To Reviewer 1 (R1)
–Neural Information Processing Systems
We thank all three reviewers for their constructive comments. We address them below one by one. Q1: what makes it nontrivial to extend the regularity condition and proof technique in [11] to Riemmanian optimization. The Grassmannian manifold is nonconvex, making the analysis more complex. We will incorporate these into a revised version of the manuscript.
Neural Information Processing Systems
Aug-19-2025, 23:04:42 GMT
- Technology: