Researcher, Facebook
1 paper at NeurIPS 2025
We develop an optimizer ASGO that can provably exploit the low-rank gradients and block-wise diagonal Hessians in training.