Using AI translation services in learning technology to optimize results

    Applying Artificial intelligence(AI) within a LCMS solution

    This article aims to provide a very brief overview on how AI supports automatic translations, focusing more on AI translation services, describing the capabilities of a translation management system (TMS). Later this piece goes one step further in outlining greater efficiencies that can be achieved with translation services that are integrated with localization functionality within learning technology – namely a Learning Content Management System (LCMS).

    The question ‘What is Artificial Intelligence (AI)?’, can be answered most simply as ‘Artificial intelligence (AI) is the simulation of human intelligence processes by machines’. How can this be beneficial to the learning community? We will discuss some applications here.

     

    Machine learning and AI in translations

    We have discussed elsewhere in articles and presentations how the branch of machine learning that relates to automatic translations Neural Machine Translation (NMT) works, in relation to automatic translations. Essentially, NMT translates a word based on the entire sentence, instead of a couple of surrounding words. This improves accuracy enormously. Here we display some research findings, by way of example.

     

    Traditional machine translations vs Neural Machine Translations (NMT)

     

    Training of the model & Training of the dictionary: How does it work?

    Creating an ecosystem model contains a minimum of 10,000 sentence pairs of the specific targeted languages. The custom translator will create a model from the sentence pairs. This model will contain your company’s ‘lingo’ as well as your standard of translation. It will also provide more examples on complete sentence structures. With the model you can add a company dictionary containing specific ‘end-to-end’ translation definitions. This model will be referred back to with each translation.

     

    Data on how translation can be enhanced for firms

    In one real example from a client, we stepped in with our AI-enhanced automatic translations functionality. The multi-national organisation was spending six months creating, translating and ‘rolling out’ a course to their global sales teams. Further, the company provided training in multiple languages, in a variety of different character sets.  As displayed in the graphic, automatic translations reduced time and cost considerably. This amounted to a 80% reduction in time and between 30% and 50%, on average, in cost. You can read more about this here.

     

    Date on improvements to be made with automatic translations in a LCMS

     

    One step further…

    The data above shows that there are clear efficiencies to be made by applying AI to translation services, including an LCMS. However, there are even greater improvements that can be made with integration of extra functionality – namely, with localization and translation.

    Localization

    So, briefly, ‘What does content localization mean?’ it is ‘adapting content to any local cultural sensitivities’.

     

    In the elearning content technology process, localization is the final step in translation.  Localization means aligning the translated document to the specific region or cultural groups that the content will reach.

     

    Localization benefits

    There are a number of benefits to ‘localizing’ your content, as outlined (in this example) by E-learning Industry, which includes:

    – increasing your reach;

    – achieving consistency in messaging to your internal and external audience(s); and

    – tailoring an accurately translated document to your region(s).

     

    However, perhaps critical for elearning professionals is the potential for an increase in acquisition, when targeting external audiences, made possible by localization of content. Where a native language can only reach a certain geography, tailoring to a wider array of local language nuances and customs can improve the relevance to the reader or listener. Moreover, localization of content to a relevant scenario in this way can also vastly improve the retention (or ‘recall’) of content by your target audience, when it is tailored to local cultures.

     

    How is it even better within a LCMS?

     

    With the added functionality of a digital repository (within a LCMS), then localization is a streamlined process, making it possible to change content ‘chunks’, including (for example), videos, images, audio, etc, to meet any local customs and traditions.

     

    You can gain peace of mind across your organization by having translated content that is already centrally stored in one system. The localization feature allows you to translate, then send your translated documents in a xliff package to your localization company. Then, you can track the progress of the localized documents.

     

    The eXact learning solutions localization feature helps you version control your translated documents. If your master documents contain a new version you will be notified to update your variations, this helps with consistency of messaging across your documents.

     

     

    AI Translation and Localization in the LCMS

    Displayed below is an example of how the Localization page looks in the eXact learning solutions LCMS. The page allows you to translate and download your Xliff package containing your translated documents. You can then re-upload your Xliff package when the localization process is completed.

     

    Displays how Localization jobs appear in the eXact LCMS

     

    Here, the below graphic provides a further overview of the localization management process within the LCMS. This gives you a partial insight into some of the project management and workflow aspects of localization within the technology.

     

    Overview of the Localization process in the LCMS

     

     

     

    The Localization process page gives you an overview of your translation. Here you can see which of your translations are going out for external translation (Localization), the number of translation units, and which documents are out of date. A document becomes out of date when the master document is in a newer version than the variations that have been translated from it.

     

     

    LCMS technology can improve accuracy and efficiency in translation and localization

     

    We have discussed that there is significant RoI to be gained by organizations from investing in the right technology. Further, technology can streamline processes and improve accuracy in translation and localization. We have also discussed how the process can be made more seamless when it is supported by a digital repository, within a LCMS. Clearly, implementation of the most suitable AI translation technology can demonstrate vast improvements in your cost outlay. However, perhaps it is most pertinent to note the critical benefits of combining both (translation and localization) functionality for optimum impact. In addition, you need to ensure that you are connecting to your key audiences and stakeholders in the best way possible to maintain and build those relationships worldwide.

     

    Contact us for more information on our LCMS and how we integrate this with our AI translation technology services or how our localization processes work within our LCMS. For a free demo then get in touch here.