中国科学院机构知识库网格
Chinese Academy of Sciences Institutional Repositories Grid
首页 机构 成果 学者
  
  1. CAS IR Grid
  2. 计算技术研究所
  3. 中国科学院计算技术研究所
  4. 中国科学院计算技术研究所期刊论文
  5. 英文
MaxMatch: Semi-Supervised Learning With Worst-Case Consistency

文献类型：期刊论文

作者Jiang, Yangbangyan3,5; Li, Xiaodan4; Chen, Yuefeng4; He, Yuan4; Xu, Qianqian6; Yang, Zhiyong7; Cao, Xiaochun8; Huang, Qingming1,2,6,7
刊名IEEE TRANSACTIONS ON PATTERN ANALYSIS AND MACHINE INTELLIGENCE
出版日期2023-05-01
卷号45期号:5页码:5970-5987
关键词Predictive models Training Data models Semantics Perturbation methods Computational modeling Benchmark testing Semi-supervised learning consistency regularization worst-case consistency image classification
ISSN号0162-8828
DOI10.1109/TPAMI.2022.3208419
英文摘要In recent years, great progress has been made to incorporate unlabeled data to overcome the inefficiently supervised problem via semi-supervised learning (SSL). Most state-of-the-art models are based on the idea of pursuing consistent model predictions over unlabeled data toward the input noise, which is called consistency regularization. Nonetheless, there is a lack of theoretical insights into the reason behind its success. To bridge the gap between theoretical and practical results, we propose a worst-case consistency regularization technique for SSL in this article. Specifically, we first present a generalization bound for SSL consisting of the empirical loss terms observed on labeled and unlabeled training data separately. Motivated by this bound, we derive an SSL objective that minimizes the largest inconsistency between an original unlabeled sample and its multiple augmented variants. We then provide a simple but effective algorithm to solve the proposed minimax problem, and theoretically prove that it converges to a stationary point. Experiments on five popular benchmark datasets validate the effectiveness of our proposed method.
资助项目National Key R&D Program of China[2018AAA0102000] ; National Natural Science Foundation of China[U21B2038] ; National Natural Science Foundation of China[61931008] ; National Natural Science Foundation of China[62025604] ; National Natural Science Foundation of China[U1936208] ; National Natural Science Foundation of China[6212200758] ; National Natural Science Foundation of China[61976202] ; Fundamental Research Funds for the Central Universities ; Youth Innovation Promotion Association CAS ; Strategic Priority Research Program of Chinese Academy of Sciences[XDB28000000] ; China National Postdoctoral Program for Innovative Talents[BX2021298] ; China Postdoctoral Science Foundation[2022M713101]
WOS研究方向Computer Science ; Engineering
语种英语
WOS记录号WOS:000964792800040
出版者IEEE COMPUTER SOC
源URL[http://119.78.100.204/handle/2XEOYT63/21397]  
专题中国科学院计算技术研究所期刊论文_英文
通讯作者Xu, Qianqian; Huang, Qingming
作者单位1.Univ Chinese Acad Sci, Key Lab Big Data Min & Knowledge Management BDKM, Beijing 101408, Peoples R China
2.Peng Cheng Lab, Shenzhen 518055, Peoples R China
3.Chinese Acad Sci, Inst Informat Engn, State Key Lab Informat Secur, Beijing 100093, Peoples R China
4.Alibaba Grp, Secur Dept, Hangzhou 311121, Peoples R China
5.Univ Chinese Acad Sci, Sch Cyber Secur, Beijing 100049, Peoples R China
6.Chinese Acad Sci, Inst Comp Technol, Key Lab Intelligent Informat Proc, Beijing 100190, Peoples R China
7.Univ Chinese Acad Sci, Sch Comp Sci & Technol, Beijing 101408, Peoples R China
8.Sun Yat Sen Univ, Sch Cyber Sci & Technol, Shenzhen Campus, Shenzhen 518107, Peoples R China
推荐引用方式
GB/T 7714		 Jiang, Yangbangyan,Li, Xiaodan,Chen, Yuefeng,et al. MaxMatch: Semi-Supervised Learning With Worst-Case Consistency[J]. IEEE TRANSACTIONS ON PATTERN ANALYSIS AND MACHINE INTELLIGENCE,2023,45(5):5970-5987.
APA Jiang, Yangbangyan.,Li, Xiaodan.,Chen, Yuefeng.,He, Yuan.,Xu, Qianqian.,...&Huang, Qingming.(2023).MaxMatch: Semi-Supervised Learning With Worst-Case Consistency.IEEE TRANSACTIONS ON PATTERN ANALYSIS AND MACHINE INTELLIGENCE,45(5),5970-5987.
MLA Jiang, Yangbangyan,et al."MaxMatch: Semi-Supervised Learning With Worst-Case Consistency".IEEE TRANSACTIONS ON PATTERN ANALYSIS AND MACHINE INTELLIGENCE 45.5(2023):5970-5987.

入库方式： OAI收割

来源：计算技术研究所

浏览0
下载0
收藏0
其他版本

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。