Learning Fair Policies in Multi-Objective (Deep) Reinforcement Learning with Average and Discounted Rewards

Umer Siddique, Paul Weng, Matthieu Zimmer
Proceedings of the 37th International Conference on Machine Learning, PMLR 119:8905-8915, 2020.

Abstract

As the operations of autonomous systems generally affect simultaneously several users, it is crucial that their designs account for fairness considerations. In contrast to standard (deep) reinforcement learning (RL), we investigate the problem of learning a policy that treats its users equitably. In this paper, we formulate this novel RL problem, in which an objective function, which encodes a notion of fairness that we formally define, is optimized. For this problem, we provide a theoretical discussion where we examine the case of discounted rewards and that of average rewards. During this analysis, we notably derive a new result in the standard RL setting, which is of independent interest: it states a novel bound on the approximation error with respect to the optimal average reward of that of a policy optimal for the discounted reward. Since learning with discounted rewards is generally easier, this discussion further justifies finding a fair policy for the average reward by learning a fair policy for the discounted reward. Thus, we describe how several classic deep RL algorithms can be adapted to our fair optimization problem, and we validate our approach with extensive experiments in three different domains.

Cite this Paper


BibTeX
@InProceedings{pmlr-v119-siddique20a, title = {Learning Fair Policies in Multi-Objective ({D}eep) Reinforcement Learning with Average and Discounted Rewards}, author = {Siddique, Umer and Weng, Paul and Zimmer, Matthieu}, booktitle = {Proceedings of the 37th International Conference on Machine Learning}, pages = {8905--8915}, year = {2020}, editor = {Hal Daumé III and Aarti Singh}, volume = {119}, series = {Proceedings of Machine Learning Research}, month = {13--18 Jul}, publisher = {PMLR}, pdf = {http://proceedings.mlr.press/v119/siddique20a/siddique20a.pdf}, url = { http://proceedings.mlr.press/v119/siddique20a.html }, abstract = {As the operations of autonomous systems generally affect simultaneously several users, it is crucial that their designs account for fairness considerations. In contrast to standard (deep) reinforcement learning (RL), we investigate the problem of learning a policy that treats its users equitably. In this paper, we formulate this novel RL problem, in which an objective function, which encodes a notion of fairness that we formally define, is optimized. For this problem, we provide a theoretical discussion where we examine the case of discounted rewards and that of average rewards. During this analysis, we notably derive a new result in the standard RL setting, which is of independent interest: it states a novel bound on the approximation error with respect to the optimal average reward of that of a policy optimal for the discounted reward. Since learning with discounted rewards is generally easier, this discussion further justifies finding a fair policy for the average reward by learning a fair policy for the discounted reward. Thus, we describe how several classic deep RL algorithms can be adapted to our fair optimization problem, and we validate our approach with extensive experiments in three different domains.} }
Endnote
%0 Conference Paper %T Learning Fair Policies in Multi-Objective (Deep) Reinforcement Learning with Average and Discounted Rewards %A Umer Siddique %A Paul Weng %A Matthieu Zimmer %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-siddique20a %I PMLR %P 8905--8915 %U http://proceedings.mlr.press/v119/siddique20a.html %V 119 %X As the operations of autonomous systems generally affect simultaneously several users, it is crucial that their designs account for fairness considerations. In contrast to standard (deep) reinforcement learning (RL), we investigate the problem of learning a policy that treats its users equitably. In this paper, we formulate this novel RL problem, in which an objective function, which encodes a notion of fairness that we formally define, is optimized. For this problem, we provide a theoretical discussion where we examine the case of discounted rewards and that of average rewards. During this analysis, we notably derive a new result in the standard RL setting, which is of independent interest: it states a novel bound on the approximation error with respect to the optimal average reward of that of a policy optimal for the discounted reward. Since learning with discounted rewards is generally easier, this discussion further justifies finding a fair policy for the average reward by learning a fair policy for the discounted reward. Thus, we describe how several classic deep RL algorithms can be adapted to our fair optimization problem, and we validate our approach with extensive experiments in three different domains.
APA
Siddique, U., Weng, P. & Zimmer, M.. (2020). Learning Fair Policies in Multi-Objective (Deep) Reinforcement Learning with Average and Discounted Rewards. Proceedings of the 37th International Conference on Machine Learning, in Proceedings of Machine Learning Research 119:8905-8915 Available from http://proceedings.mlr.press/v119/siddique20a.html .

Related Material