Machine Learning Introduction and Motivation ADAPTED FROM WORK

Machine Learning: Introduction and Motivation ADAPTED FROM WORK BY RAYID GHANI WITH ADDITIONAL INFORMATION FROM DAVID KAUCHAK AND LEE GILES Julia Lane, Coleridge Initiative

What is Machine Learning ? Rayid Ghani “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. ” “changes in [a] system that. . . enable [it] to do the same task or tasks drawn from the same population more efficiently and more effectively the next time. '' (Simon 1983) What is Machine Learning What does it do? How is it used in public policy?

What is Machine Learning? Machine learning is a field that develops algorithms designed to be applied to datasets, with the main areas of focus being prediction (regression), classification, and clustering or grouping tasks - Susan Athey Machine learning is programming computers to optimize a performance criterion using example data or past experience. -- Ethem Alpaydin The goal of machine learning is to develop methods that can automatically detect patterns in data, and then to use the uncovered patterns to predict future data or other outcomes of interest. -- Kevin P. Murphy The field of pattern recognition is concerned with the automatic discovery of regularities in data through the use of computer algorithms and with the use of these regularities to take actions. -- Christopher M. Bishop Data mining: machine learning applied to “databases”, i. e. collections of data; Statistics: inference and/or estimation; engineering: pattern recognition; electrical engineering: signal processing What is Machine Learning What does it do? How is it used in public policy?

What is Machine Learning? Machine learning is about predicting the future based on the past. -- Hal Daume III past Training Data What is Machine Learning future r a le n model/ predictor Testing Data What does it do? pr t c i d e model/ predictor How is it used in public policy?

Machine Learning is different from Regression Builds systems instead of analysis - Algorithmic - Focusses on out of sample/future performance - Applies to a broader set of tasks - Allows more variables to be used Learns by: 1. Acquiring new knowledge ◦ acquiring new facts ◦ acquiring new skills 2. Adapting its behavior ◦ solving problems more accurately ◦ solving problems more efficiently What is Machine Learning What does it do? How is it used in public policy?

Examples of Machine Learning generally What is Machine Learning What does it do? How is it used in public policy?

Machine Learning uses for public policy Government program management • Existing systems are rigid, inflexible and difficult to modify • Machine learning can automate, scale, and adapt to new information. ◦ Lots of data is often available to “train” the system ◦ Business experts are better at “training” the system than “building” the system (i. e. giving feedback about results of model rather than have to create a set of rules) What is Machine Learning What does it do? How is it used in public policy?

Machine Learning uses for public policy Government policy development Description (Understand the past) ◦ Can I cluster firms into different sectors? Detection (Anomalies, Events, Patterns) ◦ Can I predict fraud? Prediction (Predict the Future) ◦ Can I detect who’s going to get lead poisoning early so that I can determine which home inspections to prioritize? Behavior Change (Causal Inference) ◦ What happens when I change options on a website? What is Machine Learning What does it do? How is it used in public policy?

How to use Machine Learning in Public Policy Practical steps ● Formulate your policy problem ● Map the problem to a Machine Learning problem ● Get (and integrate) the data you need to work on the problem ● Explore, process and clean the data ● Create “Features” you want to include in the model ● Select which methods to test ● Evaluate methods ● Deploy, Maintain, Update What is Machine Learning What does it do? How is it used in public policy?

Where Machine Learning fits in to this class Source: Big Data and Social Science, by Foster, Ghani, Jarmin, Kreuter, Lane What is Machine Learning What does it do? How is it used in public policy?