AI is changing how we study bird migration

A multi-university research team led by Juan Pablo Bello – who has a joint appointment in the NYU Steinhardt School of Culture, Education and Human Development – has developed BirdVoxDetect, an AI system that identifies migrating birds by their nocturnal flight calls. The software helps track bird migration patterns in a warming world filled with hazardous human infrastructure. The neural network can detect and classify bird species from brief, 50-millisecond calls. Through "few-shot" learning, the system can be trained to recognize new bird species with minimal examples. According to Bello, it's "like learning a language similar to one you already speak," making the technology adaptable for different species worldwide with just a few training samples.