AI Paper Atlas

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

CAFE: Catastrophic Data Leakage in Vertical Federated Learning

Xiao Jin, Pin-Yu Chen, Chia-Yi Hsu, Chia-Mu Yu, Tianyi Chen

NeurIPS 2021 · Federated learning and privacy › Privacy attacks and secure learning · approach: Gradient inversion attacks · 15 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 CAFE rather than prior gradient-leakage attacksVFL data recovery · CIFAR-10, MNIST, and Linnaeus 5PSNR: CIFAR-10: 31.83 vs 18.94 (best baseline); MNIST: 43.15 vs 13.38; Linnaeus 5: 33.22 vs 12.32Improved
Add the gradient-matching loss term (α) (ablation)CAFE · Linnaeus 5PSNR: 28.62 -> 33.22Improved
Add the gradient-matching loss term (α) (ablation)CAFE · CIFAR-10 and MNISTPSNR: CIFAR-10: 33.93 -> 31.83; MNIST: 31.93 -> 43.15Mixed
Use a positive truncated-TV threshold ξ rather than ξ = 0 (ablation)CAFE · CIFAR-10, Linnaeus 5, and MNISTPSNR: CIFAR-10: 25.57 -> 31.83; Linnaeus 5: 25.29 -> 33.22; MNIST: 34.51 -> 43.15Improved
Add the TV-norm regularizer (β) (ablation)CAFE · CIFAR-10 and Linnaeus 5PSNR: CIFAR-10: 18.25 -> 31.83; Linnaeus 5: 23.22 -> 33.22Improved
Add the TV-norm regularizer (β) (ablation)CAFE · MNISTPSNR: 31.98 -> 43.15No clear effect
Add the internal-representation regularizer (γ) (ablation)CAFE · CIFAR-10, Linnaeus 5, and MNISTPSNR: CIFAR-10: 12.51 -> 31.83; Linnaeus 5: 12.37 -> 33.22; MNIST: 6.34 -> 43.15Improved
Increase batch size K from 10 through 100 (ablation)CAFE · CIFAR-10, MNIST, and Linnaeus 5PSNR: CIFAR-10: 30.83, 35.70, 31.83, 36.87, 38.94; MNIST: 32.60, 39.00, 43.15, 47.05, 47.50; Linnaeus 5: 28.00, 30.53, 33.22, No clear effect
Use the single-loop update protocol instead of nested loops (ablation)CAFE · CIFAR-10, MNIST, and Linnaeus 5iterations to stated PSNR stopping point: Single loop: 7300 (8000), 6600 (8000), 12400 (20000)Improved
Use fake gradients instead of ordinary gradientsCAFE attack · CIFAR-10PSNR: 28.68 -> 7.67Hurt
Use fake gradients as the training update signalVFL model training · CIFAR-10, Linnaeus 5, and MNISTtraining performanceNo clear effect
Increase the learning rate while attacking during training (ablation)CAFE · CIFAR-10, MNIST, and Linnaeus 5PSNR: CIFAR-10: 31.24 -> 27.62 -> 25.22; MNIST: 31.82 -> 28.42 -> 23.60; Linnaeus 5: 30.74 -> 21.45 -> 20.68Hurt

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