How to train and publish direct custom model using Azure AI Translator | Custom Translator


Custom Translator has been a key translation enabler capability in Azure Translator for almost a decade, with customers across industries such as transportation, healthcare and financial services using it to develop custom translator models. Custom Translator launched the Neural Machine Translation (NMT) back in 2016 and has made it available through the Microsoft Translator Text REST API ever since, incrementally adding support for languages over the years. 

Today, Microsoft is thrilled to announce the release of direct model customization to all customers. Customers have been creating a model where one language had to be English. That changes today. This new capability allows customers to build custom translator models to translate documents from one non-English language to another non-English language. This method allows for greater accuracy, maintaining cultural sensitivity in the translation, as well as a better understanding of the nuances and idiomatic expressions of the source language. Direct custom translation engines can also save time and resources by eliminating the need for additional editing and proofreading, resulting in a more efficient and cost-effective translation process.
“Transitioning through English as an intermediary in translations has shown its limitations, particularly when nuances and context, such as ‘sex', become lost in translation. This not only risks accuracy but accumulates mistakes throughout the process. By adopting the direct model, we address these translation concerns.  For Belgium, a nation with official languages being Dutch, French and German, the business value of such an approach is huge. The tests we performed at CrossLang with the direct model have demonstrated that with only a handful of additional segments, the new engine surpasses previous benchmarks in performance,” said Tom Francis, CrossLang CEO.
The direct model customization feature is easy to use and available to all users signed up for Translator in an Azure subscription. Users can select from a list of supported languages published here.
To learn more about Custom Translator and how it can help your business thrive in the global marketplace, start with the Custom Translator beginner's guide.

What you can do with the Custom Translator feature in Azure Translator

Build custom models with your domain specific terminology and translate real-time using the Translator API.
Use Custom Translator with your translation solutions to help globalize your business and improve customer interactions.
For more information on Microsoft Translator business solutions and Custom Translator release notes.


This article was originally published by Microsoft's Azure AI Services Blog. You can find the original article here.