National Science Foundation to help build AI-enabled labs

A robotic arm places a tray into a laboratory machine, illustrating an automated research process in a high-tech lab. — SPEED’s robots coordinating AI-driven synthesis and characterization of small molecules. (Image credit: Milad Abolhasani/NC State)

UNC-Chapel Hill, NC State University and the Massachusetts Institute of Technology will work together as part of a new National Science Foundation-funded effort to  establish a nationwide network of artificial-intelligence-enabled automated laboratories .

The project is part of the NSF’s Directorate of Technology, Innovation and Partnership’s Programmable Cloud Laboratories program.

NC State will lead the effort with a four-year, $20 million award from NSF to support the Self-Driving Platforms for Experimental co-Design in Chemistry and Materials Science. The SPEED lab is led by NC State engineering professor Milad Abolhasani. Carolina will be a partner institution in the NSF grant-funded project, with professor Alex Miller of UNC College of Arts and Science’s chemistry department leading a team that includes professor Jillian Dempsey and professor Erik Alexanian, also in chemistry, and professor Ron Alterovitz of the College’s computer science department.

The effort will be part of the research portfolio of UNC’s Sustainable Energy Research Consortium, based in the College and led by Miller.

A key tenet of SPEED is advancing “self-driving labs” where human scientists are directing research goals, while smart robotics carry out the procedures. A major goal of the project (and the broader PCL Test Beds initiative) is to enable remote control of automated laboratory instrumentation.

One of Carolina’s major roles is to help build those capabilities by developing new access/control interfaces and by piloting a broad range of different chemical reactions. The promise of SPEED, according to Miller, is making cutting-edge research questions more accessible to researchers all across the country.

“The development of hardware and software tools that enable users that don’t have access to advanced instrumentation to test their research ideas could be revolutionary,” Miller said. “The individual self-driving lab modules within SPEED can accelerate the process of moving from lead discovery to optimized outcome dramatically — by 100 times or more — by performing experiments in parallel and using machine learning methods to predict the most promising next set of experiments. This combination of accelerated discovery and broad accessibility is really exciting.”

The Carolina team will be instrumental in implementing the project’s science drivers on the automation infrastructure at NC State. These science drivers include enabling more efficient synthetic pathways to produce high-value specialty/fine chemicals and accelerating the translation of next-generation materials into manufacturing for more energy efficient digital displays and other technologies.

“The automated robotic experimentation capabilities that have been developed at NC State over the past few years provide an extraordinary opportunity for accelerating the discovery and improvement of chemical reactions that can benefit society,” Miller added. “Our UNC team will expand the scope of reactions being studied in the self-driving labs and bring new tools for robust remote communication with the instruments.”