Zusammenfassung
Bei der Verarbeitung kontinuierlich gesprochener Sprache entsteht eine sehr große Anzahl von Worthypothesen. Wir stellen einen Ansatz vor, der domänenspezifisches Wissen ausnutzt, um semantisch plausible Worthypothesen zu bevorzugen. Anwendung ist das Diktieren radiologischer Befundungstexte. Für die jeweils nächste Äußerung werden Erwartungen generiert, die aus einem Modell der radiologischen Befundung abgeleitet werden. Die Kontrollstruktur kann, im Gegensatz zu traditionellen sequentiellen Architekturen, als zyklisch und erwartungsgesteuert beschrieben werden. Mehrstufige Erwartungen ermöglichen im Falle eines Scheiterns die nochmalige Verarbeitung einer Äußerung. Erwartungen äußern sich in einer Erhöhung des Konfidenzwertes von Worthypothesen.
Abstract
Continuous speech processing has to face the problem of dealing with a huge number of word hypotheses. We present an approach that uses domain-specific knowledge to prefer semantically plausible word hypotheses. The approach is applied to the dictation of radiological reports. Expectations, which are based on a model of radiological reporting, are generated for the next incoming utterance. The control structure can be characterized as cyclic and expectation-driven in contrast to a traditional sequential architecture. Multi-layered expectations allow the reprocessing of an utterance in case of failure. Expectations are realized by increasing the confidence value of word hypotheses.
This is a preview of subscription content, log in via an institution.
Buying options
Tax calculation will be finalised at checkout
Purchases are for personal use only
Learn about institutional subscriptionsPreview
Unable to display preview. Download preview PDF.
References
Baud, R. H., Rassinoux, A.-M. and Scherrer, J.-R. Knowledge Representation of Discharge Summaries. In: AIME 91, Third Conference on Artificial Intelligence in Medicine, Maastricht, June 1991, pp. 173–182.
Billi, R., Buttafava, P., De Stefani, P., Gamba, M., and Voltolini, D. Computer-Aided, Voice-based, Medical Report Preparation: An Application To Radiology. In EUROSPEECH-91, pp. 961,1991.
Cerf-Danon, H., DeGennaro, S., Ferreti, M., Gonzales, J., and Keppel, E. TANGORA — A Large Vocabulary Speech Recognition System For Five Languages. In: EUROSPEECH-91, pp. 183–192.
DeJong, G. Prediction and Substantiation: A New Approach to Natural Language Processing. In: Cognitive Science, 3, pp. 251–273, 1979.
Ehrlich, U. Multilevel semantic analysis in an automatic speech understanding and dialog system. In: Proc. of the Third Conf. of the European Chapter ofthe Association for Computational Linguistics, Copenhagen, 1987.
Hayes, P.J., Hauptmann, A.G., Carbonell, J.G., and Tomita, M. Parsing spoken language: a semantic caseframe approach. In: Proceedings of COLING-86, Bonn, pp. 587–592.
Kingsland, L. C. (ed) The 13th Arnual Symposium on Computer Applications in Medical Care. IEEE Computer Society Press, November 1989.
Kurzweil, R., and Steingart, R. The Application of Large Vocabulary Speech Recognition and Knowledge Engineering to the Creation of Written Documents. In Speech Tech ’88, pp. 56 – 60, 1988.
Lehrberger, J. Sublanguage Analysis. In: Grishman, R., and Kittredge, R. (eds) Analyzing Language in Restricted Domains: Sublanguage Description and Processing, pp. 19 – 38, Lawrence Erlbaum, Hilldale (NJ), 1986.
Möller, T.B. Röntgennormalbefunde. Thieme, Stuttgart, 1987.
Ney, H. and Billi, R. Prototype systems for laige-vocabulary speech recognition: POLYGLOTT and SPICOS. In: Proc. of EUROSPEECH-91, pp. 193–200, 1991.
Pyka, C. Management of Hypotheses in an Integrated Speech-Language Architecture. In: ECAI-92, Wien, 1992.
Ranum, D.L. Knowledge Based Understanding of Radiology Text. In The 12th Arnual Symposium on Computer Applications in Medical Care, Greenes, R.A., IEEE Computer Society Press, November 1988, pp. 141–145.
Sager, N., Friedman, C., and Lyman, M.S. Medical Language Processing: Computer Management of Narrative Data, Addison-Wesley, Reading, MA, 1987.
Scherrer, J. R., Côté, R. A., and Mandil, S. D. (eds) Computerized Natural Medical Language Processing for Knowledge Representation. North Holland, Amsterdam, 1989.
Schröder, M. Ein semantisch-gesteuerter Bottom-Up-Parser in Prolog. Tech. Rept. FBI-HH-M-195/91, Mitteilung, University of Hamburg, Computer Science Department, May, 1991.
Schröder, M. Knowledge-based Processing of Medical Language: A Language Engineering Approach. In: GWAI-92, 16. Fachtagung für Künstliche Intelligenz, Springer-Verlag, 1992a.
Schröder, M. Knowledge-based Analysis of Radiological Reports Using Conceptual Graphs. In: Proceedings of the 7th Annual Workshop on Conceptual Graphs, Las Cruces, New Mexico, July 1992b.
Sowa, J.F. Conceptual Structures: Information Processing in Mind and Machine, Addison-Wesley, 1984.
Young, S.R., Hauptmann, A.G., Ward, W.H., Smith, E.T., and Werner, P. High Level Knowledge Sources in Usable Speech Recognition Systems. In: Communications of the ACM, Vol. 32, No. 2, February 1989a.
Young, S.R., Ward, W.H., and Hauptmann, A.G. Layering Predictions: Flexible Use of Dialog Expectation in Speech Recognition. In: IJCAI-89, pp. 1543–1549, 1989b.
Author information
Authors and Affiliations
Editor information
Editors and Affiliations
Rights and permissions
Copyright information
© 1992 Springer-Verlag Berlin Heidelberg
About this paper
Cite this paper
Schröder, M. (1992). Supporting Speech Processing By Expectations: A Conceptual Model Of Radiological Reports To Guide The Selection Of Word Hypotheses. In: Görz, G. (eds) Konvens 92. Informatik aktuell. Springer, Berlin, Heidelberg. https://doi.org/10.1007/978-3-642-77809-4_13
Download citation
DOI: https://doi.org/10.1007/978-3-642-77809-4_13
Publisher Name: Springer, Berlin, Heidelberg
Print ISBN: 978-3-540-55959-7
Online ISBN: 978-3-642-77809-4
eBook Packages: Springer Book Archive