Wikidata

📢 The WikiProjects Days 2026 is fast-approaching! WikiProjects are what happens when a group of contributors decide to work together as a team on a specific theme or topic. They enhance coordination, collaborative efforts, decision-making on data-modelling and offer mutual assistance to its members and cross-projects. ⏰ Call for Proposals now ends May 31 - Share your WikiProject, discuss best practices, present tools or discuss cross-wiki work: https://lnkd.in/gFjS7xj2 🔗 Register for updates: https://lnkd.in/dKRiWGV9

🌐 The AI-BRIDGES Symposium | 📅May 28-29, 2026 “How do open knowledge ecosystems remain foundational in the age of LLMs?” Tomorrow at the AI-BRIDGES Symposium, we can discuss this question and others as we learn more about open data, Wikidata & knowledge graphs, digital public infrastructure, AI governance, GLAM & cultural heritage, semantic web, open-source AI, retrieval-augmented generation (RAG), AI grounding & provenance, and community-governed datasets during the upcoming 2-day event! 🔍 Topics include: Wikidata as AI infrastructure · Knowledge graphs for hallucination reduction · Embeddings + Model Context Protocol (MCP) for Wikidata · AI-ready institutional archives · Linked open data for agentic systems · Semantic interoperability + digital commons & data sovereignty · Human-in-the-loop curation · Public-interest AI ecosystems · AI-enhanced metadata pipelines May 28 - Training day: Woburn Suite, G22/26, Senate House – up to 50 onsite · online via Zoom May 29 - Symposium and Discussion day: Beveridge Hall, Senate House – 300 participants · onsite only 🎤 Featuring: Jimmy Wales (Wikipedia Co-Founder), Denny Vrandečić (Wikidata Founder), Renata Avila (CEO Open Knowledge Foundation) + more 📍Senate House, University of London, Malet Street, London WC1E 7HU 🎟️ Free attendance - register here: https://lnkd.in/gqvHtQh2 🎥 Most sessions will be recorded and shared publicly (roundtables excluded) #OpenData #AI #Wikidata #KnowledgeGraphs #SemanticWeb #OpenSourceAI

This Week on Wikidata – 18 May 2026: Catch up on the latest updates and community highlights. 👉 Read the full report here: https://lnkd.in/d_MwKgP9

Embedding Project, Vectors, Vector search… if you've been nodding along quietly while wondering what these terms mean, this post is for you! What is a Vector? - A vector is when a concept (a word, an image, a sentence, a Wikidata item) is represented as a long list of numbers. Think of it like a set of coordinates on a gigantic map of meaning. Concepts that are similar get plotted close together — "cat" and "kitten" will be grouped nearby, while "skyscraper" ends up somewhere very different. What is an Embedding? - The process of turning something into a vector. AI models read the content and produce that list of numbers, capturing its meaning. In this way, different words, terms and concepts can be recognised as semantically similar or related — even if they use different words, or even different languages. What is a Vector Search Database? - If traditional search uses name-matching and keywords to find results, a vector database looks for items with similar vector values — things that mean roughly the same thing, rather than things that say exactly the same words. Still confused? Let's try an example: a search for "renewable energy in small island nations". Each word gets embedded and turned into a vector. The database is then queried for Wikidata items whose vectors are closest in meaning to my query — and those are returned as results. No exact keyword match needed. Read more: https://lnkd.in/gdNHTZBF Project Newsletter: https://zcmp.eu/X5wG Philippe Saadé (AI/Machine Learning Project Manager, WMDE) was kind enough to explain these terms. To hear more about AI in structured open knowledge, Philippe is giving a session on ‘Connecting AI to Wkidata’ at the AI-Bridges Symposium; May 28–29, University of London. Free to register: https://lnkd.in/gqvHtQh2

🤖 AI + Wikidata working together: how does that look in practice? The open knowledge community has good reason to be cautious about AI. Large-language models (LLMs) hallucinate, they scrape data without attribution, and they’re limited and biased by their training data. For projects built on verifiability, that is a big problem. But the Pandora’s box of AI has already been opened, and open-data communities like Wikidata are realising they will need to work cautiously with AI when integrating it into their tools and workflows. Here are two examples of AI being used to serve open knowledge, rather than replace it: PoliLoom is a tool created by EveryPolitician.org that uses LLMs to seek out and extract data on politicians from articles, then matches that information against entities from a dump of Wikidata’s 100-million+ entities. PoliLoom identifies potential gaps and missing information and presents bite-sized chunks to a human for verification. AI can handle the scale while humans provide good judgment. Read more about PoliLoom here: https://lnkd.in/gHT5F8VZ Wikidata Embedding Project: SPARQL queries have a high barrier to entry, which keeps SPARQL’s power out of reach for most users. The Embedding Project provides keyword search with context with the help of a vector search database, combining natural-language search with context-aware, semantic results. With these, you can explore the knowledge graph in the same way you think. Read more: https://lnkd.in/gdNHTZBF AI’s challenges, risks and opportunities are just some of the topics slated for discussion at the AI Bridges Symposium (May 28-29, University of London) among leaders and experts from open knowledge communities, AI researchers, cultural heritage and policymakers. Registration is free. For more details: https://lnkd.in/gqvHtQh2

This Week on Wikidata – 11 May 2026: Catch up on the latest updates and community highlights. 👉 Read the full report here: https://lnkd.in/gwYSm7f8

📢 WikiProject Days 2026 is coming. June 19 to 21 online. We will talk about how WikiProjects can be used as a space to connect with others and inform data modelling and quality decisions. Sessions will include how to make Wikiprojects more visible, create guidelines for good WikiProjects, and bring life back into neglected WikiProjects. 💡 Propose a session by May 26 or register to attend. Everyone is welcome. 🔗 https://lnkd.in/gFjS7xj2

This Week on Wikidata – 4 May 2026: Catch up on the latest updates and community highlights. 👉 Read the full report here: https://lnkd.in/dVQKajAS

What did the Wikidata community get up to this week? Let's find out. https://lnkd.in/dTYw5uSg

Catch up on everything that happened in the Wikidata world this past week. https://lnkd.in/dxuh9BHa