VASE: Variational Assorted Surprise Exploration for Reinforcement Learning.
Haitao XuLech SzymanskiBrendan McCanePublished in: IEEE Trans. Neural Networks Learn. Syst. (2023)
Keyphrases
- reinforcement learning
- active exploration
- exploration strategy
- action selection
- model based reinforcement learning
- function approximation
- exploration exploitation
- reinforcement learning algorithms
- markov decision processes
- image segmentation
- state space
- exploration exploitation tradeoff
- machine learning
- learning algorithm
- optical flow
- learning process
- model free
- dynamic programming
- autonomous learning
- data sets
- optimal policy
- robotic control
- computer vision
- multi agent
- policy search
- multi agent reinforcement learning
- stochastic approximation
- free energy
- markov decision process
- information visualization
- optimal control