Machine Translation Post Editing — How You Can Get it Right
Machine Translation (MT) is slowly gaining traction due to its speed and cost-effectiveness. However, technology alone can’t achieve the best results, and translations using them often have an inorganic feel. Human translators still need to do machine translation post editing to ensure accuracy and context!
The basic rules of machine translation post editing
MT post editing combines the speed of machine translation with the skill of professional translators. The result is a more accurate text with better flow, greater cultural sensitivity, and a solid grasp of slang and local terminology.
A good human editor can:
- Make quick decisions
- Translate the entire text
- Correct incoherent sentences
- Delete inaccurate sentences if they are irrelevant
- Focus on semantic and syntactic mistakes
If there are substantial errors on the machine-translated text, an editor can re-translate the problematic the text, correct stylistic mistakes, and replace recurring terms with synonyms.
Strategies for machine translation postediting
In addition to the rules above, good editors also take time to do some pre-editing of the source text. After all, pre-editing is just as crucial as post-editing. Human editors should check their source text thoroughly for the following:
- Grammatical and spelling errors
- Long and complex sentences
- Sentences and phrases with nuance and sarcasm
- Sentences and phrases with double-negatives
It’s also essential for an editor to understand how the machine translator works before feeding text into the system so that they can prepare the source text specifically for that machine.
Technology has never been more advanced, and Neural Machine Translation (NMT) systems can now process millions of words in a single hour. The resulting text is relatively accurate but still feels unnatural. The nuances, tone, and sensitivity are lost.
This is why the human component can never be taken away in the translation process. There is certainly a place for software in the translation field, but machine translation post editing must always be done for better results!