Amazon SageMaker continues to expand the use of machine learning (ML) in the cloud

Amazon SageMakerstarted operating 5 years ago, is one of the most widely used machine learning (ML) services in existence.

Back in 2017, Sagemaker was a unique service designed to help organizations use the cloud to train ML models. Just as Amazon Web Services (AWS) has grown significantly over the past 5 years, so has the number of ML services in the Sagemaker portfolio.

In 2018, Amazon SageMaker GroundTruth added data labeling capabilities. In 2019, AWS expanded SageMaker with a number of services including SageMaker Studio, which provides an integrated developer environment (IDE) for data scientists to build ML application workflows. The SageMaker Data Wrangler service announced in 2020 for data preparation and in 2021 new capabilities include Clarifying explainability and ML feature store services.

AWS is continuing to add services to SageMaker, including a pair of announcements made yesterday, with new support for AWS Graviton cloud instances and multi-model endpoint support. During an AWS event on October 26, Bratin Saha, VP and general manager of AI/ML at AWS, said that more than 100,000 customers from virtually every industry use AWS cloud ML services.

“Machine learning is not the future that we need to plan for, but the present that we need to exploit right now,” Bratin said.

AWS scales SageMaker with multi-model endpoints (MME)

One of the things that has happened in the last 5 years with SageMaker adoption is an increase in scale to the way models are trained and deployed.

To help organizations cope with the scaling challenge, Bratin said that AWS has released SageMaker’s multi-model endpoint (MME) capability.

“This allows a single GPU to store thousands of models,” says Bratin. “Many of the most common use cases for machine learning, such as personalization, require you to manage anywhere from a few hundred to hundreds of thousands of models.”

For example, Bratin says that in the case of taxi services, an organization can have customized models based on each city’s traffic patterns. He noted that in a traditional machine learning system, customers would have to deploy one model per instance, which means they would have to deploy hundreds or thousands of instances.

The SageMaker MME changes needs, giving organizations the ability to host multiple models on a single instance, reducing overall costs. Bratin says the MME service also handles all the work of orchestrating ML model traffic and uses sophisticated caching algorithms to understand which model should be in memory at a particular time. .

How a company continues to benefit from SageMaker

Among the many users of the Amazon SageMaker service is Mueller Water Products.

Mueller Water Products is using Amazon SageMaker on a mission to limit water loss. Use ML service together with EchoShoe-DX . System for leak detection, the company was able to improve accuracy by 40%.

“AWS was really able to merge the disparate needs in a machine learning environment into a really powerful set of tools for our team to use,” said Dave Johnston, director of intelligent infrastructure at Mueller. Water Products, told VentureBeat.

Johnston says that many organizations, including utility organizations, have more data than they know what to do. In his view, with the ML tools that AWS has developed in SageMaker, there are many opportunities, not only for the water industry but for many different industries.

“There is a lot of hidden value in the data that has been collected, and there will be plenty of opportunities to unlock that value,” says Johnston. “I think [SageMaker is] a low-cost approach to unlocking hidden value without having to deploy a bunch of new, expensive infrastructure, and you can do it with the data you’ve already collected. “