INTERSPEECH 2010

We describe an algorithm that performs regularized nonnegative matrix factorization (NMF) to find independent components in non negative data. Previous techniques proposed for this purpose require the data to be grounded, with support that goes down to 0 along each dimension. In our work, this requirement is eliminated. Based on it, we present a technique to find a lowdimensional decomposition of spectrograms by casting it as a problem of discovering independent nonnegative components from it. The algorithm itself is implemented as regularized nonnegative matrix factorization (NMF). Unlike other ICA algorithms, this algorithm computes the mixing matrix rather than an unmixing matrix. This algorithm provides a better decomposition than standard NMF when the underlying sources are independent. It makes better use of additional observation streams than previous nonnegative ICA algorithms.
Bibliographic reference. Raj, Bhiksha / Wilson, Kevin W. / Krueger, Alexander / HaebUmbach, Reinhold (2010): "Ungrounded independent nonnegative factor analysis", In INTERSPEECH2010, 330333.