Controlling the Fidelity and Diversity of Deep Generative Models via Pseudo Density.
Shuangqi LiChen LiuTong ZhangHieu LeSabine SüsstrunkMathieu SalzmannPublished in: CoRR (2024)
Keyphrases
- generative model
- probabilistic model
- deep belief networks
- mixture model
- discriminative learning
- discriminative models
- prior knowledge
- em algorithm
- semi supervised
- expectation maximization
- topic models
- object categories
- hidden variables
- conditional random fields
- data sets
- hierarchical hidden markov models
- deep learning
- maximum entropy principle
- graphical models
- computer vision
- markov random field
- image segmentation
- discriminative methods
- machine learning