The Bottleneck Simulator: A Model-Based Deep Reinforcement Learning Approach.
Iulian Vlad SerbanChinnadhurai SankarMichael PieperJoelle PineauYoshua BengioPublished in: J. Artif. Intell. Res. (2020)
Keyphrases
- reinforcement learning
- model free
- real robot
- reinforcement learning algorithms
- function approximation
- state space
- simulation model
- markov decision processes
- optimal control
- data driven
- learning capabilities
- multi agent
- temporal difference
- fully unsupervised
- reinforcement learning methods
- simulation tool
- optimal policy
- neural network
- least squares
- learning process
- artificial neural networks
- knowledge base
- e learning
- learning algorithm
- machine learning