analiza aiqbench
This research addresses a critical gap in federated learning by providing an auditable framework that ensures differential privacy even when user participation streams are dynamically modified. By introducing a randomized buffering wrapper, the authors demonstrate how to lift standard non-adaptive privacy proofs to handle adaptive single-edit scenarios without sacrificing security guarantees. The work establishes a concrete trade-off between privacy parameters and system latency, offering a practical recipe for deploying robust continual learning systems in sensitive environments.
Continual Learning With Participation Privacy: An Auditable Buffering-Aggregation Recipe
arXiv:2607.07209v2 Announce Type: replace-cross Abstract: Modern federated and streaming learning systems often release intermediate models, so privacy must hold for the full trajectory under adaptive interaction. Motivated by participation privacy, we study single-edit neighboring user streams, where one insertion/deletion shifts all subsequent updates and defeats standard Hamming-neighbor continual-release analyses. We give an auditable modular recipe. A randomized buffering wrapper emits bins of size $[U,2U]$, reducing single-edit streams to a Hamming-style per-bin update stream with explicit backlog/delay guarantees, where $U$ is calibrated by the privacy parameters $(\varepsilon,\delta)$. We then prove a certification theorem identifying when a non-adaptive Hamming-neighbor DP proof for a continual primitive lifts to adaptive inputs: the primitive must use fresh per-round randomness and have a stable one-round privacy profile under common adaptive context. Together, these ingredients yield trajectory-level $(\varepsilon,\delta)$-DP for single-edit streams using standard primitives (e.g., tree prefix sums), with an explicit privacy--latency link via $U$.
Abstract:Modern federated and streaming learning systems often release intermediate models, so privacy must hold for the full trajectory under adaptive interaction. Motivated by participation privacy, we study single-edit neighboring user streams, where one insertion/deletion shifts all subsequent updates and defeats standard Hamming-neighbor continual-release analyses. We give an auditable modular recipe. A randomized buffering wrapper emits bins of size $[U,2U]$, reducing single-edit streams to a Hamming-style per-bin update stream with explicit backlog/delay guarantees, where $U$ is calibrated by the privacy parameters $(\varepsilon,\delta)$. We then prove a certification theorem identifying when a non-adaptive Hamming-neighbor DP proof for a continual primitive lifts to adaptive inputs: the primitive must use fresh per-round randomness and have a stable one-round privacy profile under common adaptive context. Together, these ingredients yield trajectory-level $(\varepsilon,\delta)$-DP for single-edit streams using standard primitives (e.g., tree prefix sums), with an explicit privacy--latency link via $U$.
Cryptography and Security (cs.CR); Machine Learning (cs.LG)
arXiv:2607.07209 [cs.CR]
(or arXiv:2607.07209v2 [cs.CR] for this version)
https://doi.org/10.48550/arXiv.2607.07209 Focus to learn more arXiv-issued DOI via DataCite
arXivLabs is a framework that allows collaborators to develop and share new arXiv features directly on our website.
Both individuals and organizations that work with arXivLabs have embraced and accepted our values of openness, community, excellence, and user data privacy. arXiv is committed to these values and only works with partners that adhere to them.
Have an idea for a project that will add value for arXiv's community? Learn more about arXivLabs.
Text integral sub licența deschisă a publicației · 239 cuvinte
Subiecte conexe
Echelon: Auditable Aggregate-Only Language-Model Adaptation Across Privacy Boundaries
arXiv · 3 iun. 2026
The Shibboleth Effect: Auditing the Cross-Lingual Distributional Skew of Large Language Models
arXiv · 10 iun. 2026
PoCoFL: POlicy-COmpliant Federated Learning
arXiv · 5 oct. 2026
Digital States 2026: AI Moves From Experiment to Infrastructure
govtech.com · 24 sept. 2026