Using phone apps and software, it’s become easier than ever for bird watchers and ecologists to spot birds in different environments. However, spotting and monitoring rare species in the wild is still not an easy task for conservationists.
Now, researchers have developed a new deep-learning artificial intelligence (AI) tool that generates birdsongs to train bird identification tools that can help keep an eye on underrepresented avian species.
The tool, called ECOGEN, developed by researchers at the University of Moncton, Canada, is a first-of-its-kind deep learning AI tool, a press statement by the British Ecological Society explains. The tool generates lifelike bird sounds of rare bird species that can be used to train audio identification tools, which are part of ecological monitoring and often have disproportionately more accessible information on common species.
The researchers also found that adding birdsong samples generated by ECOGEN to a birdsong identifier improved the song classification accuracy by 12% on average. These findings were published recently in the journal Methods in Ecology and Evolution.
According to the researchers, there is an urgent need for automated tools such as acoustic monitoring to track changes in biodiversity and significant shifts in animal populations across the world. However, the AI models used to identify species in acoustic monitoring lack comprehensive reference libraries, Nicolas Lecomte, one of the lead researchers, said in the statement.
“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,” he adds.
The researchers believe that creating artificial bird songs can also provide valuable insight into endangered bird species’ vocalisations, behaviours and habitat preferences. Moreover, the ECOGEN AI tool can also be used to help protect extremely rare species, such as the critically endangered regent honeyeaters, where young birds are unable to learn their species’ songs because there aren’t enough adult birds to learn from, the statement added.
The main advantage of the tool is its accessibility, as it is open source and can be used on basic computers.
The tool could prove to be helpful for other types of animals as well, Lecomte said in the statement: “While ECOGEN was developed for birds, we’re confident that it could be applied to mammals, fish (yes they can produce sounds!), insects and amphibians.”