Ricci-NN : Ricci flow-like process for Neural-Network

This repository contains both the R and Python codes to run the analysis decribed in the following article:

arXiv link to the preprint: https://arxiv.org/abs/2404.14265

The R codes have been created by Christopher Banerji and the Python codes by Anthony Baptista.

R version

fMNIST_DNN_training.r: Code to train the different Neural Network architectures
fMNIST_kNN_RCoef_eval.r: Exploration of the k value for the k-nearest-neighbours (knn) graph construction. The graphs construct based on the knn are used to computed the Ricci flow-low like process

Python version

fmnist_extraction.py: Code to extract from the FMNIST dataset the test and train sub-dataset for the cloths labelled 5 (Sandal) and 9 (Ankle Boot). The raw data can be found at the following link: https://www.kaggle.com/datasets/zalando-research/fashionmnist
training.py: Code to train the different Neural Network architectures
knn.py: Exploration of the k value for the k-nearest-neighbours (knn) graph construction. The graphs construct based on the knn are used to computed the Ricci flow-low like process

Applications on FMINST