The Randomized Block Coordinate Descent Method in the H\"older Smooth Setting

Abstract

This work provides the first convergence analysis for the Randomized Block Coordinate Descent method for minimizing a function that is both H\"older smooth and block H\"older smooth. Our analysis applies to objective functions that are non-convex, convex, and strongly convex. For non-convex functions, we show that the expected gradient norm reduces at an $O\left(k^{\frac{\gamma}{1+\gamma}}\right)$ rate, where $k$ is the iteration count and $\gamma$ is the H\"older exponent. For convex functions, we show that the expected suboptimality gap reduces at the rate $O\left(k^{-\gamma}\right)$. In the strongly convex setting, we show this rate for the expected suboptimality gap improves to $O\left(k^{-\frac{2\gamma}{1-\gamma}}\right)$ when $\gamma>1$ and to a linear rate when $\gamma=1$. Notably, these new convergence rates coincide with those furnished in the existing literature for the Lipschitz smooth setting.