The Devil Is in the GAN: Backdoor Attacks and Defenses in Deep Generative Models.
Ambrish RawatKillian LevacherMathieu SinnPublished in: ESORICS (3) (2022)
Keyphrases
- generative model
- denial of service attacks
- probabilistic model
- dos attacks
- mixture model
- deep belief networks
- semi supervised
- em algorithm
- discriminative learning
- conditional random fields
- object categories
- hierarchical hidden markov models
- discriminative models
- prior knowledge
- network security
- ddos attacks
- maximum entropy principle
- hidden variables
- topic models
- deep learning
- expectation maximization
- multiscale
- machine learning
- object recognition
- discriminatively trained
- image processing