Federated learning improves robot manipulation across different tasks
Federated Subspace Guided Vision-Language-Action Policy Distillation for Non-IID Multi-Robot Manipulation
RoboticsComputer Vision and Pattern RecognitionDistributed, Parallel, and Cluster ComputingMachine Learning
Summary
Training multiple robots to work together on different tasks is tricky because their experiences and environments vary a lot. The authors designed a method called FedDRMan that helps robots share useful skills without needing to pool all their data in one place. Their system groups robots with similar updates and carefully blends their knowledge to keep important behaviors intact. Tests showed FedDRMan helped robots learn better manipulation skills than other federation methods, boosting success rates by a noticeable margin.
What this means in practice
- •For industrial automation teams: Improve collaborative robot training in factories with diverse tasks and equipment without centralizing data.
- •For autonomous warehouse operators: Enhance policy sharing across robot fleets performing varied manipulation tasks under different conditions.
Authors
Biprodip Pal, Kaushik Roy, Yanming Zhu, Brendan Tidd, Alan Wee-Chung Liew, Peyman Moghadam
Abstract
Federated learning offers a natural way for multiple robots to jointly improve manipulation policies without requiring centralized access to training demonstrations. However, non-IID task and environment distributions can induce representation drift and mutually incompatible robot-policy updates, making naive parameter aggregation destructive. We present FedDRMan, a federated subspace-guided distillation framework for heterogeneous robot manipulation. At each communication round, the server model provides a frozen teacher for local behavior cloning, while low-rank multimodal subspace and action-distribution distillation preserve globally useful representation geometry and policy behavior. To address heterogeneous aggregation, FedDRMan groups clients by update compatibility and maintains a persistent model for each cluster. The server then spectrally rebalances each compatible aggregate to mitigate attenuation of weaker task-relevant robot-policy update directions. Extensive experiments on LIBERO across diverse non-IID settings, heterogeneity levels, client participation variation, together with ablations and aggregation analyses, show that FedDRMan substantially improves knowledge transfer and consistently outperforms strong federated baselines achieving a peak mean success rate of 80.7%, 11.6 percentage points above the strongest evaluated federated baseline.