Learning from Human Directional Corrections

Wanxin Jin, Todd D. Murphey, Zehui Lu +1

This paper proposes a novel approach that enables a robot to learn an objective function incrementally from human directional corrections. Existing methods learn from human magnitu…