The Evolution of Fidelity and Intelligibility of Machine Translation Quality Over Time
Monday, 25 March, 2019
The use of machine translation, together with translation memories is becoming increasingly prominent, and companies rely heavily on them for services like post-editing. But just how cost-effective is their use? Wagner Consulting did a study on this topic. Four different texts from the fields of Medicine, Law and Science with the addition of Web content were selected and run through Google Translate and SDL Language Cloud, and translated into German by certified translators for this purpose. The translated texts were reviewed by a team of ten independent translators who, following the guidelines presented by the company, completed questionnaires and forms providing information on the quality of the translations. Among the parameters analyzed were translation fidelity, intelligibility, reading time required for clear understanding, formatting legibility, term consistency, production costs and time, among other factors.
For this study, we decided to use several papers by Dutch documentarist Charles Van Slype and update them to analyze the effectiveness of MT in the different markets in the translation industry, inviting members of GALA to participate and help us assess the information obtained and the diverse ways to study the reasons they rely on MT. Our aim is to provide tools to allow our colleagues to take a closer look at their practices and take back home not only hard data in figures but hopefully also interest them in taking the initiative to start analyzing our current business and new opportunities in a different way.
Gino Toti is an Argentine-born certified and sworn English-Spanish and vice versa translator. After completing his studies and obtaining his degree, he began his career in Wagner Consulting, starting as a translator and training to become a Project Manager, later being promoted to US Division Head. Member of Gala, alumni of UMSA in Buenos Aires, he strives for constant learning and brainstorming of new ideas to make the company which has placed its trust in him stand out and live up to the constantly changing standards of the industry.
Most machine translation output is fine for basic Western LSP needs, as they usually translate within European...
Machine translation is great when it works, but it fails often and in unpredictable ways. What causes bad machine...