DeHiB: Deep Hidden Backdoor Attack on Semi-supervised Learning via Adversarial Perturbation.
Zhicong YanGaolei LiYuan TIanJun WuShenghong LiMingzhe ChenH. Vincent PoorPublished in: AAAI (2021)
Keyphrases
- semi supervised learning
- semi supervised
- unlabeled data
- labeled data
- supervised learning
- machine learning
- unsupervised learning
- training data
- semi supervised classification
- learning problems
- manifold regularization
- improve the classification accuracy
- co training
- learning models
- transfer learning
- label propagation
- regularization framework
- domain adaptation
- labeled examples
- semi supervised learning algorithms
- metric learning
- active learning
- graph based semi supervised learning
- maximum likelihood
- text classification
- data sets
- transductive support vector machine
- learning algorithm
- information retrieval
- real world