[comp.ai.neural-nets] Workshop announcement - York

janet@minster.york.ac.uk (06/17/91)

		Human Interface Club
Models and Methods for Interactive Systems Design Special Interest Group
			and
	  British Neural Networks Society

		   Workshop on
	Pattern Recognition and Neural Networks
	    in Human-Computer Interaction

		University of York, UK
		   19th August 1991

The ability to recognise and manipulate patterns is a prerequisite of a
number of different applications in HCI.  Some, such as user or student
modelling and plan recognition, rely on identifying recurrent patterns in
user behaviour in order to establish, for example, a user's skill level,
expertise, or current activity.  Protocol analysis, whether of audio or
video tape or of usage logs, seeks to recognise patterns in user behaviour
to evaluate a system or identify user difficulties.  Security applications
rely on patterns of activity to verify the identity of specific users or
user groups.  Other applications use patterns as a means of providing
novel input and output mechanisms; speech recognition and synthesis, image
recognition, and the use of gestural input all require the ability to
distinguish the particular sound, image or gesture from those possible.

Traditionally such applications have not been approached specifically as
pattern recognition problems, and diverse means have been used to tackle
them, including the use of knowledge-based methods, statistics, and manual
analysis. However more recently novel pattern recognition methodologies
such as neural networks and other example-based learning approaches have
emerged.  Research in applying such methods to HCI concerns is at an early
stage but has indicated their potential in areas such as user and student
modelling, protocol analysis, and gestural input.

The workshop will be chaired by Russell Beale and Janet Finlay of the
University of York and will include presentations by researchers in the
area from both Europe and North America, as well as opportunities for
discussion. A poster presentation of relevant research at York will be on
display throughout the day.

Papers:
=======

Tom Carey, D Calvert, D Chiu, D Edwards, A Halabieh, D Stacey (University
of Guelph/IBM Canada): Pattern Analysis of HCI Usage Data

Daniel Crow and Barbara Smith (University of Leeds): DB_Habits: comparing
pattern recognition and knowledge-based approaches to the computer
analysis of user-computer interactions

Leticia Villegas and Ray Eberts (Purdue University): Implementing a neural
network for a cognitive text-editing task

Jun Hong and Michael McTear (University of Ulster at Jordanstown):
Creating an action hierarchy to acquire plan inference knowledge

F Kozato and P de Wilde (Imperial College, London): How neural networks
reduce HCI overhead of problem solvers

David Reid (University of Liverpool): Some practical applications of neural
networks in a hypertext system

S Hanlon and R Boyle (University of Leeds): Syntactic knowledge in word
level text recognition

Ray Frank and Caroline Lyon (Hatfield Polytechnic): Improving a speech
interface by intgrating a neural net that processes syntactic information

Michael Oakes and Malcolm Taylor (University of Liverpool): Clustering of
Thesaurus terms using Adaptive Resonance Theory, Fuzzy Cognitive Maps and
traditional string matching techniques

Paul Booth (Manchester Polytechnic): Modelling misunderstandings using
artificial neural networks

Dimitrios Bairaktaris (University of Edinburgh): Discovering temporal
structure using Hebbian learning

Simon Grant: Rule Induction as a tool for comparing representations of
human task performance

Posters:
========

Janet Finlay and Russell Beale
The role of representation and classification in static and dynamic
user modelling

Russell Beale and Alistair Edwards
Gesture recognition with neural networks

Gregory Abowd and Russell Beale
A framework for understanding interaction

Alan Dix
To be announced

A book including full papers from this workshop and one held at CHI'91 will
be published by Ellis Horwood at the beginning of 1992. These can be ordered
at a special pre-publication workshop price of 40 pounds (retail price will
be 45 pounds) as part of an inclusive package for workshop attendees. The
book will be forwarded on publication.
Attendance at the workshop will cost 65 pounds (to include lunch and the book)
or 25 pounds (to include lunch only).  If you wish to attend, please fill in 
and return the attached form. Further details, including travel directions 
and a programme, will be sent on receipt of your registration.

Please direct any enquiries to:

Dr. Janet Finlay
Human Computer Interaction Group
Department of Computer Science
University of York
Heslington
York, YO1 5DD
UK
Fax: +44 904 432767
JANET:  janet-york@uk.ac.york.minster
Telephone:  +44 904 432747

______________________________________________________________________________

Registration for Workshop on Pattern Recognition and Neural Networks in
Human Computer Interaction

University of York, 19 August, 1991.

Name:

Department:

Institution:

Address:

Postcode:                                      Telephone:

Email address:

I enclose registration of 65 pounds (to include proceedings and lunch).
I enclose registration of 25 pounds (to include lunch).
(please delete as appropriate)

Please indicate any special needs of which we should be aware.

Please return this form to Janet Finlay at the address above, with payment
made out to ``The University of York''.