Updating before the big day

Machine learning for novices: AWS SageMaker update brings new collaboration features

Sarah Schlothauer
machine learning
© Shutterstock / Aleutie

Machine learning is complicated, but it’s becoming easier and easier to grasp with new tools and platforms. In preparation for AWS re:Invent 2018, AWS updated SageMaker. What are the newest features?

Las Vegas is prepping for the AWS re:Invent conference. From 26 – 30, 2018, Amazon Web Services will host an event full of seminars and keynotes out in the desert oasis. In anticipation of re:Invent, Amazon updated SageMaker– their machine learning platform. This platform promises to bring ease of learning to the complexities of machine learning.

Let’s see what the new update adds!

What’s new?

This update focuses mostly on automation, orchestration, and easier collaboration with your team. Here are the newest features:

  • SageMaker Workflows: collaborate together with your team and share feedback on your ML models
  • SageMaker Search: search through training runs and get the most relevant data faster than ever
  • IP Insights: Detect suspicious IP addresses and log-ins to avoid potential account takeovers
  • Git integration: link GitHub, AWS CodeCommit, and other Git repos
  • Step Functions: automate and orchestrate steps
  • New algorithms and frameworks: Add to the already robust amount! The update includes Horovod, Spark MLap, and scikit-learn
  • Updated compliance standards

SEE ALSO: Fitting models with AWS AI

Explore of the new additions here.

Machine learning features galore

So, now that some time has passed since SageMaker’s debut, let’s explore what it brought to the table and why it is changing the ML scene.

Some of the benefits of using SageMaker include the ability to choose frameworks and algorithms. If you have a working familiarity with Apache MXNet, Chainer, TensorFlow, Caffe2, or Torch they already come bundled in and ready for use. If the amount of bundled in algorithms aren’t enough, you can use a Docker container to bring in more.

SageMaker also promises ease of integration, fast performance, one-click training, automatic model tuning, and easy deployment.

Overall, it is best suited for research with large datasets. If you aren’t going to focus on a large amount of data, chances are that SageMaker is not the tool for your particular use case. Some of the appropriate use cases are ad targeting, quality prediction, and supply chain forecasting.

Cost is another factor that may hold you back, especially if you are just looking to explore machine learning as a novice. One more roadblock is that it is not currently available in all regions. Before you fall in love, check out the regional table and make sure you can use it first. (AWS continuously adds regions, so don’t despair if you aren’t included yet.)

Looking for inspiration? Browse these examples on GitHub. Explore the Jupyter notebooks and see how the data is used.

SEE ALSO: AI and machine learning in software development: Benefits for developers

Jump into SageMaker

The claim that AWS SageMaker is easy to learn captivates novices looking to break into the ML sphere. From their overview, it “removes the complexity that holds back developer success”. Part of this ease of learning is due to the built in scalable algorithms.

During last year’s re:Invent, DigitalGlobe discussed how their team used SageMaker even without previous machine learning experience. Novices new to the ML world will have a fairly easy time grasping how to use SageMaker and build better ML products. (In theory, of course! Perhaps in another year we will see if the claim that SageMaker is best suited for the masses is true when we have a larger user base. ML novices are welcome to share their accounts with us in the comments.)

For beginners, plenty of (free!) tutorials exist explaining how to best use SageMaker. Here are just a few of what we recommend dipping into it. Take a swim and come out with more ML knowledge!

Sarah Schlothauer

Sarah Schlothauer

All Posts by Sarah Schlothauer

Sarah Schlothauer is the editor for She received her Bachelor's degree from Monmouth University, West Long Branch, New Jersey. She currently lives in Frankfurt, Germany with her husband and cat where she enjoys reading, writing, and medieval reenactment. She is also the editor for Conditio Humana, an online magazine about ethics, AI, and technology.

Inline Feedbacks
View all comments