103
Voted
ICML
2008
IEEE
16 years 3 months ago
2008
IEEE
We consider the problem of learning to follow a desired trajectory when given a small number of demonstrations from a sub-optimal expert. We present an algorithm that (i) extracts...