Advertisement

GWAI-81 pp 69-78 | Cite as

SWYSS — A Natural Language question-answering System for Scene-Analysis

  • Peter Schefe
  • Bernd Pretschner
Part of the Informatik-Fachberichte book series (INFORMATIK, volume 47)

Abstract

The adequate representation of a universe of discourse in a computational system may be tested by introducing a sensory channel to the “real world”. SWYSS (“Say-what-you-see-system”) is a natural language system connected to a TV-camera gathering snapshots analysed by a scene analysis component. Since there is a continuum of object shapes and other property values, the uncertainty of definitions becomes a main problem of communication. It is described, how representations of natural language inputs and pictorial input are computed and tied together. At present, the system is focussed on the processing of natural language queries containing vague descriptors and imprecise quantifiers. The answers of SWYSS are in elaborate German as well as the questions.

Keywords

Natural Language Deep Structure Natural Scene Scene Analysis Vague Predicate 
These keywords were added by machine and not by the authors. This process is experimental and the keywords may be updated as the learning algorithm improves.

Preview

Unable to display preview. Download preview PDF.

Unable to display preview. Download preview PDF.

References

  1. [1]
    VON HAHN, W., HOEPPNER, W., JAMESON, A., WAHLSTER, W. (1979). The anatomy of the natural language system HAM-RPM. Project Simulation of Natural-Language understanding. Universitaet Hamburg, Germanisches Seminar, Ber. Nr. 12, 1979Google Scholar
  2. [2]
    HANSSMANN, K.-J. (1980). Sprachliche Bildinterpretation fuer ein Frage-Antwort-System. Universitaet Hamburg, Fachbereich Informatik: Mitteilung Nr. 74, IFI-HH-M-74/80, 1980.Google Scholar
  3. [3]
    JAMESON, A., HOEPPNER, W., WAHLSTER, W. (1980). The natural language system Ham-Rpm as a hotel manager: some representation prerequisites. Project Simulation of Natural-Language understanding. Universitaet Hamburg, Germanisches Seminar, Ber. Nr. 17, 1980Google Scholar
  4. [4]
    LAKOFF, G. (1975). Hedges: A study in meaning criteria and the logic of fuzzy concepts. In Hockney, D., Harper, W., Freed, B. (Eds.), Contemporal research in philosophical logic and linguistic semantics. Reidel: Dordrecht, Boston, 1975Google Scholar
  5. [5]
    LEFAIVRE, R.A. (1974). The representation of fuzzy knowledge. New Brunswick: Rutgers University, Dept. Comp. Sci., DCS-TR-33., 1974Google Scholar
  6. [6]
    LEFAIVRE, R.A. (1977). Fuzzy reference manual. New Brunswick: Rutgers University, Dept. Comp. Sci., 1977Google Scholar
  7. [7]
    MITTELSTEIN, M. , NEBEL, B. , PRETSCHNER, B. , SCHEFE, P. (1976). Hasy — ein Programm zur syntaktischen Analyse natuerlicher Sprachen. Hamburg: Universitaet, Fachbereich Informatik, Mitteilung Nr. 36, IFI-HH-M-36/76, 1976Google Scholar
  8. [8]
    NEUMANN, B. (1978). Interpretation of imperfect object contours for identification and tracking. IJCPR-78, Kyoto, Japan, 691–693, (1978)Google Scholar
  9. [9]
    PEPPER, S., PRYTULAK, L.S. (1974). Sometimes frequently means seldom: Context effects in the interpretation of quantitative expressions. Journal of Research in Personality, 8, 95–101, (1974)CrossRefGoogle Scholar
  10. [10]
    PRETSCHNER, B. (1980). Die Behandlung natuerlichsprachlicher Quantifizierungen in Frage-Antwort-Systemen. Universitaet Hamburg, Fachbereich Informatik, Diplomarbeit, 1980Google Scholar
  11. [11]
    SCHEFE, P. (1977). Theoretische und hochschuldidaktische Aspekte der Integration von Phrasenstrukturgrammatik und Dependenzgrammatik. Linguistik und Didaktik 29, 36–50, (1977)Google Scholar
  12. [12]
    SCHEFE, P. (1980) . On foundations of reasoning with uncertain facts and vague concepts. Int. J. Man-Machine Studies 12, 35–62, (1980)MathSciNetMATHCrossRefGoogle Scholar
  13. [13]
    WALTZ, D.L. (1979). Visual analog representation for natural language understanding. IJCAI 1979, 926–934.Google Scholar

Copyright information

© Springer-Verlag Berlin Heidelberg 1981

Authors and Affiliations

  • Peter Schefe
  • Bernd Pretschner

There are no affiliations available

Personalised recommendations