Class Central is learner-supported. When you buy through links on our site, we may earn an affiliate commission.

YouTube

Stephen Wright: Fundamentals of Optimization in Signal Processing

Hausdorff Center for Mathematics via YouTube

Overview

Save Big on Coursera Plus. 7,000+ courses at $160 off. Limited Time Only!
Explore the second lecture in a series on optimization fundamentals for signal processing, delivered by Stephen Wright. Delve into key topics including group sparsity, matrix inference problems, and various optimization algorithms. Learn about non-overlapping groups, tree-structured groups, and nuclear norm regularization. Examine steepest descent methods, backtracking, and convergence rates for different scenarios. Discover multistep methods like the Heavy-Ball algorithm and conjugate gradient, and gain insights into accelerated first-order methods. This comprehensive lecture provides essential knowledge for tackling complex optimization challenges in signal processing applications.

Syllabus

Group Sparsity
Three Scenarios
Non-overlapping Groups
Tree-Structured Groups
Matrix Inference Problems
Nuclear Norm Regularization
Another Matrix Inference Problem: Inverse Covariance
Atomic-Norm Regularization
References
Steepest Descent
Backtracking
Weakly convex: 1/k sublinear rate
Linear convergence without strong convexity
The slow linear rate is typical!
Multistep Methods: The Heavy-Ball
Conjugate Gradient
Accelerated First-Order Methods

Taught by

Hausdorff Center for Mathematics

Reviews

Start your review of Stephen Wright: Fundamentals of Optimization in Signal Processing

Never Stop Learning.

Get personalized course recommendations, track subjects and courses with reminders, and more.

Someone learning on their laptop while sitting on the floor.