The Extended Regularized Dual Averaging Method for Composite Optimization
Year: 2019
Type: preprint
Abstract: We present a new algorithm, extended regularized dual averaging (XRDA), for solving composite optimization problems, which are a generalization of the regularized dual averaging (RDA) method. The main novelty of the method is that it allows more flexible control of the backward step size. For instance, the backward step size for RDA grows without bound, while XRDA the backward step size can be kept bounded.
Source: arXiv (Cornell University)
Authors: Jonathan W. Siegel, Jinchao Xu
Institution: Pennsylvania State University
Language: English
Subfield: Artificial Intelligence
Field: Computer Science
Domain: Physical Sciences
Open Access status: green
OpenAlex ID: W3135471456