Learning preferences for manipulation tasks from online coactive feedback

We consider the problem of learning preferences over trajectories for mobile manipulators such as personal robots and assembly line robots. The preferences we learn are more intricate than simple g...