# Has any research been done on DNN Music?

DNNs are typically used to classify things (of course) but can we let them go wild with sounds and then tell them if we think it sounds good or not? I'd like to think after a training class has been made (perhaps comparing the output to an existing song) we could get an NN that has a basic concept of music.

Timing would be an issue; I'm not sure how feasible this is. A strongly weighted input attached to all hidden layers perhaps? Use it as the bias?

Is this even slightly feasible?

• Google recently released a package in the area of music and art its called Magenta. I think its possible to classify a sound is good or not if we have proper dataset.
– Eka
Aug 5, 2016 at 9:19
• DNN := Dynamic Neural Networks, right? Aka. Recurrent Neural Networks?
– Luis
Aug 5, 2016 at 11:49