Towards Fair Personalization by Avoiding Feedback Loops

Self-reinforcing feedback loops are both cause and effect of over and/or underpresentation of some content in interactive recommender systems. This leads to erroneous user preference estimates, namely, overestimation of over-presented content while violating the right to be presented of each alternative, contrary of which we define as a fair system. We consider two models that explicitly incorporate, or ignore the systematic and limited exposure to alternatives. By simulations, we demonstrate that ignoring the systematic presentations overestimates promoted options and underestimates censored alternatives. Simply conditioning on the limited exposure is a remedy for these biases.

[1]  Paul Covington,et al.  Deep Neural Networks for YouTube Recommendations , 2016, RecSys.

[2]  Matthew D. Hoffman,et al.  Variational Autoencoders for Collaborative Filtering , 2018, WWW.

[3]  Bamshad Mobasher,et al.  Managing Popularity Bias in Recommender Systems with Personalized Re-ranking , 2019, FLAIRS.

[4]  David M. Blei,et al.  Causal Inference for Recommendation , 2016 .

[5]  Carlos Riquelme,et al.  Human Interaction with Recommendation Systems , 2017, AISTATS.

[6]  Olfa Nasraoui,et al.  Debiasing the Human-Recommender System Feedback Loop in Collaborative Filtering , 2019, WWW.

[7]  George Kingsley Zipf,et al.  Human behavior and the principle of least effort , 1949 .

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

[9]  Eli Pariser,et al.  The Filter Bubble: What the Internet Is Hiding from You , 2011 .

[10]  Weiwen Liu,et al.  Personalizing Fairness-aware Re-ranking , 2018, ArXiv.

[11]  Karthik Ramani,et al.  Deconvolving Feedback Loops in Recommender Systems , 2016, NIPS.

[12]  Bert Huang,et al.  Beyond Parity: Fairness Objectives for Collaborative Filtering , 2017, NIPS.

[13]  Nick Craswell,et al.  An experimental comparison of click position-bias models , 2008, WSDM '08.

[14]  Ali Taylan Cemgil,et al.  A Bayesian Choice Model for Eliminating Feedback Loops , 2019, ArXiv.

[15]  Lise Getoor,et al.  A Fairness-aware Hybrid Recommender System , 2018, ArXiv.

[16]  Jonathan L. Herlocker,et al.  Evaluating collaborative filtering recommender systems , 2004, TOIS.

[17]  W. R. Thompson.  ON THE LIKELIHOOD THAT ONE UNKNOWN PROBABILITY EXCEEDS ANOTHER IN VIEW OF THE EVIDENCE OF TWO SAMPLES , 1933 .

[18]  Tor Lattimore,et al.  Degenerate Feedback Loops in Recommender Systems , 2019, AIES.

[19]  John Guiver,et al.  Bayesian inference for Plackett-Luce ranking models , 2009, ICML '09.

[20]  Tao Qin,et al.  Mechanism Learning with Mechanism Induced Data , 2015, AAAI.

[21]  Bamshad Mobasher,et al.  Controlling Popularity Bias in Learning-to-Rank Recommendation , 2017, RecSys.

[22]  Himan Abdollahpouri,et al.  Popularity Bias in Ranking and Recommendation , 2019, AIES.

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

[24]  Arnaud Doucet,et al.  Efficient Bayesian Inference for Generalized Bradley–Terry Models , 2010, 1011.1761.