Coordinate descent on the orthogonal group for recurrent neural network training

Author name not available (Why is that?)

Publication date: 30 July 2021

Abstract: We propose to use stochastic Riemannian coordinate descent on the orthogonal group for recurrent neural network training. The algorithm rotates successively two columns of the recurrent matrix, an operation that can be efficiently implemented as a multiplication by a Givens matrix. In the case when the coordinate is selected uniformly at random at each iteration, we prove the convergence of the proposed algorithm under standard assumptions on the loss function, stepsize and minibatch noise. In addition, we numerically demonstrate that the Riemannian gradient in recurrent neural network training has an approximately sparse structure. Leveraging this observation, we propose a faster variant of the proposed algorithm that relies on the Gauss-Southwell rule. Experiments on a benchmark recurrent neural network training problem are presented to demonstrate the effectiveness of the proposed algorithm.

Has companion code repository: https://github.com/EMassart/OrthCDforRNNs

This page was built for publication: Coordinate descent on the orthogonal group for recurrent neural network training

Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q6374181)