paper-with-me

Papers

Understanding Automatic Differentiation Pitfalls

2023-05-12 · Jan Hückelheim, Harshitha Menon, William Moses, Bruce Christianson, Paul Hovland, Laurent Hascoët

Automatic differentiation, also known as backpropagation, AD, autodiff, or algorithmic differentiation, is a popular technique for computing derivatives of computer programs accurately and efficiently. Sometimes, however, the derivatives computed by AD could be interpreted as incorrect. These pitfalls occur systematically across tools and approaches. In this paper we broadly categorize problematic usages of AD and illustrate each category with examples such as chaos, time-averaged oscillations, discretizations, fixed-point loops, lookup tables, and linear solvers. We also review debugging techniques and their effectiveness in these situations. With this article we hope to help readers avoid unexpected behavior, detect problems more easily when they occur, and have more realistic expectations from AD tools.

📄 PDF Abstract BibTeX arXiv:2305.07546

Code (0)

등록된 구현이 없습니다.

Similar Papers 제목 키워드 기반

A scheme for automatic differentiation of complex loss functions

2020-03-02 · Chu Guo, Dario Poletti

For a real function, automatic differentiation is such a standard algorithm used to efficiently compute its gradient, that it is integrated in various neural network frameworks. However, despite the recent advances in us…

BIG-bench Machine Learning

Automatic Differentiation of Algorithms for Machine Learning

2014-04-28 · Atilim Gunes Baydin, Barak A. Pearlmutter

Automatic differentiation---the mechanical transformation of numeric computer programs to calculate derivatives efficiently and accurately---dates to the origin of the computer age. Reverse mode automatic differentiation…

BIG-bench Machine Learning

On the Equivalence of Automatic and Symbolic Differentiation

2019-04-05 · Soeren Laue

We show that reverse mode automatic differentiation and symbolic differentiation are equivalent in the sense that they both perform the same operations when computing derivatives. This is in stark contrast to the common …

Analyzing and Enhancing the Backward-Pass Convergence of Unrolled Optimization

2023-12-28 · James Kotary, Jacob Christopher, My H Dinh, Ferdinando Fioretto

The integration of constrained optimization models as components in deep networks has led to promising advances on many specialized learning tasks. A central challenge in this setting is backpropagation through the solut…

Automatic differentiation of nonsmooth iterative algorithms

2022-05-31 · Jérôme Bolte, Edouard Pauwels, Samuel Vaiter

Differentiation along algorithms, i.e., piggyback propagation of derivatives, is now routinely used to differentiate iterative solvers in differentiable programming. Asymptotics is well understood for many smooth problem…