Longitudinal Fairness with Censorship

Recent works in artificial intelligence fairness attempt to mitigate discrimination by proposing constrained optimization programs that achieve parity for some statistic. Most assume availability of the class label, which is impractical many real-world applications such as precision medicine, actuarial analysis and recidivism prediction. Here we consider longitudinal right-censored environments, where time event might be unknown, resulting censorship label inapplicability existing studies. We devise applicable measures, propose a debiasing algorithm, provide necessary theoretical constructs bridge with without these important socially-sensitive tasks. Our experiments on four censored datasets confirm utility our approach.