Latent Space Smoothing for Individually Fair Representations

Momchil Peychev, Anian Ruoss, Mislav Balunovic, Maximilian Baader, Martin Vechev

Keywords: fairness, fair representation learning, adversarial fairness, trustworthy machine learning, randomized smoothing

Abstract: Fair representation learning encodes user data to ensure fairness and utility, regardless of the downstream application. However, learning individually fair representations, i.e., guaranteeing that similar individuals are treated similarly, remains challenging in high-dimensional settings such as computer vision. In this work, we introduce LASSI, the first representation learning method for certifying individual fairness of high-dimensional data. Our key insight is to leverage recent advances in generative modeling to capture the set of similar individuals in the generative latent space. This allows learning an individually fair representation where similar individuals are mapped close together, by using adversarial training to minimize the distance between the representations of similar individuals. Finally, we employ randomized smoothing to provably map similar individuals close together, in turn ensuring that local robustness verification of the downstream application results in end-to-end fairness certification. Our experimental evaluation on challenging real-world image data demonstrates that our method increases certified individual fairness by more than 60%, without significantly affecting task utility.

One-sentence Summary: In this work we certify individual fairness of high dimensional datasets via randomized smoothing in the latent space of a generative model.

Community Implementations: [ 1 code implementation](https://www.catalyzex.com/paper/arxiv:2111.13650/code)