Title at NITech
Professor(December 2017-January2018)
Affiliation, Institution & University
1) The Centre for Speech Technology Research, and 2) Linguistics and English Language, School of Philosophy, Psychology and Language Sciences, University of Edinburgh, UK
Introduction
Simon King is the professor of Speech Processing at the University of Edinburgh and the Director of CSTR (The Centre for Speech Technology Research), in the UK. He graduated from Cambridge University with an M.A. in Engineering and M.Phil. in Computer Speech and Language Processing, then obtained his Ph.D. from the University of Edinburgh.
In his early career, Prof. King pioneered the use of articulatory feature representations of speech for use in automatic speech recognition, and collaborated with leading scientists at the University of Washington to apply the Dynamic Bayesian Network formalism to this problem. His current research interests still include novel graphical models but, like the entire field of speech technology, his latest work mostly uses Deep Neural Networks (DNNs) of various types. His earlier work on unit selection speech synthesis has been updated to make use of DNNs as a guiding acoustic model. His research passion can be seen through his nearly 200 research publications in the areas of automatic speech recognition and text-to-speech synthesis, and Open Source toolkits, Festival and Merlin.
Prof. King’s research leadership includes being project director and coordinator for two high profile projects funded by the European Commission’s FP7 programme: EMIME (speech-to-speech translation) and Simple4All (unsupervised learning to enable low-resource speech synthesis to scale up to many languages). He was also a co-Investigator on the 5-year Natural Speech Technology Project (NST) – an EPSRC-funded Programme Grant – in which his team made advances in machine learning for speech synthesis, in both the front end and the acoustic model, using Deep Neural Networks. In his latest large project, SCRIPT, he is working with the BBC World Service to deliver speech synthesis of “broadcast quality” across many languages, using techniques developed in Simple4All and NST
In 2015 Professor King was elevated to IEEE Fellow. He has dedicated his time to support scientific organizations, as a co-organizer of the Blizzard series of speech synthesis evaluation and previously for both ISCA and the IEEE Speech and Language Technical Committee. Professor King is currently on the Editorial Board of Computer Speech and Language and is a recognized scientist in his research field.