Fast Approximation of Rotations and Hessians matrices

Abstract

A new method to represent and approximate rotation matrices is introduced. The method represents approximations of a rotation matrix $Q$ with linearithmic complexity, i.e. with $\frac{1}{2}n\lg (n)$ rotations over pairs of coordinates, arranged in an FFT-like fashion. The approximation is "learned" using gradient descent. It allows to represent symmetric matrices $H$ as $QD{Q}^T$ where $D$ is a diagonal matrix. It can be used to approximate covariance matrix of Gaussian models in order to speed up inference, or to estimate and track the inverse Hessian of an objective function by relating changes in parameters to changes in gradient along the trajectory followed by the optimization procedure. Experiments were conducted to approximate synthetic matrices, covariance matrices of real data, and Hessian matrices of objective functions involved in machine learning problems.

Cited by in corpus (6)

• Fast Orthonormal Sparsifying Transforms Based on Householder Reflectors
• Kaleidoscope: An Efficient, Learnable Representation For All Structured Linear Maps
• Pixelated Butterfly: Simple and Efficient Sparse training for Neural Network Models
• Building Compact and Robust Deep Neural Networks with Toeplitz Matrices
• Ensemble Quasi-Newton HMC
• Parallelized Computation and Backpropagation Under Angle-Parametrized Orthogonal Matrices