Your Recourse, My Loss? Algorithmic Recourse under Shared Constraints
Published in ACM Conference on Fairness, Accountability, and Transparency (FAccT 2026), 2026
Studied algorithmic recourse in multi-agent settings, where recommendations given to one individual can affect the outcomes available to others. Formulated the problem as a capacity-constrained matching problem and developed an MILP optimization framework for allocating recourse recommendations across multiple individuals, achieving 97–99% of achievable social welfare with limited changes to the system.
Also presented at the NeurIPS 2025 Workshop on Algorithmic Collective Action, and at Remarkable 2026, hosted by the Vector Institute.
Recommended citation: Zahra Khotanlou, Kate Larson, and Amir-Hossein Karimi. (2026). "Your Recourse, My Loss? Algorithmic Recourse under Shared Constraints." ACM FAccT 2026.
Download Paper
