ET-SEED: Efficient Trajectory-Level SE(3) Equivariant Diffusion Policy
This work addresses the need for more data-efficient imitation learning policies in robotics, offering a novel method that improves training efficiency and generalization with fewer demonstrations.
The paper tackles the problem of reducing demonstration reliance in imitation learning for robotic manipulation by proposing ET-SEED, an efficient trajectory-level SE(3) equivariant diffusion model, which achieves superior data efficiency and manipulation proficiency in tasks involving rigid, articulated, and deformable objects.
Imitation learning, e.g., diffusion policy, has been proven effective in various robotic manipulation tasks. However, extensive demonstrations are required for policy robustness and generalization. To reduce the demonstration reliance, we leverage spatial symmetry and propose ET-SEED, an efficient trajectory-level SE(3) equivariant diffusion model for generating action sequences in complex robot manipulation tasks. Further, previous equivariant diffusion models require the per-step equivariance in the Markov process, making it difficult to learn policy under such strong constraints. We theoretically extend equivariant Markov kernels and simplify the condition of equivariant diffusion process, thereby significantly improving training efficiency for trajectory-level SE(3) equivariant diffusion policy in an end-to-end manner. We evaluate ET-SEED on representative robotic manipulation tasks, involving rigid body, articulated and deformable object. Experiments demonstrate superior data efficiency and manipulation proficiency of our proposed method, as well as its ability to generalize to unseen configurations with only a few demonstrations. Website: https://et-seed.github.io/