Optimizing Long-term Social Welfare in Recommender Systems: A Constrained Matching Approach

Martin Mladenov, Elliot Creager, Omer Ben-Porat, Kevin Swersky, Richard Zemel, Craig Boutilier
Proceedings of the 37th International Conference on Machine Learning, PMLR 119:6987-6998, 2020.

Abstract

Most recommender systems (RS) research assumes that a user’s utility can be maximized independently of the utility of the other agents (e.g., other users, content providers). In realistic settings, this is often not true – the dynamics of an RS ecosystem couple the long-term utility of all agents. In this work, we explore settings in which content providers cannot remain viable unless they receive a certain level of user engagement. We formulate this problem as one of equilibrium selection in the induced dynamical system, and show that it can be solved as an optimal constrained matching problem. Our model ensures the system reaches an equilibrium with maximal social welfare supported by a sufficiently diverse set of viable providers. We demonstrate that even in a simple, stylized dynamical RS model, the standard myopic approach to recommendation - always matching a user to the best provider - performs poorly. We develop several scalable techniques to solve the matching problem, and also draw connections to various notions of user regret and fairness, arguing that these outcomes are fairer in a utilitarian sense.

Cite this Paper

BibTeX
@InProceedings{pmlr-v119-mladenov20a, title = {Optimizing Long-term Social Welfare in Recommender Systems: A Constrained Matching Approach}, author = {Mladenov, Martin and Creager, Elliot and Ben-Porat, Omer and Swersky, Kevin and Zemel, Richard and Boutilier, Craig}, booktitle = {Proceedings of the 37th International Conference on Machine Learning}, pages = {6987--6998}, year = {2020}, editor = {III, Hal Daumé and Singh, Aarti}, volume = {119}, series = {Proceedings of Machine Learning Research}, month = {13--18 Jul}, publisher = {PMLR}, pdf = {http://proceedings.mlr.press/v119/mladenov20a/mladenov20a.pdf}, url = {https://proceedings.mlr.press/v119/mladenov20a.html}, abstract = {Most recommender systems (RS) research assumes that a user’s utility can be maximized independently of the utility of the other agents (e.g., other users, content providers). In realistic settings, this is often not true – the dynamics of an RS ecosystem couple the long-term utility of all agents. In this work, we explore settings in which content providers cannot remain viable unless they receive a certain level of user engagement. We formulate this problem as one of equilibrium selection in the induced dynamical system, and show that it can be solved as an optimal constrained matching problem. Our model ensures the system reaches an equilibrium with maximal social welfare supported by a sufficiently diverse set of viable providers. We demonstrate that even in a simple, stylized dynamical RS model, the standard myopic approach to recommendation - always matching a user to the best provider - performs poorly. We develop several scalable techniques to solve the matching problem, and also draw connections to various notions of user regret and fairness, arguing that these outcomes are fairer in a utilitarian sense.} }
Endnote
%0 Conference Paper %T Optimizing Long-term Social Welfare in Recommender Systems: A Constrained Matching Approach %A Martin Mladenov %A Elliot Creager %A Omer Ben-Porat %A Kevin Swersky %A Richard Zemel %A Craig Boutilier %B Proceedings of the 37th International Conference on Machine Learning %C Proceedings of Machine Learning Research %D 2020 %E Hal Daumé III %E Aarti Singh %F pmlr-v119-mladenov20a %I PMLR %P 6987--6998 %U https://proceedings.mlr.press/v119/mladenov20a.html %V 119 %X Most recommender systems (RS) research assumes that a user’s utility can be maximized independently of the utility of the other agents (e.g., other users, content providers). In realistic settings, this is often not true – the dynamics of an RS ecosystem couple the long-term utility of all agents. In this work, we explore settings in which content providers cannot remain viable unless they receive a certain level of user engagement. We formulate this problem as one of equilibrium selection in the induced dynamical system, and show that it can be solved as an optimal constrained matching problem. Our model ensures the system reaches an equilibrium with maximal social welfare supported by a sufficiently diverse set of viable providers. We demonstrate that even in a simple, stylized dynamical RS model, the standard myopic approach to recommendation - always matching a user to the best provider - performs poorly. We develop several scalable techniques to solve the matching problem, and also draw connections to various notions of user regret and fairness, arguing that these outcomes are fairer in a utilitarian sense.
APA
Mladenov, M., Creager, E., Ben-Porat, O., Swersky, K., Zemel, R. & Boutilier, C.. (2020). Optimizing Long-term Social Welfare in Recommender Systems: A Constrained Matching Approach. Proceedings of the 37th International Conference on Machine Learning, in Proceedings of Machine Learning Research 119:6987-6998 Available from https://proceedings.mlr.press/v119/mladenov20a.html.

Related Material