Reset-Free Guided Policy Search: Efficient Deep Reinforcement Learning with Stochastic Initial States.
William MontgomeryAnurag AjayChelsea FinnPieter AbbeelSergey LevinePublished in: CoRR (2016)
Keyphrases
- policy search
- reinforcement learning
- reinforcement learning algorithms
- continuous state
- learning algorithm
- continuous action
- multi agent
- dynamic programming
- state space
- markov decision problems
- function approximation
- policy gradient
- hidden state
- reward function
- state variables
- neural network
- supervised learning
- mobile robot
- machine learning