The Faculty of Science Research Showcase Award

Judges at the Faculty of Science Research Showcase from the University of Auckland, New Zealand, have awarded our research poster third place this year. I would like to thank everyone involved in this research for their support, including my collaborators, lab members, supervisors and sources of funding, the Maungaharuru Tangitu, and the Department of Conservation in New Zealand.

This research has entailed a considerable effort to extract thousands of feeding calls using synchronization between RFID readers and acoustic recorders. We found that titipounamu, Acanthisitta chloris, have unique individual feeding calls and produce group vocal signatures at nests. We then revealed that machine learning techniques could accurately classify these calls to the correct individuals and nests ( respectively, with 84% and 86% accuracy).

As highlighted in our poster below, the nest vocal signature is a rather intriguing behaviour, as it suggests that titipounamu may modulate their vocalizations towards one another.

Creative work with spectrograms showing the individual vocal signature of titipounamu feeding calls