ePrints

Long-term Reproducibility for Neural Architecture Search

Lookup NU author(s): David Towers, Dr Matthew ForshawORCiD, Dr Stephen McGough, Dr Amir Atapour AbarghoueiORCiD

Downloads

• Published version [.pdf]

Licence

This work is licensed under a Creative Commons Attribution 4.0 International License (CC BY 4.0).

Abstract

It is a sad reflection of modern academia that code is often ignored after publication -- there is no academic 'kudos' for bug fixes / maintenance. Code is often unavailable or, if available, contains bugs, is incomplete, or relies on out-of-date / unavailable libraries. This has a significant impact on reproducibility and general scientific progress. Neural Architecture Search (NAS) is no exception to this, with some prior work in reproducibility. However, we argue that these do not consider long-term reproducibility issues. We therefore propose a checklist for long-term NAS reproducibility. We evaluate our checklist against common NAS approaches along with proposing how we can retrospectively make these approaches more long-term reproducible.

Publication metadata

Author(s): Towers D, Forshaw M, McGough AS, Atapour-Abarghouei A

Publication type: Article

Publication status: Published

Journal: arXiv

Year: 2022

Pages: 4

Online publication date: 18/07/2022

Acceptance date: 11/07/2022

Date deposited: 05/08/2026

Publisher: arXiv

URL: https://doi.org/10.48550/arXiv.2207.04821

DOI: 10.48550/arXiv.2207.04821

Altmetrics

Share