analiza aiqbench
This arXiv preprint proposes a learnable lossy compressor for federated data condensation, aiming to cut client-side compute and reduce reconstruction risk from shared representations. The reported gains are a large client runtime drop and low reconstruction PSNR on small benchmarks (BloodMNIST, ChestMNIST, CIFAR-10), with CIFAR-10 accuracy at 65.63%. Readers should treat the privacy claims as results from specific attacks on medical-imaging-style datasets, not as general guarantees.
FedSumUp: User-Auditable Privacy-Aware Representation Sharing for Resource-Constrained Federated Intelligence
arXiv:2610.21194v1 Announce Type: cross Abstract: Existing federated data condensation methods face privacy risks, as uploaded synthetic data may preserve sensitive information and remain vulnerable to reconstruction attacks. Moreover, their generation often incurs considerable client-side computational overhead, limiting deployment on resource-constrained devices. To address these issues, we propose FedSumUp, which employs a learnable lossy compressor on clients to generate compact and visually interpretable representations. By controlling information capacity, FedSumUp reduces privacy leakage while allowing users to directly assess the privacy-utility tradeoff. We further introduce a dual-objective optimization mechanism to preserve task-relevant semantics while suppressing reconstruction risks. The compressor is trained on the server and only performs lightweight inference on clients. Since the server cannot access raw data, we develop a data-free closed-loop optimization strategy that optimizes the compressor by minimizing the discrepancy between original and twice-compressed representations after structure-preserving upsampling. Experiments on BloodMNIST, CIFAR-10, and ChestMNIST demonstrate reconstruction PSNR as low as 8.56 dB (with synthetic PSNR as low as 5.45 dB), client runtime reductions up to 96.5% (from 778s to 27s on ChestMNIST), and competitive classification accuracy (89.72% ChestMNIST, 82.24% BloodMNIST, 65.63% CIFAR-10), while remaining inherently robust to non-IID data distributions.
Abstract:Existing federated data condensation methods face privacy risks, as uploaded synthetic data may preserve sensitive information and remain vulnerable to reconstruction attacks. Moreover, their generation often incurs considerable client-side computational overhead, limiting deployment on resource-constrained devices. To address these issues, we propose FedSumUp, which employs a learnable lossy compressor on clients to generate compact and visually interpretable representations. By controlling information capacity, FedSumUp reduces privacy leakage while allowing users to directly assess the privacy-utility tradeoff. We further introduce a dual-objective optimization mechanism to preserve task-relevant semantics while suppressing reconstruction risks. The compressor is trained on the server and only performs lightweight inference on clients. Since the server cannot access raw data, we develop a data-free closed-loop optimization strategy that optimizes the compressor by minimizing the discrepancy between original and twice-compressed representations after structure-preserving upsampling. Experiments on BloodMNIST, CIFAR-10, and ChestMNIST demonstrate reconstruction PSNR as low as 8.56 dB (with synthetic PSNR as low as 5.45 dB), client runtime reductions up to 96.5% (from 778s to 27s on ChestMNIST), and competitive classification accuracy (89.72% ChestMNIST, 82.24% BloodMNIST, 65.63% CIFAR-10), while remaining inherently robust to non-IID data distributions.
21 pages, including appendices. Jingwei Zhao and Tiantong Sun contributed equally. Corresponding author: Xiaojin Zhang
Cryptography and Security (cs.CR); Artificial Intelligence (cs.AI)
arXiv:2610.21194 [cs.CR]
(or arXiv:2610.21194v1 [cs.CR] for this version)
https://doi.org/10.48550/arXiv.2610.21194 Focus to learn more arXiv-issued DOI via DataCite (pending registration)
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 · 299 cuvinte
Subiecte conexe
Digital States 2026: AI Moves From Experiment to Infrastructure
govtech.com · 24 sept. 2026
Drift-Aware LLM Routing with Sparse Contexts and Shared Budgets
arXiv · 2 sept. 2026
Elimination Geometry
arXiv · 20 aug. 2026
Is Fairness Truly Fair? Towards Reliable Lipschitz Fairness in Multi-Task Learning via Fixed-\texorpdfstring{$\delta$}{delta} Alignment
arXiv · 10 iun. 2026