Fairness in Forecasting of Observations of Linear Dynamical Systems
This work addresses fairness in forecasting for subgroups affected by under-representation bias, offering a novel extension of predictive parity to dynamical systems.
The paper tackles under-representation bias in time-series forecasting by introducing subgroup and instantaneous fairness notions, and presents globally convergent methods for fairness-constrained learning with reduced run times. Empirical results on biased insurance and COMPAS datasets demonstrate the efficacy of these methods.
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.