Fast Streaming Euclidean Clustering with Constant Space
Offered By: Simons Institute via YouTube
Course Description
Overview
Explore a cutting-edge streaming algorithm for k-median and k-means clustering in this 31-minute lecture by Samson Zhou from Texas A&M University. Delve into the innovative approach that achieves constant space complexity independent of data stream size and aspect ratio. Discover how this algorithm optimizes processing time to poly(log log (n*Delta)) per stream item in the unit cost RAM model. Learn about the novel compression technique for merge and reduce trees and its applications, including improved space and update time for approximate subspace embeddings in streaming scenarios. Gain insights into sublinear graph simplification and its implications for efficient data processing in high-dimensional spaces.
Syllabus
Fast Streaming Euclidean Clustering with Constant Space
Taught by
Simons Institute
Related Courses
First Steps in Linear Algebra for Machine LearningHigher School of Economics via Coursera Vanessa Robins - The Extended Persistent Homology Transform for Manifolds with Boundary
Applied Algebraic Topology Network via YouTube Johnathan Bush - Maps of Čech and Vietoris–Rips Complexes into Euclidean Spaces
Applied Algebraic Topology Network via YouTube Convex Sunflower Theorems and Neural Codes
Applied Algebraic Topology Network via YouTube Borsuk-Ulam Theorems into Higher-Dimensional Codomains
Applied Algebraic Topology Network via YouTube