<div dir="ltr"><div><p style="box-sizing:border-box;margin:0px;padding:0px 0px 1em;border:0px;outline:0px;font-size:14px;vertical-align:baseline;background-image:initial;background-position:initial;background-size:initial;background-repeat:initial;background-origin:initial;background-clip:initial;color:rgb(102,102,102);font-family:&quot;Open Sans&quot;,Arial,sans-serif"><img class="gmail-alignright gmail-wp-image-14101" src="https://codata.org/wp-content/uploads/2025/10/Croissant-899x1024.jpg" alt="" width="179" height="205" style="box-sizing: border-box; margin: 0px 0px 0px 15px; padding: 0px; border: 0px; outline: 0px; text-size-adjust: 100%; vertical-align: baseline; background: transparent; float: right; display: inline;">There is currently a lot of interest in how we maximise increase the effectiveness of datasets for AI training and fine-tuning through metadata. This led to the development of the <a href="https://mlcommons.org/working-groups/data/croissant/" target="_blank" rel="noopener" style="box-sizing:border-box;margin:0px;padding:0px;border:0px;outline:0px;vertical-align:baseline;background:transparent;color:rgb(46,163,242);text-decoration-line:none">ML Croissant</a> metadata specification, \u201can open community-built standardized metadata vocabulary for ML datasets, including key attributes and properties of datasets, as well as information required to load these datasets in ML tools. Croissant enables data interoperability between ML frameworks and beyond, which makes ML work easier to reproduce and replicate.\u201d</p><p style="box-sizing:border-box;margin:0px;padding:0px 0px 1em;border:0px;outline:0px;font-size:14px;vertical-align:baseline;background-image:initial;background-position:initial;background-size:initial;background-repeat:initial;background-origin:initial;background-clip:initial;color:rgb(102,102,102);font-family:&quot;Open Sans&quot;,Arial,sans-serif">In turn, there is also great potential for AI tools to enhance metadata and to assist in establishing semantic mappings.  A session at IDW recently explored these issues: <a href="https://scidatacon.org/event/9/contributions/36/" target="_blank" rel="noopener" style="box-sizing:border-box;margin:0px;padding:0px;border:0px;outline:0px;vertical-align:baseline;background:transparent;color:rgb(46,163,242);text-decoration-line:none">\u2018AI for Metadata Enhancement, Metadata for AI Readiness: how do we ensure a virtuous rather than a vicious circle?\u2019</a> In this session, and in a plenary session on <a href="https://scidatacon.org/event/9/contributions/233/" target="_blank" rel="noopener" style="box-sizing:border-box;margin:0px;padding:0px;border:0px;outline:0px;vertical-align:baseline;background:transparent;color:rgb(46,163,242);text-decoration-line:none">AI for Science</a>, Slava Tykhonov, Head of Interoperability and AI at CODATA presented his work on Semantic Croissant, an extension of ML Croissant that is powered by the <a href="https://cdif.codata.org/" target="_blank" rel="noopener" style="box-sizing:border-box;margin:0px;padding:0px;border:0px;outline:0px;vertical-align:baseline;background:transparent;color:rgb(46,163,242);text-decoration-line:none">CODATA Cross-Domain Interoperability Framework (CDIF)</a>. This knowledge graph, maintained at the variable level, is designed to guide and navigate AI through structured expert knowledge. The crucially important steps is to incorporate CDIF\u2019s use of the <a href="https://ddialliance.org/ddi-cdi" target="_blank" rel="noopener" style="box-sizing:border-box;margin:0px;padding:0px;border:0px;outline:0px;vertical-align:baseline;background:transparent;color:rgb(46,163,242);text-decoration-line:none">DDI-CDI</a> variable description providing a rich <img class="gmail-alignright gmail-wp-image-13178" src="https://codata.org/wp-content/uploads/2025/03/cdif640-transparent.png" alt="" width="185" height="92" style="box-sizing: border-box; margin: 0px 0px 0px 15px; padding: 0px; border: 0px; outline: 0px; text-size-adjust: 100%; vertical-align: baseline; background: transparent; float: right; display: inline;">semantic description of the core feature of the dataset: the observable property that was measured or described. This has benefits for interoperability and reuse of data and for its effectiveness in the training of AI models. It also helps situate the variable description as a first-class semantic object, which is one of the key purposes of CDIF.</p><p style="box-sizing:border-box;margin:0px;padding:0px 0px 1em;border:0px;outline:0px;font-size:14px;vertical-align:baseline;background-image:initial;background-position:initial;background-size:initial;background-repeat:initial;background-origin:initial;background-clip:initial;color:rgb(102,102,102);font-family:&quot;Open Sans&quot;,Arial,sans-serif">There is considerable interest in this work. Slava has been invited to give a keynote to the <a href="https://www.nfdi4datascience.de/news/2025/202504_conference2025/" target="_blank" rel="noopener" style="box-sizing:border-box;margin:0px;padding:0px;border:0px;outline:0px;vertical-align:baseline;background:transparent;color:rgb(46,163,242);text-decoration-line:none">NFDI4DataScience Conference</a>, taking place on 25\u201326 November 2025 at <a href="https://www.fokus.fraunhofer.de/en.html" target="_blank" rel="noopener" style="box-sizing:border-box;margin:0px;padding:0px;border:0px;outline:0px;vertical-align:baseline;background:transparent;color:rgb(46,163,242);text-decoration-line:none">Fraunhofer FOKUS</a> in Berlin.  The <a href="https://www.nfdi4datascience.de/" target="_blank" rel="noopener" style="box-sizing:border-box;margin:0px;padding:0px;border:0px;outline:0px;vertical-align:baseline;background:transparent;color:rgb(46,163,242);text-decoration-line:none">NFDI4DS initiative</a> aims to build and sustain a national research data infrastructure for the Data Science and Artificial Intelligence community in Germany \u2013 an exciting step toward more interoperable, transparent, responsible and FAIR AI.</p><p style="box-sizing:border-box;margin:0px;padding:0px;border:0px;outline:0px;font-size:14px;vertical-align:baseline;background-image:initial;background-position:initial;background-size:initial;background-repeat:initial;background-origin:initial;background-clip:initial;color:rgb(102,102,102);font-family:&quot;Open Sans&quot;,Arial,sans-serif"><img class="gmail-alignright gmail-wp-image-14097" src="https://codata.org/wp-content/uploads/2025/10/IDW_2025_15th_Oct_DAY_THREE-72.jpg" alt="" width="349" height="233" style="box-sizing: border-box; margin: 0px 0px 0px 15px; padding: 0px; border: 0px; outline: 0px; text-size-adjust: 100%; vertical-align: baseline; background: transparent; float: right; display: inline;">Slava has also been invited to speak at as CESSDA AI Workshop, as part of the 4-day \u201cCESSDA at 50\u201d conference in Bergen, 15-18 June 2026. This 50th-anniversary event will bring together CESSDA Service Providers, researchers, policy actors, partner organisations, and international networks to share knowledge and address the evolving landscape of research and innovation.</p><p style="box-sizing:border-box;margin:0px;padding:0px;border:0px;outline:0px;font-size:14px;vertical-align:baseline;background-image:initial;background-position:initial;background-size:initial;background-repeat:initial;background-origin:initial;background-clip:initial;color:rgb(102,102,102);font-family:&quot;Open Sans&quot;,Arial,sans-serif"><br></p><p style="box-sizing:border-box;margin:0px;padding:0px;border:0px;outline:0px;font-size:14px;vertical-align:baseline;background-image:initial;background-position:initial;background-size:initial;background-repeat:initial;background-origin:initial;background-clip:initial;color:rgb(102,102,102);font-family:&quot;Open Sans&quot;,Arial,sans-serif">Thanks,</p><p style="box-sizing:border-box;margin:0px;padding:0px;border:0px;outline:0px;font-size:14px;vertical-align:baseline;background-image:initial;background-position:initial;background-size:initial;background-repeat:initial;background-origin:initial;background-clip:initial;color:rgb(102,102,102);font-family:&quot;Open Sans&quot;,Arial,sans-serif">Asha</p><div style="color:rgb(0,0,0)"><div>___________________________</div><div><div><br></div><div><a href="https://codata.org/initiatives/making-data-work/cdif/cdif-at-idw2025/" target="_blank"><b>CDIF and AI Sessions at IDW2025</b></a></div><div><b><br></b></div><div><b>CODATA and the Australian Research Data Commons (ARDC) announce the updated </b><a href="https://codata.org/codata-and-the-australian-research-data-commons-ardc-announce-the-updated-2025-codata-research-data-management-terminology-rdmt/" target="_blank"><b>2025 CODATA Research Data Management Terminology (RDMT)</b></a><b><br><br>From launch to action:</b><a href="https://codata.org/from-launch-to-action-operationalising-unescos-open-science-data-policies-guidance-for-crises/" target="_blank"><b> operationalising UNESCO\u2019s open science data policies guidance for crises</b></a><b><br><br>First Climate-Adapt for EOSC Deliverable D1.1: </b><a href="https://doi.org/10.5281/zenodo.17244500" target="_blank"><b>Requirement Analysis and CLIMATE-ADAPT4EOSC potentialities</b></a></div><div><b></b></div><div><b><br></b></div><div><a href="https://codata.org/read-now-september-2025-publications-in-the-data-science-journal/" target="_blank">September 2025 publications</a> in the <a href="https://datascience.codata.org/" target="_blank">CODATA Data Science Journal</a></div></div></div><div style="color:rgb(0,0,0)"><div><div><br></div><div><b>Stay in touch with CODATA:</b></div><div></div><div><br></div><div>Stay up to date with CODATA activities: <a href="http://lists.codata.org/mailman/listinfo/codata-international_lists.codata.org" target="_blank">join the CODATA International News list</a></div><div><br></div><div>Looking for training and career opportunities in data science and data stewardship?  <a href="http://lists.codata.org/mailman/listinfo/data_science_training_lists.codata.org" target="_blank">Sign up to the CODATA early career community-run data science training and careers list</a></div><div><br></div><div>Follow us on social media! <a href="https://bsky.app/profile/codata-isc.bsky.social" target="_blank">Bluesky</a> - <a href="https://www.linkedin.com/in/simon-hodson-b3711a11/" target="_blank">LinkedIn</a></div></div></div></div><div dir="ltr" class="gmail_signature" data-smartmail="gmail_signature"><div dir="ltr"><div><div dir="ltr"><div dir="ltr"><div dir="ltr"><div dir="ltr"><div dir="ltr"><div dir="ltr"><div dir="ltr"><div dir="ltr" style="font-size:small"><div style="font-size:12.8px"><div style="color:rgb(0,0,0)">___________________________<br></div></div></div><div dir="ltr" style="font-size:small"><br></div><div dir="ltr" style="font-size:small">Asha Law | Program Assistant, CODATA | <a href="http://www.codata.org/" style="color:rgb(17,85,204)" target="_blank">http://www.codata.org</a></div><div dir="ltr" style="font-size:small"><br></div><div dir="ltr" style="font-size:small">E-Mail: <a href="mailto:asha@codata.org" style="color:rgb(17,85,204)" target="_blank">asha@codata.org</a><br></div><div dir="ltr" style="font-size:small">Tel (Office): +33 1 45 25 04 96</div><div dir="ltr" style="font-size:small"><br></div><div dir="ltr" style="font-size:small">CODATA (Committee on Data of the International Science Council), 5 rue Auguste Vacquerie, 75016 Paris, FRANCE</div></div></div></div></div></div></div></div></div></div></div></div>