Many public policies depend on the results of predictions such as how groups or individuals will respond to policy changes, or even which groups or individuals are most likely to be affected. Despite this need for predictions, much of the research on the effects of public policies relies on causal inference, focusing on understanding a causal effect. On the other hand, several recent research projects focus on using machine learning to generate predictions that are relevant to improving public policy making. In this talk, I will give a brief introduction and overview of the recent types of public policy research projects that use machine learning to make policy-relevant predictions.