Note on usage and theory of hyper-dual numbers for first and second order automatic differentiation

01/11/2018
by   Martin Neuenhofen, et al.
0

Similar to dual numbers, hyper-dual numbers are extended numbers that can be used for automatic differentiation. In contrast to dual numbers, hyper-dual numbers can be directly used to compute derivatives of second order. In this note we motivate hyper-dual numbers from finite differences and dual numbers. In this note we explain in an illustrative way how to use hyper-dual numbers. We convey our exposition with mathematical theorems, that validate the correctness of the derivatives as computed with hyper-dual numbers. Eventually we refer to freely available implementations of hyper-dual numbers.

READ FULL TEXT

page 1

page 2

page 3

page 4

research
01/11/2018

Review of theory and implementation of hyper-dual numbers for first and second order automatic differentiation

In this review we present hyper-dual numbers as a tool for the automatic...
research
06/13/2023

A dual number formulation to efficiently compute higher order directional derivatives

This contribution proposes a new formulation to efficiently compute dire...
research
07/05/2023

Using Rewrite Strategies for Efficient Functional Automatic Differentiation

Automatic Differentiation (AD) has become a dominant technique in ML. AD...
research
07/07/2022

Efficient Dual-Numbers Reverse AD via Well-Known Program Transformations

Where dual-numbers forward-mode automatic differentiation (AD) pairs eac...
research
10/12/2022

A language for evaluating derivatives of functionals using automatic differentiation

We present a simple functional programming language, called Dual PCF, th...
research
09/03/2019

Hyper-Pairing Network for Multi-Phase Pancreatic Ductal Adenocarcinoma Segmentation

Pancreatic ductal adenocarcinoma (PDAC) is one of the most lethal cancer...
research
01/04/2023

Higher order Bernstein-Bézier and Nédélec finite elements for the relaxed micromorphic model

The relaxed micromorphic model is a generalized continuum model that is ...

Please sign up or login with your details

Forgot password? Click here to reset