AWS Comprehend gets a new improvement that allows developers to create lists of specialized words and phrases without knowledge of machine learning. The company is sharing this in an update. Introduced last year, Comprehend is a natural language processing tool that allows companies to extract commonly used words and phrases from texts.

With the new features, developers can expand Comprehend to identify natural language terms and classify text that specializes for their team, company or industry. Amazon handles the complex tasks, allowing developers to add their own lists without having a strong background in machine learning or natural language processing.

Meanwhile, Comprehend takes care of building, training and hosting its own machine learning models, and makes these models available via a private API.

Two parts

Developers first create a list of their own entities, for example jargon. Amazon then learns to identify the language and builds a private, customized model based on the list.

The second part consists of classification. Once the list has been created, it is possible to create logical lists of where the terms occur. With only fifty examples, Comprehend automatically trains a classification model that can be used to categorize all documents. For example, it is possible to sort support emails by department, group social media messages by product or group analyst reports by department.

This can be useful to send the right documents to the right people on the work floor, or to deploy them in a specific application. This way, Amazon offers a way to create your own machine learning models, while the system handles the details behind the scenes. The company thereby simplifies the complex things that might normally be too complicated for developers to do on their own.

The Comprehend function is now available.

This news article was automatically translated from Dutch to give Techzine.eu a head start. All news articles after September 1, 2019 are written in native English and NOT translated. All our background stories are written in native English as well. For more information read our launch article.