No abstract

In this article, a mathematical framework that jointly optimizes the parameters of classifier and feature extractor is presented. In this approach, feature extraction is formulated as a process of projecting the signals onto a smaller subspace in which the statistical properties of the signal can be efficiently modeled. An algorithm, called statistical matching pursuit (SMP), is proposed to learn from the training data the optimal projection dimensions and the extent of signal reduction. The algorithm is designed to achieve unconditional convergence and can be seamlessly incorporated into the expectation-maximization (EM) algorithm employed to train the classifier. Finally, we report some experimental results on speech recognition and elaborate the potential of the proposed method.

In this paper, a mathematical framework for learning the acoustic features from a central auditory representation is presented. We adopt a statistical approach that models the learning process as to achieve a maximum likelihood estimation of the signal distribution. An algorithm, called statistical matching pursuit (SMP), is introduced to identify regions on the cortical surface where the features for each sound class are most prominent. We model the features with distributions of Gaussian mixture densities, and employ the expectation-maximization (EM) procedure to both improve the parameterization and refine iteratively the selection of cortical regions from which the features are extracted. The learning algorithm is applied to vowel classification on TIMIT database where all the vowels (excluding diphthongs, nine in total) are regarded as individual classes. Experimental results show that models trained under SMP/EM algorithm achieve a comparable recognition accuracy to that of conventional recognizers.

No abstract