Back

Former Lawrence fellow Amanda Randles awarded 2023 ACM Prize in Computing

Amanda Randles ACM Prize (Download Image)

Amanda Randles, recipient of the 2023 ACM Prize in Computing.

 

The Association for Computing Machinery (ACM) named Amanda Randles as the recipient of the 2023 ACM Prize in Computing. Randles, a former Lawrence fellow and current Lawrence Livermore National Laboratory (LLNL) collaborator, was recognized for “groundbreaking contributions to computational health through innovative algorithms, tools and high-performance computing methods for diagnosing and treating a variety of human diseases.”

Randles is the Alfred Winborne and Victoria Stover Mordecai associate professor of biomedical sciences at Duke University’s Pratt School of Engineering. Her lab at Duke focuses on modeling the blood flow through the human circulatory system using supercomputers, including LLNL’s Lassen and Quartz machines. By speeding up simulations, they can determine what medical interventions, like stents, also can inform the mechanisms driving cancer metastasis by examining how tumor cells interact with red blood cells and adhere to surfaces.

View video showing the adaptive physics refinement method. (Video: Liam Krauss, LLNL)

“Receiving this award and getting buy-in from the supercomputing community about this kind of interdisciplinary, applied work is exciting,” Randles said. “This is not conventional computer science and it’s great to see the community supporting such applied research.”

Supercomputing allows Randles’ team to model the circulatory system at long timescales and fine spatial resolution. Her group has two major efforts underway, both active collaborations with LLNL. In the first, they aim to extend their simulations to capture long-term data from wearable devices like smart watches. For the second, they are diving down spatially with a new method called adaptive physics refinement.

“Say you have a cancer cell moving through the body. With adaptive physics refinement, we basically put a window around the cancer cell that includes all the individual red blood cell interactions,” Randles said. “Outside the window, we have the overall fluid mechanics at much coarser resolution, saving us a lot of wasted computational power.”

Randles began this work as a graduate student at Harvard University and continued it as a Lawrence fellow at LLNL. It was then that she connected with Erik Draeger, group leader of Livermore’s Scientific Computing Group in the Center for Applied Scientific Computing.

“I first met Amanda when she interviewed for the Lawrence Fellowship in 2013. I expected to have a brief, polite chat where I’d talk up research life at Livermore, maybe hype our supercomputers a bit, then go about my day,” Draeger said. “To my surprise, Amanda was keenly aware of the power of supercomputing and had an ambitious and exciting research agenda centered around leveraging Livermore’s computing power to the maximum possible extent.”

Model of a patient's aorta. Each square is handled by a different computer processor. (Image: Liam Krauss, LLNL)

At LLNL, Randles continued to improve her circulatory system code, HARVEY, which was written from the ground up to run on large-scale supercomputers and be parallelizable. Laboratory Directed Research and Development funding also allowed her to experimentally validate the code by 3D printing models.

“The Lawrence Fellowship lets you drive your research,” Randles said. “I knew what I wanted to do, and it gave me the support and freedom to be able to do that.”

The Lawrence Fellowship is a highly competitive postdoctoral position at LLNL that is open to all technical disciplines. Fellows are free to pursue their own independent research agenda within an environment of subject matter experts who provide guidance and mentorship.

“Lawrence Fellowships are awarded to candidates with exceptional talent, scientific track records and potential for significant achievements,” said Eric Schwegler, the director of LLNL’s Academic Engagement Office. “We were proud to have Amanda as a fellow, and this award is yet another indication of her fascinating and inspiring work.”

Randles emphasized that her time at LLNL resulted in many of her current research directions and collaborations. She continues to work with Draeger on designing new algorithms and methods to focus computational horsepower where it is most needed.

“Amanda is incredibly deserving of this award, not just because of how hard she works but because of the creativity and passion she brings out in others,” Draeger said. “I’m grateful that she was part of LLNL for a time and that she has continued to apply what she learned here to such important problems.”

—Ashley Piccone