Efficient Differentiable Programming in a Functional Array-Processing Language
Mathematical Software
2018-06-07 v1 Machine Learning
Programming Languages
Symbolic Computation
Machine Learning
Abstract
We present a system for the automatic differentiation of a higher-order functional array-processing language. The core functional language underlying this system simultaneously supports both source-to-source automatic differentiation and global optimizations such as loop transformations. Thanks to this feature, we demonstrate how for some real-world machine learning and computer vision benchmarks, the system outperforms the state-of-the-art automatic differentiation tools.
Cite
@article{arxiv.1806.02136,
title = {Efficient Differentiable Programming in a Functional Array-Processing Language},
author = {Amir Shaikhha and Andrew Fitzgibbon and Dimitrios Vytiniotis and Simon Peyton Jones and Christoph Koch},
journal= {arXiv preprint arXiv:1806.02136},
year = {2018}
}