What if Machine Translation Could Understand the Context?

We’ve done a lot of MTPE projects – as required by clients, and machine translation does really good in some fields of translation, like internet technologies, but basically, it’s still not beyond human in any aspects, and it’s only an auxiliary tool.

For example, machine translation could translate a long article in seconds, but the translated file would still need to be proofread and edited by a human translator, and that would still take some time.

But what if machine translation would “understand” the context of any translation projects? In my opinion, that would be overwhelming.

A fundamental problem of human translation is that human translators aren’t always able to do too much research on the contexts and background of any texts to be translated.

Although a lot of reference materials could be provided from the client, and there is numerous relevant information on the internet, human translators wouldn’t spend too much time on studying the background and contexts.

However, time wouldn’t be a problem for machine translation as long as it could “understand” the contexts and use them for relevant translation tasks.

In that case, the more texts you input in Google Translate, the much more accurate translation of a certain sentence you could get from this MT tool.

And even before you input any texts in a translation machine, it could have studied relevant materials in advance, in order to be ready for the task.

This assumption requires better abilities to understand and study of the machines, but once they are ready, human translators would have to step aside from the dominant place in any fields of translation.