Abstract: With the increased use of black-box models, analytics management needs to better understand the options available to them and also their teams. The two primary options are building transparent from the start (interpretable machine learning) or adding an explanatory layer to a black- box model (Explainable AI, also called XAI). We will discuss the advantages of each, and some of the reasons that black-box models are prevalent, even in regulated industries. We will also discuss the two primary kinds of model explanations: local and global. Each will be defined, with examples. Several popular and recent techniques will be overviewed, including SHAP, LIME, counterfactuals, and surrogate models. Despite the brevity of the talk, attendees will leave with a solid overview and a basic taxonomy of popular model explainability options.
Bio: Keith McCormick serves as CloudFactory’s Chief Data Science Advisor. He's also an author, LinkedIn Learning contributor, university instructor, and conference speaker. Keith has been building predictive analytics models since the late 90s. More recently his focus has shifted to helping organizations build and manage their data science teams.