Datasets can be biased due to societal inequities, human biases, under-representation of minorities, etc. Our goal is to prove that models produced by a learning algorithm are robust in their predictions to potential dataset biases. This is a challenging problem: it entails learning models for a large, or even infinite, number of datasets, ensuring that they all produce the same prediction.
In this talk, I will show how we can adapt ideas from program analysis to prove robustness of a decision-tree learner on a large, or infinite, number of datasets, certifying that each and every dataset produces the same prediction for a specific test point. We evaluate our approach on datasets that are commonly used in the fairness literature, and demonstrate our approach’s viability on a range of bias models.