COPF: An Online Framework for Deployment-Stable Counterfactual Fairness in Evolving Graphs

2026-05-30

AI Summary

This work addresses the challenge of fairness estimation drift in online graph-based link recommendation, where system interventions destabilize the monitoring of group opportunity gaps. To tackle this, the authors propose the COPF framework, which uniquely integrates counterfactual fairness with performative decision-making in dynamic graphs. COPF enables real-time monitoring and control of exposure-based group opportunity disparities through a graph-aware doubly robust (GA-DR) estimator, an online multi-calibration auditor, and a primal-dual controller. The approach incorporates explicit exploration, propensity score logging, and residual outcome indistinguishability, yielding theoretical guarantees on fairness bounds under distributional shifts. Experiments on both the TGB benchmark and synthetic graph streams demonstrate that COPF significantly reduces worst-case group disparity peaks while preserving ranking utility nearly unchanged.

Abstract

Online link recommendation on evolving graphs is performative: by choosing which candidate links to show users, the system changes which links form and what feedback it later observes. Consequently, fairness estimates from logged outcomes can be misleading and may drift after deployment when the recommendation policy is updated. We introduce COPF (Counterfactual Online Performative Fairness), a decision-layer framework for deployment-stable fairness monitoring and control in online link recommendation. COPF (i) defines group-level opportunity gaps over exposure (shown vs. not shown) counterfactuals, (ii) makes them estimable by explicit exploration and by logging the probability (propensity) that each candidate is shown, and (iii) audits and controls fairness using residual outcome indistinguishability (OI) over a configurable auditor family with graph-aware doubly robust (GA-DR) estimators. We provide a noisy transfer theorem showing that Residual-OI on estimated GA-DR residuals implies bounds on exposure-counterfactual group gaps under temporal mixing and bounded local interference, and we instantiate an online multicalibration auditor together with a primal-dual controller. Experiments on two TGB streams and a controlled synthetic bipartite stream show that COPF reduces worst-case spikes in exposure-counterfactual group disparities with modest impact on ranking utility. Our code is available at https://github.com/lsnnnnnnnn/COPF.

Problem

• counterfactual fairness
• online link recommendation
• evolving graphs
• performative prediction
• fairness drift

Innovation

• counterfactual fairness
• online recommendation
• performative prediction
• doubly robust estimation
• graph-aware fairness