Welcome to the Fall 2019 Edition of Science Highlights.
Peace of Mind Through AI
DEEP LEARNING” ARTIFICIAL INTELLIGENCE ON PSC SUPERCOMPUTERS PROMISES FEWER FALSE ALARMS AND EARLY PREDICTION OF BREAST CANCER DEVELOPMENT
Screening mammography is an important tool against breast cancer. But it has its limitations. A “normal” screening mammogram means that a woman doesn’t have cancer now. But doctors wonder whether “normal” images contain clues about a woman’s risk of developing breast cancer in the future. Also, most women “recalled” for more tests when their mammograms show suspicious masses don’t have cancer. Scientists at the University of Pittsburgh Medical Center are using PSC’s Bridges and Bridges-AI supercomputers to run artificial intelligence programs meant to determine the risk of developing breast cancer and to prevent false recalls.
He Who Hesitates
Anton 2 Simulations Identify Critical Pause in Flu-Virus Protein Motion; May Be a Target for Future Therapy
Viruses such as influenza and HIV take a heavy toll, both in human life and in dollars. But these viruses are shape-shifters, changing their outer proteins via mutation so human immunity and antiviral drugs have a hard time keeping up. A team from Rice University and Baylor College of Medicine has used the Anton 2 supercomputer at PSC to simulate the role that a portion of a critical protein called “hemagglutinin (HA)” plays in the flu virus merging with host cells and injecting its genes. They found that the protein’s “stem domain,” HA2, pauses in the process—a critical step that may offer a target for future therapy that can outsmart the viral transitions. The work offers a completely new approach to therapy since it focuses on the critical injection mechanism rather than the virus’s outward structures.
Building Better Batteries
CMU Scientists Use Bridges to Simulate Improved Battery Component
The move toward cleaner, cheaper energy would be much easier if we had more powerful, safer battery technologies. Carnegie Mellon University (CMU) scientists are using PSC’s Bridges supercomputer to simulate new battery component materials that are inherently safer and more powerful than currently possible.
Life Isn’t Chess – Times Six
In a First, a CMU AI Program Running on Bridges Overcomes Multiple Human Poker Champions
Artificial intelligence (AI) research took a great leap forward when a Carnegie Mellon University computer program overcame the world’s best professional players in a series of six-player poker games. Running on PSC’s Bridges supercomputer, the Pluribus AI was the first to surpass humanity’s best at a multiplayer, “incomplete information” game. Experimenting with such games offers more useful lessons for real-world problems such as security, business negotiations and cancer therapy than “complete information” games such as chess or Go.
Going With The Flow
Bridges Calculations Shed Light on Heat Transfer Between 2D Electronic Components
Smaller electronic components offer us more power in our pockets. But thinner and thinner components pose engineering problems. Anisotropic materials—those with properties that vary in direction—hold promise for being unusually versatile. Still, their properties, particularly as they become “two dimensional” or ultra-thin, are not well understood. Using a combination of calculations on PSC’s Bridges and lab experiments, a UCLA-led group showed that an atomistic model can explain and predict the transfer of heat between aluminum and black phosphorous, a highly anisotropic material with possible applications in future devices.
PSC News in Brief
AIDR 2019 Conference Brings AI Researchers to Pittsburgh
• PSC-Affiliated Cybersecurity Center of Excellence Receives $12.5-Million Renewal Grant
• Pittsburgh Girls Code
• BEST Summer Institute Trains High School Teachers in Computational Biology
• Computer Simulation: Antibody-Resistant Infection Registry Can Reduce Prevalence
• PSC Team Wins Best Student Paper Award
• Bridges Powers AI Poker Demonstration
• PSC Interns Explore Data, Biology, Computing Architecture