We introduce Knowledge-Free Correlated Agreement (KFCA) to reward client contributions in federated learning (FL) without relying on ground truth, a public test set, or distribution knowledge. Under categorical reports and an honest majority, KFCA is strictly truthful, addressing the label-flipping vulnerability of Correlated Agreement (CA). We evaluate KFCA on federated LLM adapter tuning and a real-world PCB inspection task, showing efficient real-time reward computation suitable for decentralized and blockchain-based incentive designs.
Cite
@article{arxiv.2605.04747,
title = {Knowledge-Free Correlated Agreement for Incentivizing Federated Learning},
author = {Leon Witt and Togrul Abbasli and Kentaroh Toyoda and Wojciech Samek and Lucy Klinger},
journal= {arXiv preprint arXiv:2605.04747},
year = {2026}
}