A Study of Condition Numbers for First-Order Optimization
arXiv:2012.05782
Abstract
The study of first-order optimization algorithms (FOA) typically starts with assumptions on the objective functions, most commonly smoothness and strong convexity. These metrics are used to tune the hyperparameters of FOA. We introduce a class of perturbations quantified via a new norm, called *-norm. We show that adding a small perturbation to the objective function has an equivalently small impact on the behavior of any FOA, which suggests that it should have a minor impact on the tuning of the algorithm. However, we show that smoothness and strong convexity can be heavily impacted by arbitrarily small perturbations, leading to excessively conservative tunings and convergence issues. In view of these observations, we propose a notion of continuity of the metrics, which is essential for a robust tuning strategy. Since smoothness and strong convexity are not continuous, we propose a comprehensive study of existing alternative metrics which we prove to be continuous. We describe their mutual relations and provide their guaranteed convergence rates for the Gradient Descent algorithm accordingly tuned. Finally we discuss how our work impacts the theoretical understanding of FOA and their performances.
References in corpus (9)
- Non-strongly-convex smooth stochastic approximation with convergence rate O(1/n)
- Gradient Descent Learns Linear Dynamical Systems
- YellowFin and the Art of Momentum Tuning
- Gradient methods for convex minimization: better rates under weaker conditions
- SGD: General Analysis and Improved Rates
- Accelerated Bregman Proximal Gradient Methods for Relatively Smooth Convex Optimization
- Linear Convergence of Variance-Reduced Stochastic Gradient without Strong Convexity
- Potential-Function Proofs for First-Order Methods
- Characterizations of Lojasiewicz inequalities and applications