AI Paper Atlas

Research map / Federated learning and privacy / Differential privacy

Differentially Private Image Classification by Learning Priors from Random Processes

Xinyu Tang, Ashwinee Panda, Vikash Sehwag, Prateek Mittal

NeurIPS 2023 · Federated learning and privacy › Differential privacy · approach: DP-SGD refinements · 2 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 Phase I synthetic-data prior and Phase II linear probing to DP-SGDDP-SGD WRN-16-4 · CIFAR10 eps=1test accuracy (%): 56.8 -> 72.32Improved
Add DP-RandP phasesDP-SGD · CIFAR100 eps=3test accuracy (%): 30.73 -> 43.33Improved
Add DP-RandP phasesDP-SGD · DermaMNIST val eps=1validation accuracy (%): 69.00 -> 71.78Improved
Remove Phase II linear probing (all budget to Phase III) (ablation)DP-RandP · CIFAR10 eps=1test accuracy (%): 72.32 -> 69.03Hurt
Remove Phase II (ablation)DP-RandP · CIFAR10 eps=8test accuracy (%): 85.26 -> 84.45Hurt
Allocate all privacy budget to Phase II (linear probing only) (ablation)DP-RandP · CIFAR10 eps=3test accuracy (%): 79.99 -> 71.08Hurt
Vary epochs of Phase II from 0 to 94 at eps=2 (ablation)DP-RandP · CIFAR10 eps=2test accuracy (%): 75.30 -> 77.35 -> 63.10Mixed
Vary Phase II epochs 4,8,12,16 at eps=3 (ablation)DP-RandP · CIFAR10 eps=3test accuracy (%): 79.71 / 79.74 / 79.48 / 80.27No clear effect
Use MoCo instead of alignment/uniformity loss in Phase I (ablation)DP-RandP · CIFAR10 eps=8test accuracy (%): 85.26 -> 82.84Hurt
Use MoCo instead of alignment/uniformity loss in Phase I (ablation)DP-RandP · CIFAR10 eps=1test accuracy (%): 72.32 -> 72.79No clear effect
Use WRN-40-4 instead of WRN-16-4 (ablation)DP-RandP · CIFAR10 eps=1test accuracy (%): 72.32 -> 73.15Improved
Different synthetic datasets for Phase I (Dead leaves vs StyleGAN-Oriented) (ablation)private linear probing · CIFAR10 eps=1test accuracy (%): 59.92 -> 67.79Improved

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