FAIR for Machine Learning (FAIR4ML) IG Charter TAB Review

Interest Group Title:  FAIR for Machine Learning (FAIR4ML) IG 

Charter: https://rd-alliance.org/group/fair-machine-learning-fair4ml-ig/case-stat...

Proposers: Fotis Psomopoulos, Daniel Katz

Date Received by TAB: 17th August 2022

Date Review Completed:  22 September 2022


Focus and Fit:  

(Are the Interest Group objectives aligned with the RDA mission ?  Is the scope too large for effective progress, too small for an RDA effort, or not appropriate for the RDA?  Overall, is this a worthwhile effort for the RDA to take on?  Is this an effort that adds value over and above what is currently being done within the community?)

The proposed interest group seeks to address the adoption/implementation of FAIR for machine learning.

The topic is relevant, timely and important in the context of open and accessible research assets to support reproducibility. The proposed outputs of the FAIR4ML IG are comprehensive and fulfil the stated aims of the proposed Interest Group.

The objectives of the group are clear and concise, and focused on extending existing efforts addressing adoption of FAIR for data and software to include aspects of machine learning. The proposed use cases, which includes both domain specific and domain agnostic applications, will ensure the relevance of this activity for the ML community is as wide as possible. In addition, the proposed use cases recognise the importance of augmenting existing RDA use cases.


(Does the initial membership list include sufficient expertise, and disciplinary and international representation?  Are the people involved in the Interest Group sufficient to make tangible progress?  What individuals or organizations are missing?) 

The co-chairs have clearly invested significant time and effort to establish there is a genuine interest in this topic, which is confirmed by the extensive list of potential members.  One point of concern is the geographic diversity of the co-chairs, which are primarily from the Global North. It would be desirable for the co-chairs to identify additional co-chairs form other geographic areas.

Impact and Engagement:

(Is it likely that the Interest Group will engage the intended community?  Is there evidence that the research community wants this?  Will the outcome(s) of the Interest Group foster data sharing and/or exchange?)

The topic of FAIR for machine learning is a topic being widely and actively discussed across a number of different organisations, initiatives and projects. An IG within RDA would provide the community coordination necessary to achieve a consistent approach to  implementation of FAIR for machine learning. The case statement clearly shows an awareness of ongoing discussions around the topic of FAIR both within other RDA Interest and Working Groups and beyond, especially with respect to its application to machine learning. In addition, the proposed consultation with other relevant RDA Interest and Working Groups will facilitate exchange of expertise and minimise any duplication of effort. The case statement demonstrates that the co-chairs recognise where the group can leverage other relevant RDA IG/WG activities and outputs as well as those of external relevant initiatives. The proposed designated liaisons to support the interactions with selected RDA IG/WGs will ensure a consistent flow of knowledge and expertise.


Accepted _X__      Rejected  ___  Needs Review __

Overall, the case statement is well developed and very comprehensive. It demonstrates the relevance of the topic within RDA and highlights the potential opportunities for extending the body of knowledge related to adoption and implementation of the FAIR principles to include applications for machine learning.

Additional Comments:

The authors refer to "Task forces" in their case statement. RDA terminology makes use of the term "working group". We recommend the authors solely adopt the RDA terminology to ensure consistency and clarity across RDA groups.