FAIR for Machine Learning (FAIR4ML) IG

The idea of FAIR (findable, accessible, interoperable, and reusable) in the context of scientific data management and stewardship was developed in 2014 and turned into specific principles in 2016[1]. Along the way, the idea was generalized in concept to apply to both data and other digital scholarly objects, but it has become clear in practice that what works for data may not directly work for all other digital objects. For example, both previous and ongoing work show that many of the guiding FAIR principles need to either be re-written or reinterpreted for software, resulting in the FAIR principles for Research Software[2], already with an adoption commitment from different communities and institutes[3]. The FAIR principles also can apply to machine learning tools and models, though a direct application is not always possible as machine learning combines aspects of data, software and computational workflows.

This Interest Group will enable community members to discuss the various aspects of FAIR as applied to Machine Learning, looking both at domain specific and domain-agnostic use cases, and creating task forces and working groups as needed for specific guidance documents, recommendations, definitions and technical specification to that effect. The overall aim is to foster collaborations among researchers and developers who are interested in making machine learning (data, models, workflows, etc.) FAIR, along with those who contribute to the infrastructure and policies that support this. It will work closely with other FAIR RDA Groups (such as the FAIR for Research Software Working Group), as machine learning combines aspects of data and software, but is distinctly different from both.

Specifically, objectives of this IG are to:

1. Discuss where FAIR should apply to ML, considering the work in other working groups and focusing on gaps
2. Define and prioritize cases for new Task Forces and Working Groups
3. Ultimately, build a community of practice for information sharing about ML and FAIR pertaining to ML

In order to ensure that these scenarios are valid across domains (e.g., health, earth science, physics, agriculture, materials science, energy, biology), individual Task Forces (TFs) may be initiated from within the IG that may be focused on particular domains could be initiated, each working in parallel on distinct topics.

References
[1] Wilkinson, M., Dumontier, M., Aalbersberg, I. et al. The FAIR Guiding Principles for scientific data management and stewardship. Sci Data 3, 160018 (2016). https://doi.org/10.1038/sdata.2016.18

[2] Chue Hong, N. P., Katz, D. S., Barker, M., Lamprecht, A.-L., Martinez, C., Psomopoulos, F. E., Harrow, J., Castro, L. J., Gruenpeter, M., Martinez, P. A., Honeyman, T., et al. (2021). FAIR Principles for Research Software (FAIR4RS Principles). Research Data Alliance. https://doi.org/10.15497/RDA00065

[3] Martinez-Ortiz, Carlos, Katz, Daniel S., Lamprecht, Anna-Lena, Barker, Michelle, Loewe, Axel, Fouilloux, Anne, Wyngaard, Jane, Garijo, Daniel, Moldon, Javier, Castro, Leyla Jael, Wheeler, Daniel, Albers, Joost Rutger Demian, & Lee, Allen. (2022). FAIR4RS: Adoption support. Zenodo. https://doi.org/10.5281/zenodo.6258366