Paper Code

This repository contains the code for the paper "GraphFSA: A Finite State Automaton Framework for Algorithmic Learning on Graphs".

Prerequisites

• Anaconda or Miniconda installed. If not installed, you can download and install it from here.

Setup and Installation

To run the code, it's recommended to use the provided conda environment. Here's how to set it up:

1. Clone the repository

2. Create a new conda environment from the provided environment.yml file conda env create -f environment.yml

3. Activate the environment conda activate graphfsa

Reproducability

For reproducability please check the reproducability.txt in the folders of the specific experiment categrories.

Folder Structure

Our experiments can be divided into three experiment categories

1. Cellular Automata
2. Dataset Generator
3. Algorithms

The dataset generator uses a custom generator package located in graph-generator/fsm_dataset. This graph generator allows us to create random datasets. It is important that we additionally install this dataset generator by calling

cd graph-generator/fsm_dataset
pip install -e .

To train with the configurations used for the paper, we refer to the commands in experiments.md