The nn2poly package implements the NN2Poly method that allows to transform an already trained deep feed-forward fully connected neural network into a polynomial representation that predicts as similar as possible to the original neural network. 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).
Related Papers:
Pablo Morala, J. Alexandra Cifuentes, Rosa E. Lillo, Iñaki Ucar (2021). “Towards a mathematical framework to inform neural network modelling via polynomial regression.” Neural Networks, 142, 57-72. doi: 10.1016/j.neunet.2021.04.036
Pablo Morala, J. Alexandra Cifuentes, Rosa E. Lillo, Iñaki Ucar (2023). “NN2Poly: A Polynomial Representation for Deep Feed-Forward Artificial Neural Networks.” IEEE Transactions on Neural Networks and Learning Systems, (Early Access). doi: 10.1109/TNNLS.2023.3330328
Installation
The installation from GitHub requires the remotes package.
# install.packages("remotes")
remotes::install_github("IBiDat/nn2poly")