Like many disciplines, ecology is in the midst of a “data revolution” driven by recent availability of massive amounts of data. One approach, automated acoustic surveying, shows particular promise for studying sound-producing animals like birds, wolves, and frogs. These surveys combine inexpensive, autonomous sound recorders with artificial intelligence capable of identifying the recorded sounds.
As in other fields, the data revolution in ecology presents novel challenges and opportunities. This seminar discusses three of these challenges – developing machine learning methods, interpreting uncertain machine learning outputs, and communicating and collaborating across disciplines – through the lens of the Kitzes Lab’s research developing automated acoustic survey methods.
Presented for “Boundary-spanning Seminar Series,” Internet of Catalysis, National Science Foundation Research Traineeship, University of Kansas