Complementary computing: policies for transferring callers from dialog systems to human receptionists

Original Paper

Abstract

We describe a study of the use of decision-theoretic policies for optimally joining human and automated problem-solving efforts. We focus specifically on the challenge of determining when it is best to transfer callers from an automated dialog system to human receptionists. We demonstrate the sensitivities of transfer actions to both the inferred competency of the spoken-dialog models and the current sensed load on human receptionists. The policies draw upon probabilistic models constructed via machine learning from cases that were logged by a call routing service deployed at our organization. We describe the learning of models that predict outcomes and interaction times and show how these models can be used to generate expected-utility policies that identify when it is best to transfer callers to human operators. We explore the behavior of the policies with simulations constructed from real-world call data.

Keywords

Spoken dialog systems Machine learning Human-machine systems Probabilistic user modeling Complementary computing 

Preview

Unable to display preview. Download preview PDF.

Unable to display preview. Download preview PDF.

References

  1. 1.
    Biermann, A.W., Inouye, R.B., McKenzie, A.: Methodologies for automated telephone answering. In: Proceedings of ISMIS, Saratoga Springs, NY, pp. 1–13 (2005)Google Scholar
  2. 2.
    Boger, J., Poupart, P., Hoey, J., Boutilier, C., Fernie, G., Mihailidis, A.: A decision-theoretic approach to task assistance for persons with Dementia. In: Proceedings of IJCAI, Edinburgh, Scotland, pp. 1293–1299 (2005)Google Scholar
  3. 3.
    Chickering, D.M., Heckerman, D., Meek, C.A: Bayesian approach to learning Bayesian networks with local structure. In: Proceedings of UAI, Providence, RI, Morgan Kaufmann, pp. 80–89 (1997)Google Scholar
  4. 4.
    D’Agostino, D.: Weak speech recognition leaves customers cold, CIO Insight, Ziff-Davis. http://www.cioinsight.com/print_article2/0,1217,a=168124,00.asp. cited 29 Dec. 2005Google Scholar
  5. 5.
    Friedman, N., Goldszmidt, M.: Learning Bayesian networks with local structure. In: Proceedings of UAI, Portland, OR, Morgan Kaufmann, pp. 252–262 (1996)Google Scholar
  6. 6.
    Hearst, M.A., Trends and Controversies.: Mixed-initiative interaction IEEE Intelligent Systems vol. 14(5), pp. 14–23 IEEE Computer Society, Silver Spring, MD (1999)Google Scholar
  7. 7.
    Horvitz, E.: Principles of mixed-initiative user interfaces. In: Proceedings of the SIGCHI, Pittsburgh, PA, ACM Press, pp. 159–166 (1999)Google Scholar
  8. 8.
    Kaelbling L.P., Littman M.L. and Moore A.P. (1996). Reinforcement learning: a survey. J. Artif. Intell. Res. 4: 237–285 Google Scholar
  9. 9.
    Langkilde, I., Walker, M., Wright, J., Gorin, A., Litman, D.: Automatic prediction of problematic human-computer dialogs in how may I help you? In: Proceedings of the IEEE Workshop on Automatic Speech Recognition and Understanding, Keystone, CO, pp. 369–372 (1999)Google Scholar
  10. 10.
    Litman, D., Walker, M., Kearns, M.: Automatic detection of poor speech recognition at the dialog level. In: Proceedings of ACL 1999, College Park, MD, pp. 309–316 (1999)Google Scholar
  11. 11.
    Litman D. and Pan S. (2002). Designing and evaluating an adaptive spoken dialog system. User Model. User-Adapt. Interact. 12(2/3): 111–137 MATHCrossRefGoogle Scholar
  12. 12.
    Paek, T., Horvitz, E.: Optimizing call routing by integrating queuing models with spoken dialog models. In: Proceedings of HLT/NAACL 2004, Boston, MA, pp. 41–48 (2004)Google Scholar
  13. 13.
    Suhm, B., Bers, J., McCarthy, D., Freeman, B., Getty, D., Godfrey, K., Peterson, P.: A comparative study of speech in the call center: natural language call routing vs. Touch-tone menus. In: Proceedings of SIGCHI, MN, pp. 283–290 (2002)Google Scholar
  14. 14.
    Tatchell, G.R.: Problems with the existing telephony customer interface: The pending eclipse of touch-tone and dial-tone. In: Proceedings of SIGCHI, Vancouver, BC, pp. 242–243 (1996)Google Scholar
  15. 15.
    Walker, M., Langkilde, I., Wright, J., Gorin, A., Litman, D.: Learning to predict problematic situations in an automated dialog system: experiments with HMIHY? In: Proceedings of ANLP-NAACL-2000, Seattle, WA, pp. 210–217 (2000)Google Scholar
  16. 16.
    Walker M., Langkilde-Geary I., Hastie H., Wright J. and Gorin A. (2002). Automatically training a problematic dialog predictor for the HMIHY spoken dialog system. J. Artif. Intell. Res. 16: 293–319 MATHGoogle Scholar
  17. 17.
    Xu, W., Rudnicky, A.: Task-based dialog management using an agenda. In: Proceedings of ANLP-NAACL 2000 Workshop on Conversational Systems, Seattle, WA, pp. 42–47 (2000)Google Scholar

Copyright information

© Springer Science+Business Media B.V. 2007

Authors and Affiliations

  1. 1.Microsoft ResearchRedmondUSA

Personalised recommendations