Two approaches to aspect assignment in an English-Polish machine translation system

Authors:
Anna Kupść
Affiliations:
Carnegie Mellon University
Venue:
EAMT '03 Proceedings of the 7th International EAMT workshop on MT and other Language Technology Tools, Improving MT through other Language Technology Tools: Resources and Tools for Building MT
Year:
2003

Citing 3
Cited 0

C4.5: programs for machine learning

C4.5: programs for machine learning
Induction of Decision Trees

Machine Learning
The KANT system: fast, accurate, high-quality translation in practical domains

COLING '92 Proceedings of the 14th conference on Computational linguistics - Volume 3

Quantified Score

Hi-index	0.00

Visualization

Abstract

The paper presents two approaches to aspect assignment in a knowledge-based English-Polish machine translation (MT) system. The first method uses a set of heuristic rules based on interlingua (IR) representation provided by the system, whereas the other employs machine learning techniques. Both methods have similar performance and obtain high accuracy of over 88% on test data. The crucial difference, however, is the development effort: the machine learning technique is fully automatic, whereas heuristic rules are derived manually.