Machine learning is useful for automating all or part of a business process. The ease with which networks can be trained to solve new problems increases the speed at which your organization can adapt to new situations. The Mathematical Neuro-Oncology Lab at the Mayo Clinic recently added a deep neural net trained to recognize the modality of an magnetic resonance image (MRI) to its data intake process. We found performance improved by three orders of magnitude and the error rate dropped by over 80%. This presentation will describe our use of machine learning and why it fits well with our task, as well as other problems we are planning to solve with it. Machine learning is not a silver bullet but it works well under the right conditions.

Learning Objectives:

• Learn the incredible results that can be achieved using machine learning
• Learn the concerns that need to be addressed to use machine learning
• Come away with realistic expectations about process automation using machine learning