Fairness in Forecasting of Observations of Linear Dynamical Systems

TL;DR

This paper introduces fairness notions for forecasting in dynamical systems, proposes globally convergent algorithms, and demonstrates their effectiveness on biased datasets including insurance and COMPAS data.

Contribution

It extends fairness concepts to time-series forecasting of dynamical systems and develops scalable convex optimization methods for fair learning.

Findings

Effective fairness-constrained forecasting methods demonstrated on real datasets.

Sparsity exploitation reduces computational complexity.

Methods achieve improved fairness without sacrificing accuracy.

Abstract

In machine learning, training data often capture the behaviour of multiple subgroups of some underlying human population. This behaviour can often be modelled as observations of an unknown dynamical system with an unobserved state. When the training data for the subgroups are not controlled carefully, however, under-representation bias arises. To counter under-representation bias, we introduce two natural notions of fairness in time-series forecasting problems: subgroup fairness and instantaneous fairness. These notions extend predictive parity to the learning of dynamical systems. We also show globally convergent methods for the fairness-constrained learning problems using hierarchies of convexifications of non-commutative polynomial optimisation problems. We also show that by exploiting sparsity in the convexifications, we can reduce the run time of our methods considerably. Our empirical results on a biased data set motivated by insurance applications and the well-known COMPAS data set demonstrate the efficacy of our methods.