Learning to Be Fair: A Consequentialist Approach to Equitable Decision Making

References

• Abbasi-Yadkori Y, Pál D, Szepesvári C (2011) Improved algorithms for linear stochastic bandits. Shawe-Taylor J, Zemel R, Bartlett P, Pereira F, Weinberger KQ, eds. NIPS’11 Proc. 24th Internat. Neural Inform. Processing Systems, vol. 24 (Curran Associates, Red Hook, NY), 2312–2320.Google Scholar
• Agrawal S, Devanur NR, Li L (2016b) An efficient algorithm for contextual bandits with knapsacks, and an extension to concave objectives. Feldman V, Rakhlin A, Shamir O, eds. 29th Annual Conf. Learn. Theory, Proceedings of Machine Learning Research, vol. 49 (PMLR, New York), 4–18.Google Scholar
• Agrawal S, Avadhanula V, Goyal V, Zeevi A (2016a) A near-optimal exploration-exploitation approach for assortment selection. Proc. 2016 ACM Conf. Econom. Comput. (Association for Computing Machinery, New York), 599–600.Google Scholar
• Ali M, Sapiezynski P, Bogen M, Korolova A, Mislove A, Rieke A (2019) Discrimination through optimization: How Facebook’s ad delivery can lead to biased outcomes. Proc. ACM Human-Comput. Interaction (CSCW), vol. 3 (Association for Computing Machinery, New York), 1–30.Google Scholar
• Allen S (2024) Interview research with people in jail: Challenges and possibilities. Handbook on Prisons and Jails (Routledge, London), 387–398.Google Scholar
• Athey S, Chetty R, Imbens GW, Kang H (2016) Estimating treatment effects using multiple surrogates: The role of the surrogate score and the surrogate Index. Preprint, submitted March 30, https://arxiv.org/abs/1603.09326.Google Scholar
• Auer P, Cesa-Bianchi N, Fischer P (2002) Finite-time analysis of the multiarmed bandit problem. Machine Learn. 47(2):235–256.Crossref, Google Scholar
• Badanidiyuru A, Langford J, Slivkins A (2014) Resourceful contextual bandits. Balcan MF, Feldman V, Szepesvári C, eds. Proc. 27th Conf. Learn. Theory, Proceedings of Machine Learning Research, vol. 35 (PMLR, New York), 1109–1134.Google Scholar
• Barabas C, Virza M, Dinakar K, Ito J, Zittrain J (2018) Interventions over predictions: Reframing the ethical debate for actuarial risk assessment. Friedler SA, Wilson C, eds. Proc. 1st Conf. Fairness, Accountability Transparency, Proceedings of Machine Learning Research, vol. 81 (PMLR, New York), 62–76.Google Scholar
• Barocas S, Hardt M, Narayanan A (2023) Fairness and Machine Learning: Limitations and Opportunities (MIT Press, Cambridge, MA).Google Scholar
• Bertsimas D, Farias VF, Trichakis N (2011) The price of fairness. Oper. Res. 59(1):17–31.Link, Google Scholar
• Blodgett SL, O’Connor B (2017) Racial disparity in natural language processing: A case study of social media African-American English. Preprint, submitted June 30, https://arxiv.org/abs/1707.00061.Google Scholar
• Brams SJ, Brams SJ, Taylor AD (1996) Fair Division: From Cake-Cutting to Dispute Resolution (Cambridge University Press, Cambridge, UK).Crossref, Google Scholar
• Brough R, Freedman M, Ho DE, Phillips DC (2022) Can transportation subsidies reduce failures to appear in criminal court? Evidence from a pilot randomized controlled trial. Econom. Lett. 216:110540.Crossref, Google Scholar
• Brown A, Chouldechova A, Putnam-Hornstein E, Tobin A, Vaithianathan R (2019) Toward algorithmic accountability in public services: A qualitative study of affected community perspectives on algorithmic decision-making in child welfare services. Proc. 2019 CHI Conf. Human Factors Comput. Systems (Association for Computing Machinery, New York), 1–12.Google Scholar
• Buolamwini J, Gebru T (2018) Gender shades: Intersectional accuracy disparities in commercial gender classification. Friedler SA, Wilson C, eds. Proc. 1st Conf. Fairness, Accountability Transparency, Proceedings of Machine Learning Research, vol. 81 (PMLR, New York).Google Scholar
• Cadigan TP, Lowenkamp CT (2011) Implementing risk assessment in the federal pretrial services system. Federal Probation 75(2):30–34.Google Scholar
• Cai W, Gaebler J, Garg N, Goel S (2020) Fair allocation through selective information acquisition. Proc. AAAI/ACM Conf. AI Ethics Soc. (Association for Computing Machinery, New York), 22–28.Google Scholar
• Caliskan A, Bryson JJ, Narayanan A (2017) Semantics derived automatically from language corpora contain human-like biases. Science 356(6334):183–186.Crossref, Google Scholar
• Caragiannis I, Kaklamanis C, Kanellopoulos P, Kyropoulou M (2012) The efficiency of fair division. Theory Comput. Systems 50(4):589–610.Crossref, Google Scholar
• Card D, Smith NA (2020) On consequentialism and fairness. Frontiers Artificial Intelligence 3:34.Crossref, Google Scholar
• Chaiyachati KH, Hubbard RA, Yeager A, Mugo B, Shea JA, Rosin R, Grande D (2018) Rideshare-based medical transportation for Medicaid patients and primary care show rates: A difference-in-difference analysis of a pilot program. J. General Internal Medicine 33(6):863–868.Crossref, Google Scholar
• Chapelle O, Li L (2011) An empirical evaluation of Thompson sampling. Shawe-Taylor J, Zemel R, Bartlett P, Pereira F, Weinberger KQ, eds. NIPS’11 Proc. 24th Internat. Neural Inform. Processing Systems, vol. 24 (Curran Associates, Red Hook, NY).Google Scholar
• Chiappa S, Isaac WS (2018) A causal Bayesian networks viewpoint on fairness. Kosta E, Pierson J, Slamanig D, Fischer-Hübner S, Krenn S, eds. Privacy and Identity Management. Fairness Accountability Transparency in the Age of Big Data. Privacy and Identity 2018, IFIP Advances in Information and Communication Technology, vol. 547 (Springer, Cham, Switzerland), 3–20.Google Scholar
• Chohlas-Wood A, Coots M, Goel S, Nyarko J (2023a) Designing equitable algorithms. Nature Comput. Sci. 3(7):601–610.Crossref, Google Scholar
• Chohlas-Wood A, Coots M, Nudell J, Nyarko J, Brunskill E, Rogers T, Goel S (2023b) Automated reminders reduce incarceration for missed court dates: Evidence from a text message experiment. Preprint, submitted June 21, https://arxiv.org/abs/2306.12389.Google Scholar
• Chouldechova A, Roth A (2020) A snapshot of the frontiers of fairness in machine learning. Comm. ACM 63(5):82–89.Crossref, Google Scholar
• Chouldechova A, Benavides-Prado D, Fialko O, Vaithianathan R (2018) A case study of algorithm-assisted decision making in child maltreatment hotline screening decisions. Friedler SA, Wilson C, eds. Proc. 1st Conf. Fairness, Accountability Transparency, Proceedings of Machine Learning Research, vol. 81 (PMLR, New York), 134–148.Google Scholar
• Chu W, Ghahramani Z (2005) Preference learning with Gaussian processes. Proc. 22nd Internat. Conf. Machine Learning (Association for Computing Machinery, New York), 137–144.Google Scholar
• Cohler YJ, Lai JK, Parkes DC, Procaccia AD (2011) Optimal envy-free cake cutting. 25th AAAI Conf. Artificial Intelligence (AAAI Press, Washington, DC), 626–631.Google Scholar
• Corbett-Davies S, Gaebler J, Nilforoshan H, Shroff R, Goel S (2023) The measure and mismeasure of fairness. J. Machine Learning Res. 24(1):14730–14846.Google Scholar
• Corbett-Davies S, Pierson E, Feller A, Goel S, Huq A (2017) Algorithmic decision making and the cost of fairness. Proc. 23rd ACM SIGKDD Internat. Conf. Knowledge Discovery Data Mining (Association for Computing Machinery, New York), 797–806.Google Scholar
• Coston A, Mishler A, Kennedy EH, Chouldechova A (2020) Counterfactual risk assessments, evaluation, and fairness. Proc. 2020 Conf. Fairness Accountability Transparency (Association for Computing Machinery, New York), 582–593.Google Scholar
• Cowgill B, Tucker CE (2020) Algorithmic fairness and economics. Preprint, submitted April 4, 2019, http://dx.doi.org/10.2139/ssrn.3361280.Google Scholar
• Datta A, Datta A, Makagon J, Mulligan DK, Tschantz MC (2018) Discrimination in online advertising: A multidisciplinary inquiry. Friedler SA, Wilson C, eds. Proc. 1st Conf. Fairness, Accountability Transparency, Proceedings of Machine Learning Research, vol. 81 (PMLR, New York), 20–34.Google Scholar
• De-Arteaga M, Fogliato R, Chouldechova A (2020) A case for humans-in-the-loop: Decisions in the presence of erroneous algorithmic scores. Proc. 2020 CHI Conf. Human Factors Comput. Systems (Association for Computing Machinery, New York), 1–12.Google Scholar
• De-Arteaga M, Romanov A, Wallach H, Chayes J, Borgs C, Chouldechova A, Geyik S, Kenthapadi K, Kalai AT (2019) Bias in bios: A case study of semantic representation bias in a high-stakes setting. Proc. Conf. Fairness Accountability Transparency (Association for Computing Machinery, New York), 120–128.Google Scholar
• Donahue K, Kleinberg J (2020) Fairness and utilization in allocating resources with uncertain demand. Proc. 2020 Conf. Fairness Accountability Transparency (Association for Computing Machinery, New York), 658–668.Google Scholar
• Dong S, Ma T, Van Roy B (2019) On the performance of Thompson sampling on logistic bandits. Beygelzimer A, Hsu D, eds. Proc. Thirty-Second Conf. Learn. Theory, Proceedings of Machine Learning Research, vol. 99 (PMLR, New York), 1158–1160.Google Scholar
• Fang EX, Wang Z, Wang L (2022) Fairness-oriented learning for optimal individualized treatment rules. J. Amer. Statist. Assoc. 118(543):1733–1746.Crossref, Google Scholar
• Feldman M, Friedler SA, Moeller J, Scheidegger C, Venkatasubramanian S (2015) Certifying and removing disparate impact. Proc. 21th ACM SIGKDD Internat. Conf. Knowledge Discovery Data Mining (Association for Computing Machinery, New York), 259–268.Google Scholar
• Fishbane A, Ouss A, Shah AK (2020) Behavioral nudges reduce failure to appear for court. Science 370(6517):eabb6591.Crossref, Google Scholar
• Fraade-Blanar L, Koo T, Whaley CM (2021) Going to the doctor: Rideshare as nonemergency medical transportation. Technical report, RAND Corporation, Santa Monica, CA.Google Scholar
• Friedewald JJ, Samana CJ, Kasiske BL, Israni AK, Stewart D, Cherikh W, Formica RN (2013) The kidney allocation system. Surgical Clinics 93(6):1395–1406.Google Scholar
• Fürnkranz J, Hüllermeier E (2010) Preference learning and ranking by pairwise comparison. Fürnkranz J, Hüllermeier E, eds. Preference Learning (Springer, Berlin, Heidelberg), 65–82.Crossref, Google Scholar
• Gal Y, Mash M, Procaccia AD, Zick Y (2017) Which is the fairest (rent division) of them all? J. ACM 64(6):1–22.Crossref, Google Scholar
• Goel S, Shroff R, Skeem JL, Slobogin C (2018) The accuracy, equity, and jurisprudence of criminal risk assessment. Vogl R, ed. Research Handbook on Big Data Law (Edward Elgar Publishing, Cheltenham, UK), 9–28.Google Scholar
• Goodman SN, Goel S, Cullen MR (2018) Machine learning, health disparities, and causal reasoning. Ann. Intern. Med. 169(12):883–884.Crossref, Google Scholar
• Grgić-Hlača N, Lima G, Weller A, Redmiles EM (2022) Dimensions of diversity in human perceptions of algorithmic fairness. EAAMO ‘22 Proc. 2nd ACM Conf. Equity Access Algorithms Mechanisms Optim. (Association for Computing Machinery, New York), 1–12.Google Scholar
• Gupta S, Jalan A, Ranade G, Yang H, Zhuang S (2020) Too many fairness metrics: Is there a solution? Preprint, submitted April 9, http://dx.doi.org/10.2139/ssrn.3554829.Google Scholar
• Hadad V, Hirshberg DA, Zhan R, Wager S, Athey S (2021) Confidence intervals for policy evaluation in adaptive experiments. Proc. Natl. Acad. Sci. USA 118(15):e2014602118.Crossref, Google Scholar
• Hardt M, Price E, Srebro N (2016) Equality of opportunity in supervised learning. NIPS’16 Proc. 30th Internat. Conf. Neural Inform. Processing Systems (Curran Associates, Red Hook, NY).Google Scholar
• Hsu D, Kakade SM, Zhang T (2014) Random design analysis of ridge regression. Foundations Comput. Math. 14:569–600.Crossref, Google Scholar
• Jin C, Allen-Zhu Z, Bubeck S, Jordan MI (2018) Is Q-learning provably efficient? NIPS’18 Proc. 32nd Internat. Conf. Neural Inform. Processing Systems (Curran Associates, Red Hook, NY), 3323–3331.Google Scholar
• Jun KS, Jain L, Mason B, Nassif H (2021) Improved confidence bounds for the linear logistic model and applications to bandits. Meila M, Zhang T, eds. Proc. 38th Internat. Conf. Machine Learn., Proceedings of Machine Learning Research, vol. 139 (PMLR, New York), 5148–5157.Google Scholar
• Jung C, Kearns M, Neel S, Roth A, Stapleton L, Wu ZS (2019) An algorithmic framework for fairness elicitation. Preprint, submitted May 25, https://arxiv.org/abs/1905.10660.Google Scholar
• Kasy M, Abebe R (2021) Fairness, equality, and power in algorithmic decision-making. Proc. 2021 ACM Conf. Fairness Accountability Transparency (Association for Computing Machinery, New York), 576–586.Google Scholar
• Kilbertus N, Rojas Carulla M, Parascandolo G, Hardt M, Janzing D, Schölkopf B (2017) Avoiding discrimination through causal reasoning. NIPS’17 Proc. 31st Internat. Conf. Neural Inform. Processing Systems (Curran Associates, Red Hook, NY), 656–666.Google Scholar
• Koenecke A, Giannella E, Willer R, Goel S (2023) Popular support for balancing equity and efficiency in resource allocation: A case study in online advertising to increase welfare program awareness. Proc. Internat. AAAI Conf. Web Social Media, vol. 17, (AAAI Press, Washington, DC), 494–506.Google Scholar
• Koenecke A, Nam A, Lake E, Nudell J, Quartey M, Mengesha Z, Toups C, Rickford JR, Jurafsky D, Goel S (2020) Racial disparities in automated speech recognition. Proc. Natl. Acad. Sci. USA 117(14):7684–7689.Crossref, Google Scholar
• Kusner MJ, Loftus J, Russell C, Silva R (2017) Counterfactual fairness. NIPS’17 Proc. 31st Internat. Conf. Neural Inform. Processing Systems (Curran Associates, Red Hook, NY), 4069–4079.Google Scholar
• Latessa EJ, Lemke R, Makarios M, Smith P (2010) The creation and validation of the Ohio risk assessment system (ORAS) Federal Probation 74(1):16–22.Google Scholar
• Lattimore T, Szepesvári C (2020) Bandit Algorithms (Cambridge University Press, Cambridge, UK).Crossref, Google Scholar
• Leo M, Sharma S, Maddulety K (2019) Machine learning in banking risk management: A literature review. Risks 7(1):29.Crossref, Google Scholar
• Li L, Lu Y, Zhou D (2017) Provably optimal algorithms for generalized linear contextual bandits. Precup D, Teh YW, eds. Proc. 34th Internat. Conf. Machine Learn., Proceedings of Machine Learning Research, vol. 70 (PMLR, New York), 2071–2080.Google Scholar
• Li Z, Ratliff L, Nassif H, Jamieson KG, Jain L (2022) Instance-optimal PAC algorithms for contextual bandits. NIPS’22 Proc. 36th Internat. Conf. Neural Inform. Processing Systems (Curran Associates, Red Hook, NY), 37590–37603.Google Scholar
• Lin ZJ, Obeng A, Bakshy E (2020) Preference learning for real-world multi-objective decision making. Bogunovic I, Neiswanger W, Yue Y, eds. Workshop Real World Experiment Design Active Learn. (ICML).Google Scholar
• Liu LT, Dean S, Rolf E, Simchowitz M, Hardt M (2018) Delayed impact of fair machine learning. Dy J, Krause A, eds. Proc. 35th Internat. Conf. Machine Learn., Proceedings of Machine Learning Research, vol. 80 (PMLR, New York), 3150–3158.Google Scholar
• Lyft (2020) Modernizing medical transportation with rideshare. Technical report, FierceHealthcare, Washington, DC.Google Scholar
• Mahoney B, Beaudin BD, Carver JA III, Ryan DB, Hoffman RB (2001) Pretrial Services Programs: Responsibilities and Potential (National Institute of Justice, Washington, DC).Google Scholar
• Mannor S, Tsitsiklis JN (2004) The sample complexity of exploration in the multi-armed bandit problem. J. Machine Learn. Res. 5:623–648.Google Scholar
• Metevier B, Giguere S, Brockman S, Kobren A, Brun Y, Brunskill E, Thomas P (2019) Offline contextual bandits with high probability fairness guarantees. NIPS’19 Proc. 33rd Internat. Neural Inform. Processing Systems (Curran Associates, Red Hook, NY), 14922–14933.Google Scholar
• Milgram A, Holsinger AM, Vannostrand M, Alsdorf MW (2014) Pretrial risk assessment: Improving public safety and fairness in pretrial decision making. Federal Sentencing Reporter 27(4):216–221.Crossref, Google Scholar
• Nabi R, Shpitser I (2018) Fair inference on outcomes. Proc. AAAI Conf. Artificial Intelligence, vol. 32 (AAAI Press, Washington, DC), 1931–1940.Google Scholar
• Nilforoshan H, Gaebler JD, Shroff R, Goel S (2022) Causal conceptions of fairness and their consequences. Chaudhuri K, Jegelka S, Song L, Szepesvari C, Niu G, Sabato S, eds. Proc. 39th Internat. Conf. Machine Learn., Proceedings of Machine Learning Research, vol. 162 (PMLR, New York), 16848–16887.Google Scholar
• Nyarko J, Goel S, Sommers R (2021) Breaking taboos in fair machine learning: An experimental study. EAAMO ‘21 Proc. 1st ACM Conf. Equity Access Algorithms Mechanisms Optim. (Association for Computing Machinery, New York), 1–11.Google Scholar
• Obermeyer Z, Powers B, Vogeli C, Mullainathan S (2019) Dissecting racial bias in an algorithm used to manage the health of populations. Science 366(6464):447–453.Crossref, Google Scholar
• Pacchiano A, Lee J, Brunskill E (2023) Experiment planning with function approximation. NIPS’23 Proc. 37th Internat. Neural Inform. Processing Systems (Curran Associates, Red Hook, NY), 9409–9421.Google Scholar
• Patil V, Ghalme G, Nair V, Narahari Y (2021) Achieving fairness in the stochastic multi-armed bandit problem. J. Machine Learn. Res. 22:1–31.Google Scholar
• Procaccia AD (2013) Cake cutting: Not just child’s play. Comm. ACM 56(7):78–87.Crossref, Google Scholar
• Raji ID, Buolamwini J (2019) Actionable auditing: Investigating the impact of publicly naming biased performance results of commercial AI products. Proc. 2019 AAAI/ACM Conf. AI Ethics Soc. (Association for Computing Machinery, New York), 429–435.Google Scholar
• Rolf E, Simchowitz M, Dean S, Liu LT, Bjorkegren D, Hardt M, Blumenstock J (2020) Balancing competing objectives with noisy data: Score-based classifiers for welfare-aware machine learning. Daumé H III, Singh A, eds. Proc. 37th Internat. Conf. Machine Learn., Proceedings of Machine Learning Research, vol. 119 (PMLR, New York), 8158–8168.Google Scholar
• Shroff R (2017) Predictive analytics for city agencies: Lessons from children’s services. Big Data 5(3):189–196.Crossref, Google Scholar
• Slivkins A (2019) Introduction to Multi-Armed Bandits, Foundations and Trends in Machine Learning, vol. 12 (now Publishers, Hanover, MA), 1–286.Crossref, Google Scholar
• Thomas PS, da Silva BC, Barto AG, Giguere S, Brun Y, Brunskill E (2019) Preventing undesirable behavior of intelligent machines. Science 366(6468):999–1004.Crossref, Google Scholar
• Vais S, Siu J, Maru S, Abbott J, Hill IS, Achilike C, Wu WJ, Adegoke TM, Steer-Massaro C (2020) Rides for refugees: A transportation assistance pilot for women’s health. J. Immigrant Minority Health 22(1):74–81.Crossref, Google Scholar
• Viviano D, Bradic J (2024) Fair policy targeting. J. Amer. Statist. Assoc. 119(545):730–743.Crossref, Google Scholar
• Wang Y, Sridhar D, Blei DM (2019) Equal opportunity and affirmative action via counterfactual predictions. Preprint, submitted May 26, https://arxiv.org/abs/1905.10870.Google Scholar
• Wilder B, Onasch-Vera L, Diguiseppi G, Petering R, Hill C, Yadav A, Rice E, Tambe M (2021) Clinical trial of an AI-augmented intervention for HIV prevention in youth experiencing homelessness. Proc. AAAI Conf. Artificial Intelligence, vol. 35 (AAAI Press, Washington, DC), 14948–14956.Google Scholar
• Wu H, Srikant R, Liu X, Jiang C (2015) Algorithms with logarithmic or sublinear regret for constrained contextual bandits. NIPS’15 Proc. 28th Internat. Neural Inform. Processing Systems (Curran Associates, Red Hook, NY), 433–441.Google Scholar
• Wu Y, Zhang L, Wu X, Tong H (2019) PC-fairness: A unified framework for measuring causality-based fairness. Proc. 33rd Internat. Neural Inform. Processing Systems (Curran Associates, Red Hook, NY), 3404–3414.Google Scholar
• Zanette A, Dong K, Lee JN, Brunskill E (2021) Design of experiments for stochastic contextual linear bandits. NIPS’21 Proc. 35th Internat. Neural Inform. Processing Systems, vol. 34 (Curran Associates, Red Hook, NY), 22720–22731.Google Scholar
• Zhang J, Bareinboim E (2018) Fairness in decision-making—The causal explanation formula. Proc. AAAI Conf. Artificial Intelligence, vol. 32 (AAAI Press, Washington, DC), 2037–2045.Google Scholar
• Zhang K, Janson L, Murphy S (2021) Statistical inference with m-estimators on adaptively collected data. NIPS’21 Proc. 35th Internat. Neural Inform. Processing Systems (Curran Associates, Red Hook, NY), 7460–7471.Google Scholar
• Zottola SA, Crozier WE, Ariturk D, Desmarais SL (2023) Court date reminders reduce court nonappearance: A meta-analysis. Criminology Public Policy 22(1):97–123.Crossref, Google Scholar
• Zuluaga M, Sergent G, Krause A, Püschel M (2013) Active learning for multi-objective optimization. Dasgupta S, McAllester D, eds. Proc. 30th Internat. Conf. Machine Learn., Proceedings of Machine Learning Research, vol. 28 (PMLR, New York).Google Scholar