🤖Et si l’#IA pouvait rimer avec sobriété et qualité ?
Une étude menée par les chercheurs d’@Orange montre que des modèles plus frugaux peuvent rivaliser avec, voire dépasser, certains LLMs sur des tâches ciblées, tout en consommant jusqu’à 1500 fois moins d’énergie.
Une piste concrète pour une IA plus responsable.
👇https://t.co/aK7sCTy9Hx
The Filters feature is now LIVE!
https://t.co/u0oA7Xtonb
Navigate 700+ Star Wars planets through films, series, games, size...
Hope this makes navigating the galaxy easier🫶
Next up: Trade routes! Stay tuned as I work on data/bug fixes
#Wikidata#Wookieepedia#StarWars#Atlas
🤔 Open source, open weight or closed #AI models: what’s the difference?
Each type addresses different needs: control, transparency, IP. There’s no universal model, the right choice depends on use.
👇https://t.co/eglcrbYHxz
@stevejarrett@baldursgate3 Couldn’t agree more! @baldursgate3 set a new bar for player freedom and replayability. Rarely have I felt so immersed in a game world. The only thing keeping me from another playthrough is my hobby projects and my ever-growing Coursera course list/Arxiv reading pile 😅
A useful perspective on adaptation in response to monitoring as the key element that distinguishes high quality agents from ‘automation with better marketing’
3 years working on data
700+ planets
Countless details
Our Galactic Map is finally in BETA!✨
Explore planets through media appearances, regions, size...
🔗Try it
https://t.co/XjTx5aiFAz
Feedback welcome! More features (filters, trade routes) coming soon #StarWars#Wikidata
A great evolution of AI architectures in the tabular data domain. Exciting to see so much innovation outside of the auto-regressive Transformer LLM world.
Ping @yoan_chabot
🤖What is the future of #AI agents?
With @stevejarrett, Chief Data & AI Officer at Orange, sharing his vision of smarter, more autonomous and more useful agents than ever.
A fascinating conversation at the heart of @Orange#OpenTech
This week, the #KnowledgeGraph community is buzzing with #WikidataCon2025 and #ISWC2025! The perfect time to discover the Holonet Encyclopedia🌌
@Wikidata meets #StarWars lore, connecting planets and characters in an interactive atlas. Check it out: https://t.co/gKelj12DEI
New Update Alert! 🌌
The #Holonet graph viewer just got a major upgrade! Explore Star Wars characters, movies, games, planets, and more in a stunning interactive network.
🔗 https://t.co/iSPaAu3p5N
#StarWars#DataVisualization#InteractiveMap
Thrilled to announce that our team has had 3 papers accepted at #KCAP2025 (https://t.co/fdTtH8KIfo) incl. an approach for knowledge extraction from text + test bed for uncertain knowledge reconciliation
@lucas_jarnac@vp_huynh @Univ_Lorraine #OrangeResearch@Orange_Future
Transactions on Graph Data and Knowledge, Volume 3, Issue 1 is now published! This issue includes two research articles, one survey article and one resource article, all available under Open Access for anyone to read at the following link: https://t.co/5G2xgnY0ka
New paper 'Uncertainty Management in the Construction of Knowledge Graphs: A Survey' now available in TGDK!
https://t.co/OCtoC4IdNm
This paper explores how to handle uncertainty/knowledge deltas in KG construction.
#OrangeResearch
Word done w/ @yoan_chabot and Miguel Couceiro
Our latest research paper, "Survey on Question Answering over Visually Rich Documents: Methods, Challenges, and Trends" is now available on Arxiv
https://t.co/lkkY4IN3Pp
Collaboration between @orange and @mlia_isir
Built or know a Wikidata-powered tool, bot, or vizualisation? Share it in #WikidataStories!
Like https://t.co/iCDSzNwJhs for book lovers, your submission could spark new ideas!
Takes minutes.
https://t.co/pKYFACEknz
@Wikimedia