Doubly Adaptive Scaled Algorithm for Machine Learning Using Second-Order Information
arXiv:2109.05198
Abstract
We present a novel adaptive optimization algorithm for large-scale machine learning problems. Equipped with a low-cost estimate of local curvature and Lipschitz smoothness, our method dynamically adapts the search direction and step-size. The search direction contains gradient information preconditioned by a well-scaled diagonal preconditioning matrix that captures the local curvature information. Our methodology does not require the tedious task of learning rate tuning, as the learning rate is updated automatically without adding an extra hyperparameter. We provide convergence guarantees on a comprehensive collection of optimization problems, including convex, strongly convex, and nonconvex problems, in both deterministic and stochastic regimes. We also conduct an extensive empirical evaluation on standard machine learning problems, justifying our algorithm's versatility and demonstrating its strong performance compared to other start-of-the-art first-order and second-order methods.
References in corpus (9)
- Decoupled Weight Decay Regularization
- SGDR: Stochastic Gradient Descent with Warm Restarts
- On the Convergence of Adam and Beyond
- Global Convergence of Online Limited Memory BFGS
- Online Learning Rate Adaptation with Hypergradient Descent
- Stochastic Polyak Step-size for SGD: An Adaptive Learning Rate for Fast Convergence
- PyHessian: Neural Networks Through the Lens of the Hessian
- Inexact Non-Convex Newton-Type Methods
- Gradient Descent: The Ultimate Optimizer