My PhD has been carried out here at the Cambridge University Engineering Department between January 1994 and June 1997 under the supervision of Phil Woodland. This research focussed on the use of part-of-speech word-categories for improving statistical language models. In particular, I developed a part-of-speech based variable-length n-gram language model, a means of combing it with conventional word-based n-grams within a backoff framework, and a method of incorporating long-range word relationships into the category-based language model.
Following my doctoral studies, I became a Research Associate at CUED from August 97 - August 98. In this time I continued my own research but also became involved more generally in the large-vocabulary work carried out in the Speech, Vision and Robotics group. In particular, I was involved in a project investigating techniques for noise-robustness in large-vocabulary speech recognition,
Online information on my publications and research reports may be obtained by choosing "research information" from my home page. For any further information, please contact me by email.