Timezone: »
The Equalized Odds (for short, EO) is one of the most popular measures of discrimination used in the supervised learning setting. It ascertains fairness through the balance of the misclassification rates (false positive and negative) across the protected groups -- e.g., in the context of law enforcement, an African-American defendant who would not commit a future crime will have an equal opportunity of being released, compared to a non-recidivating Caucasian defendant. Despite this noble goal, it has been acknowledged in the literature that statistical tests based on the EO are oblivious to the underlying causal mechanisms that generated the disparity in the first place (Hardt et al. 2016). This leads to a critical disconnect between statistical measures readable from the data and the meaning of discrimination in the legal system, where compelling evidence that the observed disparity is tied to a specific causal process deemed unfair by society is required to characterize discrimination. The goal of this paper is to develop a principled approach to connect the statistical disparities characterized by the EO and the underlying, elusive, and frequently unobserved, causal mechanisms that generated such inequality. We start by introducing a new family of counterfactual measures that allows one to explain the misclassification disparities in terms of the underlying mechanisms in an arbitrary, non-parametric structural causal model. This will, in turn, allow legal and data analysts to interpret currently deployed classifiers through causal lens, linking the statistical disparities found in the data to the corresponding causal processes. Leveraging the new family of counterfactual measures, we develop a learning procedure to construct a classifier that is statistically efficient, interpretable, and compatible with the basic human intuition of fairness. We demonstrate our results through experiments in both real (COMPAS) and synthetic datasets.
Author Information
Junzhe Zhang (Purdue University)
Elias Bareinboim (Purdue)
More from the Same Authors
-
2019 Poster: Near-Optimal Reinforcement Learning in Dynamic Treatment Regimes »
Junzhe Zhang · Elias Bareinboim -
2019 Poster: Efficient Identification in Linear Structural Causal Models with Instrumental Cutsets »
Daniel Kumor · Bryant Chen · Elias Bareinboim -
2019 Poster: Characterization and Learning of Causal Graphs with Latent Variables from Soft Interventions »
Murat Kocaoglu · Amin Jaber · Karthikeyan Shanmugam · Elias Bareinboim -
2019 Poster: Identification of Conditional Causal Effects under Markov Equivalence »
Amin Jaber · Jiji Zhang · Elias Bareinboim -
2019 Spotlight: Identification of Conditional Causal Effects under Markov Equivalence »
Amin Jaber · Jiji Zhang · Elias Bareinboim -
2018 : Datasets and Benchmarks for Causal Learning »
Csaba Szepesvari · Isabelle Guyon · Nicolai Meinshausen · David Blei · Elias Bareinboim · Bernhard Schölkopf · Pietro Perona -
2018 : Causality and Transfer Learning »
Elias Bareinboim -
2018 Poster: Structural Causal Bandits: Where to Intervene? »
Sanghack Lee · Elias Bareinboim -
2017 Poster: Experimental Design for Learning Causal Graphs with Latent Variables »
Murat Kocaoglu · Karthikeyan Shanmugam · Elias Bareinboim -
2016 : The Data-Fusion Problem: Causal Inference and Reinforcement Learning »
Elias Bareinboim -
2015 Poster: Bandits with Unobserved Confounders: A Causal Approach »
Elias Bareinboim · Andrew Forney · Judea Pearl -
2014 Poster: Transportability from Multiple Environments with Limited Experiments: Completeness Results »
Elias Bareinboim · Judea Pearl -
2014 Spotlight: Transportability from Multiple Environments with Limited Experiments: Completeness Results »
Elias Bareinboim · Judea Pearl -
2013 Poster: Transportability from Multiple Environments with Limited Experiments »
Elias Bareinboim · Sanghack Lee · Vasant Honavar · Judea Pearl -
2013 Tutorial: Causes and Counterfactuals: Concepts, Principles and Tools. »
Judea Pearl · Elias Bareinboim