TranslationRecall: A web-based translation toolset

Abstract

Towards the formation of the ASEAN Economic Community, there is a need to develop translation of documents among languages in those AEC countries. While it is hard to create a machine translation engine that can perfectly translates a source language (Thai) to a target language (English), as a more manual method the Translation Memory-based method just provides a translation of an input text by listing the translations of sentences that are similar to that input text. In this work, we propose a method to combine the fully automatic machine translation approach and the Translation Memory-based approach. Our proposed computer-assisted translation system retrieves the translation pairs which are similar to the current input and then flexibly modifies some parts in the translations using a dictionary or translations in a corpus. The system we have constructed is comprised of various tools that each has its own specialty function including GIZA++, Lucene, and MongoDB in which each tool's functions in the system are provided. We also show the overall components, process, and the techniques we use in our system.