Sponsors

Smart Apps with Azure ML CHRIS MCHENRY VP OF TECHNOLOGY, INTEGRO

“Machine learning is a way of getting computers to know things when they see them by producing for themselves the rules their programmers cannot specify. The machines do this with heavy-duty statistical analysis of lots and lots of data.” “Machine Learning: Field of study that gives computers the ability to learn without being explicitly programmed.” Arthur Samuel (1959) “A computer program is said to learn from experience E with respect to some task T and some performance measure P, if its performance on T, as measured by P, improves with experience E.” Tom Mitchell (1998) “A breakthrough in Machine Learning would be worth ten Microsoft’s” Bill Gates

ML Examples FROM THE PRESS Spam Filtering Google/Bing Ad Targeting Postal Service Mail Sorting Cortana Amazon/Netflix Recommendations Credit Card Fraud Detection Deep Blue/Watson How-Old.net BUSINESS APPS SMART APPS Automated Workflow Routing Automated Filing User Suggestions Customers Likely to Buy Customers Likely to Leave Product Pricing Order Anomalies

Applied ML – Skills Needed BYOD ◦Bring Your Own Development skills ◦REST Data Processing/Cleansing ◦SQL/NoSQL ◦R and/or Python ◦Hadoop/HD Insight/Azure Stream Analytics The Right Attitude ◦Persistence and confidence to understand a complex subject ◦Unbridled curiosity to explore and iterate and possibly fail ◦Creativity to find alternatives when you are blocked

Process

ML Studio Workspace Experiment - Modules ◦Training ◦Scoring DataSet ◦Direct Upload – 10GB Limit ◦Reader – Azure Blob, Web Page, Odata, SQL Azure, Hive, etc ◦R or Python Module Web Services

Regression

Classification

Clustering

Demo 1.Create a Training Experiment – Select a Model 2.Create a Scoring Experiment – Prep Selected Model for Runtime 3.Publish as a Web Service – Operationalize a Web Service 4.Consume a Web Service – Get Predictions from your App

Common ML Challenges UNDERFITTING - BIASOVERFITTING - VARIANCE 1.Add more features 2.Generate features 3.Evaluate training data 1.Reduce features – dimensionality reduction 2.Add more training data 3.Evaluate training data

Ecosystem Site/ML Studio/Docs: Gallery: Azure Marketplace: Blog: Forum: Stack Overflow: Webinars:

Books Predictive Analytics with Microsoft Azure Machine Learning: Build and Deploy Actionable Solutions in Minutes– Barga, Tok, and Fontama, Apress, 2014 Azure Machine Learning – Jeff Barnes, Microsoft Press, 2015 Data Science in the Cloud with Microsoft Azure Machine Learning and R – Stephen Elston, O’Reilly, 2015

Questions Contact