Enregistré dans:
Détails bibliographiques
Auteurs principaux: Witt, Leon, Abbasli, Togrul, Toyoda, Kentaroh, Samek, Wojciech, Klinger, Lucy
Format: Preprint
Publié: 2026
Sujets:
Accès en ligne:https://arxiv.org/abs/2605.04747
Tags: Ajouter un tag
Pas de tags, Soyez le premier à ajouter un tag!
_version_ 1866918486035398656
author Witt, Leon
Abbasli, Togrul
Toyoda, Kentaroh
Samek, Wojciech
Klinger, Lucy
author_facet Witt, Leon
Abbasli, Togrul
Toyoda, Kentaroh
Samek, Wojciech
Klinger, Lucy
contents 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.
format Preprint
id arxiv_https___arxiv_org_abs_2605_04747
institution arXiv
publishDate 2026
record_format arxiv
spellingShingle Knowledge-Free Correlated Agreement for Incentivizing Federated Learning
Witt, Leon
Abbasli, Togrul
Toyoda, Kentaroh
Samek, Wojciech
Klinger, Lucy
Machine Learning
Artificial Intelligence
Computer Science and Game Theory
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.
title Knowledge-Free Correlated Agreement for Incentivizing Federated Learning
topic Machine Learning
Artificial Intelligence
Computer Science and Game Theory
url https://arxiv.org/abs/2605.04747