ABSTRACT
Computer scientists have made great strides in characterizing different measures of algorithmic fairness, and showing that certain measures of fairness cannot be jointly satisfied. In this paper, I argue that the three most popular families of measures - unconditional independence, target-conditional independence and classification-conditional independence - make assumptions that are unsustainable in the context of an unjust world. I begin by introducing the measures and the implicit idealizations they make about the underlying causal structure of the contexts in which they are deployed. I then discuss how these idealizations fall apart in the context of historical injustice, ongoing unmodeled oppression, and the permissibility of using sensitive attributes to rectify injustice. In the final section, I suggest an alternative framework for measuring fairness in the context of existing injustice: distributive fairness.
- Sabina Alkire. 2002. Valuing Freedoms: Sen's capability approach and poverty reduction. Oxford University Press, Oxford.Google ScholarCross Ref
- Elizabeth Anderson. 2010. The Imperative of Integration. Princeton University Press, Princeton, NJ.Google Scholar
- Elizabeth S. Anderson. 1999. What Is the Point of Equality? Ethics 109, 2 (January 1999), 287--337. DOI:https://doi.org/10.1086/et.1999.109.issue-2Google ScholarCross Ref
- Richard J. Arneson. 1989. Equality and Equal Opportunity for Welfare. Philos. Stud. Int. J. Philos. Anal. Tradit. 56, 1 (May 1989), 77--93.Google Scholar
- Solon Barocas, Moritz Hardt, and Arvind Narayanan. 2018. Fairness and Machine Learning. Retrieved from fairmlbook.orgGoogle Scholar
- Solon Barocas and Andrew D. Selbst. 2016. Big data's disparate impact. Calif. Law Rev. 104, (2016), 671--732.Google Scholar
- Richard Berk, Hoda Heidari, Shahin Jabbari, Michael Kearns, and Aaron Roth. 2017. Fairness in Criminal Justice Risk Assessments: The State of the Art. ArXiv170309207 Stat (March 2017). Retrieved November 2, 2017 from http://arxiv.org/abs/1703.09207Google Scholar
- Reuben Binns. 2017. Fairness in Machine Learning: Lessons from Political Philosophy. ArXiv171203586 Cs (December 2017). Retrieved July 12, 2018 from http://arxiv.org/abs/1712.03586Google Scholar
- Alexandra Chouldechova. 2016. Fair prediction with disparate impact: A study of bias in recidivism prediction instruments. In arXiv:1610.07524 [cs, stat]. Retrieved November 7, 2017 from http://arxiv.org/abs/1610.07524Google Scholar
- Sam Corbett-Davies and Sharad Goel. 2018. The Measure and Mismeasure of Fairness: A Critical Review of Fair Machine Learning. ArXiv180800023 Cs (July 2018). Retrieved August 22, 2018 from http://arxiv.org/abs/1808.00023Google Scholar
- Sam Corbett-Davies, Emma Pierson, Avi Feller, Sharad Goel, and Aziz Huq. 2017. Algorithmic decision making and the cost of fairness. ArXiv170108230 Cs Stat (January 2017). DOI:https://doi.org/10.1145/3097983.309809Google Scholar
- Cynthia Dwork, Moritz Hardt, Toniann Pitassi, Omer Reingold, and Rich Zemel. 2011. Fairness Through Awareness. ArXiv11043913 Cs (April 2011). Retrieved November 16, 2018 from http://arxiv.org/abs/1104.3913Google Scholar
- Michael Feldman, Sorelle A. Friedler, John Moeller, Carlos Scheidegger, and Suresh Venkatasubramanian. 2015. Certifying and Removing Disparate Impact. In Proceedings of the 21th ACM SIGKDD International Conference on Knowledge Discovery and Data Mining (KDD '15), ACM, New York, NY, USA, 259--268. DOI:https://doi.org/10.1145/2783258.2783311Google ScholarCross Ref
- Bruce Glymour and Jonathan Herington. 2019. Measuring the Biases That Matter: The Ethical and Causal Foundations for Measures of Fairness in Algorithms. In Proceedings of the Conference on Fairness, Accountability, and Transparency (FAT* '19), ACM, New York, NY, USA, 269--278. DOI:https://doi.org/10.1145/3287560.3287573Google ScholarDigital Library
- Vivek Gupta, Pegah Nokhiz, Chitradeep Dutta Roy, and Suresh Venkatasubramanian. 2019. Equalizing Recourse across Groups. ArXiv190903166 Cs Stat (September 2019). Retrieved December 18, 2019 from http://arxiv.org/abs/1909.03166Google Scholar
- Hoda Heidari, Michele Loi, Krishna P. Gummadi, and Andreas Krause. 2019. A Moral Framework for Understanding Fair ML through Economic Models of Equality of Opportunity. In Proceedings of the Conference on Fairness, Accountability, and Transparency - FAT* '19, ACM Press, Atlanta, GA, USA, 181--190. DOI:https://doi.org/10.1145/3287560.3287584Google ScholarDigital Library
- Deborah Hellman. 2019. Measuring Algorithmic Fairness. Social Science Research Network, Rochester, NY. Retrieved July 22, 2019 from https://papers.ssrn.com/abstract=3418528Google Scholar
- Niki Kilbertus, Mateo Rojas Carulla, Giambattista Parascandolo, Moritz Hardt, Dominik Janzing, and Bernhard Schölkopf. 2017. Avoiding Discrimination through Causal Reasoning. In Advances in Neural Information Processing Systems 30, I. Guyon, U. V. Luxburg, S. Bengio, H. Wallach, R. Fergus, S. Vishwanathan and R. Garnett (eds.). Curran Associates, Inc., 656--666. Retrieved November 16, 2018 from http://papers.nips.cc/paper/6668-avoiding-discrimination-through-causal-reasoning.pdfGoogle ScholarDigital Library
- Jon Kleinberg, Sendhil Mullainathan, and Manish Raghavan. 2016. Inherent Trade-Offs in the Fair Determination of Risk Scores. In Proceedings of Innovations in Theoretical Computer Science (ITCS). Retrieved November 7, 2017 from http://arxiv.org/abs/1609.05807Google Scholar
- Julian Lamont and Christi Favor. 2017. Distributive Justice. In The Stanford Encyclopedia of Philosophy (Winter 2017), Edward N. Zalta (ed.). Metaphysics Research Lab, Stanford University. Retrieved December 17, 2019 from https://plato.stanford.edu/archives/win2017/entries/justice-distributive/Google ScholarCross Ref
- Kasper Lippert-Rasmussen. 2006. The badness of discrimination. Ethical Theory Moral Pract. 9, 2 (April 2006), 167--185. DOI:https://doi.org/10.1007/s10677-006--9014-xGoogle ScholarCross Ref
- Martha C. Nussbaum. 2000. Women and Human Development: The Capabilities Approach. Cambridge University Press, Cambridge.Google ScholarCross Ref
- Martha C. Nussbaum. 2000. Sex and Social Justice (First Edition edition ed.). Oxford University Press, Oxford New York Athens.Google ScholarCross Ref
- Ziad Obermeyer, Brian Powers, Christine Vogeli, and Sendhil Mullainathan. 2019. Dissecting racial bias in an algorithm used to manage the health of populations. Science 366, 6464 (October 2019), 447. DOI:https://doi.org/10.1126/science.aax2342Google ScholarCross Ref
- Judea Pearl. 2009. Causality: Models, Reasoning and Inference (2nd ed.). Cambridge University Press, New York.Google ScholarDigital Library
- John Rawls. 2001. Justice as Fairness: A Restatement. Belknap Press, Cambridge, MA.Google Scholar
- Amartya Sen. 1983. Poverty and Famines: An Essay on Entitlement and Deprivation. Oxford University Press, Oxford. Retrieved from http://books.google.com.au/books?id=FVC9eqGkMr8CGoogle Scholar
- Amartya Sen. 1993. Capability and Wellbeing. In The Quality of Life, Amartya Sen and Martha C. Nussbaum (eds.). Clarendon Press, Oxford, 31--53.Google Scholar
- Peter Spirtes, Clark Glymour, and Richard Scheines. 2001. Causation, Prediction, and Search (2nd ed.). MIT Press, Cambridge, Mass.Google Scholar
- Larry S. Temkin. 1993. Inequality. Oxford University Press, New York, NY, USA.Google Scholar
Index Terms
- Measuring Fairness in an Unfair World
Recommendations
Measuring the Biases that Matter: The Ethical and Casual Foundations for Measures of Fairness in Algorithms
FAT* '19: Proceedings of the Conference on Fairness, Accountability, and TransparencyMeasures of algorithmic bias can be roughly classified into four categories, distinguished by the conditional probabilistic dependencies to which they are sensitive. First, measures of "procedural bias" diagnose bias when the score returned by an ...
Weapons of moral construction? On the value of fairness in algorithmic decision-making
AbstractFairness is one of the most prominent values in the Ethics and Artificial Intelligence (AI) debate and, specifically, in the discussion on algorithmic decision-making (ADM). However, while the need for fairness in ADM is widely acknowledged, the ...
Measuring justice in machine learning
FAT* '20: Proceedings of the 2020 Conference on Fairness, Accountability, and TransparencyHow can we build more just machine learning systems? To answer this question, we need to know both what justice is and how to tell whether one system is more or less just than another. That is, we need both a definition and a measure of justice. ...
Comments