.png){kind=icon}
.png){kind=icon}
Return
MSG: Multi-Stream Generative Policies for Sample-Efficient Robotic Manipulation
.png){kind=icon}
PRE
AI
DOI：10.1109/LRA.2026.3682566.png)
.png){kind=icon}
Original request for help.png){kind=icon}
Share.png){kind=icon}
Save
Abstract
En 中文
Generative robot policies such as Flow Matching offer flexible, multi-modal policy learning but are sample-inefficient. Although object-centric policies improve sample efficiency, it does not resolve this limitation. In this work, we propose Multi-Stream Generative Policy (MSG), an inference-time composition framework that trains multiple object-centric policies and combines them at inference to improve generalization and sample efficiency. MSG is model-agnostic and inference-only, hence widely applicable to various generative policies and training paradigms. We perform extensive experiments both in simulation and on a real robot, demonstrating that our approach learns high-quality generative policies from as few as five demonstrations, resulting in a 95% reduction in demonstrations, and improves policy performance by 89 percent compared to single-stream approaches. Furthermore, we present comprehensive ablation studies on various composition strategies and provide practical recommendations for deployment. Finally, MSG enables zero-shot object instance transfer.
Keywords:
Imitation learning
learning from demonstration
deep learning in grasping and manipulation
Journal
I
IF:
5.3
Papers:
1.6K
Citations:
3.9W