Composite Convex Minimization

This lecture covers solution methods for composite convex minimization, including the proximal-gradient algorithm and the fast proximal-gradient algorithm. The instructor explains the basic schemes, convergence theorems, and complexity per iteration. The lecture also delves into examples such as ₁-regularized least squares and theoretical bounds versus practical performance. Additionally, it explores the stochastic convex composite minimization problem and the gradient mapping operator. The content includes discussions on the proximal-gradient method, Frank-Wolfe's approach, and phase retrieval. The lecture concludes with a focus on non-convex problems, stationary points, and stochastic constrained problems.