Optimization in Signal Processing and Machine Learning

Explore advanced signal processing concepts in this IEEE Signal Processing Society webinar featuring Prof. John Wright. Delve into optimization techniques, overcomplete dictionaries, and nonconvex optimization as applied to signal processing. Gain insights into dictionary learning, blind deconvolution, and local geometry. Examine practical applications through MATLAB experiments and simulation results. Understand the theoretical foundations with a focus on geometric theorems. Engage with complex topics such as sign permutation symmetry and objective functions in dictionary learning. Benefit from Prof. Wright's expertise as he bridges the gap between theory and practice in modern signal processing optimization techniques.

Introduction
Welcome
What is optimization
Experiment
Setup
Optimization Problem
Questions
Sign permutation symmetry
Matlab
Overcomplete dictionaries
Practice gap
How can we understand
Nonconvex optimization
Local geometry
Dictionary learning
Objective function
Geometric intuition
Blind Deconvolution
Local perturbations
Simulation results
Theory
Geometric Theorem