Online Algorithms for Spectral Hypergraph Sparsification
Offered By: Simons Institute via YouTube
Course Description
Overview
Explore a 32-minute lecture on online algorithms for spectral hypergraph sparsification presented by Yuichi Yoshida from the National Institute of Informatics at the Simons Institute. Delve into the first online algorithm for spectral hypergraph sparsification, where hyperedges with positive weights arrive in a stream, requiring immediate decisions on inclusion in the sparsifier. Discover how this algorithm produces an (ϵ,δ)-spectral sparsifier with multiplicative error ϵ and additive error δ, achieving O(ϵ^{-2} n log n log r log(1+ϵW/δn)) hyperedges with high probability. Learn about the significant space complexity improvement, from Ω(m) in previous algorithms to O(n2), offering an exponential reduction since m can be exponential in n. Gain insights into this groundbreaking work, based on collaborative research with Tasuku Soma and Kam Chuen Tung, as part of the Sublinear Graph Simplification series.
Syllabus
Online Algorithms for Spectral Hypergraph Sparsification
Taught by
Simons Institute
Related Courses
Sampling-Based Sublinear Low-Rank Matrix Arithmetic Framework for Dequantizing Quantum Machine LearningAssociation for Computing Machinery (ACM) via YouTube Sublinear Algorithms for Gap Edit Distance
IEEE via YouTube High Dimensional Robust Sparse Regression
Simons Institute via YouTube Learning-Augmented Sketches for Frequency Estimation
Simons Institute via YouTube Adaptive Sparse Recovery with Limited Adaptivity
Simons Institute via YouTube