PAC-inspired Option Discovery in Lifelong Reinforcement Learning.
Emma BrunskillLihong LiPublished in: ICML (2014)
Keyphrases
- reinforcement learning
- function approximation
- state space
- knowledge discovery
- markov decision processes
- temporal difference
- data mining
- robotic control
- policy search
- pac learning
- reinforcement learning algorithms
- sample complexity
- st century
- optimal control
- learning problems
- sample size
- multi agent
- upper bound
- learning algorithm
- discovery process
- scientific discovery
- learning theory
- theoretical analysis
- dynamic programming
- active learning
- lifelong learning
- function approximators
- temporal difference learning
- machine learning
- neural network