Conceptual
Browse Instruct Curate

Regularized Proportional Fairness Network for Payment-Free Resource Allocation

A neural mechanism for dividing divisible resources among strategic agents without monetary payments. It regularizes the proportional-fairness allocation with a learned approximator of the most-exploitable allocation, trading off Nash social welfare against incentive compatibility (exploitability), and comes with a subgradient method for computing PF exploitability and generalization bounds under finite and out-of-distribution samples.