Professor Mark Gales

Mark is one of our Official Fellows and Fellows' Steward, and is Professor of Information Engineering at the University. He was a Research Fellow at Emmanuel in 1995 and was elected an Official Fellow in 1999. He is on sabbatical leave during academic year 2024/25.

Biography

Mark studied for a BA in Electrical and Information Sciences at Cambridge between 1985 and 1988. After graduation, he worked as a consultant at Roke Manor Research Ltd, developing novel radar systems for governments and industry. In 1991, Mark took up a position as a Research Associate in the Speech Vision and Robotics group at Cambridge's Engineering Department. He then completed his doctoral thesis in 1995: 'Model-Based Techniques for Robust Speech Recognition', supervised by Professor Steve Young, a (now Life) Fellow at Emmanuel. After his PhD, Mark was elected a Research Fellow at the College, followed by a position as a Research Staff Member in the Speech group at the IBM Thomas J. Watson Research Centre, New York. In 1999, he returned to Cambridge's Engineering Department as a University Lecturer, when he was appointed to the Emmanuel Fellowship, and is currently a Professor of Information Engineering. Mark is a Fellow of both the Institute of Electrical and Electronics Engineers (IEEE) and the International Speech Communication Association (ISCA).

Teaching

Mark lectures on the Engineering undergraduate course primarily to third and fourth–year undergraduates taking information engineering related papers. He supervises second–year students at Emmanuel on related subjects.

Research

Mark's research interests are in speech and language processing, broadly aiming to improve access and performance in these areas. On returning to Cambridge, he originally focused on speech recognition, contributing to a variety of key developments in the application of Hidden Markov Models to this area. These models underpinned the first wave of widespread deployment of automatic speech recognition systems. In more recent years he has focused on language learning and assessment, associated with the Automated Language Teaching and Assessment (ALTA) Institute funded by Cambridge University Press & Assessment (CUP&A), and low–resource speech processing.

There is an increasing demand for English as a second language, and many students require, for example, official qualifications to demonstrate language proficiency. To cope with this demand, computer aided learning and automatic assessment have become increasingly important. The ALTA Institute at the University works on these problems, aiming to enable automatic assessment and feedback across the 4 key skills of writing, speaking, reading and listening.

Mark’s particular interest is on automating this process for spoken tests. The technology developed within his group has been commercially deployed by CUP&A for tests such as Linguaskill, which have been taken by hundreds of thousands of candidates, and a free practice research platform, Speak and Improve, that is used by tens of thousands of people worldwide a day.

He is also interested in low–resource speech processing. It is estimated that there are up to 7,000 different languages spoken around the world. Of these 90% are used by less than 100,000 people. It is not economically feasible to develop speech technology for all languages. This has led to an interest in developing approaches that allow the same speech processing system to be applied to many, preferably all, languages. This has led to Mark working on a wide range of languages from Dholuo to Paraguayan Guarani to Tok Pisin, developing approaches that are widely applicable with limited knowledge of the target language.