Deep Variational Reinforcement Learning for POMDPs.
Maximilian IglLuisa M. ZintgrafTuan Anh LeFrank WoodShimon WhitesonPublished in: ICML (2018)
Keyphrases
- reinforcement learning
- partially observable markov decision processes
- state space
- partially observable
- policy search
- function approximation
- markov decision processes
- optimal policy
- continuous state
- learning algorithm
- model free
- machine learning
- optical flow
- temporal difference
- multi agent
- learning problems
- reinforcement learning algorithms
- image segmentation
- variational methods
- learning process
- dynamic programming
- action selection
- optimal control
- supervised learning
- reinforcement learning methods
- model based reinforcement learning
- actor critic
- policy iteration algorithm
- partial observability
- markov decision problems
- function approximators
- search space
- reward function
- heuristic search