Poisoning attacks and countermeasures in intelligent networks: Status quo and prospects

• DOI: 10.1016/j.dcan.2021.07.009
• Publication Date: 2021-07-30T03:32:48Z
• Authors: Chen Wang; Jian Chen; Yang Yang; Xiaoqiang Ma; Jiangchuan Liu
• Keywords: Security threat; Poisoning attack; Machine learning; 0202 electrical engineering, electronic engineering, information engineering; Information technology; 02 engineering and technology; Intelligent networks; T58.5-58.64

Over the past years, emergence of intelligent networks empowered by machine learning techniques has brought great facilitates to different aspects human life. However, using in also presents potential security and privacy threats. A common practice is so-called poisoning attacks where malicious users inject fake training data with aim corrupting learned model. In this survey, we comprehensively review existing as well countermeasures for first time. We emphasize compare principles formal employed categories algorithms, analyze strengths limitations corresponding defense methods a compact form. highlight some remaining challenges future directions attack-defense confrontation promote further research emerging yet promising area.