Pseudo-rehearsal: Achieving deep reinforcement learning without catastrophic forgetting.
Craig AtkinsonBrendan McCaneLech SzymanskiAnthony V. RobinsPublished in: Neurocomputing (2021)
Keyphrases
- reinforcement learning
- function approximation
- model free
- state space
- machine learning
- reinforcement learning algorithms
- markov decision processes
- learning algorithm
- incremental learning
- robotic control
- temporal difference learning
- control problems
- optimal control
- learning process
- unsupervised learning
- semi supervised
- hidden markov models
- action selection
- temporal difference
- multi agent
- learning agents
- data mining
- policy search
- database