Bijaya Adhikari and Guanpeng Li, both assistant professors in the Department of Computer Science in the College of Liberal Arts and Sciences, recently earned CAREER awards from the National Science Foundation.
The award, the organization’s most prestigious honor in support of early-career faculty, includes a five-year research grant for each faculty member.
Adhikari aims to stop spread of hospital infections
Adhikari will receive $599,783 for his efforts to develop new computational methods to identify the most effective intervention strategies to prevent and stop the spread of healthcare-associated infections (HAIs). HAIs spread primarily in healthcare settings where patients are often exposed to harmful germs and bacteria and develop infections that are unrelated to why they were admitted to the hospital. According to the CDC, about one in 31 patients has an HAI at any given time, and the US spends billions of dollars each year to combat the problem.
Using medical records, Adhikari will build algorithms to assess HAI risk and help healthcare providers make better decisions in real time. Once developed, the algorithms can be deployed as software or an online platform that healthcare workers can access as they move throughout a hospital or other healthcare facility
“For example, healthcare workers need to decide which patient to move to an isolation room to reduce the total number of infections in the ward, or they might need to cluster patients into bubbles, similar to what was done during COVID-19.” Adhikari said. “We pose these questions as computational problems and look for algorithmic solutions with some guarantees. We aim to design these algorithms to work in real-time as patients are admitted and discharged.”
Adhikari plans for his innovative tool to also predict future infections and detect hidden or symptomless infections through machine learning.
“If a healthcare provider touches a contaminated surface and then touches a patient, the pathogen can spread,” Adhikari said. “By predicting such events, we can design better intervention strategies.”
Li tackles vulnerabilities in high-performance computing
Li will receive $590,074 for his project, which aims to make high-performance computing (HPC) more resilient to soft errors, or temporary faults in the hardware.
High-performance computing uses powerful computers and advanced algorithms to solve extremely complex problems, to the tune of trillions of calculations per second. HPC allows researchers to process vast amounts of data and is essential to an array of fields.
HPC systems can experience soft errors when external factors such as cosmic rays or electrical noise disrupt calculations. Today’s HPC systems experience soft errors “at least hourly” according to Li, and these can lead to costly mistakes.
“Imagine using a GPS navigation system for a long road trip,” Li said. “If a cosmic ray causes a soft error, it could alter a single calculation, causing the GPS to incorrectly estimate the remaining distance or suggest a wrong turn. In the context of HPC, this could mean an error in a weather prediction model that leads to incorrect forecasts or a mistake in drug simulation data that could delay the discovery of a life-saving treatment.”
Li’s team will develop software tools to detect and manage soft errors before they disrupt calculations. By analyzing how these errors spread, the researchers will identify and reinforce the most vulnerable part of the computer code in HPC systems.
By reducing errors in HPC systems, Li’s research has the potential to improve weather forecasting, speed up the drug discovery process in pharmaceutical research, and enhance engineering designs with more reliable simulations in the aerospace and automotive industries.
“By making HPC systems more dependable, our work ensures that society continues to benefit from advancements in science, technology, and everyday applications,” Li said.