National Institute for Japanese Language and Linguistics

Project Leader	:	DEN Yasuharu Invited Professor, NINJAL
Project Period	:	November 2011 - October 2014

Summary

In this project, we explore a methodology for sharing existing conversation corpora that have been developed by various organizations/researchers. In particular, we aim at collecting corpora with diverse styles and settings in order to utilize them in elucidating the universality and the variety of conversational phenomena. The agenda includes:

1. Survey of the corpora developed by the project members,
2. Designing schemes for common annotations,
3. Annotation and sharing of the members' corpora,
4. Analysis of the shared corpora, showing the efficacy of sharing diverse corpora,
5. Survey of corpora external to the project that can be shared using our methodology.

We aim to establish a foothold for future development of large-scale conversation corpus through this activity.