ACORNS: An Easy-To-Use Code Generator for Gradients and Hessians

TL;DR

ACORNS is a user-friendly code generator that automatically computes gradients and Hessians for algorithms written in C99, enhancing efficiency and reliability in scientific computing and machine learning applications.

Contribution

It introduces an algorithm and implementation for automatic differentiation of C99 code, facilitating easier and more reliable derivative computations.

Findings

Enables automatic differentiation of C99 algorithms.

Improves efficiency and reliability in derivative calculations.

Applicable to physical simulation and geometry processing.

Abstract

The computation of first and second-order derivatives is a staple in many computing applications, ranging from machine learning to scientific computing. We propose an algorithm to automatically differentiate algorithms written in a subset of C99 code and its efficient implementation as a Python script. We demonstrate that our algorithm enables automatic, reliable, and efficient differentiation of common algorithms used in physical simulation and geometry processing.