ML Hands On: Optical Character Recognition

In this workshop, you will create a notebook that does optical character recognition on hand-written digits. We will be using Logistic Regression, which is a very basic general purpose classification algorithm. It has been a popular choice for more than half a century. Even in our current era with the explosive advances in neural networks, Logistic Regression remains a mainstay for its simplicity and speed. Applying this very simple algorithm to the complex task of optical character recognition will help you see how approachable machine learning can be. To prepare for the class, create an Amazon AWS account and go to Amazon SageMaker. Create a notebook instance with a name you will remember and instance type c5.xlarge. All the rest of the settings can be left as defaults. Once it is working, make sure you shut it down so the notebook instance status says "Stopped". That way you will not be charged for it while you are not using it. Bring a laptop with you.