Backdoor Vulnerabilities in Normally Trained Deep Learning Models.
Guanhong TaoZhenting WangSiyuan ChengShiqing MaShengwei AnYingqi LiuGuangyu ShenZhuo ZhangYunshu MaoXiangyu ZhangPublished in: CoRR (2022)
Keyphrases
- learning models
- machine learning
- semi supervised learning
- loss function
- learning tasks
- learning algorithm
- learning paradigms
- machine learning algorithms
- learning problems
- conditional random fields
- machine learning models
- classification models
- svm classifier
- training set
- image classification
- image features
- pairwise
- genetic algorithm