| Title | : | Fast Algorithms and Data Structures for Regression |
| Speaker | : | Deeksha Adil (ETH Zurich) |
| Details | : | Thu, 1 May, 2025 11:00 AM @ SSB 334 |
| Abstract: | : | Recent breakthroughs in fast algorithms have efficiently combined tools from optimization, data structures, and sketching. In this talk I will present these ideas via some of my recent works on the $ell_p$-norm regression, $p in [2,infty]$. I will introduce novel acceleration frameworks and demonstrate how they can be integrated with advanced data structures. Notably, our work is the first to combine acceleration with matrix inverse maintenance data structures efficiently, making the two most important building blocks of modern structured convex optimization compatible. Additionally, I will also present some algorithms that are fast in practice. Finally, I will present some open problems that are in the intersection of these domains. |
