Assistant Professor, Northeastern University
3 papers at NeurIPS 2025
We propose practical methods to incorporate symmetry into Diffusion Policy, improving performance while maintaining simplicity.
This paper introduces and evaluates G2Sphere, a general method for mapping object geometries to spherical signals using equivariant neural networks and the Fourier Transform.
We introduce Image-to-Sphere Policy (ISP), a novel SO(3)-equivariant policy learning framework with single eye-in-hand RGB inputs.
