Defining FAIR for Machine Learning (ML)
Submitted by Daniel S. Katz
Discuss:
- Current projects (both research and infrastructure) in machine learning (ML) that are considering FAIR,
- If there's value in and a need for defining FAIR for ML, and if so,
- How to move forward to do so, ideally under the RDA umbrella based on the current role of RDA in FAIR activities
Collaborative Meeting Notes (main session): https://docs.google.com/document/d/1ZGOeqU3xZB6Gadfga8q59NuGp9nogYGPgy9Qsdm5kVc/edit?usp=sharing
Collaborative Meeting Notes (repeat session): https://docs.google.com/document/d/1qgX7BXgvBvgP-jfN-yLvqGH_uDnnwLjR4lDgthkwjVI/edit?usp=sharing
Introduction(s) (10 min):
- What is FAIR? - Fotis E. Psomopoulos, CERTH
- Quick update on FAIR for Research Software - Fotis E. Psomopoulos, CERTH
- Intro to FAIR 4 ML - Daniel S. Katz, University of Illinois
(slides for the three introduction talks)
Talks (5 + 2 min each):
- DLHub - Aristana Scourtas, University of Chicago (slides, video)
- OpenML - Bernd Bischl and Joaquin Vanschoren
- Pistoia Alliance - Vladimir Makarov, Pistoia Alliance (slides)
- ELIXIR - Jennifer Harrow, ELIXIR-Europe (slides)
- FAIR4HEP - Eliu Huerta, University of Illinois (slides, video)
- CLAIRE - Tom Lenaerts, Université Libre de Bruxelles & Vrije Universiteit Brussel (slides, video)
- kipoi - Julien Gagneur, Technische Universität München (video)
Discussion & next steps (30 min)
- Does continued discussion make sense?
- Should we propose an RDA IG or WG?
- Who else should be involved (projects & people)
- Who wants to co-lead such a proposal?
There is a large amount of FAIR work, both in RDA and elsewhere, initially focused on data and now software and other products but generally not ML models. Some of the speakers in this session are involved in projects where FAIR for ML models is a topic of discussion. Additionally, we presented poster 31b (FAIR principles for ML models - https://doi.org/10.5281/zenodo.4271995) at RDA VP16 to start discussion on this at RDA, and this BoF is the next step.
- 2209 reads