DemoGen: Synthetic Demonstration Generation for Data-Efficient Visuomotor Policy Learning
Zhengrong Xue, Shuying Deng, Zhenyang Chen, Yixuan Wang, Zhecheng Yuan, Huazhe Xu · 2025
100x Synthetic Demos "One-Shot" Imitation Versatile Skills & Platforms Extended O.O.D. CapabilitiesFig. 1: DemoGen is a fully synthetic approach for automatic demonstration generation.DemoGen promotes the spatial generalization ability of visuomotor policies and can facilitate one-shot imitation by adapting one human-collected demonstration into novel object configurations.DemoGen applies to various manipulation tasks and platforms and can be extended to enable additional out-of-distribution capabilities.Abstract-Visuomotor policies have shown great promise in robotic manipulation but often require substantial humancollected data for effective performance.A key factor driving the high data demands is their limited spatial generalization capability, which necessitates extensive data collection across different object configurations.In this work, we present DemoGen, a low-cost, fully synthetic approach for automatic demonstration generation.Using only one human-collected demonstration per task, DemoGen generates spatially augmented demonstrations by adapting the demonstrated action trajectory to novel object configurations.Visual observations are synthesized by leveraging 3D point clouds as the modality and rearranging the subjects in the scene via 3D editing.Empirically, DemoGen significantly enhances policy performance across a diverse range of real-world manipulation tasks, showing its applicability even in challenging scenarios involving deformable objects, dexterous hand endeffectors, and bimanual platforms.Furthermore, DemoGen can be extended to enable additional out-of-distribution capabilities, including disturbance resistance and obstacle avoidance.