SYNTOLOGY HomeExplorerAtlasCodeMethodologyAboutDevelopersFeedPricing
Paper · 2510.19934 · NeurIPS · 2025

Mitigating Privacy-Utility Trade-off in Decentralized Federated Learning via f -Differential Privacy

Weijie Su, Qi Long, Xiang Li, Chendi Wang, Buxin Su

arXiv · PDF · Open in the Atlas

Code that ran

We lifted 17 functions out of this paper's own repositories and ran 1 of them in a sandbox. "Ran" means the function executed on a synthesized input and returned a value. It is not a reproduction of the paper's results.

RepositoryRoleRan
lx10077/PN-f-DP canonical 1 of 17
FunctionStatusWhere it lives
rdp_account Ran lx10077/PN-f-DP/corelated_noises/utils/dp_account.py
pointer only (licence: NONE) · get_code("43752e988febb856")
binary_search_eps Not yet run lx10077/PN-f-DP/corelated_noises/utils/param_search.py
pointer only (licence: NONE) · get_code("60f6e28a813edde5")
clip_vector Not yet run lx10077/PN-f-DP/corelated_noises/misc.py
pointer only (licence: NONE) · get_code("6d9cb7162352c6f2")
compute_rho Not yet run lx10077/PN-f-DP/corelated_noises/findSigma_original.py
pointer only (licence: NONE) · get_code("6b3f873641382c7b")
compute_rho_all Not yet run lx10077/PN-f-DP/corelated_noises/findSigma_original.py
pointer only (licence: NONE) · get_code("4ed12b184697158a")
consensus_distance Not yet run lx10077/PN-f-DP/corelated_noises/utils/optimizers.py
pointer only (licence: NONE) · get_code("ce162fc138403b5b")
consensus_error Not yet run lx10077/PN-f-DP/corelated_noises/utils/avg_consensus_algorithms.py
pointer only (licence: NONE) · get_code("c583562f71041a41")
find_sigma_cor_eps Not yet run lx10077/PN-f-DP/corelated_noises/utils/param_search.py
pointer only (licence: NONE) · get_code("cc0147c5cbc2ac6a")
finite_time_consensus Not yet run lx10077/PN-f-DP/corelated_noises/utils/avg_consensus_algorithms.py
pointer only (licence: NONE) · get_code("a8cd575c56e64b43")
flatten Not yet run lx10077/PN-f-DP/corelated_noises/misc.py
pointer only (licence: NONE) · get_code("3f3febe09cc2eb46")
minimize_alpha Not yet run lx10077/PN-f-DP/corelated_noises/utils/dp_account.py
pointer only (licence: NONE) · get_code("aaefe1981ef8fb18")
optimize_averaging Not yet run lx10077/PN-f-DP/corelated_noises/utils/optimizers.py
pointer only (licence: NONE) · get_code("ca1b2f43e04dedc6")
optimize_decentralized Not yet run lx10077/PN-f-DP/corelated_noises/utils/optimizers.py
pointer only (licence: NONE) · get_code("f5e6837a5fd81bcc")
second_smallest_eigenvalue Not yet run lx10077/PN-f-DP/corelated_noises/findSigma_original.py
pointer only (licence: NONE) · get_code("65dde04c867913a7")
standard_algo Not yet run lx10077/PN-f-DP/corelated_noises/utils/avg_consensus_algorithms.py
pointer only (licence: NONE) · get_code("8b361b258de1a8c9")
unflatten Not yet run lx10077/PN-f-DP/corelated_noises/misc.py
pointer only (licence: NONE) · get_code("e706424e4a18d7dd")
user_level_rdp Not yet run lx10077/PN-f-DP/corelated_noises/utils/dp_account.py
pointer only (licence: NONE) · get_code("0c9f55048e70292c")

Repositories linked to this paper

Some links come from the archived Papers with Code dataset (CC BY-SA 4.0): attribution and licence.

Abstract

Differentially private (DP) decentralized Federated Learning (FL) allows local users to collaborate without sharing their data with a central server. However, accurately quantifying the privacy budget of private FL algorithms is challenging due to the co-existence of complex algorithmic components such as decentralized communication and local updates. This paper addresses privacy accounting for two decentralized FL algorithms within the f -differential privacy (f -DP) framework. We develop two new f -DP-based accounting methods tailored to decentralized settings: Pairwise Network f -DP (PN-f -DP), which quantifies privacy leakage between user pairs under random-walk communication, and Secret-based f -Local DP (Sec-f -LDP), which supports structured noise injection via shared secrets. By combining tools from f -DP theory and Markov chain concentration, our accounting framework captures privacy amplification arising from sparse communication, local iterations, and correlated noise. Experiments on synthetic and real datasets demonstrate that our methods yield consistently tighter (ϵ, δ) bounds and improved utility compared to Rényi DP-based approaches, illustrating the benefits of f -DP in decentralized privacy accounting.

For agents

The same record, over MCP at https://syntology.ai/mcp:

get_harvested_code_for_paper("2510.19934")
get_code_for_paper("2510.19934")
have("2510.19934")

Connect an agent — have() is free.