An object-oriented representation for efficient reinforcement learning.
Carlos DiukAndre CohenMichael L. LittmanPublished in: ICML (2008)
Keyphrases
- reinforcement learning
- object oriented
- function approximation
- computationally efficient
- artificial intelligence
- cost effective
- state space
- representation scheme
- optimal control
- optimal policy
- lightweight
- neural network
- dynamic programming
- object recognition
- search algorithm
- data structure
- image sequences
- learning algorithm