Babelscape

✨ Meet the AtLaS Consortium! We’re excited to kick off our partner‑spotlight series, highlighting the organisations driving innovation across the AtLaS project 🤝 🔍 Today we present: Babelscape 🌍 Country: Italy 🏢 Profile: Babelscape is a research-driven Sapienza University spin-off that translates leading academic research into innovative multilingual NLP and knowledge graph solutions. 🎯 Key contribution to AtLaS: Innovative multilingual NLP and knowledge‑graph technologies. 📘 About Babelscape Founded in 2016 through tech transfer from Sapienza University of Rome and its NLP research group, Babelscape bridges the gap between Large Language Models and explicit semantic knowledge. Leveraging its deep expertise, exemplified by its role in developing Minerva, the first family of LLMs built from scratch for Italian and English, the company grounds generative AI in robust knowledge graphs. This neuro-symbolic approach empowers systems to process, distill, and aggregate complex text with high accuracy across multiple languages, ensuring reliable, high-level linguistic intelligence for mission-critical applications. 🔗 Learn more: https://babelscape.com/ #AtLaS #EUDefence #DefenceInnovation #NationalSecurity #EDF

📚 𝗥𝗲𝗮𝗱𝗶𝗻𝗴 𝘀𝘂𝗴𝗴𝗲𝘀𝘁𝗶𝗼𝗻𝘀 | 𝗔𝗜 & 𝗛𝘂𝗺𝗮𝗻𝗶𝘁𝗶𝗲𝘀 What happens when literature meets Artificial Intelligence? We’re happy to highlight “𝗟𝗲𝘁𝘁𝗲𝗿𝗮𝘁𝘂𝗿𝗮 𝗲 𝗶𝗻𝘁𝗲𝗹𝗹𝗶𝗴𝗲𝗻𝘇𝗮 𝗮𝗿𝘁𝗶𝗳𝗶𝗰𝗶𝗮𝗹𝗲. 𝗨𝗻 𝗱𝗶𝗮𝗹𝗼𝗴𝗼 𝗶𝗻𝘁𝗲𝗿𝗱𝗶𝘀𝗰𝗶𝗽𝗹𝗶𝗻𝗮𝗿𝗲”, an edited volume by Tiziana Catarci, 𝗔𝗴𝗻𝗲𝘀𝗲 𝗠𝗮𝗰𝗼𝗿𝗶, and 𝗗𝗮𝗻𝗶𝗲𝗹 𝗥𝗮𝗳𝗳𝗶𝗻𝗶 (𝗟𝗶𝘁𝗵𝗼𝘀, 2026), exploring the intersection between computational methods and literary studies, from analysis to interpretation and beyond. Among its contributions, a chapter by Roberto Navigli reflects on the evolving role of language technologies:  “𝘛𝘩𝘦𝘳𝘦 𝘪𝘴 𝘢 𝘨𝘳𝘰𝘸𝘪𝘯𝘨 𝘯𝘦𝘦𝘥 𝘵𝘰 𝘮𝘰𝘷𝘦 𝘣𝘦𝘺𝘰𝘯𝘥 𝘵𝘩𝘦 𝘵𝘳𝘢𝘥𝘪𝘵𝘪𝘰𝘯𝘢𝘭 𝘱𝘢𝘳𝘢𝘥𝘪𝘨𝘮 𝘰𝘧 𝘭𝘢𝘯𝘨𝘶𝘢𝘨𝘦 𝘮𝘰𝘥𝘦𝘭𝘴 𝘴𝘦𝘦𝘯 𝘢𝘴 𝘮𝘦𝘳𝘦 𝘵𝘦𝘹𝘵 𝘨𝘦𝘯𝘦𝘳𝘢𝘵𝘰𝘳𝘴. 𝘛𝘩𝘦 𝘨𝘰𝘢𝘭 𝘪𝘴 𝘵𝘰 𝘥𝘦𝘷𝘦𝘭𝘰𝘱 𝘴𝘺𝘴𝘵𝘦𝘮𝘴 𝘸𝘪𝘵𝘩 𝘨𝘳𝘦𝘢𝘵𝘦𝘳 𝘪𝘯𝘵𝘦𝘳𝘱𝘳𝘦𝘵𝘢𝘣𝘪𝘭𝘪𝘵𝘺, 𝘵𝘳𝘢𝘯𝘴𝘱𝘢𝘳𝘦𝘯𝘤𝘺, 𝘤𝘰𝘯𝘵𝘳𝘰𝘭𝘭𝘢𝘣𝘪𝘭𝘪𝘵𝘺, 𝘢𝘯𝘥 𝘵𝘩𝘦 𝘤𝘢𝘱𝘢𝘤𝘪𝘵𝘺 𝘧𝘰𝘳 𝘢𝘣𝘴𝘵𝘳𝘢𝘤𝘵𝘪𝘰𝘯 𝘢𝘯𝘥 𝘳𝘦𝘢𝘴𝘰𝘯𝘪𝘯𝘨. 𝘛𝘩𝘪𝘴 𝘪𝘴, 𝘪𝘯 𝘧𝘢𝘤𝘵, 𝘵𝘩𝘦 𝘥𝘪𝘳𝘦𝘤𝘵𝘪𝘰𝘯 𝘪𝘯 𝘸𝘩𝘪𝘤𝘩 𝘤𝘶𝘳𝘳𝘦𝘯𝘵 𝘳𝘦𝘴𝘦𝘢𝘳𝘤𝘩 𝘪𝘯 𝘢𝘳𝘵𝘪𝘧𝘪𝘤𝘪𝘢𝘭 𝘪𝘯𝘵𝘦𝘭𝘭𝘪𝘨𝘦𝘯𝘤𝘦 𝘢𝘱𝘱𝘭𝘪𝘦𝘥 𝘵𝘰 𝘭𝘢𝘯𝘨𝘶𝘢𝘨𝘦 𝘪𝘴 𝘮𝘰𝘷𝘪𝘯𝘨.” This perspective resonates strongly with ongoing work in AI, where the focus is shifting from surface-level generation to deeper understanding and reasoning. 🔎 A recommended read for anyone interested in the dialogue between AI and the humanities.

This week we would like to highlight an interesting event on journalism and artificial intelligence, organised by Associazione Iscom: “𝙅𝙤𝙪𝙧𝙣𝙖𝙡𝙞𝙨𝙢 𝙖𝙣𝙙 𝘼𝙄: 𝙥𝙧𝙤𝙛𝙚𝙨𝙨𝙞𝙤𝙣𝙖𝙡 𝙩𝙧𝙖𝙣𝙨𝙛𝙤𝙧𝙢𝙖𝙩𝙞𝙤𝙣𝙨 𝙖𝙣𝙙 𝙚𝙩𝙝𝙞𝙘𝙖𝙡 𝙥𝙧𝙞𝙣𝙘𝙞𝙥𝙡𝙚𝙨.” The discussion will focus on how AI is reshaping professional practices and the ethical framework of journalism, in a context where questions around responsibility, transparency, and human oversight are becoming increasingly central. Among the speakers, Prof. Roberto Navigli will contribute to a session dedicated to: “𝙉𝙚𝙬 𝙛𝙧𝙤𝙣𝙩𝙞𝙚𝙧𝙨 𝙞𝙣 𝙞𝙣𝙛𝙤𝙧𝙢𝙖𝙩𝙞𝙤𝙣 𝙚𝙩𝙝𝙞𝙘𝙨: 𝙛𝙧𝙤𝙢 𝙛𝙖𝙘𝙩-𝙘𝙝𝙚𝙘𝙠𝙞𝙣𝙜 𝙩𝙤 𝙘𝙤𝙥𝙮𝙧𝙞𝙜𝙝𝙩 𝙖𝙣𝙙 𝙞𝙣𝙩𝙚𝙡𝙡𝙚𝙘𝙩𝙪𝙖𝙡 𝙥𝙧𝙤𝙥𝙚𝙧𝙩𝙮.” 📅 April 15, 2026 📍 Pontifical University of the Holy Cross, Rome • Álvaro del Portillo Hall 🔗 More details here: https://lnkd.in/dvjJXseE

🚀𝗪𝗲 𝗮𝗿𝗲 𝗻𝗼𝘄 𝘁𝗲𝗻! We have spent the past decade working at the intersection of scientific excellence and scalable innovation, turning advances in 𝗡𝗟𝗣 into concrete technologies, from 𝗦𝗲𝗺𝗮𝗻𝘁𝗶𝗰 𝗔𝗜 to 𝗻𝗲𝘅𝘁-𝗴𝗲𝗻 𝗟𝗟𝗠𝘀 and 𝗞𝗻𝗼𝘄𝗹𝗲𝗱𝗴𝗲 𝗚𝗿𝗮𝗽𝗵𝘀  🧠 In this newsletter, we reflect on the journey that brought us here. We are proud of what we have achieved, and even more excited about what is ahead ✨ Enjoy the reading!

🎉 𝗧𝗲𝗻 𝘆𝗲𝗮𝗿𝘀 𝗼𝗳 𝗿𝗲𝘀𝗲𝗮𝗿𝗰𝗵 𝗮𝗻𝗱 𝗶𝗻𝗻𝗼𝘃𝗮𝘁𝗶𝗼𝗻 𝗶𝗻 𝗡𝗟𝗣 A decade of redefining the state of the art in 𝗡𝗮𝘁𝘂𝗿𝗮𝗹 𝗟𝗮𝗻𝗴𝘂𝗮𝗴𝗲 𝗣𝗿𝗼𝗰𝗲𝘀𝘀𝗶𝗻𝗴 through innovative research and real-world impact. A Sapienza Università di Roma spin-off that is now a pillar of the Rome Technopole ecosystem. Our track record speaks for itself: • ~𝟱𝟬 𝗛𝗶𝗴𝗵-𝗶𝗺𝗽𝗮𝗰𝘁 𝗿𝗲𝘀𝗲𝗮𝗿𝗰𝗵 𝗽𝗮𝗽𝗲𝗿𝘀 • 𝟲 𝗘𝗨 𝗿𝗲𝘀𝗲𝗮𝗿𝗰𝗵 𝗮𝗻𝗱 𝗶𝗻𝗻𝗼𝘃𝗮𝘁𝗶𝗼𝗻 𝗽𝗿𝗼𝗷𝗲𝗰𝘁𝘀 (3 ongoing, 2 about to start!) • Our upcoming 2026 release of 𝗠𝗶𝗻𝗲𝗿𝘃𝗮 𝗟𝗟𝗠 joint with Sapienza NLP We do not just use AI: we write its future! 🥂 #Babelscape #NLP #LLM #AI #Innovation #StateOfTheArt #MinervaLLM #10Years #RomeTechnopole

𝗕𝗮𝗯𝗲𝗹𝘀𝗰𝗮𝗽𝗲 𝘁𝘂𝗿𝗻𝘀 𝟭𝟬 🎉 Born as a spin-off of Sapienza Università di Roma and now part of the Rome Technopole ecosystem, Babelscape has grown thanks to the people who made it possible: 𝗼𝘂𝗿 𝗮𝗺𝗮𝘇𝗶𝗻𝗴 𝘁𝗲𝗮𝗺, whose work and curiosity drive the research and technology behind everything we build, and 𝘁𝗵𝗲 𝗽𝗮𝗿𝘁𝗻𝗲𝗿𝘀 who have trusted us and chosen to build with us along the way 🤝 𝗧𝗵𝗮𝗻𝗸 𝘆𝗼𝘂 𝗳𝗼𝗿 𝗯𝗲𝗶𝗻𝗴 𝗽𝗮𝗿𝘁 𝗼𝗳 𝘁𝗵𝗲𝘀𝗲 𝗳𝗶𝗿𝘀𝘁 𝘁𝗲𝗻 𝘆𝗲𝗮𝗿𝘀! In the next few days we’ll share a short newsletter reflecting on the journey so far. If you’d like to read it: 📩 Subscribe on our website, https://lnkd.in/dSizx2fw 🔔 Or follow the newsletter on LinkedIn, https://lnkd.in/de8SQYyD

A new episode of Grande Giove, the podcast by Wired Italia, is now online, exploring a timely question: which language is best for training Artificial Intelligence? The discussion looks at how AI translates between languages, how training data shapes model performance, and how academic research and industrial application increasingly inform one another. Featuring our Managing Director Enrico Giannotti, and Roberto Navigli, Professor of Natural Language Processing at Sapienza Università di Roma University of Rome and Scientific Director of Babelscape, the episode offers a thoughtful exchange on the evolving relationship between linguistic research, innovation, and real-world AI solutions. From Enrico’s perspective, this is especially relevant for solutions such as CABOLO, where the quality, reliability, and contextual understanding of language technologies have a direct impact on how spoken content is captured, processed, and transformed into usable information. Watch the full interview here: https://lnkd.in/d6UyQq_b #offlineAI #WiredItalia #CABOLO #podcast

🎥 Roberto Navigli | 𝘿𝙤 𝙈𝙖𝙘𝙝𝙞𝙣𝙚𝙨 “𝙐𝙣𝙙𝙚𝙧𝙨𝙩𝙖𝙣𝙙” 𝙇𝙖𝙣𝙜𝙪𝙖𝙜𝙚? In this lecture at the 𝗔𝗰𝗰𝗮𝗱𝗲𝗺𝗶𝗮 𝗡𝗮𝘇𝗶𝗼𝗻𝗮𝗹𝗲 𝗱𝗲𝗶 𝗟𝗶𝗻𝗰𝗲𝗶, Prof. 𝗥𝗼𝗯𝗲𝗿𝘁𝗼 𝗡𝗮𝘃𝗶𝗴𝗹𝗶 (Sapienza Università di Roma) takes us on a fascinating journey through the evolution of 𝗔𝗿𝘁𝗶𝗳𝗶𝗰𝗶𝗮𝗹 𝗜𝗻𝘁𝗲𝗹𝗹𝗶𝗴𝗲𝗻𝗰𝗲 applied to language. From the pioneering insights of 1949 to the rise of 𝗟𝗮𝗿𝗴𝗲 𝗟𝗮𝗻𝗴𝘂𝗮𝗴𝗲 𝗠𝗼𝗱𝗲𝗹𝘀 (LLMs) such as ChatGPT and the Italian project 𝗠𝗶𝗻𝗲𝗿𝘃𝗮, the talk explores how machines process texts and addresses the crucial question: do machines truly “understand” what they write? The lecture examines technological advances, the limits of automatic understanding, and the ethical and cultural challenges associated with language models. For those who missed it, it is available on the Accademia dei Lincei’s YouTube channel: https://lnkd.in/dEU7iA6N

📢 IntelligentIA | A new intelligence for a new era • Rome, March 6-7, 2026 On March 6 at 3:30 PM, Prof. Roberto Navigli will be speaking at IntelligentIA, the artificial intelligence event organized by SCAI Comunicazione, taking place at the Hotel Cristoforo Colombo in Rome. In his talk, "𝙃𝙤𝙬 𝙙𝙤 𝙮𝙤𝙪 𝙗𝙪𝙞𝙡𝙙 𝙖𝙣 𝙇𝙇𝙈?", our Scientific Director will discuss how 𝗹𝗮𝗿𝗴𝗲 𝗹𝗮𝗻𝗴𝘂𝗮𝗴𝗲 𝗺𝗼𝗱𝗲𝗹𝘀 (LLMs) are built, starting from the experience of 𝗠𝗶𝗻𝗲𝗿𝘃𝗮, the first LLM trained from scratch in Italian, and will explain how these models can be applied in practice.
 📅 Full program: https://lnkd.in/dYQNQ2eZ 🎟️ Tickets: https://lnkd.in/dFydNShG

On February 27, Quante storie (Rai - Radiotelevisione Italiana) aired an episode on the transformation of the #Italian language in the digital age. With linguist Giuseppe Antonelli, hosted by giorgio zanchini, the discussion explored how Italian is evolving, from informal emails and chats to AI. Professor Roberto Navigli and his R&D team also joined the program to present #Minerva, the first #LLM trained entirely in Italian, developed by Sapienza NLP and Babelscape with support from CINECA and the PNRR Future AI Research project, highlighting the strong link between academic research and industry. 🎬 Here’s an excerpt and the link to the full episode: https://lnkd.in/d_gXBHEn