Safe Exploration Method for Reinforcement Learning under Existence of Disturbance.
Yoshihiro OkawaTomotake SasakiHitoshi YanamiToru NamerikawaPublished in: CoRR (2022)
Keyphrases
- detection method
- high accuracy
- clustering method
- similarity measure
- reinforcement learning
- synthetic data
- prior knowledge
- significant improvement
- cost function
- objective function
- neural network
- computational cost
- input data
- model selection
- high precision
- optimization algorithm
- computationally efficient
- probabilistic model
- feature vectors
- preprocessing