Context is crucial for both written and spoken language as it sets the natural flow and consistency of any text or speech. Context is defined by the words or sentences that precede or follow a certain piece of written or spoken text. While humans are able to intuitively keep context in mind when they write a piece of text, this is not the case with Machine Translation systems.
The problem with most Machine Translation (MT) systems is that they learn to translate sentence by sentence. So, when a user asks an MT system to translate any piece of text, the MT engine breaks the paragraph into sentences and then it translates them separately, without keeping any grammatical connection between them. This is why users very often observe weird pronoun mismatches and other inconsistencies between adjacent machine-translated sentences that make them scratch their heads.
Correcting such mistakes seems quite straightforward for a human editor, but still they remain another source of inconsistency, it’s always possible to creep in the final edited text, and they also make human editors lose trust in the MT system.
NeuralDesktop is the only Neural Machine Translation system that provides context-aware output. This feature works with all the translation facilities NeuralDesktop offers: Clipboard translation, file translation, and even inside CAT tools (such as Trados and memoQ), through plugins. This ground-breaking feature helps minimize the human editor’s cognitive effort and makes post-editing easier, thus boosting productivity.
Context in NeuralDesktop works seamlessly while you translate and it is provided in all language combinations currently available.
The two screenshots below show the output from NeuralDesktop and two other neural machine translation providers while working inside Trados Studio. It becomes obvious that NeuralDesktop is the only system that maintains grammatical consistency among segments. At this point, we should remind our readers that every feature in NeuralDesktop, including context, works offline, locally!

We invite you to give it a try yourself!