Policy Reuse in Deep Reinforcement Learning.
Ruben GlattAnna Helena Reali CostaPublished in: AAAI (2017)
Keyphrases
- reinforcement learning
- optimal policy
- policy search
- markov decision process
- action selection
- reinforcement learning problems
- reinforcement learning algorithms
- reward function
- function approximators
- partially observable environments
- partially observable
- markov decision processes
- state space
- control policy
- policy gradient
- policy iteration
- actor critic
- function approximation
- control policies
- state and action spaces
- dynamic programming
- markov decision problems
- action space
- approximate dynamic programming
- rl algorithms
- policy evaluation
- model free reinforcement learning
- finite state
- temporal difference
- approximate policy iteration
- average reward
- learning objects
- multi agent
- continuous state spaces
- model free
- machine learning
- transition model
- long run
- state dependent
- control problems
- partially observable markov decision processes
- reinforcement learning methods
- temporal difference learning
- state action
- deep learning
- software reuse
- policy makers
- supervised learning
- learning process
- policy gradient methods