Convex Training Algorithms for Hard Machine Learning Problems

Learn about convex optimization approaches for challenging machine learning problems in this 78-minute plenary lecture delivered by Dale Schuurmans from the University of Alberta. Explore advanced training algorithms that leverage convex optimization principles to tackle computationally difficult machine learning challenges. Discover how convex methods can provide theoretical guarantees and practical solutions for problems that are traditionally considered intractable. Gain insights into the mathematical foundations and algorithmic techniques that enable effective training procedures for complex learning scenarios. Examine real-world applications and case studies demonstrating the effectiveness of convex training approaches across various machine learning domains.