The Providence Journal named Brandon Marshall, assistant professor of epidemiology at the School of Public Health, and Sherief Reda, associate professor of engineering and computer science, to its list of “Trailblazers to follow in 2017” for their cutting-edge research in their respective fields.
Marshall, who came to Brown in 2012, uses statistical models to predict the spread of epidemics, including overdose and infectious diseases such as Hepatitis C, he said. “We do some mathematical modeling and … surveillance using existing data to understand where overdoses are occurring.”
“(Marshall) is addressing very important questions about the interface between injection drug use and infectious diseases,” said Stephen Buka, professor of epidemiology who chaired the department when Marshall was hired.
Lynn Taylor, a physician at Miriam Hospital working under the “Rhode Island Defeats Hep C” program reached out to Marshall for help understanding the spread of the disease in the state. The project aims to provide policy makers with a more accurate picture of the epidemic, including data on how many people need treatment so the state can more effectively fight the disease, Marshall said. He estimated that 12,000 to 16,000 people in Rhode Island have Hepatitis C.
“(Hepatitis C) is very tricky — it’s called a silent killer because people get infected and they don’t know that they have it for decades,” Marshall said.
Marshall also directs Rhode Island’s overdose surveillance program, which monitors the state’s opioid epidemic, he said. “We have the fifth highest rate of overdose in the nation,” he added. In 2015, 290 Rhode Islanders died of overdose — a figure higher than the combined mortalities from car accidents, gun violence and suicide.
Marshall collaborates with the Rhode Island Department of Health, which provides data on the number of overdose deaths, emergency visits and addiction treatment programs available in the state and uses this data to model the epidemic. “The goal is to reduce the number of deaths by a third by 2018, so we need to work really quickly and effectively in order to make that happen,” he said.
While many public health researchers hope to influence public policy decisions and clinical actions, “few are accomplishing it with the speed and impact that (Marshall) is,” Buka said.
In the computer science department, Reda has been researching ways to configure computer circuits to make them more adaptive and efficient, he said. By using programmable circuits called field-programmable gate arrays, Reda can adapt computers to run certain algorithms faster than a standard computer could.
The Journal reached out to Reda after he received the Rhode Island Commerce Corporation’s Innovation Voucher, a $50,000 grant, he said. Reda won the grant with Luis Camacho ScM’07, one of his former students, who reached out for help developing iris recognition cameras.
These cameras have several potential security applications, such as identity confirmation. But those cameras would need computers to process and recognize images of different irises, and “we want to embed those computers in the camera,” Reda said. Doing so would require a small computer capable of high-level processing. “Our goal is to develop custom silicon processors that would deliver that high performance,” he said.
Working with companies, Reda said he is forced to consider real-world constraints that he would not usually confront in academia.
Outside of his work with iris recognition cameras, Reda and his students have earned a variety of other external grants, including one from the Defense Advanced Research Projects Agency for the development of chips in drones and unmanned aerial vehicles.
Reda’s research plays into the larger emerging field of deep neural networks, which are systems that allow computers to recognize complex inputs such as speech or images. “If you look at most of the advancements that happened in the last … five years” in speech recognition products such as Apple’s Siri and Amazon’s Alexa, “these have come from deep neural networks,” Reda said.
These deep neural networks will eventually be constrained, as in the case of iris recognition, by the physical size of the circuits. When those circuits can no longer get any smaller, Reda’s work will allow them to continue improving through reconfiguration instead of resizing.
“Professor Reda is one of our most innovative young faculty (members),” said Lawrence Larson, dean of the School of Engineering. “(The award) is great confirmation of all the great work that our faculty do.”