Perturbation confusion in forward automatic differentiation of higher-order functions
Explore this paper's citation graph
Summary
A subtle bug is exhibited, present in fielded implementations which support derivatives of higher-order functions, in which perturbations are confused despite the tagging machinery, leading to incorrect results.
- Type
- article
- Published
- 2012-11-20
- Cited by
- 17
- References
- 47
- Access
- Open access
- OpenAlex
- https://openalex.org/W1533682968
- Semantic Scholar
- https://api.semanticscholar.org/CorpusID:9403678
Keywords
Computer science, Automatic differentiation, Invocation, Derivative (finance), Function (biology)
References
- TAPENADE 2.1 user's guide
- Functional Differential Geometry
- Compiling Fast Partial Derivatives of Functions Given by Algorithms
- Evaluating derivatives - principles and techniques of algorithmic differentiation, Second Edition
- The calculi of lambda-conversion
- FADBAD, a flexible C++ package for automatic differentiation - using the forward and backward method
- Opticks, or a Treatise of the Reflexions, Refractions, Inflexions and Colours of Light: Also Two Treatises of the Species and Magnitude of Curvilinear Figures
- Structure and interpretation of classical mechanics
- Nesting forward-mode AD in a functional framework
- ADIFOR-Generating Derivative Codes from Fortran Programs
- Basic Concepts of Synthetic Differential Geometry
- A Simply Typed λ-Calculus of Forward Automatic Differentiation
- A simple automatic derivative evaluation program
- A dependent nominal type theory
- Lazy multivariate higher-order forward-mode AD
- Automatic Differentiation: Techniques and Applications
- First-class nonstandard interpretations by opening closures
- Functional Differentiation of Computer Programs
- Beautiful differentiation
- Evolving the Incremental Lambda Calculus into a Model of Forward Automatic Differentiation (AD)
Cited by
- Automatic differentiation in machine learning: a survey
- Getting to the point: index sets and parallelism-preserving autodiff for pointful array programming
- A parameter refinement method for Ptychography based on Deep Learning concepts
- Forward- or Reverse-Mode Automatic Differentiation: What's the Difference?
- dO: A Differentiable Engine for Deep Lens Design of Computational Imaging Systems
- AD for an Array Language with Nested Parallelism
- A language for evaluating derivatives of functionals using automatic differentiation
- Optimising Hyperparameters of Artificial Neural Network Topology for SHM Damage Detection and Identification
- Composing Automatic Differentiation with Custom Derivatives of Higher-Order Functions
- Optimized visual simulation of 3D light field display based on differentiable ray tracing
- Rose: Composable Autodiff for the Interactive Web (Artifact)
- AbstractDifferentiation.jl: Backend-Agnostic Differentiable Programming in Julia
- On the Versatility of Open Logical Relations
- Forward- or reverse-mode automatic differentiation: What's the difference?
- On the Versatility of Open Logical Relations: Continuity, Automatic Differentiation, and a Containment Theorem
- Automatic differentiation for performing the Cauchy-Kovalevskaya procedure in Lax-Wendroff type discretizations
- Automatic differentiation in machine learning: a survey
- Relevant Text Extraction Using Enhanced Artificial Neural Network Fuzzy Inference System
- ATMAD: Agile Transistor Compact Modeling with Parameter Extraction Based on Automatic Differentiation