In Madrid, Spain, the Centro de Investigaciones Energéticas, Medioambientales y Tecnológicas (CIEMAT), IBM, and Aggity have joined forces. This partnership, announced recently, aims to integrate generative artificial intelligence into experiments on the TJ-II device at the National Fusion Laboratory (LNF).
The TJ-II, a stellarator operational since 1998, is crucial for exploring plasma behavior and optimizing nuclear fusion parameters. The collaboration utilizes IBM watsonx and cloud technologies to analyze vast datasets, identify patterns, and develop predictive models. This will enhance experimental operations and contribute to the ITER project.
“Connecting specific databases with generative language models and machine learning will enable us to implement a virtual assistant that recommends effective configurations,” said Augusto Pereira, head of the AI project at LNF-CIEMAT. The project will create a specialized generative AI platform using Large Language Models trained with TJ-II data. This will help formulate scientific hypotheses and identify new correlations.
This initiative is vital for fusion energy, a sustainable alternative for future electricity generation. The LNF is also working on diagnostics for ITER, integrating AI to improve safety and efficiency. This collaboration sets new standards in scientific instrumentation and opens opportunities for the technology industry in nuclear fusion.