ChatterBaby (interpretation of crying of the child)

2018: Release of the application

In May, 2018 the University of California, Los Angeles (UCLA) released the free application decrypting crying of the child. It received the name ChatterBaby.

The algorithm tested more than on 2 thousand fragments of children's crying is capable to distinguish the sounds made by children on three main categories: pain, hunger and fidget. In a case with the first category identification happens to 90 percent accuracy, and here shouts when the child wants to eat and is in fidget, differ hardly, developers are recognized.

ChatterBaby application screenshots

Shouts of children differ therefore in the application machine learning thanks to which ChatterBaby can adapt to sounds of the specific child is applied.

The application analyzes frequencies at which there is a crying and also different models of sounds and silence — the leading author of a research Ariana Anderson from UCLA which is mother of four children says. — For example, when you hear shout with long pauses, it is the most probable that the child is capricious. But when babies are ill, shouts is normal more loudly, anguishes are more lingering, and between sounds there are almost no breaks.

ChatterBaby initially thought to help deaf parents to distinguish crying of the child from other loud sounds as the equipment for control of noise level is not capable to separate a baby's cry from loud babble. Further developers decided to expand functionality of the software. The application is placed in directories of Apple App Store and Google Play Store.^[1]

Ariana Andreson together with  the colleagues also works on using the application for early diagnosis of autism. It is supposed that the database will help to reveal anomalies in  crying of the child which demonstrate  violations in  development.^[2]

Notes