AWS Introduces SageMaker Machine Learning Platform. Amazon Web Services (AWS), which is a leader in artificial intelligence/machine-learning, has been busy expanding SageMaker’s platform for creating and managing ML models. Amazon SageMaker, which leverages the scaling capabilities of AWS cloud, is described by the company to be an end-to-end service that helps data scientists, developers, and machine learning experts quickly create, train, and host machine-learning models for production applications. AWS has recently increased the SageMaker platform in a variety of ways to demonstrate its commitment to the ML space.

• Working with AWS glue. On Friday, SageMaker was integrated with AWS Glue. This fully managed extract, transform and load (ETL), service from AWS cloud allows customers to prepare and load data for analysis. AWS announced that you can now create an Amazon SageMaker Notebook from the AWS Glue console and connect it with an AWS Glue Development endpoint. These endpoints can be used by customers to interactively create, debug, and test AWS Glue ETL programs. AWS stated that this integration allows customers to use Amazon SageMaker’s fully managed notebooks rather than provisioning and managing their own notebook servers. This makes it easier and quicker to begin developing your AWS Glue scripts.
• Image classification update. AWS announced last week an enhancement to SageMaker’s built-in image classifier algorithm. This is intended to address one of the most common uses cases for ML modeling. The new functionality includes support for multi-label inputs as well as mixed-precision mode to speed up training. AWS stated that the Image Classification algorithm now accepts an image as an input, and can add multiple labels to it. This is useful because images often contain multiple elements. Multiple labels can help produce better and more useful results. This can be used to classify and detect different elements in medical images for better treatment and detection. “The mixed-precision mode also provides faster model training and lower cost, since accuracy is maintained even though less memory is used during modeling.
• AWS GovCloud availability. SageMaker is now available in the GovCloud (US Region) region, according to the company. AWS GovCloud is designed to host sensitive data and regulated workloads, and meet strict compliance and security requirements of the U.S. government. It targets U.S. government customers as well as contractors and educational institutions who need to run sensitive workloads on the cloud.

We reported earlier on SageMaker’s debut in the company’s re-Invent conference and noted a refreshment of SageMaker frameworks and algorithms in July. Keep checking back for more AWS updates on SageMaker as AWS seeks to keep pace in the rapidly growing AI/ML arena.