Paper: Disambiguating Between Generic and Referential You in Dialog

ACL ID P07-2027
Title Disambiguating Between Generic and Referential You in Dialog
Venue Annual Meeting of the Association of Computational Linguistics
Session System Demonstration
Year 2007
Authors

We describe an algorithm for a novel task: disam- biguating the pronoun you in conversation. You can be generic or referential; finding referential you is im- portant for tasks such as addressee identification or extracting ‘owners’ of action items. Our classifier achieves 84% accuracy in two-person conversations; an initial study shows promising performance even on more complex multi-party meetings.