The Value-Improvement Path: Towards Better Representations for Reinforcement Learning.
Will DabneyAndré BarretoMark RowlandRobert DadashiJohn QuanMarc G. BellemareDavid SilverPublished in: AAAI (2021)
Keyphrases
- reinforcement learning
- function approximation
- higher level
- state space
- shortest path
- function approximators
- model free
- learning algorithm
- reinforcement learning algorithms
- markov decision processes
- multiple representations
- robot control
- temporal difference learning
- reinforcement learning methods
- learning capabilities
- data mining
- stochastic approximation
- endpoints
- optimal policy
- dynamic programming
- multi agent
- case study
- decision trees
- machine learning