Stochastic Gradient Descent for SVM - Lecture 21

Explore a comprehensive lecture on stochastic sub-gradient descent as an optimization strategy for Support Vector Machine (SVM) loss functions. This 1 hour 20 minute session from UofU Data Science examines the effectiveness of this approach and investigates its relationship with the perceptron algorithm. Gain valuable insights into machine learning optimization techniques through practical explanations and theoretical connections. Additional resources are available through the accompanying lecture materials webpage.