Data-Free Contribution Estimation in Federated Learning using Gradient von Neumann Entropy

Fuente: arXiv
Saved in:
Bibliographic Details
Main Authors: Ukaye, Asim, Abdu-Aguye, Mubarak, Tastan, Nurbek, Nandakumar, Karthik
Format: Preprint
Published: 2026
Subjects:
Online Access:
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866908991097929728
author Ukaye, Asim
Abdu-Aguye, Mubarak
Tastan, Nurbek
Nandakumar, Karthik
author_facet Ukaye, Asim
Abdu-Aguye, Mubarak
Tastan, Nurbek
Nandakumar, Karthik
contents Client contribution estimation in Federated Learning is necessary for identifying clients' importance and for providing fair rewards. Current methods often rely on server-side validation data or self-reported client information, which can compromise privacy or be susceptible to manipulation. We introduce a data-free signal based on the matrix von Neumann (spectral) entropy of the final-layer updates, which measures the diversity of the information contributed. We instantiate two practical schemes: (i) SpectralFed, which uses normalized entropy as aggregation weights, and (ii) SpectralFuse, which fuses entropy with class-specific alignment via a rank-adaptive Kalman filter for per-round stability. Across CIFAR-10/100 and the naturally partitioned FEMNIST and FedISIC benchmarks, entropy-derived scores show a consistently high correlation with standalone client accuracy under diverse non-IID regimes - without validation data or client metadata. We compare our results with data-free contribution estimation baselines and show that spectral entropy serves as a useful indicator of client contribution.
format Preprint
id arxiv_https___arxiv_org_abs_2604_22562
institution arXiv
publishDate 2026
record_format arxiv
spellingShingle Data-Free Contribution Estimation in Federated Learning using Gradient von Neumann Entropy
Ukaye, Asim
Abdu-Aguye, Mubarak
Tastan, Nurbek
Nandakumar, Karthik
Machine Learning
Artificial Intelligence
Computer Vision and Pattern Recognition
Distributed, Parallel, and Cluster Computing
Client contribution estimation in Federated Learning is necessary for identifying clients' importance and for providing fair rewards. Current methods often rely on server-side validation data or self-reported client information, which can compromise privacy or be susceptible to manipulation. We introduce a data-free signal based on the matrix von Neumann (spectral) entropy of the final-layer updates, which measures the diversity of the information contributed. We instantiate two practical schemes: (i) SpectralFed, which uses normalized entropy as aggregation weights, and (ii) SpectralFuse, which fuses entropy with class-specific alignment via a rank-adaptive Kalman filter for per-round stability. Across CIFAR-10/100 and the naturally partitioned FEMNIST and FedISIC benchmarks, entropy-derived scores show a consistently high correlation with standalone client accuracy under diverse non-IID regimes - without validation data or client metadata. We compare our results with data-free contribution estimation baselines and show that spectral entropy serves as a useful indicator of client contribution.
title Data-Free Contribution Estimation in Federated Learning using Gradient von Neumann Entropy
topic Machine Learning
Artificial Intelligence
Computer Vision and Pattern Recognition
Distributed, Parallel, and Cluster Computing
url https://arxiv.org/abs/2604.22562