Safe Reinforcement Learning via Shielding.
Mohammed AlshiekhRoderick BloemRüdiger EhlersBettina KönighoferScott NiekumUfuk TopcuPublished in: AAAI (2018)
Keyphrases
- reinforcement learning
- function approximation
- markov decision processes
- state space
- learning algorithm
- temporal difference
- optimal policy
- reinforcement learning algorithms
- multi agent
- dynamic programming
- continuous state
- control system
- social networks
- learning capabilities
- supervised learning
- machine learning
- model free
- control problems
- reinforcement learning methods
- direct policy search
- autonomous learning
- evolutionary learning
- multi agent reinforcement learning
- policy search
- perceptual aliasing
- robotic control
- real time
- temporal difference learning
- action selection
- transfer learning
- search space
- decision trees
- artificial intelligence
- information retrieval
- neural network
- data sets