Supporting multiple user types with a multimodal dialog agent

Michael Groble; Will Thompson

doi:10.1109/WIIATW.2007.4427600

Supporting multiple user types with a multimodal dialog agent

Michael Groble^*, Will Thompson

^*Corresponding author for this work

Medicine, General Medicine Division

Research output: Chapter in Book/Report/Conference proceeding › Conference contribution

Abstract

Recent research has addressed the problem of formulating a dialog agent as a partially observable Markov decision process (POMDP), and learning a dialog policy that is optimal given the particular characteristics of the transition, observation and reward functions of the POMDP. This paper addresses the problem of trying to learn a small set of dialog agent policies that provide near-optimal behavior over a wide range of variations in POMDPs, reflecting different user preferences and environment characteristics. We show for a very simple dialog, we can cover a large number of simulated users to within 10% of their optimal return using fewer than 5% of the individual optimal policies.

Original language	English (US)
Title of host publication	Proceedings - 2007 IEEE/WIC/ACM International Conference on Web Intelligence and Intelligent Agent Technology - Workshops, WI-IAT Workshops 2007
Pages	329-332
Number of pages	4
DOIs	https://doi.org/10.1109/WIIATW.2007.4427600
State	Published - 2007
Event	2007 IEEE/WIC/ACM International Conference on Web Intelligence and Intelligent Agent Technology - Workshops, WI-IAT Workshops 2007 - Silicon Valley, CA, United States Duration: Nov 2 2007 → Nov 5 2007

Publication series

Name	Proceedings - 2007 IEEE/WIC/ACM International Conference on Web Intelligence and Intelligent Agent Technology - Workshops, WI-IAT Workshops 2007

Other

Other	2007 IEEE/WIC/ACM International Conference on Web Intelligence and Intelligent Agent Technology - Workshops, WI-IAT Workshops 2007
Country/Territory	United States
City	Silicon Valley, CA
Period	11/2/07 → 11/5/07

ASJC Scopus subject areas

Artificial Intelligence
Computer Networks and Communications

Access to Document

10.1109/WIIATW.2007.4427600

Cite this

Groble, M., & Thompson, W. (2007). Supporting multiple user types with a multimodal dialog agent. In Proceedings - 2007 IEEE/WIC/ACM International Conference on Web Intelligence and Intelligent Agent Technology - Workshops, WI-IAT Workshops 2007 (pp. 329-332). Article 4427600 (Proceedings - 2007 IEEE/WIC/ACM International Conference on Web Intelligence and Intelligent Agent Technology - Workshops, WI-IAT Workshops 2007). https://doi.org/10.1109/WIIATW.2007.4427600

Groble, Michael ; Thompson, Will. / Supporting multiple user types with a multimodal dialog agent. Proceedings - 2007 IEEE/WIC/ACM International Conference on Web Intelligence and Intelligent Agent Technology - Workshops, WI-IAT Workshops 2007. 2007. pp. 329-332 (Proceedings - 2007 IEEE/WIC/ACM International Conference on Web Intelligence and Intelligent Agent Technology - Workshops, WI-IAT Workshops 2007).

@inproceedings{f8dc42da10114e7d9524094e3f2da165,

title = "Supporting multiple user types with a multimodal dialog agent",

abstract = "Recent research has addressed the problem of formulating a dialog agent as a partially observable Markov decision process (POMDP), and learning a dialog policy that is optimal given the particular characteristics of the transition, observation and reward functions of the POMDP. This paper addresses the problem of trying to learn a small set of dialog agent policies that provide near-optimal behavior over a wide range of variations in POMDPs, reflecting different user preferences and environment characteristics. We show for a very simple dialog, we can cover a large number of simulated users to within 10% of their optimal return using fewer than 5% of the individual optimal policies.",

author = "Michael Groble and Will Thompson",

year = "2007",

doi = "10.1109/WIIATW.2007.4427600",

language = "English (US)",

isbn = "0769530281",

series = "Proceedings - 2007 IEEE/WIC/ACM International Conference on Web Intelligence and Intelligent Agent Technology - Workshops, WI-IAT Workshops 2007",

pages = "329--332",

booktitle = "Proceedings - 2007 IEEE/WIC/ACM International Conference on Web Intelligence and Intelligent Agent Technology - Workshops, WI-IAT Workshops 2007",

note = "2007 IEEE/WIC/ACM International Conference on Web Intelligence and Intelligent Agent Technology - Workshops, WI-IAT Workshops 2007 ; Conference date: 02-11-2007 Through 05-11-2007",

}

Groble, M & Thompson, W 2007, Supporting multiple user types with a multimodal dialog agent. in Proceedings - 2007 IEEE/WIC/ACM International Conference on Web Intelligence and Intelligent Agent Technology - Workshops, WI-IAT Workshops 2007., 4427600, Proceedings - 2007 IEEE/WIC/ACM International Conference on Web Intelligence and Intelligent Agent Technology - Workshops, WI-IAT Workshops 2007, pp. 329-332, 2007 IEEE/WIC/ACM International Conference on Web Intelligence and Intelligent Agent Technology - Workshops, WI-IAT Workshops 2007, Silicon Valley, CA, United States, 11/2/07. https://doi.org/10.1109/WIIATW.2007.4427600

Supporting multiple user types with a multimodal dialog agent. / Groble, Michael; Thompson, Will.
Proceedings - 2007 IEEE/WIC/ACM International Conference on Web Intelligence and Intelligent Agent Technology - Workshops, WI-IAT Workshops 2007. 2007. p. 329-332 4427600 (Proceedings - 2007 IEEE/WIC/ACM International Conference on Web Intelligence and Intelligent Agent Technology - Workshops, WI-IAT Workshops 2007).

Research output: Chapter in Book/Report/Conference proceeding › Conference contribution

TY - GEN

T1 - Supporting multiple user types with a multimodal dialog agent

AU - Groble, Michael

AU - Thompson, Will

PY - 2007

Y1 - 2007

N2 - Recent research has addressed the problem of formulating a dialog agent as a partially observable Markov decision process (POMDP), and learning a dialog policy that is optimal given the particular characteristics of the transition, observation and reward functions of the POMDP. This paper addresses the problem of trying to learn a small set of dialog agent policies that provide near-optimal behavior over a wide range of variations in POMDPs, reflecting different user preferences and environment characteristics. We show for a very simple dialog, we can cover a large number of simulated users to within 10% of their optimal return using fewer than 5% of the individual optimal policies.

AB - Recent research has addressed the problem of formulating a dialog agent as a partially observable Markov decision process (POMDP), and learning a dialog policy that is optimal given the particular characteristics of the transition, observation and reward functions of the POMDP. This paper addresses the problem of trying to learn a small set of dialog agent policies that provide near-optimal behavior over a wide range of variations in POMDPs, reflecting different user preferences and environment characteristics. We show for a very simple dialog, we can cover a large number of simulated users to within 10% of their optimal return using fewer than 5% of the individual optimal policies.

UR - http://www.scopus.com/inward/record.url?scp=48149106380&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=48149106380&partnerID=8YFLogxK

U2 - 10.1109/WIIATW.2007.4427600

DO - 10.1109/WIIATW.2007.4427600

M3 - Conference contribution

AN - SCOPUS:48149106380

SN - 0769530281

SN - 9780769530284

T3 - Proceedings - 2007 IEEE/WIC/ACM International Conference on Web Intelligence and Intelligent Agent Technology - Workshops, WI-IAT Workshops 2007

SP - 329

EP - 332

BT - Proceedings - 2007 IEEE/WIC/ACM International Conference on Web Intelligence and Intelligent Agent Technology - Workshops, WI-IAT Workshops 2007

T2 - 2007 IEEE/WIC/ACM International Conference on Web Intelligence and Intelligent Agent Technology - Workshops, WI-IAT Workshops 2007

Y2 - 2 November 2007 through 5 November 2007

ER -

Groble M, Thompson W. Supporting multiple user types with a multimodal dialog agent. In Proceedings - 2007 IEEE/WIC/ACM International Conference on Web Intelligence and Intelligent Agent Technology - Workshops, WI-IAT Workshops 2007. 2007. p. 329-332. 4427600. (Proceedings - 2007 IEEE/WIC/ACM International Conference on Web Intelligence and Intelligent Agent Technology - Workshops, WI-IAT Workshops 2007). doi: 10.1109/WIIATW.2007.4427600

Supporting multiple user types with a multimodal dialog agent

Abstract

Publication series

Other

ASJC Scopus subject areas

Access to Document

Other files and links

Fingerprint

Cite this