Two methods for learning ALT-J/E translation rules from examples and a semantic hierarchy

Authors:
Hussein Almuallim;Yasuhiro Akiba;Takefumi Yamazaki;Akio Yokoo;Shigeo Kaneda
Affiliations:
King Fahd University of Petroleum and Minerals, Dhahran, Saudi Arabia;NTT Communication Science Labs., Kanagawa-ken, Japan;NTT Communication Science Labs., Kanagawa-ken, Japan;NTT Communication Science Labs., Kanagawa-ken, Japan;NTT Communication Science Labs., Kanagawa-ken, Japan
Venue:
COLING '94 Proceedings of the 15th conference on Computational linguistics - Volume 1
Year:
1994

Citing 6
Cited 4

Quantifying inductive bias: AI learning algorithms and Valiant's learning framework

Artificial Intelligence
Concept learning and heuristic classification in weak-theory domains

Artificial Intelligence
Knowledge base refinement as improving an incorrect and incomplete domain theory

Machine learning
Readings in knowledge acquisition and learning: automating the construction and improvement of expert systems

Readings in knowledge acquisition and learning: automating the construction and improvement of expert systems
Induction of Decision Trees

Machine Learning
Lexical knowledge acquisition from bilingual corpora

COLING '92 Proceedings of the 14th conference on Computational linguistics - Volume 2

Review Article: Example-based Machine Translation

Machine Translation
Generalizing case frames using a thesaurus and the MDL principle

Computational Linguistics
Decision Tree Learning Algorithm with structured attributes: application to verbal case frame acquisition

COLING '96 Proceedings of the 16th conference on Computational linguistics - Volume 2
Panning for EBMT gold, or "Remembering not to forget"

Machine Translation

Quantified Score

Hi-index	0.02

Visualization

Abstract

This paper presents our work towards the automatic acquisition of translation rules from Japanese-English translation examples for NTT's ALT-J/E machine translation system. We apply two machine learning algorithms: Haussler's algorithm for learning internal disjunctive concept and Quinlan's ID3 algorithm. Experimental results show that our approach yields rules that are highly accurate compared to the manually created rules.