On the role of planning in model-based deep reinforcement learning.
Jessica B. HamrickAbram L. FriesenFeryal BehbahaniArthur GuezFabio ViolaSims WitherspoonThomas AnthonyLars Holger BuesingPetar VelickovicTheophane WeberPublished in: ICLR (2021)
Keyphrases
- reinforcement learning
- model free
- state space
- planning problems
- partial observability
- function approximation
- action selection
- deterministic domains
- macro actions
- stochastic domains
- partially observable
- goal oriented
- reinforcement learning algorithms
- markov decision processes
- heuristic search
- planning process
- machine learning
- multi agent reinforcement learning
- motion planning
- planning systems
- partially observable markov decision processes
- deep learning
- optimal control
- blocks world
- reinforcement learning methods
- optimal policy
- ai planning
- reinforcement learning problems
- dynamic programming
- multi agent
- temporal difference