Home Page

This course will explore the fundamentals of speech recognition and statistical language modeling. The course will begin by providing a brief introduction to acoustic phonetics, signal processing, and probability theory for speech recognition.  The problem of automatic speech recognition will be explored and lead to the presentation Hidden Markov Models (HMMs) and their practical use for acoustic modeling, language modeling and search.  The course will cover in detail both the frame-synchronous and stack-based approaches to automatic speech recognition.  Statistical Language Modeling methods such as n-gram language models, class-based language models, and probabilistic context-free grammars will also be introduced.  Issues related to speaker and environment adaptation will be presented at the end of the course.

© James Martin 2011