While Neural Machine Translation (NMT) allows us to translate content at scale with great accuracy and efficiency, there is no replacing a human review by a subject expert. At Exfluency™, the human, post-editing step is carried out by Subject Matter Experts (SME) and skilled linguists.
A strategic combination of machine and human input is beneficial to any translation project. Sometimes referred to as post-editing machine translation, we like to call it ‘hybrid intelligence’: bringing the best of both methods together to create the optimal translation experience.
A machine-human collaboration
We’re not quite there yet, but we aim to improve the value of hybrid intelligence using a continuous, human-to-machine feedback loop.
This would mean, when a human detects an error in the machine-translated text, data relating to the error’s cause and solution would be added to the system’s training data. Subsequent translations will then benefit from this knowledge as the machine won’t make the same error again.