California-based startup Atomic Canyon has announced a new project with the US Department of Energy’s (DOE’s) Oak Ridge National Laboratory (ORNL), to build a safe, efficient open-source AI (artificial intelligence) model for the nuclear sector. Atomic Canyon will use ORNL’s Frontier supercomputer – the world’s fastest supercomputer – to train the company’s AI model to understand complex nuclear terminology. Frontier was the first to achieve “exascale” computing performance (a quintillion calculations per second).
Atomic Canyon will train its AI-powered search tool on Frontier after being awarded time on the computer by DOE. Scientists around the world can apply to use Frontier’s unprecedented processing power for free to solve a range of scientific problems, as long as the results are available to the public.
Using Frontier, Atomic Canyon says it aims to open-source nuclear sector terminologies, helping to develop a nuclear AI infrastructure to build upon and to provide a powerful tool for efficient data navigation, document indexing, performance improvement and cost savings.
The technology, developed by Atomic Canyon was trained on millions of publicly available documents from the US Nuclear Regulatory Commission’s (NRC’s) Agency-wide Documents Access & Management System (ADAMS). It will use sentence-embedding models. These algorithms convert written sentences into numerical representations to allow AI to grasp nuclear terminology and accurately discern information. This deeper understanding makes the AI more effective and helps prevent AI “hallucinations” and biases during search and data analysis. Atomic Canyon says its technology will provide innovators with trusted, accurate assistance in understanding the broad spectrum of nuclear safety and security requirements.
The model will not be trained on sensitive information on the design or operation of nuclear technologies, but will be able to assist with analysing a vast array of public regulatory guides, inspection reports, and other documents, with the goal of increasing safety and public accountability.
“Our leading-edge technology is tailored to the nuclear sector,” said Trey Lauderdale, CEO & founder of Atomic Canyon. “Collaborating with ORNL’s premier supercomputing facilities such as Frontier will help our platform understand complex nuclear concepts, removing a major barrier to AI implementation in the nuclear sector. Moreover, our open-source approach will encourage transparency and collaboration among nuclear stakeholders and innovators focused on safety and security. Our goal is to help the world’s best and brightest have the information and technology they need to help navigate the climate and energy crisis.”
“Open-source AI search, trained using the lab’s Frontier supercomputer, will play an essential role in the research, engineering, and development of nuclear energy and the nuclear sector, now and into the future,” said ORNL’s Thomas M Evans, Group Lead and Distinguished R&D Staff, HPC Methods for Nuclear Applications. “This is crucial to advance nuclear innovation and move toward a future powered by sustainable energy.”
Atomic Canyon was established in late 2023. In March, the company launched Neutron, its first free search tool trained on complex nuclear terms.