Why is machine translation important?
Machine translation has increasingly become an integral part of the translation process. With customers expecting content in their native language at the “speed of now”, it is becoming difficult to manage the demands of customers through traditional human translation processes.
Depending on the nature of your content, machine translation technology can be utilised to speed up the translation process, changing the role of the translator to a post-editor, whilst still ensuring high-quality multilingual content.
We can provide you with custom-built, cloud-hosted neural machine translation engines which integrate with your current translation workflow. Our powerful productivity tool can increase your translation capacity and quality, and is designed for companies producing large volumes of multilingual content.
We currently offer three tiers of engines and will work with you to find the best fit for your scenario:
- Master engine: We offer this to clients with a very large (at least 15,000 segments, preferably 200,000 segments or ideally 1,000,000 segments) translation memory for the languages they wish to translate, who do not wish to use any stock corpora in their content. This is the ideal engine to use due to its focus on producing content that is entirely unique to you, even before any form of review. Anything that the engine would typically require corpora to fill in will need to be finalised during a post-editing stage.
- Domain adapted: We offer this to clients with a large translation memory (at least 15,000 segments, preferably 100,000 segments) for the languages they wish to translate, who do wish to use stock corpora in their content. This engine is best for clients whose translation memories are not large enough or as suited for a master engine, or are happy to use stock corpora for any areas that the engine would be unable to translate.
- Stock+: We offer this to clients whose translation memories are relatively smaller (at least 1,000 segments, but less than 15,000 segments) but want to start with machine translation to increase the volume of their content. We start by taking a stock engine and training it with your translation memory to start tailoring it to your content. Over time as you translate more, and with more engine training, we will be able to convert it to a master or domain adapted engine, where we can edit the engine to rely on stock content to a lesser degree if necessary.