The Robust Manifold Defense: Adversarial Training using Generative Models.
Andrew IlyasAjil JalalEirini AsteriConstantinos DaskalakisAlexandros G. DimakisPublished in: CoRR (2017)
Keyphrases
- generative model
- probabilistic model
- mixture model
- discriminative learning
- hierarchical hidden markov models
- discriminative models
- discriminatively trained
- em algorithm
- conditional random fields
- prior knowledge
- maximum entropy principle
- sparse coding
- representational power
- supervised learning
- topic models
- training set
- information retrieval
- deep belief networks
- similarity measure
- object categories
- semi supervised
- expectation maximization
- text classification