DARLA: Improving Zero-Shot Transfer in Reinforcement Learning.
Irina HigginsArka PalAndrei A. RusuLoïc MattheyChristopher BurgessAlexander PritzelMatthew BotvinickCharles BlundellAlexander LerchnerPublished in: ICML (2017)
Keyphrases
- reinforcement learning
- transfer learning
- function approximation
- model free
- knowledge transfer
- markov decision processes
- optimal policy
- learning process
- reinforcement learning algorithms
- dynamic programming
- state space
- artificial intelligence
- learning algorithm
- real time
- temporal difference
- multi agent reinforcement learning
- robot control
- stochastic approximation
- support vector
- website
- feature selection
- genetic algorithm
- information retrieval
- neural network