Long-term Reproducibility for Neural Architecture Search
This work addresses reproducibility challenges for researchers in NAS, but it is incremental as it builds on prior reproducibility efforts.
The paper tackles the problem of code decay and lack of long-term reproducibility in Neural Architecture Search (NAS) by proposing a checklist to address these issues, and it evaluates this checklist against common NAS approaches while suggesting retrospective improvements.
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 Code is often unavailable or, if available, contains bugs, is incomplete, or relies on out-of-date / unavailable libraries.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.