Backtracking Gradient Descent allowing unbounded learning rates

TL;DR

This paper extends convergence analysis of Gradient Descent by allowing unbounded learning rates through a backtracking method, potentially leading to better minima in unconstrained optimization.

Contribution

It introduces a novel unbounded backtracking approach for learning rates in Gradient Descent, proving convergence under general conditions and demonstrating its optimal growth rate.

Findings

Unbounded learning rates can improve convergence to better minima.

The proposed method generalizes previous backtracking algorithms.

The growth rate of the unbounded learning rates is shown to be optimal.

Abstract

In unconstrained optimisation on an Euclidean space, to prove convergence in Gradient Descent processes (GD) $x_{n+1}=x_n-\delta _n \nabla f(x_n)$ it usually is required that the learning rates $\delta _n$'s are bounded: $\delta _n\leq \delta $ for some positive $\delta $. Under this assumption, if the sequence $x_n$ converges to a critical point $z$, then with large values of $n$ the update will be small because $||x_{n+1}-x_n||\lesssim ||\nabla f(x_n)||$. This may also force the sequence to converge to a bad minimum. If we can allow, at least theoretically, that the learning rates $\delta _n$'s are not bounded, then we may have better convergence to better minima. A previous joint paper by the author showed convergence for the usual version of Backtracking GD under very general assumptions on the cost function $f$. In this paper, we allow the learning rates $\delta _n$ to be unbounded, in the sense that there is a function $h:(0,\infty)\rightarrow (0,\infty )$ such that $\lim _{t\rightarrow 0}th(t)=0$ and $\delta _n\lesssim \max \{h(x_n),\delta \}$ satisfies Armijo's condition for all $n$, and prove convergence under the same assumptions as in the mentioned paper. It will be shown that this growth rate of $h$ is best possible if one wants convergence of the sequence $\{x_n\}$. A specific way for choosing $\delta _n$ in a discrete way connects to Two-way Backtracking GD defined in the mentioned paper. We provide some results which either improve or are implicitly contained in those in the mentioned paper and another recent paper on avoidance of saddle points.