We study the problem of fitting task-specific learning rate schedules from the perspective of hyperparameter optimization. This allows us to explicitly search for schedules that achieve good generalization. We describe the structure of the gradient of a validation error w.r.t. the learning rate, the hypergradient, and based on this we introduce a novel online algorithm. Our method adaptively interpolates between the recently proposed techniques of Franceschi et al. (2017) and Baydin et al. (2017), featuring increased stability and faster convergence. We show empirically that the proposed method compares favourably with baselines and related methods in terms of final test accuracy.

Scheduling the Learning Rate via Hypergradients: New Insights and a New Algorithm