A paper on deep-learning speech features of PPAOS published in CLP.