Discriminatively estimated joint acoustic, duration, and language model for speech recognition
We introduce a discriminative model for speech recognition that integrates acoustic, duration and language components. In the framework of finite state machines, a general model...