Path Length Bounds for Gradient Descent and Flow
Chirag Gupta, Sivaraman Balakrishnan, Aaditya Ramdas.
Year: 2021, Volume: 22, Issue: 68, Pages: 1−63
Abstract
We derive bounds on the path length $\zeta$ of gradient descent (GD) and gradient flow (GF) curves for various classes of smooth convex and nonconvex functions. Among other results, we prove that: (a) if the iterates are linearly convergent with factor $(1-c)$, then $\zeta$ is at most $\mathcal{O}(1/c)$; (b) under the Polyak-Kurdyka-\L ojasiewicz (PKL) condition, $\zeta$ is at most $\mathcal{O}(\sqrt{\kappa})$, where $\kappa$ is the condition number, and at least $\widetilde\Omega(\sqrt{d} \wedge \kappa^{1/4})$; (c) for quadratics, $\zeta$ is $\Theta(\min\{\sqrt{d},\sqrt{\log \kappa}\})$ and in some cases can be independent of $\kappa$; (d) assuming just convexity, $\zeta$ can be at most $2^{4d\log d}$; (e) for separable quasiconvex functions, $\zeta$ is ${\Theta}(\sqrt{d})$. Thus, we advance current understanding of the properties of GD and GF curves beyond rates of convergence. We expect our techniques to facilitate future studies for other algorithms.