The Value Function Polytope in Reinforcement Learning.
Robert DadashiMarc G. BellemareAdrien Ali TaïgaNicolas Le RouxDale SchuurmansPublished in: ICML (2019)
Keyphrases
- reinforcement learning
- function approximators
- function approximation
- machine learning
- information systems
- temporal difference learning
- supervised learning
- markov decision processes
- convex hull
- lattice points
- policy gradient
- state action
- reinforcement learning algorithms
- action selection
- state space
- dynamic programming
- learning process
- objective function
- learning algorithm