Amazon Web Services Inc. today augmented its SageMaker artificial intelligence service with new features that will enable developers to automate core aspects of their work.

The cloud giant launched SageMaker last year to reduce the amount of work involved in building AI algorithms.

The tool is offered as a fully managed platform, meaning that users don’t have to worry about managing hardware resources, and also provides capabilities for simplifying the development workflow itself. One of SageMaker’s standout features is a set of ready-made AI algorithms that can be quickly integrated into applications.

Today’s update expands the algorithm lineup. AWS has added a model for detecting suspicious IP addresses and an implementation of k-means clustering, a method commonly used in AI software to sort images and other files based on similarity. The update also includes a general-purpose algorithm dubbed Object2Vec that’s useful for tasks such as sentiment analysis.

The models are joined by an integration with AWS’ Step Functions service and the AirFlow open-source project. Both are built to automate the kind of multistep data management workflows commonly involved in machine learning projects.

The integration should help AI developers streamline some of the more time-consuming aspects of their work. An engineer could, for example, use Step Functions to collate sample records in one AWS service, use them to train an AI model on SageMaker and then deploy the model to their company’s cloud environment. Automation workflows created with these integrations that have the added benefit of being shareable.

The two other major enhancements introduced today continue the theme of easing users’ work. The most significant of the pair is a capability dubbed SageMaker Search that will enable developers to navigate a project’s code base faster.

“Developing a successful ML model requires continuous experimentation, trying new algorithms and model hyperparameters, all the while observing the impact of potentially small changes on performance and accuracy,” Matt Woods, AWS’ general manager of artificial intelligence, explained in a blog post. “SageMaker Search lets you quickly find and evaluate the most relevant model training runs from the potentially thousands of Amazon SageMaker model training runs, right from the AWS console.”

The search function is rolling out alongside support for the popular Git code management system. Companies can now host SageMaker projects in GitHub and AWS’ competing CodeCommit service, which is also based on Git, or use their self-hosted code repositories.

AWS also today announced a new version of its Snowball Edge appliance that features an onboard graphics processing unit. Snowball Edge is a data transport device aimed at the enterprise market. Companies can load up the system with up to 50 terabytes of information and then physically ship it to an AWS data center. They can also perform processing on the information while it’s still on their premises.

The new GPU-powered option, which AWS calls Edge Compute Optimized, is geared toward video analysis and machine learning. It can be expected to make an appearance at the provider’s annual re:Invent conference next week. Last year’s event saw AWS introduce dozens of new services and features for its cloud platform, among them SageMaker.

Image: Tomas Cloer/Flickr