Translation German-English - Birgit Bode-Welch

Computer programs that translate texts from a source into a target text are an integral part of a translator’s universe. However, to use online-tools such as Google-Translate to name but one poses a major risk in regard to data security. Why? Because nobody really knows where the servers, we send our data to, are located potentially leading to considerable consequences due to violations of applicable Data Protection laws such as GDPR or looking at the worst-case disclosure of company secrets.

With the implementation of Neuronal Machine Translation (NMT) in the area of Machine Translation technology temporarily peaked and me as a translator have to be honest: Results of such NMT really speak for themselves!

Such automated competition of the human translator has its advantages as well as disadvantages but is generally speaking an alternative for any text where quality is secondary, which have to be available at short notice despite large volumes and/or when no publication of such text is intended.
Not every text is suitable for Machine Translation. There are specifications, requirements as to quality and risk which simply do not allow for utilisation of NMT. For this reason every text must be assessed as to its suitability for NMT processing.

Which advantages does the utilisation of an NMT have for my customers?

Speed
By use of various NMTs in connection with post-editing a tremendous amount of data can be processed such ensuring short delivery times. Thus accelerating a product launch in foreign markets considerably.

Savings
With suitable texts such as product information, operator's manual, general business correspondence as well as internet data arising in forums, social networks or online-shops an enormous potential for savings may be realised.

Learning effect of the maschine
NMT-Engines learn autonomously and continually improve themselves. This means, that target texts are continually getting better..

Data security
Translation of texts with an NMT by a specialist translator takes place in a secure environment thus ensuring that any data processed does not, as in the case of Google Translate & Co., does not end up in the wrong hands.

However, such (rawly) translated text should be assessed by a specialist translator. This type of assessment, the so-called Post-Editing is an important service to assure text quality. Sometime NMT is unable to translate context appropriate or expressions, terminology and so on correctly. This is where the machine lacks "linguistic finesse" which results in misinterpretations, ambiguous or even completely entirely false translations. Here the post-editor steps in.

Depending on requirement and demand there are different type of post-editing:

Pre-Editing: The linguistic/formal preparation of the source text for a better NMT result.
Light Post-Editing: Review and correction for better textual readability and comprehensibility. Light post-editing is also suitable for target text prepared by non-native speakers.
Full Post-Editing: Review and correction of grammar, punctuation, syntax, customer terminology etc., localization, adaptation to the target market or target audience and corporate style guides. Full post-editing is also suitable for sophisticated target text prepared by non-native speakers.

Maybe you are interested in integrating an MT-tool into your process chain but your are not sure how such a tool works or what you will need in particular, how the translation processes are to be set up, document and data security is ensured. I am happy to advise an provide support from the first ideas right through to implementation.