nn2poly: Neural Network Weights Transformation into Polynomial Coefficients

Implements a method that builds the coefficients of a polynomial model that performs almost equivalently as a given neural network (densely connected). This is achieved using Taylor expansion at the activation functions. The obtained polynomial coefficients can be used to explain features (and their interactions) importance in the neural network, therefore working as a tool for interpretability or eXplainable Artificial Intelligence (XAI). See Morala et al. 2021 <doi:10.1016/j.neunet.2021.04.036>, and 2023 <doi:10.1109/TNNLS.2023.3330328>.

Version: 0.1.1
Depends: R (≥ 3.5.0)
Imports: Rcpp, generics, matrixStats, pracma
LinkingTo: Rcpp, RcppArmadillo
Suggests: keras, tensorflow, reticulate, luz, torch, cowplot, ggplot2, patchwork, testthat (≥ 3.0.0), vdiffr, knitr, rmarkdown
Published: 2024-01-30
Author: Pablo Morala ORCID iD [aut, cre], Iñaki Ucar ORCID iD [aut], Jose Ignacio Diez [ctr]
Maintainer: Pablo Morala <moralapablo at gmail.com>
License: MIT + file LICENSE
URL: https://ibidat.github.io/nn2poly/
NeedsCompilation: yes
Citation: nn2poly citation info
Materials: README NEWS
CRAN checks: nn2poly results

Documentation:

Reference manual: nn2poly.pdf
Vignettes: Introduction to nn2poly
Supported DL frameworks
Classification example using tensorflow

Downloads:

Package source: nn2poly_0.1.1.tar.gz
Windows binaries: r-prerel: nn2poly_0.1.1.zip, r-release: nn2poly_0.1.1.zip, r-oldrel: nn2poly_0.1.1.zip
macOS binaries: r-prerel (arm64): nn2poly_0.1.1.tgz, r-release (arm64): nn2poly_0.1.1.tgz, r-oldrel (arm64): nn2poly_0.1.1.tgz, r-prerel (x86_64): nn2poly_0.1.1.tgz, r-release (x86_64): nn2poly_0.1.1.tgz
Old sources: nn2poly archive

Linking:

Please use the canonical form https://CRAN.R-project.org/package=nn2poly to link to this page.