AI Paper Atlas

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

Exploring the Vulnerabilities of Federated Learning: A Deep Dive Into Gradient Inversion Attacks

Pengxin Guo, Runxi Wang, Shuang Zeng, Jinjing Zhu, Haoning Jiang, Yanran Wang, Yuyin Zhou, Feifei Wang, Hui Xiong, Liangqiong Qu

TPAMI 2025 · Federated learning and privacy › Privacy attacks and secure learning · approach: Gradient inversion attacks · 20 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
add server-side simulation with known local training info (strong simulation) versus noneIG under FedAvg · CIFAR-100, E=2, B=16PSNR (dB): 11.62 -> 15.17Improved
use FedAvg with multiple local epochs instead of single stepIG · CIFAR-100reconstruction qualityHurt
add more identical labels in one batch (ablation)IG · CIFAR-10, batch of 4gradient cosine similarity: -0.0305 -> 0.9541Hurt
increase batch size (ablation)IG · CIFAR-10/100, ImageNetattack performanceHurt
increase image resolution (ablation)IG · ImageNet vs CIFARattack performanceHurt
use a more complicated network (ResNet-18 to ResNet-101) (ablation)IG · CIFAR-100attack performanceHurt
attack a well-trained model instead of an untrained one (ablation)IG · CIFAR-100attack performanceHurt
use Sigmoid instead of other activations (ReLU, Tanh) (ablation)CI-Net · CIFAR/ImageNetreconstructionImproved
use random Gaussian noise in place of real gradients (ablation)GGL · ImageNetsemantic reconstructionNo clear effect
increase batch size from 1 to 64Robbing the Fed (1000 bins) · CIFAR-10images exactly reconstructed: 64/64 -> 60/64Hurt
increase batch size from 1 to 64Robbing the Fed (1000 bins) · ImageNetimages exactly reconstructed: 64/64 -> 61/64Hurt
increase image resolution (ablation)Fishing · ImageNetattack performanceHurt

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