AWS Adds Governance, New Capabilities to Amazon SageMaker
- By DSAITrends editors
- December 07, 2022
Amazon Web Services (AWS) has announced a slew of new capabilities for its managed machine learning service Amazon SageMaker, adding improved governance, support for Model Cards, a new Model Dashboard, and new capabilities in Amazon SageMaker Studio Notebook.
The process of building, training, and deploying models is often tedious and requires continuous iteration by data scientists – sometimes for months – before a model is production-ready.
Amazon SageMaker was launched in 2017 to help practitioners simplify and accelerate ML deployments in the cloud, and has been regularly upgraded to meet evolving data science needs.
Multiple new capabilities
This latest announcement sees Amazon SageMaker enhanced with new capabilities to help customers more easily scale governance across the ML model lifecycle. This is crucial as the growing number of models and users make it harder to set least-privilege access controls and establish governance processes to document key model information, says AWS.
Specifically, Amazon SageMaker Role Manager makes it easier for administrators to control access and define permissions for users. Prebuilt templates based on various user roles and responsibilities can be created and populated with the relevant access policies and permissions to reduce the time and effort to onboard and manage users over time.
Importantly, the new Amazon SageMaker Model Cards now allow data science users to auto-populate training details like input datasets, training environment, and training results directly to a single location in the AWS console. Amazon SageMaker Model Dashboard also provides a central interface to track ML models over time.
Finally, data scientists can now access a fully managed notebook experience within Amazon SageMaker itself. According to AWS, Studio Notebook now gives practitioners a fully managed notebook experience that includes support for collaboration across data science teams, simplified data preparation with the ability to import datasets directly into notebooks, and automatic conversion of notebook code to production-ready jobs.
“The new Amazon SageMaker capabilities announced today make it even easier for teams to expedite the end-to-end development and deployment of ML models,” said Bratin Saha, vice president of Artificial Intelligence and Machine Learning at AWS.
“From purpose-built governance tools to a next-generation notebook experience and streamlined model testing to enhanced support for geospatial data, we are building on Amazon SageMaker’s success to help customers take advantage of ML at scale.”
Image credit: iStockphoto/urfinguss