An abstract model of man-machine interaction based on concepts from NL dialog processing

  • Authors:
  • Helmut Horacek

  • Affiliations:
  • Fachrichtung Informatik, Universität des Saarlandes, Saarbrücken, Germany

  • Venue:
  • NLDB'06 Proceedings of the 11th international conference on Applications of Natural Language to Information Systems
  • Year:
  • 2006

Quantified Score

Hi-index 0.00

Visualization

Abstract

Developing human-computer (man-machine) interaction system components is time consuming, error-prone, and it is hard to produce high-quality interfaces with good usability. A fundamental reason for this unsatisfactory situation lies in the way the development process is organized, which widely works on a syntactic level in terms of sets of widgets, rather than on a semantic level that captures the task-relevant flow of information. As a step towards the development of human-computer interaction system components on principled grounds, we present an abstract model of human-computer interaction based on concepts borrowed from natural language processing, prominently operational models of human dialogs. Major ingredients of this model are speech act specifications and information state-based update rules capturing the effects of these speech acts, adapted to particularities of human-computer communication. The model is a crucial prerequisite for automatically building man-machine interfaces on the basis of high-level specifications.