Dependable and Secure Privacy Trading in Federated Learning via MFG-RegretNet
Shanghai Jiao Tong University · Lancaster University
Privacy is a resource that participants can choose to trade. This manuscript studies privacy procurement in federated learning through a Privacy Auction Game, population-aware neural allocation and payment, and regret-constrained mechanism learning.
PDF build: · Added to this website: · 1.81 MiB
Cite the manuscript
K. Sun, J. Li, X. Chen, W. Meng, and M. Guo, “Dependable and Secure Privacy Trading in Federated Learning via MFG-RegretNet,” author manuscript, 2026.
View BibTeX
@unpublished{sun2026mfgregretnet,
author = {Sun, Kangkang and Li, Jianhua and Chen, Xiuzhen
and Meng, Weizhi and Guo, Minyi},
title = {Dependable and Secure Privacy Trading in Federated
Learning via {MFG-RegretNet}},
year = {2026},
note = {Author manuscript; PDF compiled October 15, 2026},
url = {https://privacy-as-commodity.pages.dev/papers/mfg-regretnet.pdf}
}Follow the paper through the website
See the mechanism
Explore population context, equilibrium reasoning, neural offers, and voluntary participation.
Open method explorer ↗02 / BASELINESUnderstand the alternatives
Follow the decisions made by each comparison method through a worked example.
Browse the baselines ↗03 / EVIDENCECompare the reported results
Read accuracy, incentives, and welfare alongside the resources and uncertainty reported.
Open results comparison ↗The PDF contains the manuscript's assumptions, analysis, and experimental discussion. Website interactions are teaching examples; result charts retain the separately documented 20 Oct 2026 manuscript-table snapshot.