Machine Translation

Machine Translation

During the last couple of years, machine translation post-editing has become one of the hottest, most discussed topics in the translation industry.

Post-editing Machine Translation speeds up the translation process and enables translation users to translate higher volumes of content, and content that might not otherwise be feasible to translate. However, while post-editing can be highly productive and rewarding, a deep understanding of MT and post-editing processes is required to fulfill the expectations of end users.

Link Translation utilizes our trained linguistic resources to provide post-editing of content that was machine-translated. This machine-translated content output can come from our customers or from our own machine translation engine. With our experts going through the completed MT translation for post-editing, clients can ensure their multilingual communications are highly targeted for a specific market.

Post Editing

Post-editing usually refers to the process where an entire document, or those parts where there are no exact or good matches from the translation memory, is submitted to a machine translation engine, is machine translated and then the target text is corrected by a human post-editor.

We can either automatically send the documents to machine translation engines to retrieve translations, or work with XLIFF (XML Localization Interchange File Format) or TMX (Translation Memory eXchange) files previously machine translated. Our post-editor has access to all of the usual tools for leveraging and editing existing translations, terminology management and quality control, and their work can also be quantified using the edit distance and the editing time reports.

Post-editing serves to provide the same linguistic quality as would be expected and achieved by standard human translation and serves as a productivity tool to reduce the total cost of translating.

MT Quality Management

In addition to the advanced editing environment, We can improve the quality of the MT output. We offer full version control on the post-edited document. We also offer edit distance reports that quantify the changes by the post-editor and gives our clients an indication of the quality of the machine translation engine.

We also allow our clients to track the time spent on editing individual segments or the project. This way we can compare machine translation engines. Moreover, our LQA models – error typologies – can be specially customized for MT and used to improve both the quality of the documents and the MT engine.

Depending on time constraints, cost requirements, and end-use, the machine translation process can be customized with the ideal combination of terminology glossaries, translation memory technology, optical character recognition (OCR), and post-translation proofreading by human linguists.