Publication

Posterior features applied to speech recognition tasks with user-defined vocabulary