Fair Inference on Outcomes

In this paper, we consider the problem of fair statistical inference involving outcome variables. Examples include classification and regression problems, and estimating treatment effects in randomized trials or observational data. The issue of fairness arises in such problems where some covariates or treatments are "sensitive," in the sense of having potential of creating discrimination. In this paper, we argue that the presence of discrimination can be formalized in a sensible way as the presence of an effect of a sensitive covariate on the outcome along certain causal pathways, a view which generalizes (Pearl 2009). A fair outcome model can then be learned by solving a constrained optimization problem. We discuss a number of complications that arise in classical statistical inference due to this view and provide workarounds based on recent work in causal and semi-parametric inference.

[1]  G. G. Stokes.  "J." , 1890, The New Yale Book of Quotations.

[2]  Ilya Shpitser,et al.  Counterfactual Graphical Models for Longitudinal Mediation Analysis With Unobserved Confounding , 2012, Cogn. Sci..

[3]  J. Robins.  A new approach to causal inference in mortality studies with a sustained exposure period—application to control of the healthy worker survivor effect , 1986 .

[4]  Jin Tian,et al.  On the Identification of Causal Effects , 2015 .

[5]  H. Chipman,et al.  Bayesian Additive Regression Trees , 2006 .

[6]  J. Pearl.  Causality: Models, Reasoning and Inference , 2000 .

[7]  James M. Robins,et al.  Marginal Structural Models versus Structural nested Models as Tools for Causal inference , 2000 .

[8]  Faisal Kamiran,et al.  Quantifying explainable discrimination and removing illegal discrimination in automated decision making , 2012, Knowledge and Information Systems.

[9]  Nathan Srebro,et al.  Equality of Opportunity in Supervised Learning , 2016, NIPS.

[10]  Carlos Eduardo Scheidegger,et al.  Certifying and Removing Disparate Impact , 2014, KDD.

[11]  Eric J. Tchetgen Tchetgen,et al.  Semiparametric Estimation of Models for Natural Direct and Indirect Effects , 2011 .

[12]  Jin Tian,et al.  On the Testable Implications of Causal Models with Hidden Variables , 2002, UAI.

[13]  Avi Feller,et al.  Algorithmic Decision Making and the Cost of Fairness , 2017, KDD.

[14]  Judea Pearl,et al.  Direct and Indirect Effects , 2001, UAI.

[15]  I. Shpitser,et al.  CAUSAL INFERENCE WITH A GRAPHICAL HIERARCHY OF INTERVENTIONS. , 2014, Annals of statistics.

[16]  N. Metropolis,et al.  Equation of State Calculations by Fast Computing Machines , 1953, Resonance.

[17]  Stijn Vansteelandt,et al.  Odds ratios for mediation analysis for a dichotomous outcome. , 2010, American journal of epidemiology.

[18]  J. Robins,et al.  Identifiability and Exchangeability for Direct and Indirect Effects , 1992, Epidemiology.

[19]  Ilya Shpitser,et al.  Quantifying an Adherence Path-Specific Effect of Antiretroviral Therapy in the Nigeria PEPFAR Program , 2014, Journal of the American Statistical Association.

[20]  Lu Zhang,et al.  A Causal Framework for Discovering and Removing Direct and Indirect Discrimination , 2016, IJCAI.

[21]  T. Richardson.  Single World Intervention Graphs ( SWIGs ) : A Unification of the Counterfactual and Graphical Approaches to Causality , 2013 .

[22]  T. VanderWeele,et al.  On the causal interpretation of race in regressions adjusting for confounding and mediating variables. , 2014, Epidemiology.

[23]  Franco Turini,et al.  Discrimination-aware data mining , 2008, KDD.

[24]  Eric J. Tchetgen Tchetgen,et al.  Semiparametric Estimation of Models for Natural Direct and Indirect E ¤ ects by , 2016 .

[25]  J. Pearl.  The Causal Mediation Formula—A Guide to the Assessment of Pathways and Mechanisms , 2012, Prevention Science.

[26]  Judea Pearl,et al.  Identification of Joint Interventional Distributions in Recursive Semi-Markovian Causal Models , 2006, AAAI.

[27]  Eric J. Tchetgen Tchetgen,et al.  On Partial Identification of the Pure Direct Effect , 2015, 1509.01652.

[28]  Tom Burr,et al.  Causation, Prediction, and Search , 2003, Technometrics.

[29]  Judea Pearl,et al.  Complete Identification Methods for the Causal Hierarchy , 2008, J. Mach. Learn. Res..

[31]  Teva J. Scheer.  Uniform Guidelines on Employee Selection Procedures , 2007 .

[32]  H. Chipman,et al.  BART: Bayesian Additive Regression Trees , 2008, 0806.3286.

[33]  Adrian F. M. Smith,et al.  Bayesian Analysis of Constrained Parameter and Truncated Data Problems , 1991 .

[34]  Judea Pearl,et al.  Probabilistic reasoning in intelligent systems , 1988 .

[35]  Judea Pearl,et al.  Missing Data as a Causal and Probabilistic Problem , 2015, UAI.

[36]  Ilya Shpitser,et al.  Semiparametric Theory for Causal Mediation Analysis: efficiency bounds, multiple robustness, and sensitivity analysis. , 2012, Annals of statistics.