Multi-task policy search for robotics.
Marc Peter DeisenrothPeter EnglertJan PetersDieter FoxPublished in: ICRA (2014)
Keyphrases
- multi task
- policy search
- multi task learning
- reinforcement learning
- learning tasks
- continuous state
- multi class
- learning problems
- reinforcement learning algorithms
- metric learning
- transfer learning
- dynamic programming
- feature selection
- partially observable markov decision processes
- learning algorithm
- reward function
- finite state
- function approximators
- supervised learning
- pairwise
- policy gradient
- data mining