Fair Contextual Multi-Armed Bandits: Theory and Experiments

Yifang Chen, Alex Cuellar, Haipeng Luo, Jignesh Modi, Heramb Nemlekar, Stefanos Nikolaidis
Proceedings of the 36th Conference on Uncertainty in Artificial Intelligence (UAI), PMLR 124:181-190, 2020.

Abstract

When an AI system interacts with multiple users, it frequently needs to make allocation decisions. For instance, a virtual agent decides whom to pay attention to in a group, or a factory robot selects a worker to deliver a part.Demonstrating fairness in decision making is essential for such systems to be broadly accepted. We introduce a Multi-Armed Bandit algorithm with fairness constraints, where fairness is defined as a minimum rate at which a task or a resource is assigned to a user. The proposed algorithm uses contextual information about the users and the task and makes no assumptions on how the losses capturing the performance of different users are generated. We provide theoretical guarantees of performance and empirical results from simulation and an online user study. The results highlight the benefit of accounting for contexts in fair decision making, especially when users perform better at some contexts and worse at others.

Cite this Paper

BibTeX
@InProceedings{pmlr-v124-chen20a, title = {Fair Contextual Multi-Armed Bandits: Theory and Experiments}, author = {Chen, Yifang and Cuellar, Alex and Luo, Haipeng and Modi, Jignesh and Nemlekar, Heramb and Nikolaidis, Stefanos}, booktitle = {Proceedings of the 36th Conference on Uncertainty in Artificial Intelligence (UAI)}, pages = {181--190}, year = {2020}, editor = {Peters, Jonas and Sontag, David}, volume = {124}, series = {Proceedings of Machine Learning Research}, month = {03--06 Aug}, publisher = {PMLR}, pdf = {http://proceedings.mlr.press/v124/chen20a/chen20a.pdf}, url = {https://proceedings.mlr.press/v124/chen20a.html}, abstract = {When an AI system interacts with multiple users, it frequently needs to make allocation decisions. For instance, a virtual agent decides whom to pay attention to in a group, or a factory robot selects a worker to deliver a part.Demonstrating fairness in decision making is essential for such systems to be broadly accepted. We introduce a Multi-Armed Bandit algorithm with fairness constraints, where fairness is defined as a minimum rate at which a task or a resource is assigned to a user. The proposed algorithm uses contextual information about the users and the task and makes no assumptions on how the losses capturing the performance of different users are generated. We provide theoretical guarantees of performance and empirical results from simulation and an online user study. The results highlight the benefit of accounting for contexts in fair decision making, especially when users perform better at some contexts and worse at others. } }
Endnote
%0 Conference Paper %T Fair Contextual Multi-Armed Bandits: Theory and Experiments %A Yifang Chen %A Alex Cuellar %A Haipeng Luo %A Jignesh Modi %A Heramb Nemlekar %A Stefanos Nikolaidis %B Proceedings of the 36th Conference on Uncertainty in Artificial Intelligence (UAI) %C Proceedings of Machine Learning Research %D 2020 %E Jonas Peters %E David Sontag %F pmlr-v124-chen20a %I PMLR %P 181--190 %U https://proceedings.mlr.press/v124/chen20a.html %V 124 %X When an AI system interacts with multiple users, it frequently needs to make allocation decisions. For instance, a virtual agent decides whom to pay attention to in a group, or a factory robot selects a worker to deliver a part.Demonstrating fairness in decision making is essential for such systems to be broadly accepted. We introduce a Multi-Armed Bandit algorithm with fairness constraints, where fairness is defined as a minimum rate at which a task or a resource is assigned to a user. The proposed algorithm uses contextual information about the users and the task and makes no assumptions on how the losses capturing the performance of different users are generated. We provide theoretical guarantees of performance and empirical results from simulation and an online user study. The results highlight the benefit of accounting for contexts in fair decision making, especially when users perform better at some contexts and worse at others.
APA
Chen, Y., Cuellar, A., Luo, H., Modi, J., Nemlekar, H. & Nikolaidis, S.. (2020). Fair Contextual Multi-Armed Bandits: Theory and Experiments. Proceedings of the 36th Conference on Uncertainty in Artificial Intelligence (UAI), in Proceedings of Machine Learning Research 124:181-190 Available from https://proceedings.mlr.press/v124/chen20a.html.

Related Material