Safer reinforcement learning through evolved instincts.
Djordje GrbicSebastian RisiPublished in: GECCO Companion (2020)
Keyphrases
- reinforcement learning
- function approximation
- state space
- reinforcement learning algorithms
- multi agent
- dynamic programming
- model free
- learning process
- markov decision processes
- data mining
- temporal difference learning
- optimal policy
- robotic control
- control problems
- optimal control
- search space
- learning algorithm
- real time
- supervised learning
- learning problems
- bayesian networks
- search engine
- artificial intelligence
- partially observable
- learning agents
- machine learning
- autonomous learning
- evolutionary learning
- relational reinforcement learning
- neural network