Zero-shot policy generation in lifelong reinforcement learning.
Yiming QianFangzhou XiongZhi-Yong LiuPublished in: Neurocomputing (2021)
Keyphrases
- reinforcement learning
- optimal policy
- policy search
- markov decision process
- reinforcement learning problems
- state space
- reinforcement learning algorithms
- action selection
- function approximators
- control policy
- action space
- approximate dynamic programming
- control policies
- function approximation
- actor critic
- learning process
- policy gradient
- partially observable environments
- multi agent
- partially observable
- supervised learning
- reward function
- markov decision processes
- policy iteration
- state and action spaces
- learning activities
- markov decision problems
- inverse reinforcement learning
- long run
- state dependent
- competence development
- continuous state spaces
- policy evaluation
- model free
- average reward
- control problems
- decision problems
- policy making
- reinforcement learning methods
- state action
- allocation policy
- agent receives
- model free reinforcement learning