Rosie Jones is a Director of Computational Advertising at Akamai, based in Boston, MA. Her research interests include computational advertising, information retrieval and web search.
12:00 - 12:30 arrival + lunch served
12:30 - 13:30 Rosie's talk
13:30 - 14:00 informal discussion
Machine Learning for Display Advertising with Behavioral Targeting
Deciding which banner ad to show to which person at which time opens up a variety of interesting machine learning problems. Predicting future buyers is a good strategy, but marketers also wish to expand their reach to new audiences, different from previous buyers, necessitating training with proxy labels. We show that shopping data can be used to model demographic audiences, as well as audiences with more complex profiles. Even when we have ground truth, evaluating the effectiveness of a live advertising campaign goes beyond measures of prediction accuracy, since we also wish to understand whether the advertising is having a causal effect on shopping outcomes. Finally I'll discuss incrementality modeling - the problem of learning how to target an audience in order to effect the biggest change in outcomes.