Learning to be Fair: A Consequentialist Approach to Equitable Decision-Making

In the dominant paradigm for designing equitable machine learning systems, one works to ensure that model predictions satisfy various fairness criteria, such as parity in error rates across race, gender, and other legally protected traits. That approach, however, typically divorces predictions from the downstream outcomes they ultimately affect, and, as a result, can induce unexpected harms. Here we present an alternative framework for fairness that directly anticipates the consequences of actions. Stakeholders first specify preferences over the possible outcomes of an algorithmically informed decision-making process. For example, lenders may prefer extending credit to those most likely to repay a loan, while also preferring similar lending rates across neighborhoods. One then searches the space of decision policies to maximize the specified utility. We develop and describe a method for efficiently learning these optimal policies from data for a large family of expressive utility functions, facilitating a more holistic approach to equitable decision-making.

[1]  Alexandra Chouldechova,et al.  The Frontiers of Fairness in Machine Learning , 2018, ArXiv.

[2]  R. Hubbard,et al.  Association of Rideshare-Based Transportation Services and Missed Primary Care Appointments: A Clinical Trial , 2018, JAMA internal medicine.

[3]  Michael Carl Tschantz,et al.  Discrimination in Online Advertising: A Multidisciplinary Inquiry , 2018 .

[4]  Silvia Chiappa,et al.  A Causal Bayesian Networks Viewpoint on Fairness , 2018, Privacy and Identity Management.

[5]  Ravi Shroff,et al.  Predictive Analytics for City Agencies: Lessons from Children's Services , 2017, Big Data.

[6]  Ilya Shpitser,et al.  Fair Inference on Outcomes , 2017, AAAI.

[7]  Christopher T. Lowenkamp,et al.  Gender, risk assessment, and sanctioning: The cost of treating women like men. , 2016, Law and human behavior.

[8]  Andreas Krause,et al.  Active Learning for Multi-Objective Optimization , 2013, ICML.

[9]  Timnit Gebru,et al.  Gender Shades: Intersectional Accuracy Disparities in Commercial Gender Classification , 2018, FAT.

[10]  S. Maru,et al.  Rides for Refugees: A Transportation Assistance Pilot for Women’s Health , 2019, Journal of Immigrant and Minority Health.

[11]  Dan Jurafsky,et al.  Racial disparities in automated speech recognition , 2020, Proceedings of the National Academy of Sciences.

[12]  Yuriy Brun,et al.  Preventing undesirable behavior of intelligent machines , 2019, Science.

[13]  Elias Bareinboim,et al.  Fairness in Decision-Making - The Causal Explanation Formula , 2018, AAAI.

[14]  Alexandra Chouldechova,et al.  A case study of algorithm-assisted decision making in child maltreatment hotline screening decisions , 2018, FAT.

[15]  Alexandra Chouldechova,et al.  A Case for Humans-in-the-Loop: Decisions in the Presence of Erroneous Algorithmic Scores , 2020, CHI.

[16]  Ravi Shroff,et al.  The accuracy, equity, and jurisprudence of criminal risk assessment , 2021, Research Handbook on Big Data Law.

[17]  G. Imbens,et al.  The Surrogate Index: Combining Short-Term Proxies to Estimate Long-Term Treatment Effects More Rapidly and Precisely , 2019 .

[18]  G. Andrew,et al.  arm: Data Analysis Using Regression and Multilevel/Hierarchical Models , 2014 .

[19]  Vashist Avadhanula,et al.  A Near-Optimal Exploration-Exploitation Approach for Assortment Selection , 2016, EC.

[20]  E. Bakshy,et al.  Preference Learning for Real-World Multi-Objective Decision Making , 2020 .

[21]  K. Maddulety,et al.  Machine Learning in Banking Risk Management: A Literature Review , 2019, Risks.

[22]  Esther Rolf,et al.  Balancing Competing Objectives with Noisy Data: Score-Based Classifiers for Welfare-Aware Machine Learning , 2020, ICML.

[23]  Yuriy Brun,et al.  Offline Contextual Bandits with High Probability Fairness Guarantees , 2019, NeurIPS.

[24]  John Langford,et al.  Resourceful Contextual Bandits , 2014, COLT.

[25]  Joseph Hilbe,et al.  Data Analysis Using Regression and Multilevel/Hierarchical Models , 2009 .

[26]  Brendan T. O'Connor,et al.  Racial Disparity in Natural Language Processing: A Case Study of Social Media African-American English , 2017, ArXiv.

[27]  Sharad Goel,et al.  Breaking Taboos in Fair Machine Learning: An Experimental Study , 2021, EAAMO.

[28]  Richard J. Lemke,et al.  The Creation and Validation of the Ohio Risk Assessment System ( ORAS ) , 2010 .

[29]  Aleksandrs Slivkins,et al.  Introduction to Multi-Armed Bandits , 2019, Found. Trends Mach. Learn..

[30]  Hanghang Tong,et al.  PC-Fairness: A Unified Framework for Measuring Causality-based Fairness , 2019, NeurIPS.

[31]  Nikhil R. Devanur,et al.  An efficient algorithm for contextual bandits with knapsacks, and an extension to concave objectives , 2015, COLT.

[32]  Alexandra Chouldechova,et al.  Counterfactual risk assessments, evaluation, and fairness , 2020, FAT*.

[33]  Wei Chu,et al.  Preference learning with Gaussian processes , 2005, ICML.

[34]  Yixin Wang,et al.  Equal Opportunity and Affirmative Action via Counterfactual Predictions , 2019, ArXiv.

[35]  R. Srikant,et al.  Algorithms with Logarithmic or Sublinear Regret for Constrained Contextual Bandits , 2015, NIPS.

[36]  Bernhard Schölkopf,et al.  Avoiding Discrimination through Causal Reasoning , 2017, NIPS.

[37]  Alexandra Chouldechova,et al.  Bias in Bios: A Case Study of Semantic Representation Bias in a High-Stakes Setting , 2019, FAT.

[38]  Inioluwa Deborah Raji,et al.  Actionable Auditing: Investigating the Impact of Publicly Naming Biased Performance Results of Commercial AI Products , 2019, AIES.

[39]  A. Chouldechova,et al.  Toward Algorithmic Accountability in Public Services: A Qualitative Study of Affected Community Perspectives on Algorithmic Decision-making in Child Welfare Services , 2019, CHI.

[40]  S. Goodman,et al.  Machine Learning, Health Disparities, and Causal Reasoning , 2018, Annals of Internal Medicine.

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

[42]  R. Hubbard,et al.  Rideshare-Based Medical Transportation for Medicaid Patients and Primary Care Show Rates: A Difference-in-Difference Analysis of a Pilot Program , 2018, Journal of General Internal Medicine.

[43]  Sharad Goel,et al.  The Measure and Mismeasure of Fairness: A Critical Review of Fair Machine Learning , 2018, ArXiv.

[44]  Brian W. Powers,et al.  Dissecting racial bias in an algorithm used to manage the health of populations , 2019, Science.

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

[46]  Alexander M. Holsinger,et al.  Pretrial Risk Assessment: Improving Public Safety and Fairness in Pretrial Decision Making , 2015 .

[47]  Christopher T. Lowenkamp,et al.  Special Issue: Evidence-Based Practices in Action *30IMPLEMENTING RISK ASSESSMENT IN THE FEDERAL PRETRIAL SERVICES SYSTEM , 2011 .

[48]  Eyke Hüllermeier,et al.  Preference Learning and Ranking by Pairwise Comparison , 2010, Preference Learning.

[49]  Luca Oneto,et al.  Fairness in Machine Learning , 2020, INNSBDDL.

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

[51]  Arvind Narayanan,et al.  Semantics derived automatically from language corpora contain human-like biases , 2016, Science.

[52]  A. Korolova,et al.  Discrimination through Optimization , 2019, Proc. ACM Hum. Comput. Interact..