AI Paper Atlas

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

Privacy-Preserving Collaborative Learning With Automatic Transformation Search

Wei Gao, Shangwei Guo, Tianwei Zhang, Han Qiu, Yonggang Wen, Yang Liu

CVPR 2021 · Federated learning and privacy › Privacy attacks and secure learning · approach: Encryption-based private learning · 38 citations (OpenAlex)

Check 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
Replace untransformed training with the searched hybrid policyCollaborative ResNet20 training on CIFAR100 · CIFAR100, ResNet20, Adam+Cosine attackPSNR: 13.88 -> 7.64Improved
Replace untransformed training with the searched hybrid policyCollaborative ResNet20 training on CIFAR100 · CIFAR100, ResNet20ACC: 76.88 -> 77.92Improved
Use randomly selected augmentation policies instead of searched policiesCollaborative learning · CIFAR100, ResNet20PSNR: 11.41 vs 7.64Hurt
Apply the hybrid of searched policies rather than one fixed searched policy (ablation)Collaborative ResNet20 training · CIFAR100, ResNet20ACC: 70.56 -> 77.92Improved
Apply the hybrid of searched policies rather than one fixed searched policy (ablation)Collaborative ConvNet training · CIFAR100, ConvNetACC: 66.98 -> 70.27Improved
Replace untransformed training with the searched hybrid policyCollaborative learning · Fashion MNIST, ResNet20PSNR: 10.04 -> 7.60Improved
Replace untransformed training with the searched hybrid policyCollaborative learning · Fashion MNIST, ConvNetACC: 94.25 -> 91.35Hurt
Use the searched hybrid policy instead of no defenseCIFAR100, ResNet20 collaborative training · Adam+Cosine attackPSNR: 13.88 -> 7.64Improved
Use gradient pruning at 99% compressionCIFAR100, ResNet20 collaborative training · Adam+Cosine attackACC: 58.33Hurt
Increase the number of random-search policy candidates from 1500 to 5000 (ablation)Automatic transformation search · Search candidate-count analysisprivacy score rangeNo clear effect

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