Policy Learning for Fairness in Ranking

Conventional Learning-to-Rank (LTR) methods optimize the utility of the rankings to the users, but they are oblivious to their impact on the ranked items. However, there has been a growing understanding that the latter is important to consider for a wide range of ranking applications (e.g. online marketplaces, job placement, admissions). To address this need, we propose a general LTR framework that can optimize a wide range of utility metrics (e.g. NDCG) while satisfying fairness of exposure constraints with respect to the items. This framework expands the class of learnable ranking functions to stochastic ranking policies, which provides a language for rigorously expressing fairness specifications. Furthermore, we provide a new LTR algorithm called Fair-PG-Rank for directly searching the space of fair ranking policies via a policy-gradient approach. Beyond the theoretical evidence in deriving the framework and the algorithm, we provide empirical results on simulated and real-world datasets verifying the effectiveness of the approach in individual and group-fairness settings.

[1]  R. Duncan Luce,et al.  Individual Choice Behavior , 1959 .

[2]  R. Plackett.  The Analysis of Permutations , 1975 .

[3]  Jing Peng,et al.  Function Optimization using Connectionist Reinforcement Learning Algorithms , 1991 .

[4]  R. J. Williams,et al.  Simple Statistical Gradient-Following Algorithms for Connectionist Reinforcement Learning , 2004, Machine Learning.

[5]  S. Robertson.  The probability ranking principle in IR , 1997 .

[6]  Richard S. Sutton,et al.  Introduction to Reinforcement Learning , 1998 .

[7]  Helen Nissenbaum,et al.  Shaping the Web: Why the Politics of Search Engines Matters , 2000, Inf. Soc..

[8]  Gregory N. Hullender,et al.  Learning to rank using gradient descent , 2005, ICML.

[9]  Thorsten Joachims,et al.  Training linear SVMs in linear time , 2006, KDD '06.

[10]  Filip Radlinski,et al.  Evaluating the accuracy of implicit feedback from clicks and query reformulations in Web search , 2007, TOIS.

[11]  Tie-Yan Liu,et al.  Learning to rank: from pairwise approach to listwise approach , 2007, ICML '07.

[12]  Charles L. A. Clarke,et al.  Novelty and diversity in information retrieval evaluation , 2008, SIGIR '08.

[13]  Stephen E. Robertson,et al.  SoftRank: optimizing non-smooth rank metrics , 2008, WSDM '08.

[14]  Filip Radlinski,et al.  Learning diverse rankings with multi-armed bandits , 2008, ICML '08.

[15]  Zhaohui Zheng,et al.  Stochastic gradient boosted distributed decision trees , 2009, CIKM.

[16]  Toon Calders,et al.  Building Classifiers with Independency Constraints , 2009, 2009 IEEE International Conference on Data Mining Workshops.

[17]  Filip Radlinski,et al.  Redundancy, diversity and interdependent document relevance , 2009, SIGF.

[18]  Thorsten Joachims,et al.  Cutting-plane training of structural SVMs , 2009, Machine Learning.

[19]  R. Atkinson.  The Next Digital Decade: Essays on the Future of the Internet , 2011 .

[20]  Yi Chang,et al.  Yahoo! Learning to Rank Challenge Overview , 2010, Yahoo! Learning to Rank Challenge.

[21]  Eduardo F. Morales,et al.  An Introduction to Reinforcement Learning , 2011 .

[22]  James Grimmelmann.  Some Skepticism About Search Neutrality , 2011 .

[23]  Toniann Pitassi,et al.  Fairness through awareness , 2011, ITCS '12.

[24]  Toniann Pitassi,et al.  Learning Fair Representations , 2013, ICML.

[25]  Sean A. Munson,et al.  Unequal Representation and Gender Stereotypes in Image Search Results for Occupations , 2015, CHI.

[26]  John D. Lafferty,et al.  Beyond independent relevance: methods and evaluation metrics for subtopic retrieval , 2003, SIGIR.

[27]  Indre Zliobaite,et al.  On the relation between accuracy and fairness in binary classification , 2015, ArXiv.

[28]  Andrew D. Selbst,et al.  Big Data's Disparate Impact , 2016 .

[29]  Alex Graves,et al.  Asynchronous Methods for Deep Reinforcement Learning , 2016, ICML.

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

[31]  Matt J. Kusner,et al.  Counterfactual Fairness , 2017, NIPS.

[32]  Ricardo Baeza-Yates,et al.  FA*IR: A Fair Top-k Ranking Algorithm , 2017, CIKM.

[33]  Julia Stoyanovich,et al.  Measuring Fairness in Ranked Outputs , 2016, SSDBM.

[34]  Krishna P. Gummadi,et al.  Fairness Beyond Disparate Treatment & Disparate Impact: Learning Classification without Disparate Mistreatment , 2016, WWW.

[35]  Nathan Srebro,et al.  Learning Non-Discriminatory Predictors , 2017, COLT.

[36]  Michael Luca,et al.  Racial Discrimination in the Sharing Economy: Evidence from a Field Experiment , 2016 .

[37]  Jiafeng Guo,et al.  Reinforcement Learning to Rank with Markov Decision Process , 2017, SIGIR.

[38]  Jade Goldstein-Stewart,et al.  The Use of MMR, Diversity-Based Reranking for Reordering Documents and Producing Summaries , 1998, SIGIR Forum.

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

[40]  Thorsten Joachims,et al.  Fairness of Exposure in Rankings , 2018, KDD.

[41]  Krishna P. Gummadi,et al.  Equity of Attention: Amortizing Individual Fairness in Rankings , 2018, SIGIR.

[42]  Unbiased Learning-to-Rank with Biased Feedback , 2018, IJCAI.

[43]  Fernando Diaz,et al.  Towards a Fair Marketplace: Counterfactual Evaluation of the trade-off between Relevance, Fairness & Satisfaction in Recommendation Systems , 2018, CIKM.

[44]  Nisheeth K. Vishnoi,et al.  Ranking with Fairness Constraints , 2017, ICALP.

[45]  Thorsten Joachims,et al.  Intervention Harvesting for Context-Dependent Examination-Bias Estimation , 2018, SIGIR.

[46]  Abolfazl Asudeh,et al.  Designing Fair Ranking Schemes , 2017, SIGMOD Conference.

[47]  Thorsten Joachims,et al.  Estimating Position Bias without Intrusive Interventions , 2018, WSDM.

[48]  Carlos Castillo,et al.  Reducing Disparate Exposure in Ranking: A Learning To Rank Approach , 2018, WWW.