Partially Observable Markov Decision Processes (POMDPs) are attractive for dialogue management because they are made to deal with noise and partial information. This paper addresses the problem of using them in a practical development cycle. We apply factored POMDP models to three applications. We examine our experiences with respect to design choices and issues, and compare performance with hand-crafted policies.
|Title of host publication||Proceedings of the 8th SIGdial Workshop on Discourse and Dialogue|
|Editors||S. Keizer, H. Bunt, T. Paek|
|Place of Publication||PA, USA|
|Publisher||Association for Computational Linguistics (ACL)|
|Number of pages||4|
|Publication status||Published - 1 Sep 2007|
|Publisher||The Association for Computational Linguistics|
- HMI-SLT: Speech and Language Technology
- HMI-MI: MULTIMODAL INTERACTIONS
Bui Huu Trung, B. H. T., van Schooten, B. W., & Hofs, D. H. W. (2007). Practical Dialogue Manager Development using POMDPs. In S. Keizer, H. Bunt, & T. Paek (Eds.), Proceedings of the 8th SIGdial Workshop on Discourse and Dialogue (pp. 215-218). PA, USA: Association for Computational Linguistics (ACL).