Hierarchically Decoupled Imitation for Morphological Transfer
This work addresses sample efficiency in robot learning for long-range tasks across different morphologies, presenting an incremental improvement in transfer methods.
The paper tackles the problem of learning long-range behaviors for complex high-dimensional robots by transferring learned information from simpler agents to improve sample efficiency, demonstrating applicability on navigation and manipulation environments with improved zero-shot transfer and stabilized learning.
Learning long-range behaviors on complex high-dimensional agents is a fundamental problem in robot learning. For such tasks, we argue that transferring learned information from a morphologically simpler agent can massively improve the sample efficiency of a more complex one. To this end, we propose a hierarchical decoupling of policies into two parts: an independently learned low-level policy and a transferable high-level policy. To remedy poor transfer performance due to mismatch in morphologies, we contribute two key ideas. First, we show that incentivizing a complex agent's low-level to imitate a simpler agent's low-level significantly improves zero-shot high-level transfer. Second, we show that KL-regularized training of the high level stabilizes learning and prevents mode-collapse. Finally, on a suite of publicly released navigation and manipulation environments, we demonstrate the applicability of hierarchical transfer on long-range tasks across morphologies. Our code and videos can be found at https://sites.google.com/berkeley.edu/morphology-transfer.