A neural network alternative to non-negative audio models.

Paris Smaragdis, Shrikant Venkataramani

Browse the full ICASSP paper archive.