Connections

NEW YORK — Amazon Web Services (AWS) has observed “three major trends” as its customers adopt machine learning (ML) in growing numbers, according to Dan Mbanga, AWS global lead business development manager of Amazon AI Platforms.

“AI is here to stay,” he said at the Artificial Intelligence (AI) Conference May 1, citing Amazon’s Alexa virtual assistant as one of the many ways in which the company is using AI technology.


AWS is seeing “rapid AI experimentation” by its customers that’s leading to increased deployment, he noted.

The first main trend as developers increasingly adopt ML is that “data matters at any scale,” he told attendees, adding: “Every single piece and item of data that you have is important and you want to leverage that.” Developers, meanwhile, are also embracing ML in “every size and every shape,” and they “love speed,” he said, citing the other two main trends.

Developers are looking to “iterate rapidly” after using AI and some of them want to “iterate on their own models; they want to be able to build and train their own models without having to figure out how to build a whole infrastructure behind that,” he said.

“That’s why we built Amazon SageMaker,” he said, referring to the “end-to-end machine learning platform” the company introduced last year. Developers are able to significantly decrease their “heavy lifting” by using SageMaker for ML, he said. While introducing SageMaker last year, AWS CEO Andy Jassy said that service was “an easy way to build, train and deploy machine learning models for everyday developers.”

In announcing its results for the first quarter (ended March 31), Amazon said April 26: “Tens of thousands of customers are using AWS machine learning services, with active users increasing more than 250% in the last year, spurred by the broad adoption” of SageMaker.

SageMaker’s benefits were also touted at the conference by Randall Hunt, AWS senior technical evangelist, who explained how the platform can be used to build scalable ML workflows.

“You don’t need to be an expert in any particular realm in order to start following best practices” when using SageMaker’s suite of tools because he said they “allow you, as a developer or as an engineer or as a manager, to go in and deploy machine learning models.”

AI has helped make it possible to start solving “a lot of problems that were previously outside of the reach of infrastructure,” he went on to say, noting
Amazon has “almost a 20-year history of working” with AI.

Hunt also had a suggestion for developers looking for success with AI: “You don’t need to be dedicating your life to AI research to be able to bring a unique perspective to it and be able to advance the field a little bit. What you need is creativity in a lot of these things. You don’t necessarily need hardcore data science or hardcore stats. You really need to have a business problem that no one else has attempted to solve yet – or even just a personal problem that no one else has attempted to solve yet – and you can use that kind of intuition that you have about that problem to attack it in a new way.”

He’s seen examples of this happen before, “especially in the healthcare space,” he said.

Like last year’s conference, one of the main takeaways from the first day of this year’s conference was that although we’re still in the relatively early days of AI, the technology has already been used across a wide range of enterprise sectors and by a wide range of companies and other organizations. And that stands to only continue growing in the years to come.