Recovering lost bird songs is now possible with the help of technology. Through robots and artificial intelligence, Argentine scientists managed to teach juvenile specimens of the Rufous-collared Sparrow (Zonotrichia capensis) to learn songs they no longer heard in nature. Now they are working on replicating this experience with endangered birds from Patagonia.
The project started with the handwritten musical annotations of Argentine ornithologist Fernando Nottebohm, who had recorded various bird songs in Pereyra Iraola Park (north of La Plata) in the 1960s.
Upon returning to the same area six decades later, researchers recorded the current songs of Rufous-collared Sparrows to determine how many had survived. To their surprise, they noticed that very few had been preserved.
Technology to recover bird songs
What would happen if the old songs were played again through acoustic robots? Would these artificial reproductions be considered by the juvenile Rufous-collared Sparrows as imitable models? These were questions posed by Gabriel Mindlin, director of the Institute of Interdisciplinary and Applied Physics (INFINA) at UBA (University of Buenos Aires).
During the spring, three devices were placed in the park to emit a pair of songs per minute every morning, with interspersed periods of pause.
Upon returning the following season, they noticed that the juveniles had incorporated the melodies of the mechanical tutor into their repertoire. Mindlin mentions that this finding could counteract the loss of culture in certain birds due to the reduction in the adult population.
“When we talk about biodiversity we talk about genetics, but there is also an animal culture, this was an opportunity to see if we could revive some fragments of culture,” he points out.
The contribution of Artificial Intelligence
For the reconstruction of the songs, based on the data collected by Nottebohm, scientists used interpretable science tools such as dynamic systems and algorithms.
In the case of Artificial Intelligence, this tool allows for a more personalized classification of each specimen. This technology is important for studying and monitoring endangered species. “Once we have the sound pattern, AI helps us classify and identify each individual by their song,” explains Mindlin.
The team continues to refine the methodology and set new goals. They are working with the Patagonia Program of Aves Argentinas in the province of Santa Cruz, identifying individuals of the small crake (an endangered species) through their vocalizations. Due to the difficulties of visualizing it in the field because of its rugged habitat (sparse territories of the Patagonian desert), they aim to identify and track individuals through their songs. “We are working to monitor the population through acoustic sensors,” says Mindlin.
Additionally, in Pereyra Iraola Park, they are designing the second part of the research with Rufous-collared Sparrows. “We are trying to replicate the work with ringed animals. We want to see the evolution of this communication that we introduced in the field. We want to see if the juveniles that incorporated the song transmit it to their offspring or if it disappears,” notes Mindlin.
Fact
The scientist explains that birds sing for two reasons: for territorial defense and to find a mate for reproduction.
