This paper investigates the use of recurrent surface text patterns to represent and index open-domain dialogue utterances for a retrieval system that can be embedded in a conversational agent. This approach involves both the building of a database of such patterns by mining a corpus of written dialogic interactions, and the exploitation of this database in a generalised vector space model for utterance retrieval. It is a corpus-based, unsupervised, parameterless and language independent process. Our study indicates that the proposed model performs objectively well comparatively to other retrieval models on a task of selection of dialogue examples derived from a large corpus of written dialogues.
CITATION STYLE
Dubuisson Duplessis, G., Charras, F., Letard, V., Ligozat, A. L., & Rosset, S. (2017). Utterance retrieval based on recurrent surface text patterns. In Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics) (Vol. 10193 LNCS, pp. 199–211). Springer Verlag. https://doi.org/10.1007/978-3-319-56608-5_16
Mendeley helps you to discover research relevant for your work.