Automatic construction of an English-Chinese bilingual FrameNet

Authors:
Benfeng Chen;Pascale Fung
Affiliations:
University of Science and Technology (HKUST), Clear Water Bay, Hong Kong;University of Science and Technology (HKUST), Clear Water Bay, Hong Kong
Venue:
HLT-NAACL-Short '04 Proceedings of HLT-NAACL 2004: Short Papers
Year:
2004

Citing 5
Cited 4

Automatic labeling of semantic roles

Computational Linguistics
Construction of a Chinese–English Verb Lexicon for Machine Translation and Embedded Multilingual Applications

Machine Translation
The Berkeley FrameNet Project

COLING '98 Proceedings of the 17th international conference on Computational linguistics - Volume 1
Simple features for Chinese word sense disambiguation

COLING '02 Proceedings of the 19th international conference on Computational linguistics - Volume 1
Identifying concepts across languages: a first step towards a corpus-based approach to automatic ontology alignment

COLING '02 Proceedings of the 19th international conference on Computational linguistics - Volume 1

Multilingual collection retrieving via ontology alignment

ICADL'04 Proceedings of the 7th international Conference on Digital Libraries: international collaboration and cross-fertilization
Analysis of an iterative algorithm for term-based ontology alignment

IJCNLP'05 Proceedings of the Second international joint conference on Natural Language Processing
FrameNet, current collaborations and future goals

Language Resources and Evaluation
Wikipedia-based WSD for multilingual frame annotation

Artificial Intelligence

Quantified Score

Hi-index	0.00

Visualization

Abstract

We propose a method of automatically constructing an English-Chinese bilingual FrameNet where the English FrameNet lexical entries are linked to the appropriate Chinese word senses. This resource can be used in machine translation and cross-lingual IR systems. We coerce the English FrameNet into Chinese using a bilingual lexicon, frame context in FrameNet and taxonomy structure in HowNet. Our approach does not require any manual mapping between FrameNet and HowNet semantic roles. Evaluation results show that we achieve a promising 82% average F-measure for the most ambiguous lexical entries.