Postdoc, Massachusetts Institute of Technology
1 paper at NeurIPS 2025
This paper proposes a novel geometric deep learning model to preserve the underlying mixture of symmetry transformations for sample-efficient dynamic learning
