, 0.020996094, -0.10852051, -0.095336914, 0.03564453, -0.024475098, -0.06689453, 0.00037145615, 0.042663574, 0.0039787292, -0.047607422, -0.007183075, 0.005569458, -0.011627197, -0.03543091, -0.015716553, -0.0033836365, 0...
http://arxiv.org/abs/2305.13068v1
Making Language Models Better Tool Learners with Execution Feedback
Shuofei Qiao, Honghao Gui, Huajun Chen, Ningyu Zhang
cs.CL, cs.AI, cs.HC, cs.IR, cs.LG
Tools serve as pivotal interfaces that enable humans to understand and reshape the world. With the advent of foundational models, AI systems can utilize tools to expand their capabilities and interact with the world. Existing tool learning methodologies, encompassing supervised fine-tuning and prompt engineering approa...
http://arxiv.org/pdf/2305.13068v1
2023-05-22T14:37:05
2023-05-22T14:37:05
{ "title": [ -0.049316406, -0.016448975, -0.017318726, -0.03414917, -0.019927979, 0.010498047, -0.062164307, -0.01033783, 0.0019207001, 0.0340271, -0.021530151, -0.010261536, -0.0014438629, -0.009674072, -0.059020996, -0.0038871765, -0.0070648193, ...
http://arxiv.org/abs/2305.13066v2
Biomedical Named Entity Recognition via Dictionary-based Synonym Generalization
Zihao Fu, Yixuan Su, Zaiqiao Meng, Nigel Collier
cs.CL, cs.AI
Biomedical named entity recognition is one of the core tasks in biomedical natural language processing (BioNLP). To tackle this task, numerous supervised/distantly supervised approaches have been proposed. Despite their remarkable success, these approaches inescapably demand laborious human effort. To alleviate the nee...
http://arxiv.org/pdf/2305.13066v2
2023-10-13T11:19:19
2023-05-22T14:36:32
{ "title": [ -0.031707764, 0.002281189, -0.052764893, -0.062683105, 0.018096924, 0.015136719, -0.04534912, 0.0074920654, 0.011489868, 0.03390503, -0.013496399, 0.0021076202, 0.005329132, -0.0099105835, -0.049346924, 0.0073547363, -0.0064201355, 0...
http://arxiv.org/abs/2305.13062v3
GPT4Table: Can Large Language Models Understand Structured Table Data? A Benchmark and Empirical Study
Yuan Sui, Mengyu Zhou, Mingjie Zhou, Shi Han, Dongmei Zhang
cs.CL, cs.AI, cs.IR
Large language models (LLMs) are becoming attractive as few-shot reasoners to solve Natural Language (NL)-related tasks. However, there is still much to learn about how well LLMs understand structured data, such as tables. While it is true that tables can be used as inputs to LLMs with serialization, there is a lack of...
http://arxiv.org/pdf/2305.13062v3
2023-11-15T12:18:39
2023-05-22T14:23:46
{ "title": [ -0.02961731, -0.0048332214, -0.054229736, -0.06762695, -0.000020325184, -0.0034809113, -0.077941895, -0.0013246536, 0.018188477, 0.024856567, -0.030929565, -0.0057258606, 0.0037231445, -0.007663727, -0.053894043, -0.0033226013, -0.012054...
http://arxiv.org/abs/2305.13058v1
Retrieval-augmented Multi-label Text Classification
Ilias Chalkidis, Yova Kementchedjhieva
cs.CL
Multi-label text classification (MLC) is a challenging task in settings of large label sets, where label support follows a Zipfian distribution. In this paper, we address this problem through retrieval augmentation, aiming to improve the sample efficiency of classification models. Our approach closely follows the stand...
http://arxiv.org/pdf/2305.13058v1
2023-05-22T14:16:23
2023-05-22T14:16:23
{ "title": [ -0.010231018, 0.0029239655, -0.032440186, -0.04071045, 0.0067481995, -0.012908936, -0.007659912, -0.010498047, -0.000016093254, 0.038604736, -0.02053833, 0.019104004, -0.010391235, -0.005252838, -0.04626465, 0.0009946823, -0.02432251, ...
http://arxiv.org/abs/2305.13052v1
Federated Learning of Medical Concepts Embedding using BEHRT
Ofir Ben Shoham, Nadav Rappoport
cs.LG, cs.AI, cs.CL, cs.DC
Electronic Health Records (EHR) data contains medical records such as diagnoses, medications, procedures, and treatments of patients. This data is often considered sensitive medical information. Therefore, the EHR data from the medical centers often cannot be shared, making it difficult to create prediction models usin...
http://arxiv.org/pdf/2305.13052v1
2023-05-22T14:05:39
2023-05-22T14:05:39
{ "title": [ 0.038146973, 0.011505127, -0.042419434, -0.083984375, 0.03765869, -0.019302368, -0.012756348, -0.008354187, 0.008277893, 0.045318604, -0.006641388, 0.009155273, -0.016281128, -0.018081665, -0.042633057, -0.0021820068, -0.019134521, 0...
http://arxiv.org/abs/2305.13048v1
RWKV: Reinventing RNNs for the Transformer Era
Bo Peng, Eric Alcaide, Quentin Anthony, Alon Albalak, Samuel Arcadinho, Huanqi Cao, Xin Cheng, Michael Chung, Matteo Grella, Kranthi Kiran GV, Xuzheng He, Haowen Hou, Przemyslaw Kazienko, Jan Kocon, Jiaming Kong, Bartlomiej Koptyra, Hayden Lau, Krishna Sri Ipsit Mantri, Ferdinand Mom, Atsushi Saito, Xiangru Tang, Bolun...
cs.CL, cs.AI
Transformers have revolutionized almost all natural language processing (NLP) tasks but suffer from memory and computational complexity that scales quadratically with sequence length. In contrast, recurrent neural networks (RNNs) exhibit linear scaling in memory and computational requirements but struggle to match the ...
http://arxiv.org/pdf/2305.13048v1
2023-05-22T13:57:41
2023-05-22T13:57:41
{ "title": [ -0.058441162, -0.02418518, -0.024002075, -0.03463745, -0.007610321, 0.017791748, -0.030029297, -0.015098572, -0.028335571, 0.03237915, -0.015396118, 0.005947113, 0.019744873, -0.02671814, -0.019104004, 0.020614624, -0.0063285828, 0.0...
http://arxiv.org/abs/2305.13047v1
Automated stance detection in complex topics and small languages: the challenging case of immigration in polarizing news media
Mark Mets, Andres Karjus, Indrek Ibrus, Maximilian Schich
cs.CL
Automated stance detection and related machine learning methods can provide useful insights for media monitoring and academic research. Many of these approaches require annotated training datasets, which limits their applicability for languages where these may not be readily available. This paper explores the applicabi...
http://arxiv.org/pdf/2305.13047v1
2023-05-22T13:56:35
2023-05-22T13:56:35
{ "title": [ -0.00014054775, -0.00667572, -0.0066947937, -0.029144287, -0.011924744, -0.00894165, -0.010101318, -0.033233643, -0.0019035339, 0.036346436, -0.04196167, 0.013252258, -0.003419876, -0.00038456917, -0.060150146, -0.019989014, -0.013465881...
http://arxiv.org/abs/2305.13040v4
SpokenWOZ: A Large-Scale Speech-Text Benchmark for Spoken Task-Oriented Dialogue Agents
Shuzheng Si, Wentao Ma, Haoyu Gao, Yuchuan Wu, Ting-En Lin, Yinpei Dai, Hangyu Li, Rui Yan, Fei Huang, Yongbin Li
cs.CL, cs.AI
Task-oriented dialogue (TOD) models have made significant progress in recent years. However, previous studies primarily focus on datasets written by annotators, which has resulted in a gap between academic research and real-world spoken conversation scenarios. While several small-scale spoken TOD datasets are proposed ...
http://arxiv.org/pdf/2305.13040v4
2023-10-24T15:19:39
2023-05-22T13:47:51
{ "title": [ -0.03149414, -0.0209198, -0.034332275, -0.064697266, -0.001458168, -0.0007095337, -0.043395996, 0.031280518, -0.000047147274, 0.062042236, -0.021530151, 0.020858765, 0.010192871, 0.011192322, -0.06707764, 0.007511139, -0.018753052, 0...
http://arxiv.org/abs/2305.13034v2
Nearest Neighbor Machine Translation is Meta-Optimizer on Output Projection Layer
Ruize Gao, Zhirui Zhang, Yichao Du, Lemao Liu, Rui Wang
cs.CL
Nearest Neighbor Machine Translation ($k$NN-MT) has achieved great success in domain adaptation tasks by integrating pre-trained Neural Machine Translation (NMT) models with domain-specific token-level retrieval. However, the reasons underlying its success have not been thoroughly investigated. In this paper, we compre...
http://arxiv.org/pdf/2305.13034v2
2023-10-24T10:22:05
2023-05-22T13:38:53
{ "title": [ -0.03189087, -0.023803711, 0.02658081, 0.014427185, -0.08520508, 0.019073486, -0.036865234, -0.039916992, -0.033691406, 0.07397461, 0.027175903, -0.023071289, 0.04473877, -0.04168701, -0.0064735413, 0.007904053, 0.015571594, 0.023361...
http://arxiv.org/abs/2305.13026v2
DUMB: A Benchmark for Smart Evaluation of Dutch Models
Wietse de Vries, Martijn Wieling, Malvina Nissim
cs.CL
We introduce the Dutch Model Benchmark: DUMB. The benchmark includes a diverse set of datasets for low-, medium- and high-resource tasks. The total set of nine tasks includes four tasks that were previously not available in Dutch. Instead of relying on a mean score across tasks, we propose Relative Error Reduction (RER...
http://arxiv.org/pdf/2305.13026v2
2023-10-13T10:43:05
2023-05-22T13:27:37
{ "title": [ -0.01965332, -0.0025863647, -0.0032672882, -0.040802002, -0.0042037964, 0.019851685, -0.05215454, -0.04916382, -0.0058250427, 0.041656494, 0.019989014, 0.008560181, -0.023986816, -0.028121948, -0.07122803, 0.0158844, -0.013114929, 0....
http://arxiv.org/abs/2305.13016v2
Iterative Forward Tuning Boosts In-context Learning in Language Models
Jiaxi Yang, Binyuan Hui, Min Yang, Binhua Li, Fei Huang, Yongbin Li
cs.CL
Large language models (LLMs) have exhibited an emergent in-context learning (ICL) ability. However, the ICL models that can solve ordinary cases are hardly extended to solve more complex tasks by processing the demonstration examples once. This single-turn ICL is incoordinate with the decision making process of humans ...
http://arxiv.org/pdf/2305.13016v2
2023-05-30T05:47:19
2023-05-22T13:18:17
{ "title": [ -0.0012664795, 0.0055274963, -0.039642334, -0.08087158, -0.018585205, -0.014305115, -0.075683594, 0.014816284, -0.0046920776, 0.044952393, -0.030029297, 0.013542175, 0.015586853, -0.017578125, -0.033355713, 0.0070343018, -0.0048980713, ...
http://arxiv.org/abs/2305.13015v1
3D Rotation and Translation for Hyperbolic Knowledge Graph Embedding
Yihua Zhu, Hidetoshi Shimodaira
cs.CL
The main objective of Knowledge Graph (KG) embeddings is to learn low-dimensional representations of entities and relations, enabling the prediction of missing facts. A significant challenge in achieving better KG embeddings lies in capturing relation patterns, including symmetry, antisymmetry, inversion, commutative c...
http://arxiv.org/pdf/2305.13015v1
2023-05-22T13:17:23
2023-05-22T13:17:23
{ "title": [ -0.034484863, 0.0023994446, -0.07385254, -0.045837402, -0.004043579, -0.014892578, -0.03894043, 0.0010671616, 0.016204834, 0.04928589, -0.0390625, 0.014724731, 0.001376152, 0.011810303, -0.048797607, -0.009552002, 0.03314209, 0.01146...
http://arxiv.org/abs/2305.13014v3
Can Large Language Models emulate an inductive Thematic Analysis of semi-structured interviews? An exploration and provocation on the limits of the approach and the model
Stefano De Paoli
cs.CL
Large Language Models (LLMs) have emerged as powerful generative Artificial Intelligence solutions which can be applied to several fields and areas of work. This paper presents results and reflection of an experiment done to use the model GPT 3.5-Turbo to emulate some aspects of an inductive Thematic Analysis. Previous...
http://arxiv.org/pdf/2305.13014v3
2023-09-18T07:36:55
2023-05-22T13:16:07
{ "title": [ -0.0067100525, -0.0054779053, 0.0028839111, -0.011512756, -0.03277588, -0.021743774, -0.036590576, -0.013771057, -0.0017089844, -0.010734558, -0.033935547, -0.02798462, -0.018127441, -0.039093018, 0.00051498413, -0.06262207, -0.00705719,...
http://arxiv.org/abs/2305.13009v2
Textually Pretrained Speech Language Models
Michael Hassid, Tal Remez, Tu Anh Nguyen, Itai Gat, Alexis Conneau, Felix Kreuk, Jade Copet, Alexandre Defossez, Gabriel Synnaeve, Emmanuel Dupoux, Roy Schwartz, Yossi Adi
cs.CL, cs.LG, cs.SD, eess.AS
Speech language models (SpeechLMs) process and generate acoustic data only, without textual supervision. In this work, we propose TWIST, a method for training SpeechLMs using a warm-start from a pretrained textual language models. We show using both automatic and human evaluations that TWIST outperforms a cold-start Sp...
http://arxiv.org/pdf/2305.13009v2
2023-11-02T11:29:53
2023-05-22T13:12:16
{ "title": [ -0.015258789, -0.008712769, -0.034423828, -0.07678223, -0.009262085, 0.008590698, -0.07714844, 0.014678955, -0.0058021545, 0.047027588, -0.008125305, -0.0020523071, 0.022460938, -0.023834229, -0.035186768, 0.021697998, 0.0037384033, ...
http://arxiv.org/abs/2305.13002v1
Rethinking Semi-supervised Learning with Language Models
Zhengxiang Shi, Francesco Tonolini, Nikolaos Aletras, Emine Yilmaz, Gabriella Kazai, Yunlong Jiao
cs.CL, cs.AI, cs.LG
Semi-supervised learning (SSL) is a popular setting aiming to effectively utilize unlabelled data to improve model performance in downstream natural language processing (NLP) tasks. Currently, there are two popular approaches to make use of unlabelled data: Self-training (ST) and Task-adaptive pre-training (TAPT). ST u...
http://arxiv.org/pdf/2305.13002v1
2023-05-22T13:07:35
2023-05-22T13:07:35
{ "title": [ -0.03274536, -0.0061912537, -0.038146973, -0.05307007, 0.00044298172, 0.00081539154, -0.061798096, 0.0023479462, -0.004890442, 0.016784668, -0.01876831, -0.004989624, -0.006877899, -0.018341064, -0.042785645, -0.0035171509, -0.014816284,...
http://arxiv.org/abs/2305.13000v1
Bidirectional Transformer Reranker for Grammatical Error Correction
Ying Zhang, Hidetaka Kamigaito, Manabu Okumura
cs.CL
Pre-trained seq2seq models have achieved state-of-the-art results in the grammatical error correction task. However, these models still suffer from a prediction bias due to their unidirectional decoding. Thus, we propose a bidirectional Transformer reranker (BTR), that re-estimates the probability of each candidate sen...
http://arxiv.org/pdf/2305.13000v1
2023-05-22T13:04:48
2023-05-22T13:04:48
{ "title": [ 0.026397705, 0.018936157, 0.0002040863, -0.058441162, -0.004135132, 0.0073509216, -0.032562256, -0.032287598, 0.001282692, 0.061553955, -0.023620605, -0.009025574, 0.062683105, -0.023513794, -0.026275635, 0.0022659302, -0.03149414, 0...
http://arxiv.org/abs/2305.12995v1
MaNtLE: Model-agnostic Natural Language Explainer
Rakesh R. Menon, Kerem Zaman, Shashank Srivastava
cs.CL, cs.AI, cs.LG
Understanding the internal reasoning behind the predictions of machine learning systems is increasingly vital, given their rising adoption and acceptance. While previous approaches, such as LIME, generate algorithmic explanations by attributing importance to input features for individual examples, recent research indic...
http://arxiv.org/pdf/2305.12995v1
2023-05-22T12:58:06
2023-05-22T12:58:06
{ "title": [ -0.035308838, -0.028381348, -0.028152466, -0.070129395, -0.017929077, 0.016220093, -0.052520752, 0.008087158, -0.012466431, 0.016021729, -0.023910522, 0.015594482, -0.005748749, -0.003622055, -0.02670288, -0.005302429, -0.008087158, ...
http://arxiv.org/abs/2305.12990v1
Sentence Representations via Gaussian Embedding
Shohei Yoda, Hayato Tsukagoshi, Ryohei Sasano, Koichi Takeda
cs.CL
Recent progress in sentence embedding, which represents the meaning of a sentence as a point in a vector space, has achieved high performance on tasks such as a semantic textual similarity (STS) task. However, sentence representations as a point in a vector space can express only a part of the diverse information that ...
http://arxiv.org/pdf/2305.12990v1
2023-05-22T12:51:38
2023-05-22T12:51:38
{ "title": [ 0.008522034, -0.0024280548, -0.016906738, -0.0501709, 0.009666443, -0.004306793, -0.034332275, -0.02104187, 0.025756836, 0.032165527, -0.011390686, -0.0029945374, 0.010238647, -0.007534027, -0.048736572, 0.010864258, -0.009666443, 0....
http://arxiv.org/abs/2305.12987v2
GPT-SW3: An Autoregressive Language Model for the Nordic Languages
Ariel Ekgren, Amaru Cuba Gyllensten, Felix Stollenwerk, Joey Öhman, Tim Isbister, Evangelia Gogoulou, Fredrik Carlsson, Alice Heiman, Judit Casademont, Magnus Sahlgren
cs.CL, cs.AI
This paper details the process of developing the first native large generative language model for the Nordic languages, GPT-SW3. We cover all parts of the development process, from data collection and processing, training configuration and instruction finetuning, to evaluation and considerations for release strategies....
http://arxiv.org/pdf/2305.12987v2
2023-05-23T06:59:16
2023-05-22T12:47:48
{ "title": [ -0.0052261353, -0.002614975, -0.03387451, -0.05606079, 0.01939392, -0.0010213852, -0.039916992, 0.0090408325, -0.0044021606, 0.033569336, -0.011100769, 0.02720642, 0.022109985, -0.027679443, -0.035186768, 0.021942139, -0.00617218, 0....
http://arxiv.org/abs/2305.12962v2
Distilling ChatGPT for Explainable Automated Student Answer Assessment
Jiazheng Li, Lin Gui, Yuxiang Zhou, David West, Cesare Aloisi, Yulan He
cs.CL
Providing explainable and faithful feedback is crucial for automated student answer assessment. In this paper, we introduce a novel framework that explores using ChatGPT, a cutting-edge large language model, for the concurrent tasks of student answer scoring and rationale generation. We identify the appropriate instruc...
http://arxiv.org/pdf/2305.12962v2
2023-10-24T10:59:50
2023-05-22T12:11:39
{ "title": [ -0.01625061, -0.0048332214, -0.0044174194, -0.038238525, -0.0021915436, 0.017593384, -0.001332283, -0.008834839, 0.007019043, 0.015235901, -0.031280518, 0.0071029663, -0.016448975, -0.00025129318, -0.046661377, -0.0112838745, -0.02494812...
http://arxiv.org/abs/2305.12951v1
Cross-functional Analysis of Generalisation in Behavioural Learning
Pedro Henrique Luz de Araujo, Benjamin Roth
cs.CL, cs.LG
In behavioural testing, system functionalities underrepresented in the standard evaluation setting (with a held-out test set) are validated through controlled input-output pairs. Optimising performance on the behavioural tests during training (behavioural learning) would improve coverage of phenomena not sufficiently r...
http://arxiv.org/pdf/2305.12951v1
2023-05-22T11:54:19
2023-05-22T11:54:19
{ "title": [ 0.012321472, 0.009941101, -0.023803711, -0.04336548, 0.02178955, -0.03768921, 0.010345459, -0.02217102, -0.020065308, 0.015914917, -0.02784729, 0.036193848, -0.044921875, -0.010261536, -0.032104492, -0.0049819946, -0.008956909, 0.054...
http://arxiv.org/abs/2305.12947v2
ChatGPT to Replace Crowdsourcing of Paraphrases for Intent Classification: Higher Diversity and Comparable Model Robustness
Jan Cegin, Jakub Simko, Peter Brusilovsky
cs.CL
The emergence of generative large language models (LLMs) raises the question: what will be its impact on crowdsourcing? Traditionally, crowdsourcing has been used for acquiring solutions to a wide variety of human-intelligence tasks, including ones involving text generation, modification or evaluation. For some of thes...
http://arxiv.org/pdf/2305.12947v2
2023-10-19T21:11:53
2023-05-22T11:46:32
{ "title": [ 0.03817749, 0.018875122, -0.026153564, -0.02859497, 0.0013628006, -0.037902832, -0.019729614, -0.025726318, -0.0074272156, 0.054107666, -0.027069092, 0.016296387, 0.032196045, -0.012969971, -0.027709961, -0.0012521744, -0.033691406, ...
http://arxiv.org/abs/2305.12945v2
ExplainCPE: A Free-text Explanation Benchmark of Chinese Pharmacist Examination
Dongfang Li, Jindi Yu, Baotian Hu, Zhenran Xu, Min Zhang
cs.CL
As ChatGPT and GPT-4 spearhead the development of Large Language Models (LLMs), more researchers are investigating their performance across various tasks. But more research needs to be done on the interpretability capabilities of LLMs, that is, the ability to generate reasons after an answer has been given. Existing ex...
http://arxiv.org/pdf/2305.12945v2
2023-10-26T03:15:41
2023-05-22T11:45:42
{ "title": [ -0.03778076, -0.036315918, -0.03111267, -0.055267334, -0.041534424, -0.010498047, -0.033111572, 0.015052795, -0.014671326, 0.0004274845, -0.015975952, 0.014762878, -0.017150879, -0.04547119, 0.017669678, -0.013839722, -0.021972656, 0...
http://arxiv.org/abs/2305.12941v1
On the Correspondence between Compositionality and Imitation in Emergent Neural Communication
Emily Cheng, Mathieu Rita, Thierry Poibeau
cs.CL, cs.NE
Compositionality is a hallmark of human language that not only enables linguistic generalization, but also potentially facilitates acquisition. When simulating language emergence with neural networks, compositionality has been shown to improve communication performance; however, its impact on imitation learning has yet...
http://arxiv.org/pdf/2305.12941v1
2023-05-22T11:41:29
2023-05-22T11:41:29
{ "title": [ -0.011138916, 0.0029449463, -0.013954163, -0.029006958, 0.0041618347, 0.0033035278, -0.014678955, -0.027954102, -0.025405884, 0.04559326, 0.012992859, 0.0025596619, 0.0078086853, -0.029663086, -0.042144775, -0.0013027191, -0.023483276, ...
http://arxiv.org/abs/2305.12940v1
GEST: the Graph of Events in Space and Time as a Common Representation between Vision and Language
Mihai Masala, Nicolae Cudlenco, Traian Rebedea, Marius Leordeanu
cs.CL
One of the essential human skills is the ability to seamlessly build an inner representation of the world. By exploiting this representation, humans are capable of easily finding consensus between visual, auditory and linguistic perspectives. In this work, we set out to understand and emulate this ability through an ex...
http://arxiv.org/pdf/2305.12940v1
2023-05-22T11:38:27
2023-05-22T11:38:27
{ "title": [ 0.04144287, -0.030349731, -0.019714355, -0.044769287, -0.037597656, -0.029876709, -0.033569336, -0.013824463, 0.034179688, 0.01663208, -0.016784668, -0.030532837, 0.059753418, -0.050964355, -0.015213013, -0.012794495, 0.020523071, 0....
http://arxiv.org/abs/2305.12927v1
Exploring Speaker-Related Information in Spoken Language Understanding for Better Speaker Diarization
Luyao Cheng, Siqi Zheng, Zhang Qinglin, Hui Wang, Yafeng Chen, Qian Chen
cs.CL, cs.SD, eess.AS
Speaker diarization(SD) is a classic task in speech processing and is crucial in multi-party scenarios such as meetings and conversations. Current mainstream speaker diarization approaches consider acoustic information only, which result in performance degradation when encountering adverse acoustic conditions. In this ...
http://arxiv.org/pdf/2305.12927v1
2023-05-22T11:14:19
2023-05-22T11:14:19
{ "title": [ -0.004512787, 0.0039253235, -0.03213501, -0.060455322, -0.007461548, -0.0064697266, -0.045318604, 0.00730896, -0.017211914, 0.030151367, -0.02494812, 0.007293701, 0.0007991791, -0.012619019, -0.028152466, -0.008056641, -0.0101623535, ...
http://arxiv.org/abs/2305.12924v1
EnCore: Pre-Training Entity Encoders using Coreference Chains
Frank Mtumbuka, Steven Schockaert
cs.CL
Entity typing is the task of assigning semantic types to the entities that are mentioned in a text. Since obtaining sufficient amounts of manual annotations is expensive, current state-of-the-art methods are typically trained on automatically labelled datasets, e.g. by exploiting links between Wikipedia pages. In this ...
http://arxiv.org/pdf/2305.12924v1
2023-05-22T11:11:59
2023-05-22T11:11:59
{ "title": [ -0.0067596436, 0.0039863586, -0.03656006, -0.07019043, 0.020843506, -0.00592041, -0.021209717, 0.0007452965, 0.01499939, 0.04449463, -0.008399963, 0.020904541, 0.025802612, -0.010108948, -0.033081055, 0.03652954, -0.016143799, 0.0580...
http://arxiv.org/abs/2305.12920v3
A Diachronic Analysis of Paradigm Shifts in NLP Research: When, How, and Why?
Aniket Pramanick, Yufang Hou, Saif M. Mohammad, Iryna Gurevych
cs.CL
Understanding the fundamental concepts and trends in a scientific field is crucial for keeping abreast of its continuous advancement. In this study, we propose a systematic framework for analyzing the evolution of research topics in a scientific field using causal discovery and inference techniques. We define three var...
http://arxiv.org/pdf/2305.12920v3
2023-10-25T11:56:49
2023-05-22T11:08:00
{ "title": [ -0.025497437, -0.011177063, -0.059387207, -0.040283203, 0.0063323975, -0.0209198, -0.010879517, 0.0027046204, 0.007827759, 0.009963989, -0.040252686, 0.015655518, -0.005870819, -0.0056648254, -0.05117798, -0.0022220612, -0.025482178, ...