Fairness in Reinforcement Learning

2016/11/09 by Shahin Jabbari, Matthew Joseph, Jabbari, Shahin +7 · 7 citations
Computer Science · Decision Sciences · Social Sciences · #Auction Theory and Applications #Ethics and Social Impacts of AI #FOS: Computer and information sciences #Machine Learning (cs.LG) #Reinforcement Learning in Robotics

paper · pdf · doi:10.48550/arxiv.1611.03071

openalex publication_date 2016/11/09 · openalex created_date 2025/10/10 · openalex updated_date 2026/07/28

Abstract

We initiate the study of fairness in reinforcement learning, where the actions of a learning algorithm may affect its environment and future rewards. Our fairness constraint requires that an algorithm never prefers one action over another if the long-term (discounted) reward of choosing the latter action is higher. Our first result is negative: despite the fact that fairness is consistent with the optimal policy, any learning algorithm satisfying fairness must take time exponential in the number of states to achieve non-trivial approximation to the optimal policy. We then provide a provably fair polynomial time algorithm under an approximate notion of fairness, thus establishing an exponential gap between exact and approximate fairness

Citations

• Learning Non-Discriminatory Predictors

Cited by

• FairGRPO: Fair Reinforcement Learning for Equitable Clinical Reasoning
• VORTEX: Aligning Task Utility and Human Preferences through LLM-Guided Reward Shaping
• Feasibility-Guided Fair Adaptive Offline Reinforcement Learning for Medicaid Care Management
• Learning Time-Varying Convexifications of Multiple Fairness Measures
• The Authority of "Fair" in Machine Learning
• A Causal Lens for Learning Long-term Fair Policies
• Collaborative Min-Max Regret in Grouped Multi-Armed Bandits

Related

• Survey on Fair Reinforcement Learning: Theory and Practice
• Learning Fair Policies in Multiobjective (Deep) Reinforcement Learning\n with Average and Discounted Rewards
• Fairness in Reinforcement Learning: A Survey
• What Hides behind Unfairness? Exploring Dynamics Fairness in Reinforcement Learning
• A Causal Lens for Learning Long-term Fair Policies