IEEE

Fractional Gradient Descent With Matrix Stepsizes for Non-Convex Optimization

Abstract: Fractional derivatives generalize integer-order derivatives, making them relevant for studying their convergence in descent-based optimization algorithms. However, existing convergence ...
GitHub

PolarGrad: A Class of Matrix-Gradient Optimizers from a Unifying Preconditioning Perspective

PolarGrad (Polar Gradient methods; Lau et al., 2025) is a class of matrix-gradient optimizers based on the concept of gradient-anisotropy preconditioning in optimization. It has close relation to Muon ...
IEEE

Matrix-Based Representations and Gradient-Free Algorithms for Neural Network Training

Abstract: This paper presents a compact, matrix-based representation of neural networks. Although neural networks are often understood pictorially as interconnected neurons, they are fundamentally ...
GitHub

Matrix HPP - C++11 library for matrix class container and linear algebra computations

This library provides a self-contained and easy to use implementation of matrix container class. The main features include: Full template parameterization with support for both real and complex ...