Search papers, labs, and topics across Lattice.
This paper introduces a distributionally robust imitation learning (IL) framework that tackles the challenges posed by both policy-induced and uncertainty-induced distribution shifts, which are critical for ensuring safety in decision-making tasks. By integrating Taylor Series Imitation Learning (TaSIL) with distributionally robust adaptive control, the authors formulate an IL problem that optimizes performance while adhering to safety constraints. The effectiveness of this approach is validated through a UAV case study, demonstrating enhanced safety and reliability in uncertain environments.
Safety in imitation learning can be achieved even in the face of significant distribution shifts, as shown by a UAV navigating uncertain environments while avoiding danger zones.
Imitation learning (IL) has achieved remarkable success in complex decision-making tasks. However, its performance is highly sensitive to distribution shifts, which can pose significant safety risks. We propose a distributionally robust and safe IL framework that explicitly addresses both policy-induced and uncertainty-induced distribution shifts. Our approach develops a unified framework leveraging Taylor Series Imitation Learning (TaSIL) to mitigate policy-induced shifts and distributionally robust adaptive control to handle uncertainty-induced shifts. This architecture enables the formulation of an IL problem that optimizes performance under distributional uncertainty while systematically accounting for safety constraints. We demonstrate the effectiveness of the proposed approach on an unmanned aerial vehicle (UAV) case study where the UAV performs a task in an uncertain environment while avoiding unsafe regions.
