Asymptotic Guarantees for Learning Generative Models with the Sliced-Wasserstein Distance.
Kimia NadjahiAlain DurmusUmut SimsekliRoland BadeauPublished in: NeurIPS (2019)
Keyphrases
- generative model
- discriminative learning
- boltzmann machine
- prior knowledge
- mixture model
- learning process
- parameter learning
- hidden variables
- learning tasks
- probabilistic model
- weakly supervised
- learning algorithm
- natural images
- object recognition
- supervised learning
- topic models
- information extraction
- pairwise
- reinforcement learning
- representational power