Speech technology research is showing an increasing focus on clinical and educational problems, with applications in diagnosis as well as in therapy. The most widespread examples are software programs for language learning incorporating speech synthesis and/or speech recognition that are available on the market. However, also in speech therapy and treatment of communication disorders, a number of interesting tools can be envisaged. The objective assessment of pathological voices is a problem in diagnosis and therapy evaluation. Computer-based techniques have the potential of replacing human evaluation by methods which are reproducible in space and time.

Similarly, the assessment of reading skills for early readers and patients with dyslexia can be made more objective through speech recognition techniques. In tandem with speech synthesis, novel scenarios for administrating therapy can be envisaged. Speech technology cannot only make the assessment more objective, it can also reduce cost and increase the intensity as well as the efficacy of the therapy.

The Symposium on Speech Technology for Clinical and Educational Applications provides a state-of-the-art in this challenging research field through contributions of several experts working in this field.

	Registration
Registration is free but obligatory. Send an email with name and affiliation to demessemaeker@esat.kuleuven.be before August 27, 2005 (but the sooner, the better).

	Programme (abstracts below)
09:30h Coffee and registration 10:00h Inventing Virtual Reading Teachers and Virtual Speech Therapists. Demonstration of Three Systems (Ron Cole) 10:45h Quantitative assessment of second language learners' fluency in read and spontaneous speech (Catia Cucchiarini) 11:30h The SPACE project: SPeech Algorithms for Clinical and Educational applications (Hugo Van hamme) 12:00h Lunch break 13:30h A Quantitative Approach to Dysarthria Diagnosis and Treatment (James Carmichael) 14:30h Can speech technology be useful for people with dysarthria? (Helmer Strik) 15:15h Coffee break 15:45h Inventing Virtual Humans that Teach and Conduct Therapy Research and Technical Challenges (Ron Cole) 16:30h Panel discussion 17:00h Closing

	Organising committee
Prof. dr. ir. Hugo Van hamme, ESAT, K.U.Leuven (general chair) Prof. dr. Marc De Bodt, Antwerp University Hospital (local arrangements) Prof. dr. ir. Jean-Pierre Martens, ELIS, U.Gent Prof. dr. ir. Werner Verhelst, ETRO, V.U.Brussel Prof. dr. Pol Ghesquière, Centre for Disability, Special Needs Education and Child Care, K.U.Leuven

	Abstracts
Inventing Virtual Reading Teachers and Virtual Speech Therapists. Demonstration of Three Systems Ron Cole Center for Spoken Language Research University of Colorado Boulder, USA Advances in computing and interface technologies have made possible the development of a new generation of human computer interfaces that enable natural face to face conversational interaction with virtual humans-3D computer characters that speak, emote and gesture much like human experts in specific task domains. One of the key challenges in this emerging field is designing systems that are both engaging and effective; that is, the system immerses the user in a multimedia computer environment with the virtual teacher or therapist to produce targeted learning or clinical outcomes. This talk provides the rationale for and presents demonstration of three systems; one that teaches speech and vocabulary skills to children with profound hearing loss, one that teaches children to read and understand text, and one that teaches adults with Parkinson disease to improve their speech communication abilities. Quantitative assessment of second language learners' fluency in read and spontaneous speech Catia Cucchiarini Language and Speech Radboud University, Nijmegen, The Netherlands In this presentation we report on research aimed at exploring the relationship between objective properties of speech and perceived fluency in read and spontaneous speech, with a view to determining whether such quantitative measures can be used to develop objective fluency tests. Fragments of read speech and spontaneous speech produced by non-native speakers of Dutch were scored for fluency by human raters and were analyzed by means of a continuous speech recognizer to calculate a number of objective measures of speech quality known to be related to perceived fluency. The results show that the objective measures investigated can be employed to predict fluency ratings and therefore to develop objective testing instruments of fluency in read and spontaneous speech. However, the selection of the variables to be used in such tests should be dependent on the specific type of speech material investigated and the specific task carried out by the speaker. The SPACE project: SPeech Algorithms for Clinical and Educational applications Hugo Van hamme ESAT K.U.Leuven, Belgium In the SPACE project, 4 Flemish universities (Leuven, Brussels, Ghent and Antwerp) join forces to explore two particular applications of speech assessment. A first application is a reading tutor for early readers and for children with dyslexia. The second one deals with objective assessment of ill-formed speech, such as produced by people with dysarthria. In this presentation, the goals and approaches of this project are explained. A Quantitative Approach to Dysarthria Diagnosis and Treatment James Carmichael Speech and Hearing Group, Dept. of Computer Science University of Sheffield, United Kingdom Despite the acceptance and prevalence of automated measurement systems in the monitoring and control of pathological conditions, the diagnosis and long-term treatment of dysarthria remains heavily reliant on subjective assessment and its attendant inconsistencies. We report on our efforts to design and implement various objective measurement systems - in particular diagnostic tools and small- vocabulary speech recognition applications capable of correctly interpreting dysarthric speech - to improve the quality of patient care. Can speech technology be useful for people with dysarthria? Helmer Strik Language and Speech Radboud University, Nijmegen, The Netherlands Dysarthria is a speech disorder resulting from dysfunction of the nerves and muscles that control speech production. The intelligibility of dysarthric speech is usually low, especially for unfamiliar listeners. In this presentation we will discuss whether speech technology can be employed to the benefit of people with dysarthria, e.g. to improve their communication possibilities. Inventing Virtual Humans that Teach and Conduct Therapy Research and Technical Challenges Ron Cole Center for Spoken Language Research University of Colorado Boulder, USA There are many decisions that must be made when designing a system that uses virtual human to teach, conduct therapy or sell products on a Web site. These decisions are motivated by the goals of the task and constraints imposed by limitations and costs of computer infrastructure and technologies. This talk describes research that informs the design of virtual humans and virtual human environments in different task domains and challenges posed by capabilities and limitations of state of the art human language and character animation technologies. A design process is described that is used in our system development efforts optimize user experiences and treatment or task outcomes.

	Accomodation
The hotel nearest to the symposium site is: Ter Elst Ter Elststraat 310 2650 Edegem +32 3 450 90 00 www.terelst.be Of course, you can find plenty of hotels in downtown Antwerp.