Constraint relaxation with weighted feature structures

Frederik Fouvry


Abstract
In this paper, we present a definition of unification of weighted feature structures designed to deal with constraint relaxation. The application of phrase structure rules in a unification-based Natural Language Processing system is adapted such that inconsistent values do not lead to failure, but are penalised. These penalties are based on the signature and the shape of the feature structures, and thus realise an elegant and general approach to relaxation.
Anthology ID:
W03-3010
Volume:
Proceedings of the Eighth International Conference on Parsing Technologies
Month:
April
Year:
2003
Address:
Nancy, France
Venue:
IWPT
SIG:
SIGPARSE
Publisher:
Note:
Pages:
103–114
Language:
URL:
https://aclanthology.org/W03-3010
DOI:
Bibkey:
Cite (ACL):
Frederik Fouvry. 2003. Constraint relaxation with weighted feature structures. In Proceedings of the Eighth International Conference on Parsing Technologies, pages 103–114, Nancy, France.
Cite (Informal):
Constraint relaxation with weighted feature structures (Fouvry, IWPT 2003)
Copy Citation:
PDF:
https://aclanthology.org/W03-3010.pdf