Learning Rate Annealing Improves Tuning Robustness in Stochastic Optimization

The learning rate in stochastic gradient methods is a critical hyperparameter that is notoriously costly to tune via standard grid search, especially for training modern large-scale models with billions of parameters. We identify a theoretical advantage of learning rate annealing schemes that decay the learning rate to zero at a polynomial rate, such as the widely-used cosine schedule, by demonstrating their increased robustness to initial parameter misspecification due to a coarse grid search. We present an analysis in a stochastic convex optimization setup demonstrating that the convergence rate of stochastic gradient descent with annealed schedules depends sublinearly on the multiplicative misspecification factor ρ (i.e., the grid resolution), achieving a rate of O(ρ1/(2p+1)/√T) where p is the degree of polynomial decay and T is the number of steps. This is in contrast to the O(ρ/√T) rate obtained under the inverse-square-root and fixed stepsize schedules, which depend linearly on ρ. Experiments confirm the increased robustness compared to tuning with a fixed stepsize, that has significant implications for the computational overhead of hyperparameter search in practical training scenarios.