Fixed-Sparsity Matrix Approximation from Matrix-Vector Products
Contributed Talk at Mid-Atlantic Numerical Analysis Day, Temple University
Talks and presentations
The Low-Rank Bottleneck in Attention
Contributed Talk at New York Academy of Sciences Annual Machine Learning Symposium, New York Academy of Sciences
Fixed-Sparsity Matrix Approximation from Matrix-Vector Products
Contributed Talk at SIAM Linear Algebra, Sorbonne University
Near-Optimal Approximation of Matrix Functions by the Lanczos Method
Contributed Talk at SIAM New York-New Jersey-Pennsylvania, New Jersey Institute of Technology
Lanczos-FA is Nearly Krylov-Optimal for Rational Functions
Contributed Poster at "Modern Applied and Computational Analysis" / "Synergistic Interactions between Theory and Computation" / "Mihalis Fest", ICERM / MIT / Columbia
Designing Data Center Networks Using Bottleneck Structures
Conference talk at ACM SIGCOMM 2021, Virtual
Computing Bottleneck Structures at Scale
Conference talk at IEEE/ACM INDIS 2020 at SC20 (7th International Workshop on Innovating the Network for Data Intensive Science), Virtual