Overview
Learn about stochastic sub-gradient descent (SGD) optimization algorithms for Support Vector Machines (SVMs) in this comprehensive lecture that explores the connection between SGD and the perceptron algorithm. Dive deep into the mathematical foundations and practical implementations of SVM optimization techniques, gaining valuable insights into how these fundamental machine learning algorithms work together. Master the concepts through detailed explanations and discover why SGD remains a simple yet highly effective method for training SVMs in modern machine learning applications.
Syllabus
Machine Learning: Lecture 22a: SGD for SVMs
Taught by
UofU Data Science