Hypothesis-Driven Skill Discovery for Hierarchical Deep Reinforcement Learning.
Caleb ChuckSupawit ChockchowwatScott NiekumPublished in: CoRR (2019)
Keyphrases
- reinforcement learning
- function approximation
- data driven
- multi agent
- knowledge discovery
- dynamic programming
- state space
- hierarchical reinforcement learning
- markov decision processes
- scientific discovery
- model free
- learning algorithm
- discovery process
- hierarchical model
- multiple layers
- skills needed
- robotic control
- transfer learning
- data mining
- learning classifier systems
- pattern discovery
- coarse to fine
- deep learning
- learning tasks
- hierarchical structure
- active learning