Teaching Agents When They Fail: End User Development in Goal-Oriented Conversational Agents

  • Toby Jia-Jun LiEmail author
  • Igor Labutov
  • Brad A. Myers
  • Amos Azaria
  • Alexander I. Rudnicky
  • Tom M. Mitchell
Part of the Human–Computer Interaction Series book series (HCIS)


This chapter introduces an end user development (EUD) approach for handling common types of failures encountered by goal-oriented conversational agents. We start with identifying three common sources of failures in human-agent conversations: unknown concepts, out-of-domain tasks and wrong fulfillment means or level of generalization in task execution. To handle these failures, it is useful to enable the end user to program the agent and to “teach” the agent what to do as a fallback strategy. Showing examples for this approach, we walk through our two integrated systems: Sugilite and Lia. Sugilite uses the programming by demonstration (PBD) technique, allowing the user to program the agent by demonstrating new tasks or new means for completing a task using the GUIs of third-party smartphone apps, while Lia learns new tasks from verbal instructions, enabling the user to teach the agent through breaking down the procedure verbally. Lia also enables the user to verbally define unknown concepts used in the commands and adds those concepts into the agent’s ontology. Both Sugilite and Lia can generalize what they have learned from the user across related entities and perform a task with new parameters in a different context.



This work was supported by Yahoo! through CMU’s InMind project and by Samsung GRO 2015.


  1. Allen J, Chambers N, Ferguson G, Galescu L, Jung H, Swift M, Taysom W (2007) PLOW: a collaborative task learning agent. In: Proceedings of the 22nd national conference on artificial intelligence. AAAI Press, Vancouver, British Columbia, Canada, vol 2, pp 1514–1519Google Scholar
  2. Armstrong G (2017a) Helping your baby bot learn to chat like a grown up bot. In: Chatbots MagazineGoogle Scholar
  3. Armstrong G (2017b) The UX of not making your users shout at your Chatbot. In: Chatbots Magazine, Aug 8, 2017.
  4. Azaria A, Hong J (2016) Recommender systems with personality. In: Proceedings of the 10th ACM conference on recommender systems. ACM, New York, pp 207–210Google Scholar
  5. Azaria A, Krishnamurthy J, Mitchell TM (2016) Instructable intelligent personal agent. In: Proceedings of the 30th AAAI conference on artificial intelligence (AAAI)Google Scholar
  6. Barkin J (2016) When bots fail at conversation. In: Being Janis, Aug 19, 2016.
  7. Bohus D, Rudnicky AI (2005a) Constructing accurate beliefs in spoken dialog systems. In: IEEE workshop on automatic speech recognition and understanding, pp 272–277Google Scholar
  8. Bohus D, Rudnicky AI (2005b) Error handling in the RavenClaw dialog management framework. In: Proceedings of the conference on human language technology and empirical methods in natural language processing. Association for Computational Linguistics, pp 225–232Google Scholar
  9. Bohus D, Rudnicky AI (2005c) Sorry, I didn’t catch that!—An investigation of non-understanding errors and recovery strategies. In: 6th SIGdial workshop on discourse and dialogueGoogle Scholar
  10. Chen J-H, Weld DS (2008) Recovering from errors during programming by demonstration. In: Proceedings of the 13th international conference on intelligent user interfaces. ACM, New York, pp 159–168Google Scholar
  11. Cypher A (1993) Eager: programming repetitive tasks by demonstration. In: Watch what I do. MIT Press, pp 205–217Google Scholar
  12. Cypher A, Halbert DC (1993) Watch what I do: programming by demonstration. MIT PressGoogle Scholar
  13. Dzikovska MO, Allen JF, Swift MD (2003) Integrating linguistic and domain knowledge for spoken dialogue systems in multiple domains. In: Proceedings of IJCAI-03 workshop on knowledge and reasoning in practical dialogue systemsGoogle Scholar
  14. Gajos K, Weld DS (2004) SUPPLE: automatically generating user interfaces. In: Proceedings of the 9th international conference on Intelligent user interfaces. ACM, pp 93–100Google Scholar
  15. Gruber TR, Cheyer AJ, Kittlaus D, Guzzoni DR, Brigham CD, Giuli RD, Bastea-Forte M, Saddler HJ (2017) Intelligent automated assistant. U.S. Patent 9,548,050Google Scholar
  16. Koedinger KR, Aleven V, Heffernan N, McLaren B, Hockenberry M (2004) Opening the door to non-programmers: authoring intelligent tutor behavior by demonstration. Intelligent tutoring systems. Springer, Berlin, pp 162–174CrossRefGoogle Scholar
  17. Leshed G, Haber EM, Matthews T, Lau T (2008) CoScripter: automating & sharing how-to knowledge in the enterprise. In: Proceedings of the SIGCHI conference on human factors in computing systems. ACM, New York, pp 1719–1728Google Scholar
  18. Li TJ-J (2017) End user mobile task automation using multimodal programming by demonstration. In: 2017 IEEE symposium on Visual Languages and Human-Centric Computing (VL/HCC), pp 323–324Google Scholar
  19. Li TJ-J, Azaria A, Myers BA (2017a) SUGILITE: Creating multimodal smartphone automation by demonstration. In: Proceedings of the 2017 CHI conference on human factors in computing systems. ACM, New York, pp 6038–6049Google Scholar
  20. Li TJ-J, Li Y, Chen F, Myers BA (2017b) Programming IoT devices by demonstration using mobile apps. End-user development. Springer, Cham, pp 3–17CrossRefGoogle Scholar
  21. Li TJ-J, Myers BA, Azaria A, Labutov I, Rudnicky AI, Mitchell TM (2017c) Designing a conversational interface for a multimodal smartphone programming-by-demonstration agent. In: Conversational UX Design CHI 2017 Workshop. DenverGoogle Scholar
  22. Lieberman H (2001) Your wish is my command: programming by example. Morgan KaufmannGoogle Scholar
  23. Mitman93 Customer review for Pizza Hut: Alexa Skills. Accessed 5 Oct 2017
  24. Myers BA, Ko AJ, Scaffidi C, Oney S, Yoon Y, Chang K, Kery MB, Li TJ-J (2017) Making end user development more natural. New perspectives in end-user development. Springer, Cham, pp 1–22Google Scholar
  25. Pappu A, Rudnicky A (2014) Knowledge acquisition strategies for goal-oriented dialog systems. In: Proceedings of the 15th annual meeting of the Special Interest Group on Discourse and Dialogue (SIGDIAL), pp 194–198Google Scholar
  26. Srivastava S, Labutov I, Mitchell T (2017) Joint concept learning and semantic parsing from natural language explanations. In: Proceedings of the 2017 conference on empirical methods in natural language processing, pp 1527–1536Google Scholar
  27. Witten IH (1993) A predictive calculator. In: Watch what I do. MIT Press, pp 67–76Google Scholar

Copyright information

© Springer International Publishing AG, part of Springer Nature 2018

Authors and Affiliations

  • Toby Jia-Jun Li
    • 1
    Email author
  • Igor Labutov
    • 1
  • Brad A. Myers
    • 1
  • Amos Azaria
    • 2
  • Alexander I. Rudnicky
    • 1
  • Tom M. Mitchell
    • 1
  1. 1.Carnegie Mellon UniversityPittsburghUSA
  2. 2.Ariel UniversityArielIsrael

Personalised recommendations