A Geometric Theory of Higher-Order Automatic Differentiation
M. Betancourt. (2018)cite arxiv:1812.11592Comment: 55 pages, 10 figures.
Abstract
First-order automatic differentiation is a ubiquitous tool across statistics,
machine learning, and computer science. Higher-order implementations of
automatic differentiation, however, have yet to realize the same utility. In
this paper I derive a comprehensive, differential geometric treatment of
automatic differentiation that naturally identifies the higher-order
differential operators amenable to automatic differentiation as well as
explicit procedures that provide a scaffolding for high-performance
implementations.
Description
[1812.11592] A Geometric Theory of Higher-Order Automatic Differentiation
%0 Journal Article
%1 betancourt2018geometric
%A Betancourt, Michael
%D 2018
%K autograd geometry
%T A Geometric Theory of Higher-Order Automatic Differentiation
%U http://arxiv.org/abs/1812.11592
%X First-order automatic differentiation is a ubiquitous tool across statistics,
machine learning, and computer science. Higher-order implementations of
automatic differentiation, however, have yet to realize the same utility. In
this paper I derive a comprehensive, differential geometric treatment of
automatic differentiation that naturally identifies the higher-order
differential operators amenable to automatic differentiation as well as
explicit procedures that provide a scaffolding for high-performance
implementations.
@article{betancourt2018geometric,
abstract = {First-order automatic differentiation is a ubiquitous tool across statistics,
machine learning, and computer science. Higher-order implementations of
automatic differentiation, however, have yet to realize the same utility. In
this paper I derive a comprehensive, differential geometric treatment of
automatic differentiation that naturally identifies the higher-order
differential operators amenable to automatic differentiation as well as
explicit procedures that provide a scaffolding for high-performance
implementations.},
added-at = {2019-07-20T20:34:22.000+0200},
author = {Betancourt, Michael},
biburl = {https://www.bibsonomy.org/bibtex/29d3917cd4ae025be58b1bb43378b8358/kirk86},
description = {[1812.11592] A Geometric Theory of Higher-Order Automatic Differentiation},
interhash = {529f9bac377386cf4484633fcb336773},
intrahash = {9d3917cd4ae025be58b1bb43378b8358},
keywords = {autograd geometry},
note = {cite arxiv:1812.11592Comment: 55 pages, 10 figures},
timestamp = {2019-07-20T20:34:22.000+0200},
title = {A Geometric Theory of Higher-Order Automatic Differentiation},
url = {http://arxiv.org/abs/1812.11592},
year = 2018
}