Optimising natural language generation decision making for situated dialogue

Nina Dethlefs*, Heriberto Cuayáhuitl, Jette Viethen

*Corresponding author for this work

Research output: Chapter in Book/Report/Conference proceedingConference proceeding contributionpeer-review

14 Citations (Scopus)

Abstract

Natural language generators are faced with a multitude of different decisions during their generation process. We address the joint optimisation of navigation strategies and referring expressions in a situated setting with respect to task success and human-likeness. To this end, we present a novel, comprehensive framework that combines supervised learning, Hierarchical Reinforcement Learning and a hierarchical Information State. A human evaluation shows that our learnt instructions are rated similar to human instructions, and significantly better than the supervised learning baseline.

Original languageEnglish
Title of host publicationProceedings of the SIGDIAL 2011 Conference
Subtitle of host publicationThe 12th Annual Meeting of the Special Interest Group on Discourse and Dialogue
Place of PublicationStroudsburg, PA
PublisherAssociation for Computational Linguistics (ACL)
Pages78-87
Number of pages10
ISBN (Print)9781937284107
Publication statusPublished - 2011
Event12th Annual Meeting of the Special Interest Group on Discourse and Dialogue, SIGDIAL 2011 - Portland, OR, United States
Duration: 17 Jun 201118 Jun 2011

Other

Other12th Annual Meeting of the Special Interest Group on Discourse and Dialogue, SIGDIAL 2011
CountryUnited States
CityPortland, OR
Period17/06/1118/06/11

Fingerprint Dive into the research topics of 'Optimising natural language generation decision making for situated dialogue'. Together they form a unique fingerprint.

Cite this