Research map / Federated learning and privacy / Differential privacy
FedAdOb: Privacy-Preserving Fed erated Deep Learning With Ad aptive Ob fuscation
Hanlin Gu, Jiahuan Luo, Yan Kang, Yuan Yao, Gongxi Zhu, Bowen Li, Lixin Fan, Qiang Yang
Read the paperCheck your idea against 137,217 papers
AI-extracted. The measured effects below were read by AI from the paper's full text (no summary yet). Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.
Measured effects reported in the paper
| What was changed | Where | Measured | Effect |
|---|---|---|---|
| increase passport mean range N from 1 to 200 (ablation) | FedAdOb · AlexNet-CIFAR10 VFL, CAFE | feature recovery error: 0.0589 -> 0.1205 | Improved |
| increase passport mean range N from 1 to 200 (ablation) | FedAdOb · AlexNet-CIFAR10 VFL, CAFE | main task accuracy: 0.8798 -> 0.8739 | Hurt |
| increase passport variance sigma^2 from 0 to 100 (ablation) | FedAdOb · AlexNet-CIFAR10 VFL, PMC | label recovery error: 0.368 -> 0.751 | Improved |
| increase passport variance sigma^2 from 0 to 100 (ablation) | FedAdOb · AlexNet-CIFAR10 VFL, PMC | main task accuracy: 0.877 -> 0.791 | Hurt |
| use FedAdOb instead of no defense | VFL LeNet-MNIST · PMC attack | CAP: 0.117 -> 0.506 | Improved |
| use FedAdOb instead of no defense | VFL LeNet-MNIST · CAFE attack | CAP: 0.033 -> 0.137 | Improved |
| use FedAdOb instead of SplitFed against BMI | HFL · MNIST BMI | data recovery error: 0.052 -> 0.071 | Improved |
| use FedAdOb instead of SplitFed against BGI | HFL · MNIST BGI | data recovery error: 0.111 -> 0.111 | No clear effect |
| use FedAdOb instead of SplitFed against BGI | HFL · CIFAR100 BGI | data recovery error: 0.087 -> 0.107 | Improved |
| use InstaHide defense in HFL | FedAvg · CIFAR10 and CIFAR100 | main task accuracy: 18.6% and 10.66% best-case degradation, 52.12% and 16.65% worst case | Hurt |
| use InstaHide instead of FedAdOb in VFL training | VFL LeNet-MNIST · training time per epoch | training time (s): 7.05 -> 21.76 | Hurt |
| add FedAdOb to VFL training | no-defense VFL · LeNet-MNIST | training time per epoch (s): 7.03 -> 7.05 | No clear effect |
Closest papers by meaning
- A New Federated Learning Framework Against Gradient Inversion Attacks (AAAI 2025 · similarity 0.87)
- Clients Collaborate: Flexible Differentially Private Federated Learning with Guaranteed Improvement of Utility-Privacy Trade-off (ICML 2025 · similarity 0.87)
- Dual Defense: Enhancing Privacy and Mitigating Poisoning Attacks in Federated Learning (NeurIPS 2024 · similarity 0.86)
- Concealing Sensitive Samples against Gradient Leakage in Federated Learning (AAAI 2024 · similarity 0.86)
- Little Is Enough: Boosting Privacy by Sharing Only Hard Labels in Federated Semi-Supervised Learning (AAAI 2025 · similarity 0.86)
- DictPFL: Efficient and Private Federated Learning on Encrypted Gradients (NeurIPS 2025 · similarity 0.86)
Writing a paper in this area? scime draft is a free LaTeX editor with AI included, real-time co-authors and a check of your references and venue rules before you submit. It also imports OpenAI Prism projects.
