Contrastive Examples for Addressing the Tyranny of the Majority

Abstract

Computer vision algorithms, e.g. for face recognition, favour groups of individuals that are better represented in the training data. This happens because of the generalization that classifiers have to make. It is simpler to fit the majority groups as this fit is more important to overall error. We propose to create a balanced training dataset, consisting of the original dataset plus new data points in which the group memberships are intervened, minorities become majorities and vice versa. We show that current generative adversarial networks are a powerful tool for learning these data points, called contrastive examples. We experiment with the equalized odds bias measure on tabular data as well as image data (CelebA and Diversity in Faces datasets). Contrastive examples allow us to expose correlations between group membership and other seemingly neutral features. Whenever a causal graph is available, we can put those contrastive examples in the perspective of counterfactuals.

References in corpus (6)

• Equality of Opportunity in Supervised Learning
• CyCADA: Cycle-Consistent Adversarial Domain Adaptation
• Data Decisions and Theoretical Implications when Adversarially Learning Fair Representations
• The Frontiers of Fairness in Machine Learning
• Learning Adversarially Fair and Transferable Representations
• InclusiveFaceNet: Improving Face Attribute Detection with Race and Gender Diversity

Cited by in corpus (2)

• Towards Intersectionality in Machine Learning: Including More Identities, Handling Underrepresentation, and Performing Evaluation
• Causal datasheet: An approximate guide to practically assess Bayesian networks in the real world