as a single stream and with an attendance of approximately 70 people. Each part was headed by a keynote speaker, two from the USA and two from the UK. Geoffrey Hinton, Director of the Gatsby Computational Neuroscience Unit, at University College London led off with some front-line thinking on training many small HMMs and which he saw as highly relevant for speech processing. On the first afternoon we heard Allen Gorin from AT&T Research Labs expounding on semantic information processing of spoken language, the next morning

Author: Bruce Millar, President of ASSTA.

Some while ago, I was invited to be a member of the Scientific Committee of a Workshop on Innovative Speech Processing to be held in Stratford-upon-Avon in the UK in April 2001. As innovative speech processing has been both a long-term and current interest and Stratford-upon-Avon in the spring sounded very pleasant, I decided not only to accept but to consider a contribution to the workshop. What I did not realise was that this workshop, with the delightful acronym - WISP, was in fact a revamp of the longstanding national conference of the UK speech scientists but with an international edge to it - hence the original invitation.  The workshop therefore held a dual interest for me as its structure and atmosphere were the UK analogues of what we have been aiming for with our SST series since 1986.  The two day workshop was divided into four parts all operating

"Geoffrey Hinton... ...led off with some front-line thinking on training many small HMMs..."

Chris Darwin, from the Department of Experimental Psychology at the University of Sussex spoke about auditory grouping and attention to speech, then on the final afternoon, Deb Roy, of the Cognitive Machines Group of the MIT Media Lab, spoke on the grounding of language in a multisensory context. Most of the other papers came from the UK with a sprinkling of collaborators from other European countries aswell as 4 papers from clearly commercial

(Continued on page 5)