Reinforcement Learning with Function-Valued Action Spaces for Partial Differential Equation Control.
Yangchen PanAmir-massoud FarahmandMartha WhiteSaleh NabiPiyush GroverDaniel NikovskiPublished in: CoRR (2018)
Keyphrases
- action space
- partial differential equations
- reinforcement learning
- function approximators
- control policies
- continuous state
- markov decision processes
- state space
- control policy
- state and action spaces
- state action
- control problems
- level set
- image denoising
- integral equation
- action selection
- real valued
- optimal control
- function approximation
- stochastic processes
- image enhancement
- image processing
- continuous state spaces
- skill learning
- single agent
- reinforcement learning algorithms
- energy functional
- signed distance function
- temporal difference
- model free
- optimal policy
- multiscale
- machine learning
- continuous action
- numerical algorithms
- policy gradient
- natural images
- probability distribution
- bayesian networks
- computer vision