This repository includes scripts to reproduce the consistency simulation results for p-Laplace learning from the paper below. The two main scripts are sbm.py, which reproduces the stochastic block model results, and image.py which reproduces the image classification results with MNIST and Cifar-10 data sets. Install the required packages in requirements.txt first.
J. Calder, N. Drenska. Consistency of semi-supervised learning, stochastic tug-of-war games, and the p-Laplacian, arXiv preprint: , 2024