A Hybrid Algorithm for Convex Semidefinite Optimization
This provides a scalable solution for large-scale semidefinite programs applicable to various machine learning problems, though it appears incremental as a hybrid approach.
The authors tackled the problem of convex semidefinite optimization by developing a hybrid algorithm that converges to global optimal solutions, showing it outperforms state-of-the-art methods on tasks like matrix completion, metric learning, and sparse PCA.
We present a hybrid algorithm for optimizing a convex, smooth function over the cone of positive semidefinite matrices. Our algorithm converges to the global optimal solution and can be used to solve general large-scale semidefinite programs and hence can be readily applied to a variety of machine learning problems. We show experimental results on three machine learning problems (matrix completion, metric learning, and sparse PCA) . Our approach outperforms state-of-the-art algorithms.