Never Ending Learner of Sounds

Sound waves pervade the entire universe and as humans our hearing capacity permits us to detect only a limited set of waves. Machines, however don’t have this limitation. Imagine a machine that can sense a knock on the door, someone breaking into your home, detect accidents by itself and take decisions by sensing different sounds in the environ- ment! Our project, Never Ending Learning of Sound aims to make the machine continuously learn sounds that exist, by crawling the entire web. This will help the machine to under- stand, sense, categorize and model the relationships between different sounds. It is an effort to build one of the largest struc- tured sound database. This is a difficult problem due to three aspects. 1. Amount of Data needed to be processed for sound 2. Interference of Sounds 3. Validation of the results obtained with minimum human intervention