Nello Cristianini visits Elettra Sincrotrone Trieste: understanding AI is the next frontier
             01/13     |
How do we study an artificial intelligence system capable of extraordinary performance when its inner workings remain, at least in part, a mystery? This was one of the key questions behind the visit of Nello Cristianini, one of the world's leading experts in artificial intelligence and Professor of AI at the University of Bath (UK), to Elettra Sincrotrone Trieste. The visit on August 3 2026, provided an opportunity to discuss the increasingly important role of artificial intelligence in scientific research and in the development of large-scale research infrastructures. Accompanied by Roberto Pugliese, Deputy General Coordinator and Coordinator of the IT group of Elettra Sincrotrone Trieste, Cristianini toured the FERMI free-electron laser, the Elettra synchrotron light source, the Booster and the Elettra 2.0 construction site, learning about the Centre's activities in scientific computing, artificial intelligence and quantum computing. The programme reached its peak with the public seminar 'Towards Increasingly Intelligent Machines', attended both in person and online. |
| During his lecture, Cristianini addressed one of the most pressing challenges in artificial intelligence research: understanding what happens inside AI models during training. While these systems continue to achieve remarkable results, the processes that lead them to make specific decisions remain difficult to interpret. Drawing on recent research, Cristianini described how scientists are working to uncover the internal representations developed by neural networks—from conceptual maps to implicit rules—with the goal of turning artificial intelligence into an increasingly observable and understandable object of scientific investigation. The visit highlighted the growing convergence of advanced research infrastructures, high-performance computing, artificial intelligence and quantum technologies, opening new opportunities for multidisciplinary research. Today, the challenge is no longer only to build more powerful AI systems, but also to understand how they work. It is in this context that the dialogue between artificial intelligence and large-scale research infrastructures such as Elettra Sincrotrone Trieste is becoming increasingly strategic. | |
