Evaluation of Safe Reinforcement Learning with CoMirror Algorithm in a Non-Markovian Reward Problem.
Megumi MiyashitaShiro YanoToshiyuki KondoPublished in: IAS (2022)
Keyphrases
- reinforcement learning
- dynamic programming
- learning algorithm
- optimization algorithm
- computational cost
- experimental evaluation
- preprocessing
- detection algorithm
- linear programming
- particle swarm optimization
- expectation maximization
- matching algorithm
- np hard
- computational complexity
- objective function
- probabilistic model
- segmentation algorithm
- high accuracy
- simulated annealing
- transfer learning
- bayesian networks
- worst case
- neural network
- cost function
- significant improvement
- search space