SHARP: Search-Based Adversarial Attack for Structured Prediction
2022
会议录名称FINDINGS OF THE ASSOCIATION FOR COMPUTATIONAL LINGUISTICS: NAACL 2022 - FINDINGS
页码950-961
发表状态已发表
摘要

Adversarial attack of structured prediction models faces various challenges such as the difficulty of perturbing discrete words, the sentence quality issue, and the sensitivity of outputs to small perturbations. In this work, we introduce SHARP, a new attack method that formulates the black-box adversarial attack as a search-based optimization problem with a specially designed objective function considering sentence fluency, meaning preservation and attacking effectiveness. Additionally, three different searching strategies are analyzed and compared, i.e., Beam Search, Metropolis- Hastings Sampling, and Hybrid Search. We demonstrate the effectiveness of our attacking strategies on two challenging structured prediction tasks: part-of-speech (POS) tagging and dependency parsing. Through automatic and human evaluations, we show that our method performs a more potent attack compared with pioneer arts. Moreover, the generated adversarial examples can be used to successfully boost the robustness and performance of the victim model via adversarial training. © Findings of the Association for Computational Linguistics: NAACL 2022 - Findings.

关键词Computational linguistics Forecasting Attack methods Black boxes Objective functions Optimization problems Prediction modelling Quality issues Search based optimizations Search-based Small perturbations Structured prediction
会议名称2022 Findings of the Association for Computational Linguistics: NAACL 2022
会议地点Seattle, WA, United states
会议日期July 10, 2022 - July 15, 2022
收录类别EI
语种英语
出版者Association for Computational Linguistics (ACL)
EI入藏号20223712713466
EI主题词Syntactics
EI分类号721.1 Computer Theory, Includes Formal Logic, Automata Theory, Switching Theory, Programming Theory
原始文献类型Conference article (CA)
文献类型会议论文
条目标识符https://kms.shanghaitech.edu.cn/handle/2MSLDSTB/229869
专题信息科学与技术学院_博士生
信息科学与技术学院_PI研究组_屠可伟组
作者单位
1.School of Information Science and Technology, ShanghaiTech University, China;
2.Shanghai Engineering Research Center of Intelligent Vision and Imaging, Shanghai Institute of Microsystem and Information Technology, Chinese Academy of Sciences, University of Chinese Academy of Sciences, China;
3.Beijing Institute for General Artificial Intelligence, China
第一作者单位信息科学与技术学院
第一作者的第一单位信息科学与技术学院
推荐引用方式
GB/T 7714
Zhang, Liwen,Jia, Zixia,Han, Wenjuan,et al. SHARP: Search-Based Adversarial Attack for Structured Prediction[C]:Association for Computational Linguistics (ACL),2022:950-961.
条目包含的文件
文件名称/大小 文献类型 版本类型 开放类型 使用许可
个性服务
查看访问统计
谷歌学术
谷歌学术中相似的文章
[Zhang, Liwen]的文章
[Jia, Zixia]的文章
[Han, Wenjuan]的文章
百度学术
百度学术中相似的文章
[Zhang, Liwen]的文章
[Jia, Zixia]的文章
[Han, Wenjuan]的文章
必应学术
必应学术中相似的文章
[Zhang, Liwen]的文章
[Jia, Zixia]的文章
[Han, Wenjuan]的文章
相关权益政策
暂无数据
收藏/分享
所有评论 (0)
暂无评论
 

除非特别说明,本系统中所有内容都受版权保护,并保留所有权利。