A backend for excuting machine learning models stored in the Open Neural Network Exchange format.

Abstract

This thesis investigates the effectiveness of using ONNX-based machine learning models to simulate cellular automata, specifically focusing on Conway’s Game of Life. The primary goal is to evaluate how well an ONNX-based probabilistic model can replace the traditional deterministic simulation approach used in PCraster, and to assess its ability to represent and predict the behavior of cellular automata.