Using artificial intelligence to identify bird calls is the purpose of a new competition

The laboratory already has several projects that record ambient sounds over long periods, and the analysis has been left to specialists who distinguish bird sounds from other sounds. This causes them to be time consuming and often incomplete, but now the data science can help, using large databases of bird song sounds to train Artificial Intelligence (AI) models.

With the proper classification and detection of sound, researchers will be able to learn more about the habitat of birds and their quality of life, also based on changing populations.

That is why the laboratory decided to proceed with a competition in which it is intended to identify different varieties of bird songs in recordings that are provided. Noises may be distorted by anthropogenic sounds, such as aircraft noise, other birds and the sounds of other animals. Participants are asked to bring their ideas to build ways to identify and classify these sounds in the best way.

The contest is open until September and the first prize is worth 12 thousand dollars.