Bidirectional English-Arabic Machine-Translation: A Model Enhancing Arabic Search for Cloud Services

摘要

The translation between different languages has an increasing importance due to the poor search results that do not meet some users' expectations. Specifically, Arabic users face this problem when searching for cloud services in which searching results do not match their needs. Thisproblem is challenging because of the way that these cloud services are published which is not based on well defined standards and the variety types and descriptions of such services. In this paper, we present a cloud service search engine that helps Arabic cloud service consumers to obtainbetter search results for cloud services that suit their needs. In particular, we introduce BEAM-Translation Model for bi-directional English-Arabic machine translation of scientific domain which consists of two subsystems in which each subsystem runs in one direction (either English to Arabicor Arabic to English) using transfer-based approach and two tightly related knowledge bases. The proposed Model is validated using 100 real sentences related to cloud services search words. The experimental results show the ability of this model in enhancing the result process and reducingthe translation errors of scientific texts from 14% to 10% by discovering new rules of translation based on the bi-directional try-and-error translation.