Description:
Machine Learning techniques are applied to datasets of polygons and polyhedra, examining how properties such as volume, dual volume, reflexivity, Gorenstein index, codimension can be learnt from Plucker coordinate representation using neural networks.
How to run:
~ Polygon data & investigations are available in the folder Polygon, all investigations are run from one script with an editable ML function, details on how to run are given within the .py file.
~ Polyhedra data is available at the GRDB from the link: http://www.grdb.co.uk/forms/toricf3c, scripts for the investigations can be found in the folder Polyhedra, with details of how to run given within the respective .py files.
@article{Bao:2021ofk,
author = "Bao, Jiakang and He, Yang-Hui and Hirst, Edward and Hofscheier, Johannes and Kasprzyk, Alexander and Majumder, Suvajit",
title = "{Polytopes and Machine Learning}",
eprint = "2109.09602",
archivePrefix = "arXiv",
primaryClass = "math.CO",
reportNumber = "LIMS-2021-011",
doi = "10.1142/S281093922350003X",
journal = "Math. Sci.",
volume = "01",
pages = "181--211",
year = "2023"
}