Guiding Representation Learning in Deep Generative Models with Policy Gradients.
Luca LachTimo KorthalsFrancesco FerroHelge J. RitterMalte SchillingPublished in: OLA (2021)
Keyphrases
- generative model
- discriminative learning
- prior knowledge
- representational power
- boltzmann machine
- deep belief networks
- learning process
- reinforcement learning
- unsupervised learning
- supervised learning
- deep learning
- learning tasks
- hidden variables
- discriminative methods
- parameter learning
- bayesian networks
- positive examples
- object categorization
- highly efficient
- conditional random fields
- markov random field
- e learning
- learning algorithm