What’s Fair in Fair Machine Learning?

Recent highly publicized failures in machine learning, and analytics more generally, have shown that biases or unfairness in data can sneak into, and be magnified by, our models, leading to harmful, incorrect predictions being made once the models are deployed into the real world. While the problems themselves are becoming apparent, many questions remain around how fairness should be defined in a data science context, whether it can be quantified, and how concerns for it should be integrated into existing analytics pipelines. In this talk, we will present a framework for better understanding how issues of fairness overlap with analytics as well as how we can improve our analytics pipelines to make them more interpretable, reproducible, and fair to the groups that they are intended to serve.