Decorrelation using Optimal Transport

TL;DR

This paper introduces a novel decorrelation method using Convex Neural Optimal Transport Solvers to effectively remove correlations between features and protected attributes, especially excelling in high-dimensional spaces like jet classification.

Contribution

The paper presents a new optimal transport-based decorrelation technique that outperforms existing methods in high-dimensional and multiclass scenarios.

Findings

Achieves near state-of-the-art decorrelation in binary classification.

Significantly better decorrelation performance in multiclass settings.

Demonstrates effectiveness in high energy physics jet classification.

Abstract

Being able to decorrelate a feature space from protected attributes is an area of active research and study in ethics, fairness, and also natural sciences. We introduce a novel decorrelation method using Convex Neural Optimal Transport Solvers (Cnots) that is able to decorrelate a continuous feature space against protected attributes with optimal transport. We demonstrate how well it performs in the context of jet classification in high energy physics, where classifier scores are desired to be decorrelated from the mass of a jet. The decorrelation achieved in binary classification approaches the levels achieved by the state-of-the-art using conditional normalising flows. When moving to multiclass outputs the optimal transport approach performs significantly better than the state-of-the-art, suggesting substantial gains at decorrelating multidimensional feature spaces.