This section addresses the inverse problem in robust speech processing. A problem that speaker and speech recognition systems regularly encounter in the commercialized applications is the dramatic degradation of performance due to the mismatch of the training and operating environments.