Variational Policy Chaining for Lifelong Reinforcement Learning.
Christopher DoyleMaxime GuériauIvana DusparicPublished in: ICTAI (2019)
Keyphrases
- reinforcement learning
- optimal policy
- policy search
- markov decision process
- action selection
- control policy
- partially observable environments
- function approximation
- markov decision processes
- state and action spaces
- reinforcement learning problems
- policy gradient
- function approximators
- actor critic
- control policies
- policy iteration
- partially observable
- approximate dynamic programming
- markov decision problems
- reinforcement learning algorithms
- state space
- action space
- temporal difference
- reward function
- sufficient conditions
- image segmentation
- policy evaluation
- decision problems
- learning activities
- model free
- inverse reinforcement learning
- partially observable markov decision processes
- dynamic programming
- state action
- learning algorithm
- learning process
- state dependent
- st century
- finite state
- infinite horizon
- technology enhanced learning
- learning scenarios
- rl algorithms
- learning problems
- optimal control
- learning experience
- continuous state spaces
- transition model
- long run
- lifelong learning
- multi agent