Amazon Translate

Traditionally, translation methods heavily depended on manual processes, making them prone to errors and unable to consistently deliver accurate results. Additionally, these methods incurred high costs. Amazon Translate revolutionizes the translation process by offering a cost-effective solution that enhances accuracy and efficiency. With Amazon Transcribe, businesses can now localize their content for global audiences and activate interlingual communication between customers.

How Amazon Translate works

Amazon Translate uses neural machine translation technology to convert text from one language to another without losing the context. The deep learning models allow Amazon Translate to produce accurate and fluent translations.

Amazon Translate’s working is classified into Real-time translation and Batch translation.

Real-time translation

With real-time translation, users can enter text and instantly receive translation results. This can be done either through the console interface by selecting source and target languages, then inputting or uploading text, or via APIs like Amazon Translate's TranslateText and TranslateDocument, which provide translation results directly to applications.

Batch translation

To translate multiple documents stored in Amazon Simple Storage Service (Amazon S3), asynchronous batch processing is employed. This is ideal for scenarios where real-time translation is not required, and multiple documents, such as social media posts, reviews, and user queries, have to be translated.

To initiate batch processing, a user uploads a collection of documents to an S3 bucket and initiates the batch translation job using StartTextTranslationJob. Alternatively, the job can be created directly from the AWS console.

Customize translations

With Amazon Translate, businesses can customize the translation process. The service allows tailoring the translations by specifying what and how text needs to be translated. For instance, users can specify content that does not need to be translated with Do-not-translate tags. Businesses can define how certain terms, such as domain-related jargon and brand names, must be translated.

The length of the translations can also be reduced by customization. For content sensitivity, users can filter out profane words and sentences in translation results. Amazon Translation also allows businesses to adjust the tone and formality level of translation according to their requirement and audiences.