Human Issues in the use of Pattern Recognition Techniques

Alan Dix

At time of publication: University of York
Currently: Lancaster University

Paper presented at workshop on Neural Networks and Pattern Recognition in Human Computer Interaction at King's Manor, York, October 1991

Download full paper in PDF (159K) .

Full reference:

A. Dix (1992).
Human issues in the use of pattern recognition techniques.
In Neural Networks and Pattern Recognition in Human Computer Interaction Eds. R. Beale and J. Finlay. Ellis Horwood. 429-451.

This was the first paper to introduce Query-by-Browsing. For more on QbB, see Alan's topics page on Query-by-Browsing or try out the online QbB demo.



The purpose of this chapter is to emphasise that when including neural nets or similar techniques in systems with a human component, the technological issues are far easier to address than the attendant human ones. It highlights the need for a thorough theoretical understanding of the behaviour of the computer-based techniques in order to be able to assess the human conse- quences of their use. The chapter focuses on two applications of pattern recognition. One is an innovative example based method of query construction and the other is the more established use of neural nets for routine decision making such as credit vetting. In the latter example the `user' of the system is seen as not just the operative who directly uses the computer, but also the client who is the target of the process. This wide view of human-computer interaction means we have to deal not `just' with the usability of systems but also the entailing ethical and legal responsibilities..

Alan Dix 8/7/2001