Elamur Research⌕
☰
home / papers / 1995 / Utilizing Statistical Dialogue Act Processing in Verbmobil

Utilizing Statistical Dialogue Act Processing in Verbmobil

Summary

The paper addresses dialogue act prediction in the dialogue component of VERBMOBIL, a speech-to-speech translation system for appointment scheduling. Other modules need to know which dialogue act is likely to come next. The semantic evaluation component uses predictions to narrow its choices, and the keyword spotter, which handles only a tiny vocabulary, needs keywords chosen for the expected act. The hand-built dialogue network is too permissive for this task: it allows about 5 follow-ups per state on average, and about 10 once deviation acts are included. The authors therefore annotated more than 200 dialogues with 17 dialogue acts and trained an n-gram model of act sequences, smoothed with deleted interpolation, an approach borrowed from speech recognition. With three guesses, the model's hit rate ranges from roughly 70% to 81%, depending on whether deviation acts are included. Hit rates also vary widely between individual dialogues. The model also drives a repair strategy in the plan recognizer. When an incoming act does not fit the dialogue model, the recognizer uses probabilities to propose an extra reading of an earlier utterance, which makes processing of unrestricted dialogues more robust.

Key points

  • Dialogue act inventory: 17 domain-independent dialogue acts, including five "deviation" acts that may occur anywhere in a dialogue (e.g. DELIBERATE, CLARIFY_QUERY, CLARIFY_ANSWER). Admissible sequences are defined by transition networks.
  • Experimental setup: training used 41 German dialogues (2472 acts). TS1 and TS2 were tested on 81 German dialogues, with TS1 skipping deviation acts. TS3 covered all annotated dialogues (7197 acts).
  • Hit rates for 1/2/3 predictions:
    • TS1: 44.24 / 66.47 / 81.46%
    • TS2: 37.47 / 56.50 / 69.52%
    • TS3: 40.28 / 59.62 / 71.93%
  • Comparison and variation: the rates are lower than those reported by Nagata and Morimoto (61.7 / 77.5 / 85.1%), whose act set has no deviation acts. Across 47 sampled dialogues, the three-prediction hit rate ranged from about 53% to 93%.
  • Statistical repair: candidate insertions are ranked by the product of their transition probabilities. Each candidate is then checked against a corpus-derived list of acts that can share one utterance; for example, INIT can also be read as SUGGEST, but SUGGEST and REJECT never combine. When no statistical fix exists, plan-based repair operators take over.

Our summary of the paper, in our own words — not its text. The work belongs to its authors; read it at the links above.

Fields: Linguistics & languages

Topics: Speech act theory

Cited by

13 papers of our library cite it (of the 140673 whose bibliographies we read); 1 of them have a page here, 12 not yet.

Cite

The paper itself, not this page. BibTeX from the library's record of the work. DOI 10.3115/981658.981674 · arXiv cmp-lg/9505013

@inproceedings{reithinger1995utilizing,
  title = {{Utilizing Statistical Dialogue Act Processing in Verbmobil}},
  author = {Reithinger, Norbert and Maier, Elisabeth},
  booktitle = {33rd Annual Meeting of the Association for Computational Linguistics},
  year = {1995},
  doi = {10.3115/981658.981674},
  eprint = {cmp-lg/9505013},
  archivePrefix = {arXiv},
  url = {https://doi.org/10.3115/981658.981674}
}

References

22 references, 1 of them with a card of ours. ↗ leads to the work, «find» searches for one we have no id for yet.

  1. Jan Alexandersson, Elisabeth Maier, and Norbert Reithinger. 1995. A Robust and Efficient Three-Layered Dialog Component for a Speech-to-Speech Translation System. In Proceedings of the 7th Conference of the European Chapter of the A CL (EA CL-95), Dublin, Ireland. find
  2. John Austin. 1962. How to do things with words. find
  3. Eric Bilange. 1991. A task independent oral dialogue model. In Proceedings of the Fifth Conference of the European Chapter of the Association for Computational Linguistics (EACL-91), pages 83-88, Berlin, Germany. find
  4. Johan Bos, Elsbeth Mastenbroek, Scott McGlashan, Sebastian Millies, and Manfred Pinkal. 1994. The Verbmobil Semantic Formalismus. Technical report, Computerlinguistik, Universit~it des Saarlandes, Saarbriicken. find
  5. Philip R. Cohen, Jerry Morgan, and Martha E. Pollack, editors. 1990. Intentions in Communication. MIT Press, Cambridge, MA. find
  6. Elizabeth A. Hinkelman and Stephen P. Spackman. 1994. Communicating with Multiple Agents. In Proceedings of the 15th International Conference on Computational Linguistics (COLING 94), August 5-9, 1994, Kyoto, Japan, volume 2, pages 1191-1197. find
  7. Susanne Jekat, Alexandra Klein, Elisabeth Maier, Ilona Maleck, Marion Mast, and J. Joachim Quantz. 1995. Dialogue Acts in Verbmobil. Verbmobil Report Nr. 65, Universit~it Hamburg, DFKI Saarbriicken, Universit~it Erlangen, TU Berlin. find
  8. Fred Jellinek. 1990. Self-Organized Language Modeling for Speech Recognition. In A. Waibel and K.-F. Lee, editors, Readings in Speech Recognition, pages 450-506. Morgan Kaufmann. find
  9. Martin Kay, Jean Mark Gawron, and Peter Norvig. 1994. Verbmobil. A Translation System for Face-to-Face Dialog. Chicago University Press. CSLI Lecture Notes, Vol. 33. find
  10. Roland Kuhn. 1993. Keyword Classification Trees for Speech Understanding Systems. Ph.D. thesis, School of Computer Science, McGill University, Montreal. find
  11. Elisabeth Maier and Scott McGlashan. 1994. Semantic and Dialogue Processing in the VERB-MOBIL Spoken Dialogue Translation System. In Heinrich Niemann, Renato de Mori, and Gerhard Hanrieder, editors, Progress and Prospects of Speech Research and Technology, volume 1, pages 270-273, Miinchen. find
  12. Elisabeth Maier. 1 9 9 4 . Dialogmodellierung in VERBMOBIL - Pestlegung der Sprechhandlun-gen fiir den Demonstrator. Technical Report Verbmobil Memo Nr. 31, DFKI Saarbriicken. find
  13. Marion Mast, Ralf Kompe, Franz Kummert, Heinrich Niemann, and Elmar NSth. 1992. The Dialogue Modul of the Speech Recognition and Dialog System EVAR. In Proceedings of International Conference on Spoken Language Processing (ICSLP'92), volume 2, pages 1573-1576. find
  14. Marion Mast. 1995. SchliisselwSrter zur Detek-tion yon Diskontinuit~iten und Sprechhandlun-gen. Technical Report Verbmobil Memo Nr. 57, Friedrich-Alexander-Universit~it, Erlangen-Niirnberg. find
  15. Mark T. Maybury. 1 9 9 1 . Planning Multisentential English Text Using Communicative Acts. Ph.D. thesis, University of Cambridge, Camb-dridge, GB. find
  16. Johanna Moore. 1994. Participating in Explanatory Dialogues. The MIT Press. find
  17. Masaaki Nagata and Tsuyoshi Morimoto. 1993. An experimental statistical dialogue model to predict the Speech Act Type of the next utterance. In Proceedings of the International Symposium on Spoken Dialogue (ISSD-93), pages 83-86, Waseda University, Tokyo, Japan. find
  18. Gerhard Th. Niedermair. 1992. Linguistic Modelling in the Context of Oral Dialogue. In Proceedings of International Conference on Spoken Language Processing (ICSLP'92}, volume 1, pages 635-638, Banff, Canada. find
  19. Norbert Reithinger. 1995. Some Experiments in Speech Act Prediction. In A A A I 95 Spring Symposium on Empirical Methods in Discourse Interpretation and Generation, Stanford University. find
  20. John R. Searle. 1969. Speech Acts. Cambridge: University Press. find our card
  21. Marc Vilain. 1990. Getting Serious about Parsing Plans: a Grammatical Analysis of Plan Recognition. In Proceedings of AAAI-90, pages 190-197. find
  22. Wolfgang Wahlster. 1993. Verbmobil-Translation of Pa~e-to-Pace Dialogs. Technical report, German Research Centre for Artificial Intelligence (DFKI). In Proceedings of MT Summit IV, Kobe, Japan. find

↑ back to top

This page in Markdown → https://md.research.elamur.ai/free/en/papers/reithinger-1995-utilizing-statistical-dialogue-act-processing/