Quaternionic Step Derivative: Automatic Differentiation of Holomorphic Functions using Complex Quaternions

10/19/2020
by   Martin Roelfs, et al.
0

Complex Step Derivative (CSD) allows easy and accurate differentiation up to machine precision of real functions by evaluating them a small imaginary step next to the real line. The current paper proposes that derivatives of holomorphic functions can be calculated in a similar fashion by taking a small step in a quaternionic direction instead. It is demonstrated that in so doing the properties of high accuracy and convergence are carried over to derivatives of holomorphic functions. Additionally, the extra degree of freedom means second derivatives of real functions are straightforwardly computed by taking a small step in both a quaternionic and complex direction, making the technique a powerful extension of CSD. To demonstrate the ease of implementation numerical experiments were performed using complex quaternions, the geometric algebra of space, and a 2 × 2 matrix representation thereof.

READ FULL TEXT
research
03/10/2020

Graded Automatic Differentiation

Based on a class of associative algebras with zero-divisors which are ca...
research
05/06/2021

The Complex-Step Derivative Approximation on Matrix Lie Groups

The complex-step derivative approximation is a numerical differentiation...
research
12/13/2021

On using the complex step method for the approximation of Fréchet derivatives of matrix functions in automorphism groups

We show, that the complex step approximation Im(f(A+ihE))/h to the Fréch...
research
01/01/2022

On automatic differentiation for the Matérn covariance

To target challenges in differentiable optimization we analyze and propo...
research
12/09/2020

A Differential Model of the Complex Cell

The receptive fields of simple cells in the visual cortex can be underst...
research
02/19/2021

High-order Differentiable Autoencoder for Nonlinear Model Reduction

This paper provides a new avenue for exploiting deep neural networks to ...
research
06/12/2020

On Correctness of Automatic Differentiation for Non-Differentiable Functions

Differentiation lies at the core of many machine-learning algorithms, an...

Please sign up or login with your details

Forgot password? Click here to reset