Role of Punjabi morphology in designing Punjabi-UNL enconverter

Abstract

The automated machine translation system is one of the dreams of human beings. In order to design a multilingual machine translation system Interlingua Approach plays a vital role. The Universal Network Language (UNL) is an international project of the United Nations University, with an aim to create an interlingua for all major human languages. The motivation behind UNL is to develop an interlingua representation such that semantically equivalent sentences of all languages have the same interlingua representation. UNL System has two major components. Enconverter processes input text and produces the UNL as output and Deconverter, process UNL and produce equivalent natural text in the target language. In this paper the frame work of designing Punjabi Enconverter is covered. Punjabi Enconverter has two major phases one is generation of UNL Attributes and other is resolutions of Relations. UNL attributes are generated on the basis of Punjabi morphology and UNL relations are resolved on the basis of Punjabi language case markers. A rule base database is prepared which helps to convert Punjabi Text to UNL. Our paper provides the guidelines for constructing these rules bases with help of examples. It discusses the format of enconversion rules, the Noun and verb morphology rules for generation of attributes in Punjabi Enconverter and give the overview of rules developed for resolving UNL relations for the proposed Punjabi-UNL enconverter.

DOI: 10.1145/1523103.1523215

3 Figures and Tables

Cite this paper

@inproceedings{Bhatia2009RoleOP, title={Role of Punjabi morphology in designing Punjabi-UNL enconverter}, author={Parteek Bhatia and Rohan Sharma}, booktitle={ICAC3 '09}, year={2009} }