A step towards Human-Machine unification using Translation Memory and Machine Translation System
Nishtha Jaiswal, Renu Balyan, Anuradha Sharma · 2011
It is a well known fact that development of fully generalized automatic machine translation systems is still a dream that seems too far from reality to be fulfilled. A Machine Translation (MT) system with human intervention is the most appropriate way to get translation which has greater human acceptability. The main aim of this paper is to show how semi-automatic method for translation reduces the human effort, and increases efficiency of the MT system. Most of the translation work these days relates to technical material like contracts, user manual, field reports, etc. Technical translation calls for consistency (terminological, structural, phrasal, etc.) among documents and translation accuracy. Also it is time consuming and expensive to train technical translators. So there is a need for human-machine unification and this paper focuses on an approach which can help to achieve this. This paper also focuses on the need for integration of existing MT systems with tools like (Wordfast, Trados etc.) which are widely used for translation by the translation industry. Currently the approach has been applied on AnglaMT system for English to Hindi and English to Punjabi machine translation systems. This approach can however be applied to other available MT systems as well.