Using Deep Reinforcement Learning to Learn High-Level Policies on the ATRIAS Biped.
Tianyu LiAkshara RaiHartmut GeyerChristopher G. AtkesonPublished in: CoRR (2018)
Keyphrases
- reinforcement learning
- high level
- optimal policy
- hierarchical reinforcement learning
- reinforcement learning agents
- low level
- control policies
- markov decision process
- policy search
- markov decision processes
- function approximators
- state space
- lower level
- higher level
- multi agent
- reinforcement learning algorithms
- function approximation
- control policy
- reward function
- learning algorithm
- learning agent
- model free
- state abstraction
- multiagent reinforcement learning
- policy gradient methods
- fitted q iteration
- multi agent reinforcement learning
- complex domains
- neural network
- control parameters
- humanoid robot
- dynamic programming
- machine learning