What is blended translation?

Blended translation combines the use of automated translation software with a human translator.

Translation software is used to translate millions of documents daily. Using translation software alone can only provide a gisting level of accuracy for a document. Human translators interact with the translation software to improve the gisting level. With the explosion in demand for translation, humans cannot keep up with the demand and software solutions are required. As automated translation technology continues to evolve we will see an increase in blended approach to translation. Over time it is predicted that less time will be required by the human and more accuracy will be achieved by the software.

One of the most important factors in being successful at blended translation is selecting translation software that provides tools that allow the translator to build additional dictionaries and translation memories. A translator must be able to have control over the software output.

SYSTRAN, the leader in translation technology, is the software that has been utilized the most for automated translation. SYSTRAN has two products that can be customized for better blended translations. SYSTRAN Premium Translator and SYSTRAN Enterprise Translator, both have a project manager that gives the translator a professional environment and set of tools for customizing and analyzing documents, both before and after translation, side by side viewing and alignment, color coding and more.

The real key to effectively using software in conjunction with the human translator is the workflow.

Based on Systran’s Software here are the first most important workflow steps, the process should be close if you use any other type of automated software product.

SYSTRAN SUGGESTED WORKFLOW STEPS

1. Create a New Project / Corpus

2. Import multiple documents

3. Click Translate button to translate all documents in the corpus

4. Run analysis on the translated documents in the corpus

5. Review the list of extracted terms that are suggested additions to a custom dictionary and add those terms

6. Review the list of extracted possible ambiguities to see what words need to be added to a custom dictionary

7. Click translate to retranslate all the documents in the corpus

Once the initial steps above had been completed, the translator is now ready to start manually reviewing the source and target side by side to identify more terms, ambiguities, and translation memory entries. The

The more time invested into training the software, the less time spent proofing documents later. There are a lot more advanced tools in SYSTRAN Products for customizing and training the software, in fact the tools even allow you to make contextual choices. You create rules for words. Example, if mouse is used in a sentence it should be translated meaning a rodent, but if mouse is used in the same sentence as the word computer, then it will translate relating to a computer pointer device.

Using translation software is not suggested for all types of documents but many professionals in the industry agree that some of the latest are developments are going to change the translation industry.

SYSTRAN’s latest release of the SYSTRAN Training Server is proving to deliver a higher level of accuracy and provide a platform for developing languages that not currently underdevelopment. The Training Server actually learns from previously translated documents, can build additional terminology dictionaries and automatically learn from corrections that human translators make to the target document.