Discriminative feature transformation by guided discriminative training

In this paper, we investigate guided discriminative training in the context of improving multi-class classification problems. We are interested in applications that require improvement in the classi-fication performance of only a subset of the classes at the possi-ble expense of poorer classification performance of the remain-ing classes. However, should the classification of the remaining classes deteriorate, it is guaranteed not to be worse than the extent that the user specifies. The problem is formulated as a nonlinear programming problem, which can be translated to a unconstrained nonlinear optimization problem using the barrier method that, in turn, can be solved by gradient descent method. To prove the con-cept, we apply guided discriminative training to derive an optimal linear transformation on the mel-filterbank log power spectra to improve TIMIT phoneme classification. Encouraging results are obtained. 1