Detection of Backdoors in Trained Classifiers Without Access to the Training Set
IEEE Transactions on Neural Networks and Learning Systems(2022)
摘要
With wide deployment of deep neural network (DNN) classifiers, there is great potential for harm from adversarial learning attacks. Recently, a special type of data poisoning (DP) attack, known as a backdoor (or Trojan), was proposed. These attacks do not seek to degrade classification accuracy, but rather to have the classifier learn to classify to a target class 更多
查看译文
关键词
Training,Perturbation methods,Databases,Tuning,Trojan horses,Toxicology,Testing
AI 理解论文
溯源树
样例
生成溯源树,研究论文发展脉络
Chat Paper
正在生成论文摘要