Zoltán Baracskai - Viktor Dörfler - Jolán Velencei
The ES Could Probably Know More - But Man Would Not Make Better Business Decisions
(Plenary Presentation)



Details
Conference: VIPSI 2007 Slovenia
Theme: Quality of Life
Presenter: Viktor Dörfler (plenary chair)
Date: 8-11 October 2007
Place: Bled, Slovenia
ISBN: 86-7466-117-3

presentation    full text    RIS export
Abstract
We know for a long time that expert system (ES) must not know in a different way than how man knows. We would not expect men to adapt their thinking processes to the machine, would we? Evolution would not let peoples' minds adopt criteria weighting or fuzzy logic. This is why we developed our Doctus knowledge-based expert system based on symbolic logic – that is the most understandable artificial knowledge representation to the decision taker. However, this is all history now...

When Doctus has stepped into the adulthood, we sat down to draw the line and review our experiences. In the 18 years with Doctus, we had around 140 consultancy projects, the vast majority of them for top executives. We had summarized our successes and failures on more than few occasions over these year, we wanted to do something different this time: we wanted to overview which are the features of our ES that decision takers like and make use of and to decide where to go next with the development. This paper is the story of the lessons learned.

We now understand that the ES must not have the features that the decision taker does not want to use. Decision taking is far too serious issue for men to yield it to a computer. Perhaps we could develop the m-Doctus, and perhaps it would work as we imagine it, and perhaps we would be could achieve this within the programmability of the mobile phones. But that would be again something that only we thought to be good. Previously we have spent large amounts of money to develop a data mining enabled Doctus – it was our figment that the decision taker is not interested in. Advertising Doctus is capacity to make the decisions transparent to the decision taker, and if (s)he wants also to others, was another useless idea we implemented – the decision taker does not want to get into the fine details of her/his thinking process.

Now we perhaps understand decision takers better – maybe it was worth observing them for nearly two decades. Decision takers, and we mean only the most experienced ones, want to be able to play around a bit with a particular decision and, eventually, they want some help in delegating the decisions they are bored with.