Opening book details…
Can I read A Unified Variance-reduced Accelerated Gradient Method for Convex Optimization on EtoBox?
A Unified Variance-reduced Accelerated Gradient Method for Convex Optimization by Lan, Guanghui; Li, Zhize; Zhou, Yi is a scholarly article available to read on EtoBox.
What is A Unified Variance-reduced Accelerated Gradient Method for Convex Optimization about?
We propose a novel randomized incremental gradient algorithm, namely, VAriance-Reduced Accelerated Gradient (Varag), for finite-sum optimization. Equipped with a unified step-size policy that adjusts itself to the value of the condition number, Varag exhibits the unified optimal rates of convergence for solving smooth convex finite-sum problems directly regardless of their strong convexity. Moreover, Varag is the first accelerated randomized incremental gradient method that benefits from the strong convexity of the data-fidelity term to achieve the optimal linear convergence. It also establishes an optimal linear rate of convergence for solving a wide class of problems only satisfying a certain error bound condition rather than strong convexity. Varag can also be extended to solve stochastic finite-sum problems.
- Author
- Lan, Guanghui; Li, Zhize; Zhou, Yi
- Published
- 2019
- Language
- EN