A list of machine translations for novice users, you deserve it

With the development of the times, machine translation has become an important part of translation automation. If used properly, it not only increases productivity, shortens turnaround time, but also increases the consistency and accuracy of translation.

Today, machine translation has been widely used in many fields and industries. From weather reports to travel, from technical documentation to user-generated content, from website to e-commerce, we infiltrate our work and life.

However, how to choose the right engine for the project is the focus, and we can just help you solve it. Answer the following questions and we can help you customize your exclusive translation engine.

Are you a part-time translator?

• Are you proficient in 1-3 languages?

• Have you studied 1-3 professional fields?

• Do you translate at least 20,000 words per month?

• Do you see or expect an increase in workload?

• Do you have a dedicated translation memory and vocabulary?

• Do you use CAT tools? Is the desktop version still a cloud version like Wordbee?

If you answer "yes" to all of the above questions, then you can consider using off-the-shelf general-purpose machine translation engines (such as Google Translate, Microsoft Translator or DeepL) that are easy to use and perform well, in some cases. Even beyond your expectations. After registering with these MT engine sites, you will be assigned an API key, which you can then enter into the CAT tool for configuration and start using the MT engine immediately.

If your translation memory and glossary are up-to-date, consistent, and organized, you can combine them with your chosen machine translation system. You can pre-translate the source text and then post-edit it (save the edited clip in the translation memory). Alternatively, you can use the MT engine "as-you-go on-demand" at the segment level to decide whether to keep, post-edit or reject suggestions from the MT engine.

Are you a language service provider (LSP) or a translation department?

• Do you have skilled human resources internally? For example, a software engineer or a computational linguist.

• Do you translate at least 100,000 words per month in the professional field?

• Do you have special customers who are eligible for the (dedicated) MT + PEMT workflow?

• Do you expect an increase in workload?

• Do you have a consistent and well-maintained professional translation memory (no less than 100,000 clips) and a glossary?

• Do you have a group of reliable and competent employees, both internal and freelance, who can be responsible for post-editing (or for post-editing training)?

If you have a positive answer to all of the above questions, then you can consider using a specialized machine translation system, whether you have an existing or a custom engine, you can use your existing language data. Most machine translation providers have dedicated engines that can access fast solutions through a standard translation project management system. But keep in mind that in some cases, this solution can be problematic and expensive, requiring a lot of financial investment and full-time staff support.

Some warning suggestions

In the first case, if there is a situation involving intellectual property (IP) and confidentiality. Since May 25, 2018, everyone must follow the GDPR data protection regulations, which has become a mandatory requirement in Europe. This must be taken into account both as a part-time translator and as an LSP.

• Do you handle any content that contains personal data?

• Do you handle sensitive or confidential information?

• Have you developed a program to anonymously manage the data you have?

Furthermore, from an ethical perspective, both part-time translators and LSPs should inform customers about the use of MT engines (especially public engines) or, in any case, ask them for specific requirements so that they cannot allow the use of public MT engines. .

A ready-made general-purpose or dedicated engine may be a good solution, but only if your text does not contain sensitive or personal information can it eventually appear in the engine's data training archive and in the content being translated.

The second aspect is training. As an LSP, you should first train your internal and freelancers for post-editing. You can start with some regular training and then provide your staff with project management on selected engines, post-editing projects, and even special training on how to use RegEx and other more complex features