PNNL

RICHLAND, Wash.—The Department of Energy’s Pacific Northwest National Laboratory has created the Center for AI @PNNL to coordinate the pioneering research of hundreds of scientists working on a range of projects focused on science, security and energy resilience.

Researchers at PNNL were among the first to dive into artificial intelligence decades ago. But AI has surged in the past year with the ready availability of generative AI, which allows almost anyone to produce sophisticated—though sometimes errant—text and images with just a small amount of data. At the same time, AI is a vital tool for serious researchers as well as a subject all its own for scientists to create, explore and validate new ideas. AI also presents an exciting opportunity for PNNL scientists to advance a critical area of science and chart the path forward.

“PNNL has deep expertise and decades of experience in computing and artificial intelligence that we apply to advance scientific discovery, strengthen energy resiliency and enhance national security,” said Steven Ashby, Laboratory director. “The creation of the Center for AI @PNNL will leverage and amplify these capabilities for even greater impact in service of our nation.”

The announcement from PNNL comes the same week as AI experts from around the world are gathering at the annual Conference on Neural Information Processing Systems, known as NeurIPS, Dec. 10-16 in New Orleans. PNNL researchers are making more than a dozen presentations at the highly selective meeting, a demonstration of PNNL’s expertise in AI.

“The time is right for PNNL to focus its AI-related efforts,” said Court Corley, the Laboratory’s chief scientist for AI and director of the new center. “The field is moving at light speed, and we need to move quickly to keep PNNL at the frontier.”

A priority of the Center for AI @PNNL is developing ways to keep AI secure and trustworthy, in support of President Biden’s recent executive order calling for the development of safe, secure and trustworthy AI. PNNL scientists have already contributed to the IEEE Ethics Certification Program for Autonomous and Intelligent Systems, a program to make sure algorithms are trustworthy and free of bias.

As a DOE Office of Science laboratory, the primary focus of AI at PNNL is on challenges related to science. One team is using AI to develop ways to predict which hurricanes are most likely to strengthen rapidly and unexpectedly. Another team is focused on the world of microbes, whose actions can be harnessed to produce important chemicals faster and less expensively than is possible today.

Many of the applications of AI research at PNNL are focused on energy resilience and on keeping the world secure. Laboratory scientists use AI to improve the operation of the nation’s electrical grid, keeping power flowing to homes and businesses. Others are using machine learning to explore new combinations of compounds that could power the next generation of lithium batteries. Another common application is protection against the proliferation of nuclear materials.

Most recently, PNNL’s AI capabilities have been used to help firefighters predict the paths of wildfires.

PNNL scientists don’t simply use existing artificial intelligence techniques to tackle such challenges. Its researchers are transforming the fundamentals of AI—improving or creating programs for researchers around the world. One example is the impact that PNNL has had in an area known as “few-shot learning.” When there are just a few examples for an AI program to learn from, PNNL has developed ways to use that scant data to make reliable recommendations. Laboratory researchers have also pioneered the concept of physics-informed machine learning and machine reasoning, encoding physics knowledge into AI networks to improve their accuracy for a broad set of scientific questions.

PNNL’s work in AI is enhanced by its long history of data analysis and access to large amounts of data. For example, a materials chemistry experiment might rely on processing thousands of microscope images per second, and a hunt for biomarkers for diseases like cancer will rely on millions of sensitive measurements of the body’s proteins and other molecules. PNNL researchers are experts at generating, managing and analyzing such data—and at understanding how AI can best be deployed in such situations.

Among the other areas of interest to the participants of the Center for AI @PNNL: the creation and development of an AI-ready workforce; the deployment of AI in daily operations at PNNL; exploration of how humans and AI programs can best work together; and autonomous experimentation, where AI can direct robots’ activity in the laboratory, analyze results, then map out and even direct subsequent tasks.

Much of PNNL’s work in AI and computing is done with a broad group of partners, including North Carolina State University, University of Washington, Washington State University, Microsoft, Micron, University of Texas at El Paso, Georgia Institute of Technology, Western Washington University, and other national laboratories and organizations.

PNNL works closely with DOE’s Office of Science Advanced Scientific Computing Research program, which has invested in PNNL’s leadership of the advanced memory research and development project to support AI for science. Scientists at the Center for AI @PNNL will also work closely with DOE’s newly launched Office of Critical and Emerging Technology for the development of a national AI capability to develop safe, secure and trustworthy AI.

“AI is such an integral part of the landscape at PNNL that it’s difficult to find an area that AI hasn’t touched,” said Corley. “With so much happening, it’s crucial to make collaboration easier and for people to come together to develop a Lab-wide strategy to optimize everyone’s efforts.”