AWS unveils machine learning (ML) tools for data science in the cloud

Check out the on-demand sessions from the Low-Code/No-Code Summit to learn how to successfully innovate and achieve efficiency by upskilling and scaling citizen developers. Watch now.


Artificial intelligence (AI) and machine learning (ML) workloads can run in any number of locations including on-premises, at the edge, embedded in devices and in the cloud.

Amazon Web Services (AWS) is hoping that more often than not organizations will choose the cloud, where it is offering a growing array of services. At the AWS re:invent 2022 event in Las Vegas today, the company detailed parts of its AI/ML strategy and announced a dizzying lineup of feature updates and new services to help organizations to better use the cloud for data science.

The cornerstone of the AWS AI/ML portfolio is the SageMaker suite of services. In a keynote address at AWS re:invent Swami Sivasubramanian, VP database, analytics and ML at AWS said that SageMaker enables organizations to build, train and deploy ML models for virtually any use case and has tools for every step of ML development. 

“Tens of thousands of customers are using SageMaker ML models to make more than a trillion predictions a month,” Sivasubramanian said. “Our customers are solving complex problems with SageMaker by using that data to build ML models ranging from optimizing driving routes for rideshare apps to accelerating drug discovery.”

Event

Intelligent Security Summit

Learn the critical role of AI & ML in cybersecurity and industry specific case studies on December 8. Register for your free pass today.

Register Now

Geospatial ML comes to SageMaker

One area where SageMaker’s feature set is now being improved is with enhanced geospatial ML capabilities.

Sivasubramanian said that geospatial data can be used for a wide variety of use cases. For example it can be used for helping to optimize an agricultural harvest yield, assisting with planning for sustainable urban development and can be used to identify a new location or region for a business to open.

“Accessing high-quality geospatial data to train ML models requires working with multiple data sources and multiple vendors,” he said. “These data sets are typically massive and unstructured, which needs time consuming data preparation before you can even start writing a single line of code to build your ML models.”

With the new geospatial support in SageMaker, AWS is aiming to make it easier for organizations to actually build and deploy models. Sivasubramanian said that the new support will enable users to access geospatial data in SageMaker from different data sources with just a few clicks. 

Data preparation tooling for geospatial is now integrated with SageMaker to help users process and enrich large datasets. SageMaker now also benefits from integrated visualization tools, enabling users to analyze data and explore model predictions on an interactive map using 3D accelerated graphics. 

Sivasubramanian added that SageMaker now also provides built-in pretrained neural nets to accelerate model building for many geospatial common use cases. 

ML Governance gets a boost

As organizations are increasingly bringing ML into different processes, there is a growing need for collaboration across groups. 

Building out the permissions and governance rules that enable model sharing is another area where AWS is looking to help its users with new capabilities in the Amazon SageMaker ML Governance service. The new services include SageMaker Role Manager, Model Cards and Model Dashboard.

Sivasubramanian said that SageMaker Role Manager helps organizations to define critical permissions for users, with automated policy creation tools. The Model Cards service is all about creating a central authoritative location for ML model documentation. The new Model Dashboard now provides organizations with visibility and unified monitoring for the performance of ML models. 

“These are really powerful governance capabilities that will help you build ML governance responsibly,” Sivasubramanian said.

VentureBeat’s mission is to be a digital town square for technical decision-makers to gain knowledge about transformative enterprise technology and transact. Discover our Briefings.