ORNL has a rich tradition of artificial intelligence, or AI, research dating back more than four decades and garnering more than ten patents. The laboratory’s AI Initiative is dedicated to ensuring secure, trustworthy, and energy efficient AI in the service of scientific research and national security. Through this internal research investment, subject matter experts at ORNL leverage the laboratory’s computing infrastructure and software capabilities to expedite times to solution and realize the potential of AI in projects of national and international importance. For example, the Initiative has helped multidisciplinary teams demonstrate that machine learning algorithms can be used to extract information from signals with low signal-to-noise ratios; develop algorithms capable of accelerating modeling and simulation with very little training data; a design novel biomimetic neuromorphic devices capable of detecting epileptic seizures. 

Using various tools developed by ORNL teams, researchers at the lab are adapting artificial intelligence techniques to better understand cancer and other diseases. Through a series of ongoing, interconnected research projects, they are informing health policy and improving public health outcomes. These efforts include a pilot project for the CANcer Distributed Learning Environment (CANDLE) initiative, a joint endeavor led by DOE and the National Cancer Institute that focuses on developing language processing techniques to identify connections between sources of medical data ranging from images to test results to doctors’ notes. Combing through these disparate health records simultaneously would not be feasible with conventional analytical methods.

At ORNL’s Manufacturing Demonstration Facility, researchers use additive manufacturing, or 3D printing, to develop solid parts made of metal, plastic, and other materials for a wide range of applications. By incorporating AI into this process, they aim to reliably produce more energy-efficient, durable, and customizable products at a lower cost. These high-quality designs could lead to longer-lasting vehicles, faster personal computers, better insulation in houses, and even more specialized aerospace components. Using artificial intelligence image processing techniques developed at ORNL, researchers can inspect parts in real time throughout the manufacturing process to locate cracks and other quality defects, preventing problems further into the manufacturing process and, by extension, reducing costs and time to market.

AI, in combination with the Frontier supercomputer and other resources at the lab, accelerates research across the broad spectrum of energy generation, distribution, storage, and security. For example, scientists use AI to study clean energy sources and mitigate environmental impacts by running climate simulations, developing new materials, and creating carbon-capture technologies. Additionally, AI helps ORNL staff improve certain maintenance and operation practices, extending the lifetimes and energy production capabilities of nuclear power plants and other facilities. AI techniques can also be used in studies related to the nation’s power grid, ensuring both cyber and physical security during the delivery of electricity to homes, businesses, and public spaces as part of DOE’s Grid Modernization Initiative.

Cybersecurity experts are constantly collecting data, but sifting through information to pinpoint problem areas in a timely manner can be difficult. To address this challenge, alongside the emerging threats and vulnerabilities associated with artificial intelligence-based technologies, ORNL has established the Center for Artificial Intelligence Research, or CAISER, as an expansion of the laboratory’s AI Initiative. With a particular focus on cyber, biometrics, geospatial analysis, and nonproliferation, CAISER will analyze vulnerabilities, threats, and risks related to the security and misuse of AI tools in national security domains. Through partnerships with Chattanooga’s Electric Power Board and other companies, ORNL also helps maintain energy security by deploying AI-enabled security systems capable of monitoring data streams and identifying suspicious activity. 

ORNL’s Frontier is more than the world's first exascale machine—it’s also smart. This HPE Cray EX system opened to users in 2023 and resides at the Oak Ridge Leadership Computing Facility, a DOE Office of Science User Facility located at ORNL. With computing power and architecture components optimized for AI applications, Frontier helps researchers apply machine learning and deep learning techniques to a variety of science problems to quickly obtain accurate results. Frontier's increased memory bandwidth means researchers can tackle data-intensive problems without the need to constantly move data. As a result, unimpeded deep learning algorithms run at faster speeds and reach higher levels of accuracy than would be possible on another system.