It Is Different When Items Are Older: Debiasing Recommendations When Selection Bias and User Preferences Are Dynamic

User interactions with recommender systems (RSs) are affected by user selection bias, e.g., users are more likely to rate popular items (popularity bias) or items that they expect to enjoy beforehand (positivity bias). Methods exist for mitigating the effects of selection bias in user ratings on the evaluation and optimization of RSs. However, these methods treat selection bias as static, despite the fact that the popularity of an itemmay change drastically over time and the fact that user preferences may also change over time. We focus on the age of an item and its effect on selection bias and user preferences. Our experimental analysis reveals that the rating behavior of users on the MovieLens dataset is better captured by methods that consider effects from the age of item on bias and preferences. We theoretically show that in a dynamic scenario in which both the selection bias and user preferences are dynamic, existing debiasing methods are no longer unbiased. To address this limitation, we introduce DebiAsing in the dyNamiC scEnaRio (DANCER), a novel debiasing method that extends the inverse propensity scoring debiasingmethod toaccount fordynamic selectionbias anduserpreferences. Our experimental results indicate that DANCER improves rating prediction performance compared to debiasing methods that incorrectly assume that selection bias is static in a dynamic scenario. To the best of our knowledge, DANCER is the first debiasingmethod that accounts for dynamic selection bias and user preferences in RSs.

[1]  Linas Baltrunas,et al.  Towards Time-Dependant Recommendation based on Implicit Feedback , 2009 .

[2]  Thorsten Joachims,et al.  Unbiased Learning-to-Rank with Biased Feedback , 2016, WSDM.

[3]  Harald Steck,et al.  Evaluation of recommendations: rating-prediction and ranking , 2013, RecSys.

[4]  Ed H. Chi,et al.  Measuring Recommender System Effects with Simulated Users , 2021, ArXiv.

[5]  M. de Rijke,et al.  Keeping Dataset Biases out of the Simulation: A Debiased Simulator for Reinforcement Learning based Recommender Systems , 2020, RecSys.

[6]  Alexander Tuzhilin,et al.  On over-specialization and concentration bias of recommendations: probabilistic neighborhood selection in collaborative filtering systems , 2014, RecSys '14.

[7]  Iván Cantador,et al.  Time-aware recommender systems: a comprehensive survey and analysis of existing evaluation protocols , 2013, User Modeling and User-Adapted Interaction.

[8]  Eli Pariser,et al.  The Filter Bubble: How the New Personalized Web Is Changing What We Read and How We Think , 2012 .

[9]  Yifan Zhang,et al.  Correcting for Selection Bias in Learning-to-rank Systems , 2020, WWW.

[10]  Yehuda Koren,et al.  The BellKor Solution to the Netflix Grand Prize , 2009 .

[11]  Thorsten Joachims,et al.  The Self-Normalized Estimator for Counterfactual Learning , 2015, NIPS.

[12]  Edward Y. Chang,et al.  Improving Sequential Recommendation with Knowledge-Enhanced Memory Networks , 2018, SIGIR.

[13]  Heng-Tze Cheng,et al.  Wide & Deep Learning for Recommender Systems , 2016, DLRS@RecSys.

[14]  Yongdong Zhang,et al.  LightGCN: Simplifying and Powering Graph Convolution Network for Recommendation , 2020, SIGIR.

[15]  Xing Xie,et al.  Session-based Recommendation with Graph Neural Networks , 2018, AAAI.

[16]  Xiangnan He,et al.  Disentangled Graph Collaborative Filtering , 2020, SIGIR.

[17]  Harald Steck,et al.  Item popularity and recommendation accuracy , 2011, RecSys '11.

[18]  Deborah Estrin,et al.  Unbiased offline recommender evaluation for missing-not-at-random implicit feedback , 2018, RecSys.

[19]  Patrick Gallinari,et al.  Ranking with non-random missing ratings: influence of popularity and positivity on evaluation metrics , 2012, RecSys.

[20]  Francesco Bonchi,et al.  Algorithmic Bias: From Discrimination Discovery to Fairness-aware Data Mining , 2016, KDD.

[21]  Filippo Menczer,et al.  How algorithmic popularity bias hinders or promotes quality , 2017, Scientific Reports.

[22]  Rui Zhang,et al.  Doubly Robust Joint Learning for Recommendation on Data Missing Not at Random , 2019, ICML.

[23]  Feng Yu,et al.  A Dynamic Recurrent Model for Next Basket Recommendation , 2016, SIGIR.

[24]  Yanchi Liu,et al.  Graph Contextualized Self-Attention Network for Session-based Recommendation , 2019, IJCAI.

[25]  Moshe Unger,et al.  Context-Aware Recommendations Based on Deep Learning Frameworks , 2020, ACM Trans. Manag. Inf. Syst..

[26]  Xiangnan He,et al.  Bias and Debias in Recommender System: A Survey and Future Directions , 2020, ACM Trans. Inf. Syst..

[27]  Peng Jiang,et al.  BERT4Rec: Sequential Recommendation with Bidirectional Encoder Representations from Transformer , 2019, CIKM.

[28]  Yongfeng Zhang,et al.  Sequential Recommendation with User Memory Networks , 2018, WSDM.

[29]  Richard S. Zemel,et al.  Collaborative prediction and ranking with non-random missing data , 2009, RecSys '09.

[30]  Paul R. Rosenbaum,et al.  Overt Bias in Observational Studies , 2002 .

[31]  Xiaoyu Du,et al.  Outer Product-based Neural Collaborative Filtering , 2018, IJCAI.

[32]  M. de Rijke,et al.  When People Change their Mind: Off-Policy Evaluation in Non-stationary Recommendation Environments , 2019, WSDM.

[33]  F. Maxwell Harper,et al.  The MovieLens Datasets: History and Context , 2016, TIIS.

[34]  Ji-Rong Wen,et al.  RecBole: Towards a Unified, Comprehensive and Efficient Framework for Recommendation Algorithms , 2020, CIKM.

[35]  W. Bruce Croft,et al.  Correcting for Recency Bias in Job Recommendation , 2019, CIKM.

[36]  Yehuda Koren,et al.  Matrix Factorization Techniques for Recommender Systems , 2009, Computer.

[37]  Zoubin Ghahramani,et al.  Probabilistic Matrix Factorization with Non-random Missing Data , 2014, ICML.

[38]  Sartra Wongthanavasu,et al.  Dynamic Collaborative Filtering Based on User Preference Drift and Topic Evolution , 2020, IEEE Access.

[39]  Alexander Tuzhilin,et al.  Experimental comparison of pre- vs. post-filtering approaches in context-aware recommender systems , 2009, RecSys '09.

[40]  Yehuda Koren,et al.  Collaborative filtering with temporal dynamics , 2009, KDD.

[41]  Pablo Castells,et al.  Should I Follow the Crowd?: A Probabilistic Analysis of the Effectiveness of Popularity in Recommender Systems , 2018, SIGIR.

[42]  Abbe Mowshowitz,et al.  Bias on the web , 2002, CACM.

[43]  Alex Beutel,et al.  Recurrent Recommender Networks , 2017, WSDM.

[44]  Loren G. Terveen,et al.  Exploring the filter bubble: the effect of using recommender systems on content diversity , 2014, WWW.

[45]  Panos Kalnis,et al.  AUC-MF: Point of Interest Recommendation with AUC Maximization , 2019, 2019 IEEE 35th International Conference on Data Engineering (ICDE).

[46]  Yuta Saito,et al.  Unbiased Recommender Learning from Missing-Not-At-Random Implicit Feedback , 2020, WSDM.

[47]  Alexandros Karatzoglou,et al.  Session-based Recommendations with Recurrent Neural Networks , 2015, ICLR.

[48]  Xi Chen,et al.  Temporal Collaborative Filtering with Bayesian Probabilistic Tensor Factorization , 2010, SDM.

[49]  Thorsten Joachims,et al.  Recommendations as Treatments: Debiasing Learning and Evaluation , 2016, ICML.

[50]  Harald Steck,et al.  Training and testing of recommender systems on data missing not at random , 2010, KDD.

[51]  Rishabh Mehrotra,et al.  The Music Streaming Sessions Dataset , 2018, WWW.

[52]  Nurfadhlina Mohd Sharef,et al.  Review of the temporal recommendation system with matrix factorization , 2017 .

[53]  Jie Zhang,et al.  A Re-visit of the Popularity Baseline in Recommender Systems , 2020, SIGIR.

[54]  Guohui Ling,et al.  Causal Intervention for Leveraging Popularity Bias in Recommendation , 2021, SIGIR.

[55]  Krishna P. Gummadi,et al.  Optimizing the Recency-Relevancy Trade-off in Online News Recommendations , 2017, WWW.