LAS VEGAS – Amazon Web Services (AWS) chief executive officer Andy Jassy highlighted 27 new services during his keynote at the company’s eighth annual re:Invent learning and education conference and urged customers not to procrastinate in starting their digital transformations.
He told the 65,000 attendees that not all of the key components in the process are technology elements.
First, he said, the senior leadership team needs to be aligned behind the effort. Without that, it’s easy for dissenters to block the initiatives.
Second, set top-down aggressive goals. It’s easy to go for a long time just dipping a toe into the water if those goals aren’t present and pursued.
Third, train builders.
And fourth, don’t let the organization get paralyzed by the notion of the effort involved before you even start.
Machine learning (ML) featured heavily in the new services unveiled that will help with the technology components of the transformations.
“For the first several years of ML, developers and companies were so passionate and excited to get value from data that they would use clunky tools,” he said, pointing out that there were no integrated development environments (IDEs) for ML as there are for programmers doing more traditional development.
He then announced SageMaker Studio, an IDE for Amazon SageMaker, the company’s managed service that provides developers and data scientists with the ability to build, train, and deploy machine learning models quickly. It supports the major machine learning frameworks – TensorFlow, PyTorch, Apache MXNet, Chainer, Keras, Gluon, Horovod, Scikit-learn, and Deep Graph Library – and provides a single, web-based visual interface where developers can perform all ML development steps. It gives them complete access, control, and visibility into each step required to build, train, and deploy models, including uploading data, creating new notebooks, training and tuning models.
Along with SageMaker Studio, Jassy announced a series of additional tools, all of which can be accessed through the SageMaker Studio interface. They included SageMaker Notebooks, SageMaker Experiments, SageMaker Debugger, SageMaker Model Manager, and SageMaker AutoPilot.
“We’re always going to give you the major tools to do your job,” he noted.
Added Dr. Matt Wood, vice-president, artificial intelligence at AWS, “The dirty secret of ML is that you don’t train one model, you train dozens and pick the best. For the first time, SageMaker Studio pulls it all together.”
Eric Gales, country leader, commercial sales, at AWS Canada said there’s a lot of interest around SageMaker among his customers.
“It makes machine learning easier to use,” he said. “And it’s more accessible to developers.”
François Côté, chief digital and technology officer at Montreal-based Fairstone Financial is also excited about how the technology is evolving to give more users access to it than they would have with traditional resources.
“AI is important to us as a lender,” he said, adding that it’s important that there are no biases in the models that lead to discrimination in loans. “SageMaker and a lot of automated models will put capabilities in more hands. If we have a strong governance model around the data, we will continue to do things the right way.”
Added Woods, “Our mission for ML is to take this magical technology, demystify it, and put it in as many hands as possible. We’ve invested in improving the underlying tasks, making sure AWS is the best place to run the workloads.”
