Monophonic sound source separation by non-negative sparse autoencoders | IEEE Conference Publication | IEEE Xplore