ADAPT can help you achieve greater productivity and complete transparency in estimating costs within the translation supply chain.
Integrating technology into the translation process is now standard practice. However, selecting the most appropriate system from those available is often a daunting and difficult task for providers of translation services. Human translators should be able to leverage the most appropriate technology — Machine Translation (MT) or Translation Memory (TM) — to increase productivity based on language pair and domain. However, it is difficult to identify the most effective translation method without an automated commercially viable solution. ADAPT partnered with leading language and technology companies Symantec, VistaTEC and Welocalize to find a solution that could recommend the most effective translation method after a segment-level system comparison.
The Adapt Solution
ADAPT’s language technology experts developed the TMTprime platform, a segment-based translation confidence estimation metric, which analyses and measures MT system outputs with TM matches. TMTprime uses translation quality prediction technology to recommend the most appropriate translation tool based on estimated post-editing effort and translation costs. The platform generates a recommendation score for each system, enabling the translator to increase productivity by selecting the most efficient tool for each translation project. The results, which are based on the widely recognised localisation industry standard ‘fuzzy match’ scores, are graphically represented and provide transparent cost estimations of the best performing systems.