umair-nasir14 / atep

ATEP is an open-ended learning system that introduces species-based evolution of Enhanced-POET.

ATEP

This repository contains code for Augmentative Topology Agents For Open-Ended Learning. The file structure is as follows:

├── atep                    -> The main ATEP code
├── dataframes              -> Contains csv files to reproduce the figures
├── demos                   -> Demo videos
├── figures                 -> The actual figures
├── jobscript.sh            -> Main jobscript to reproduce our experiments
├── neat                    -> Main code for neat

To create the environment, use environment.yml

conda env create -f environment.yml

To reproduce the experiments, run

qsub jobscript.sh

To visualise and recreate the plots, use visuals.ipynb.

If you use the repository, please cite:

@inproceedings{nasir2023augmentative,
  title={Augmentative topology agents for open-ended learning},
  author={Nasir, Muhammad and Beukman, Michael and James, Steven and Cleghorn, Christopher},
  booktitle={Proceedings of the Companion Conference on Genetic and Evolutionary Computation},
  pages={671--674},
  year={2023}
}