I'm thrilled and deeply honored to share that I was officially nominated as a Harvard Dataverse Ambassador by @dataverseorg @IQSS ! This recognition was announced during the annual Dataverse Community Meeting, and I couldn't be more proud to support and represent a platform that plays such important role in advancing open data, reproducibility, and collaboration across the research world. Back in the pandemic, I had the opportunity to reflect on five years of Dataverse growth in Europe, starting with DataverseNL. That journey showed what’s possible when we collaborate across institutions and countries, even continents, building open and sustainable research infrastructure. Now we’re entering a new exciting era of shaping data with AI. We need to unite across borders, not split and compete, and technology should help to support us, not break us! That's something I'm going to work on. Watch the video from back about how it all started with DataverseNL in Europe. youtube.com/watch?v=bcQk6ykx… #Dataverse2025 #Dataverse2026

And Model Context Protocol (MCP) for @dataverseorg #Dataverse is online and operational! mcp.dataverse.org #Dataverse2025

In our latest @MuseIT_EU paper titled 'A Knowledge Base for Arts and Inclusion – The Dataverse @dataverseorg Data Archival Platform as a Knowledge Base Management System Enabling Multimodal Accessibility', we present a new approach to archiving multimodal materials. Why is this important? Because it’s entirely AI-driven — from organizing and enriching content to enabling inclusive access across text, audio, video, and even haptics, in the FAIR and sustainable way. arxiv.org/pdf/2504.05976

Honored to join as a judge and contribute with @dataverseorg insights.

Our multimodalities integration for audio, video, text and haptics in @MuseIT_EU @dataverseorg repositories, with AI-powered chatbot answering any questions on its content, both on data and metadata zenodo.org/records/14919736 Voice interface is coming soon...

If you missed my presentation for @dataverseorg community on Ghostwriter, #LLM integration framework with #knowledge graphs, there are slides, recording and more materials to read zenodo.org/records/14823315

@dataverseorg team just published on Dataverse TV my recent presentation for the Dataverse community - this time it's on RAG, explaining the integration of #LLMs with knowledge graphs. harvard.zoom.us/rec/play/tqD… I've tried to make it as "light" as possible, almost without technical details, but you'll learn how Ghostwriter works: ask.dataverse.org

If you somehow missed our Croissant standard for Machine Learning #ML, the @NeurIPSConf paper has been recently published by @Meta AI. Here, you'll gain a clear understanding of how the Croissant knowledge graph will help leverage #AI and #LLMs of the future. I contributed on behalf of @DANS_knaw_nwo and @dataverseorg. ai.meta.com/research/publica…

Exciting update from @DANS_knaw_nwo during @MuseIT_EU project meeting in London: soon, no human intervention will be required to look up, select, and link terms from external vocabularies to metadata records in @kaggle, @huggingface , @dataverseorg , and other @MLCommons Croissant ML-compliant platforms. Check slides here: lnkd.in/e8wKsEq6 #AI can perform these tasks with a high level of confidence, linking all relevant terms in the knowledge graph. Additionally, the system can automatically rebuild relationships when new data arrives, ensuring the knowledge graph remains up-to-date. By leveraging concept representation in semantic space, we successfully developed a service that “predicts” the correct concept URIs based on a selected term, its surrounding words, and context. Essentially, it functions as a resolver that operates without human input, analyzing fields like titles and descriptions to select or correct terms from @wikidata , Skosmos, @GettyMuseum, @ontoportal, and other controlled vocabularies and ontologies.

Slides of our @DANS_knaw_nwo presentation for the @ODISSEI_nl conference on the next #AI generation applications of #data, including @dataverseorg ask.dataverse.org and Ghostwriters to chat with the dead people! zenodo.org/records/14364215

Original dataset includes 50 million triples and contains @dataverseorg, metadata records exported in Croissant ML format dataverse.harvard.edu/datase…