A hybrid SVM/MCE training approach for vector space topic identification of spoken audio recordings.

Timothy J. Hazen, Fred Richardson

Browse the full Interspeech paper archive.