With this new AI tool, ecologists can track rare birds by their songs!
Bird watchers, rejoice! From now on, ecologists will be able to track rare birds by their songs, thanks to a new deep-learning AI tool. Scientists have created a new AI that generates lifelike bird songs to train bird identification mechanisms, which will help ecologists monitor rare species.
Canadian scientists from the University of Moncton have developed a new deep-learning AI tool to track rare birds. The neural network can generate realistic bird songs to train tools to identify rare bird species in the wild. The results are published in the journal Methods in Ecology and Evolution.
It is easy to identify common bird species by their songs, thanks to various software and smartphone applications. The problem arises if the software has not heard a particular bird before or has only a small sample of recordings. And it is a common problem with rare birds.
To address this problem, researchers at Canada's University of Moncton have developed ECOGEN, a first-of-its-kind deep learning AI tool that can generate realistic bird sounds to improve samples of underrepresented species. Adding ECOGEN-generated artificial bird song samples improved song classification accuracy by an average of 12%.
The neural network converts real recordings of bird songs into spectrograms and then generates new AI images to increase the dataset for rare species with few recordings. These are then converted back to audio to train bird sound identifiers. Experts used 23,784 recordings of wild birds from around the world, covering 264 species, to create the algorithm.
Dr. Nicolas Lecomte, one of the lead researchers, said, "Due to significant global changes in animal populations, there is an urgent need for automated tools, such as acoustic monitoring, to track shifts in biodiversity. However, the AI models used to identify species in acoustic monitoring lack comprehensive reference libraries. With ECOGEN, you can address this gap by creating new instances of bird sounds to support AI models. Essentially, for species with limited wild recordings, such as those that are rare, elusive, or sensitive, you can expand your sound library without further disrupting the animals or conducting additional fieldwork."
Researchers say that creating "synthetic songs" of birds can help conserve endangered bird species, as well as provide valuable information about their vocals, behavior, and habitat preferences.
Thanks to the ECOGEN tool, some extremely rare species, such as the critically endangered regent honeyeaters, where young individuals cannot learn the songs of their species because there are not enough adult birds to model them, could be conserved. The tool would also be suitable for mammals, fish, insects, and amphibians. Another curious fact is that the AI tool is open source and can be used on basic computers.