Be it chirp, quack or some form of swan song, a new artificial intelligence database being created by a University of Pittsburgh professor seeks to improve avian research and conservation by linking bird calls with their respective species.
Justin Kitzes, an assistant professor in the Department of Biological Sciences, has been awarded a $91,208 Microsoft and National Geographic AI for Earth Innovation grant to develop the first free, open source models to allow academic researchers, agencies, non-profits and citizen scientists to identify bird songs in acoustic field recordings.
Kitzes will collaborate with a team that includes Steven Latta, director of conservation and field research at The National Aviary in Pittsburgh, Pennsylvania, and researchers from Powdermill Avian Research Center in Stahlstown, PA., the Smithsonian Conservation Biology Institute in Front Royal, VA., and Williams College in Williamstown, Mass. Barry Moore II, the technical lead at Pitt's Center for Research Computing, will also participate in the project.
Work is expected to kick off in early 2019, with the models and associated software to be released by the end of the year.
"We can't protect birds and other wild species unless we know how their populations are changing," Kitzes said. "Our goal is to support other scientists in monitoring bird populations using new, technologically driven approaches that have to potential to scale much larger than traditional field observation methods."
Kitzes was one of 11 scientists chosen for this year's AI for Earth Innovation Grant. In total, awardees received $1.28 million in funding.
The AI for Earth Innovation Grant program will provide award recipients with financial support, access to Microsoft Azure and AI tools, inclusion in the National Geographic Explorer community and affiliation with National Geographic Labs, an initiative launched by National Geographic to accelerate transformative change and exponential solutions to the world's biggest challenges by harnessing data, technology and innovation.
The grants will support the creation and deployment of open source trained models and algorithms that are available to other environmental researchers and innovators, and thereby have the potential to provide exponential global impact. The AI for Earth Innovation Grant program builds upon Microsoft's AI for Earth program, now providing grants to nearly 200 individuals and organizations on all seven continents and National Geographic Society's 130-year history of grantmaking, supporting more than 13,000 grant projects along the way.