Discover, Explain, Improve: An Automatic Slice Detection Benchmark for Natural Language Processing

Author:

Hua Wenyue1,Jin Lifeng2,Song Linfeng3,Mi Haitao4,Zhang Yongfeng5,Yu Dong6

Affiliation:

1. Rutgers University, New Brunswick, USA. wenyue.hua@rutgers.edu

2. Tencent America, USA. lifengjin@tencent.com

3. Tencent America, USA. lfsong@tencent.com

4. Tencent America, USA. haitaomi@tencent.com

5. Rutgers University, New Brunswick, USA. yongfeng.zhang@rutgers.edu

6. Tencent America, USA. dyu@tencent.com

Abstract

Abstract Pretrained natural language processing (NLP) models have achieved high overall performance, but they still make systematic errors. Instead of manual error analysis, research on slice detection models (SDMs), which automatically identify underperforming groups of datapoints, has caught escalated attention in Computer Vision for both understanding model behaviors and providing insights for future model training and designing. However, little research on SDMs and quantitative evaluation of their effectiveness have been conducted on NLP tasks. Our paper fills the gap by proposing a benchmark named “Discover, Explain, Improve (DEIm)” for classification NLP tasks along with a new SDM Edisa. Edisa discovers coherent and underperforming groups of datapoints; DEIm then unites them under human-understandable concepts and provides comprehensive evaluation tasks and corresponding quantitative metrics. The evaluation in DEIm shows that Edisa can accurately select error-prone datapoints with informative semantic features that summarize error patterns. Detecting difficult datapoints directly boosts model performance without tuning any original model parameters, showing that discovered slices are actionable for users.1

Publisher

MIT Press

Subject

Artificial Intelligence,Computer Science Applications,Linguistics and Language,Human-Computer Interaction,Communication

Reference33 articles.

1. Synthetic and natural noise both break neural machine translation;Belinkov;Proceedings of the 6th International Conference on Learning Representations,2017

2. Electra: Pre-training text encoders as discriminators rather than generators;Clark;Proceedings of ICLR,2020

3. Automatic error analysis for document-level information extraction;Das;Proceedings of the 60th Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers),2022

4. The spotlight: A general method for discovering systematic errors in deep learning models;d’Eon,2022

5. Learning confidence for out-of-distribution detection in neural networks;DeVries;arXiv preprint arXiv:1802.04865,2018

同舟云学术

1.学者识别学者识别

2.学术分析学术分析

3.人才评估人才评估

"同舟云学术"是以全球学者为主线,采集、加工和组织学术论文而形成的新型学术文献查询和分析系统,可以对全球学者进行文献检索和人才价值评估。用户可以通过关注某些学科领域的顶尖人物而持续追踪该领域的学科进展和研究前沿。经过近期的数据扩容,当前同舟云学术共收录了国内外主流学术期刊6万余种,收集的期刊论文及会议论文总量共计约1.5亿篇,并以每天添加12000余篇中外论文的速度递增。我们也可以为用户提供个性化、定制化的学者数据。欢迎来电咨询!咨询电话:010-8811{复制后删除}0370

www.globalauthorid.com

TOP

Copyright © 2019-2024 北京同舟云网络信息技术有限公司
京公网安备11010802033243号  京ICP备18003416号-3