Cooperative behaviour in the FIDO system

摘要

One of the demands that are put on a Natural Language interface is that it behaves in a cooperative way, that is, roughly speaking, in the most useful and natural way.This problem has been addressed in the context of the FIDO system (Flexible Interface for Database Operations), and in particular two points have been examined: the generation of answers that take into account the expectations and presuppositions the user showed in his/her question; the correction of some semantic mistakes detected during the analysis of the input question.In particular, as far as the former point is concerned, two issues have been addressed: the correction of user's misconceptions regarding the stored data; the generation of answers as suited as possible to what the user asked and to the way he/she phrased his/her question.