When a Robot is More Capable than a Human: Learning from Constrained Demonstrators
This addresses the issue of suboptimal robot learning from imperfect human demonstrations, which is incremental as it builds on imitation learning methods.
The paper tackles the problem of robots learning suboptimal policies from demonstrations by constrained experts, such as due to control interfaces, by inferring a state-only reward signal and exploring more efficient trajectories, resulting in a 10x faster task completion time compared to behavioral cloning on a real robotic arm.
Learning from demonstrations enables experts to teach robots complex tasks using interfaces such as kinesthetic teaching, joystick control, and sim-to-real transfer. However, these interfaces often constrain the expert's ability to demonstrate optimal behavior due to indirect control, setup restrictions, and hardware safety. For example, a joystick can move a robotic arm only in a 2D plane, even though the robot operates in a higher-dimensional space. As a result, the demonstrations collected by constrained experts lead to suboptimal performance of the learned policies. This raises a key question: Can a robot learn a better policy than the one demonstrated by a constrained expert? We address this by allowing the agent to go beyond direct imitation of expert actions and explore shorter and more efficient trajectories. We use the demonstrations to infer a state-only reward signal that measures task progress, and self-label reward for unknown states using temporal interpolation. Our approach outperforms common imitation learning in both sample efficiency and task completion time. On a real WidowX robotic arm, it completes the task in 12 seconds, 10x faster than behavioral cloning, as shown in real-robot videos on https://sites.google.com/view/constrainedexpert .