Learning to Factor Policies and Action-Value Functions: Factored Action Space Representations for Deep Reinforcement learning.
Sahil SharmaAravind SureshRahul RameshBalaraman RavindranPublished in: CoRR (2017)
Keyphrases
- reinforcement learning
- action space
- state space
- state action
- policy search
- action selection
- optimal policy
- reinforcement learning methods
- markov decision processes
- learning algorithm
- control policies
- continuous state
- function approximators
- markov decision process
- continuous state spaces
- function approximation
- supervised learning
- continuous action
- state and action spaces
- fitted q iteration
- learning agent
- model free
- learning problems
- learning tasks
- dynamic programming
- cooperative
- action models
- heuristic search
- prior knowledge
- decision making