AI Paper Atlas

Research map / Federated learning and privacy / Privacy attacks and secure learning

Diffusion to Obfuscation: Time-Adaptive Synthesized Generation Against Gradient Leakage Attacks in Federated Learning

Farchan Raswa, Chun-Shien Lu, Jia-Ching Wang

ECCV 2026 · Federated learning and privacy › Privacy attacks and secure learning · approach: Gradient inversion attacks · 0 citations (OpenAlex)

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 changedWhereMeasuredEffect
use Diff2Obs-I instead of no defense against GI-NASDiff2Obs · ImageNet, GI-NASLPIPS: 0.292 -> 0.838Improved
use Diff2Obs instead of CENSOR to keep accuracyDiff2Obs · CIFAR-10 test accuracytest accuracy (%): 25.79 -> 75.48Improved
use Diff2Obs instead of RefinerDiff2Obs · CIFAR-10 overheadtime to best accuracy (hour): 0.79 -> 0.05Improved
Diff2Obs-I at larger batch size 4 to 32 (ablation)Diff2Obs (Obs-I) · ImageNet HFGradInvLPIPS: 0.834 -> 0.960Improved
Diff2Obs-I instead of CENSOR at batch size 4 under GIFDDiff2Obs (Obs-I) · ImageNet GIFDLPIPS: 0.878 -> 0.883Improved
adversary adds EoT adaptive attack against Diff2Obs (ablation)Diff2Obs · ImageNet GIFDLPIPS: 0.844 -> 0.829No clear effect
adversary adds EoT adaptive attack against Diff2Obs (ablation)Diff2Obs · CIFAR-10 GIFDLPIPS: 0.742 -> 0.761No clear effect
move from IID to non-IID Dirichlet alpha=1 (ablation)Diff2Obs · CIFAR-10PMM: 98.10% -> 98.86%No clear effect
increase client samples for DDPM from 50 to 1000 (ablation)Diff2Obs · CIFAR-10test accuracy (%): 69.50 -> 75.55Improved
increase client samples for DDPM from 50 to 1000 (ablation)Diff2Obs · CIFAR-10 HFGradInvLPIPS: 0.835 -> 0.842No clear effect
replace client DDPM by pre-trained Stable Diffusion (ablation)Diff2Obs · CIFAR-10test accuracy (%): 75.48 -> 69.20Hurt
replace DDPM trained on 500 samples by Stable Diffusion in FID (ablation)Diff2Obs · CIFAR-10FID: 24.30 -> 57.30Hurt

Closest papers by meaning

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.

Open scime draft

Get the next scime Dispatch and a monthly field update

Free. We send a confirmation link first; every email has a one-click unsubscribe. Privacy