This web page was created programmatically, to learn the article in its authentic location you’ll be able to go to the hyperlink bellow:
https://www.oecd-nea.org/jcms/pl_111870/harnessing-ai-first-workshop-of-the-new-nea-joint-project-aixpertise
and if you wish to take away this text from our web site please contact us
The inaugural workshop of the NEA’s AI Platform for Nuclear Research and Education (AIxpertise), convened on-line from 15 to 16 October 2025, contributed to the additional improvement of the brand new joint challenge devoted to leveraging synthetic intelligence (AI) for the good thing about nuclear power professionals.
With AIxpertise, the NEA seeks to reinforce collaboration amongst nuclear sector stakeholders in business, security, analysis and academia to equip consultants with AI data and devices to advance their analysis and prepare the subsequent era of nuclear consultants.
The workshop introduced collectively 45 organisations from 17 nations, together with security organisations, analysis establishments, academia, business and Big Tech firms. The programme featured shows and discussions on the general challenge construction, NEA IT and Data Bank providers, the authorized framework for the AIxpertise challenge, and a draft programme of labor throughout its three focus areas: knowledge, AI algorithms’ benchmarking, and schooling & coaching.
Participants mentioned ongoing initiatives and explored methods to reinforce and align them by way of collaboration throughout the AIxpertise programme.
Data
The Jožef Stefan Institute (IJS), Slovenia, introduced the heart beat and depletion experiments being carried out on the IJS TRIGA Mark II analysis reactor and its plans to develop a brand new versatile European reactor for neutron irradiation and nuclear analysis, VERONICA. IJS has supplied to share TRIGA reactor knowledge with AIxpertise members.
The Institute for Energy Technology (IFE) of Norway introduced a overview of the Halden Reactor Project (HRP) legacy database, a key useful resource encompassing over 60 years of experimental work essential to advancing analysis and security assessments.
The staff of the Nuclear Research Data System (NRDS) at Idaho National Laboratory (INL), United States, demonstrated Retrieval-Augmented Generation (RAG) with the potential to considerably enhance the accessibility and interpretability of the HRP knowledge content material.
Oak Ridge National Laboratory (ORNL), United States, demonstrated optimised machine learning-driven exploitation of current datasets to enhance modelling and simulation (M&S) in addition to to validate M&S in domains of sparse validation knowledge like high-assay low-enriched uranium (HALEU) software circumstances, that are essential candidates for future small modular reactor deployment.
The contributors mentioned the AIxpertise challenge’s scope directed in direction of fostering knowledge that meet the FAIR precept (findability, accessibility, interoperability and reusability) as a basis for AI&ML approaches. They expressed robust help for work on AI-driven instruments fostering accessibility and interpretability of knowledge, together with the HRP knowledge. There was additionally robust curiosity in sharing analysis reactor knowledge and in operators’ options to share nuclear energy plant time collection knowledge with AIxpertise members protected by the AIxpertise authorized framework.
Benchmarking AI algorithms
The benefits of worldwide co-operation in benchmarking actions had been outlined in a presentation on the achievements of the NEA Nuclear Science Committee’s Task Force on Artificial Intelligence and Machine Learning for Scientific Computing in Nuclear Engineering. North Carolina State University (NCSU) expressed help for growing worldwide benchmarks inside AIxpertise.
Presentations by Imperial College, United Kingdom, on foundational AI for fluids, solids, particle and radiation modelling methodology, and by the INL, United States, on machine studying for nuclear fuels and supplies modelling, demonstrated new alternatives for enhancing modelling capabilities and accuracies with AI functions.
Microsoft supplied a glimpse into the way forward for Agentic AI proposing effectivity beneficial properties by totally integrating AI brokers into human-led workflows, with the potential to hurry up nuclear licensing processes.
Participants mentioned the proposed AIxpertise challenge scope, meant to benchmark the output of AI algorithms and to make sure their transparency, reliability and readiness for regulatory overview.
Hands-on coaching and greatest practices
The Institute of Science Tokyo, Japan, mentioned the event of data administration methods based mostly on supportive AI brokers based mostly on RAG expertise and the necessity for complete validation. North Carolina State University, United States, and the Polytechnical University of Milano, Italy, introduced current schooling alternatives and expressed help to develop and supply coaching to AIxpertise challenge contributors. Participants mentioned the virtues of various coaching codecs to foster nuclear workforce improvement and steady studying.
The workshop concluded with a roadmap to launch the AIxpertise challenge at first of 2026. The subsequent steps are to refine the AIxpertise programme of labor based mostly on additional suggestions from potential challenge contributors and a overview of the authorized settlement.
Interested organisations are invited to hitch the AIxpertise improvement by contacting [email protected]. The subsequent AIxpertise workshop is scheduled for 21-22 January 2026.
Learn extra on the Aixpertise web page.
This web page was created programmatically, to learn the article in its authentic location you’ll be able to go to the hyperlink bellow:
https://www.oecd-nea.org/jcms/pl_111870/harnessing-ai-first-workshop-of-the-new-nea-joint-project-aixpertise
and if you wish to take away this text from our web site please contact us
