Learning State Labels for Sparse Classification of Speech with Matrix Deconvolution

Non-negative spectral factorisation with long temporal context has been successfully used for noise robust recognition of speech in multi-source environments. Sparse classification from activations of speech atoms can be employed instead of conventional GMMs to determine speech state likelihoods. For accurate classification, cor-rect linguistic state labels must be assigned to speech atoms. We propose using non-negative matrix deconvolution for learning the labels with algorithms closely matching a framework that separates speech from additive noises. Experiments on the 1st CHiME Chal-lenge corpus show improvement in recognition accuracy over labels acquired from original atom sources or previously used least squares regression. The new approach also circumvents numerical issues en-countered in previous learning methods, and opens up possibilities for new speech basis generation algorithms. Index Terms — Automatic speech recognition, noise robust-ness, non-negative matrix factorization, sparse classification 1