A Mean-Field Optimal Control Formulation of Deep Learning.
Weinan EJiequn HanQianxiao LiPublished in: CoRR (2018)
Keyphrases
- optimal control
- deep learning
- dynamic programming
- unsupervised learning
- control problems
- unsupervised feature learning
- control strategy
- machine learning
- markov random field
- reinforcement learning
- em algorithm
- mental models
- weakly supervised
- optimal control problems
- real time
- maximum likelihood
- energy function
- mathematical model
- semi supervised
- multiscale