- MERL Seminar Series.)
(Learn more about the
Date & Time:
Tuesday, April 25, 2023; 11:00 AM
Machine learning can be used to identify animals from their sound. This could be a valuable tool for biodiversity monitoring, and for understanding animal behaviour and communication. But to get there, we need very high accuracy at fine-grained acoustic distinctions across hundreds of categories in diverse conditions. In our group we are studying how to achieve this at continental scale. I will describe aspects of bioacoustic data that challenge even the latest deep learning workflows, and our work to address this. Methods covered include adaptive feature representations, deep embeddings and few-shot learning.
Tilburg University / Naturalis Biodiversity Centre
Dan Stowell is Associate Professor of AI & Biodiversity, jointly appointed at Tilburg University and Naturalis Biodiversity Center (NL). Since 2012 he has led research on computational bioacoustics using machine learning and signal processing. He is a co-founder of the data challenge "detection and classification of sound scenes and events (DCASE)", now an annual IEEE workshop and challenge, and is an elected member of the International Bioacoustics Council (IBAC) Executive Committee, and an Associate Editor at PLOS Computational Biology and PeerJ Computer Science. He is cofounder and CTO of Warblr Ltd (UK), a phone app that automatically recognises bird sounds - the app has received awards and copious press coverage, and has tens of thousands of regular users around the UK. In the Netherlands he is involved in many funded projects to bring AI-powered biodiversity monitoring to fruition.