AI Paper Atlas

Research map / Federated learning and privacy / Personalized and heterogeneous FL

FLAIR: Federated Learning Annotated Image Repository

Congzheng Song, Filip Granqvist, Kunal Talwar

NeurIPS 2022 · Federated learning and privacy › Personalized and heterogeneous FL · approach: Federated learning with pretrained models · 3 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
federated instead of centralized training, coarse labels, random initResNet-18 · FLAIR coarseper-class AP: 60.40 -> 50.41Hurt
federated instead of centralized training, coarse labels, ImageNet initResNet-18 · FLAIR coarseper-class AP: 67.71 -> 62.09Hurt
add central differential privacy to federated training, coarse, random initResNet-18 · FLAIR coarseper-class AP: 50.41 -> 28.80Hurt
add central differential privacy to federated training, coarse, ImageNet initResNet-18 · FLAIR coarseper-class AP: 62.09 -> 44.28Hurt
ImageNet pretrained init instead of random init, centralized coarseResNet-18 · FLAIR coarseper-class AP: 60.40 -> 67.71Improved
ImageNet pretrained init instead of random init, private federated coarseResNet-18 · FLAIR coarseper-class AP: 28.80 -> 44.28Improved
federated instead of centralized training, fine labels, random initResNet-18 · FLAIR fineper-class AP: 14.90 -> 1.53Hurt
federated instead of centralized training, fine labels, ImageNet initResNet-18 · FLAIR fineper-class AP: 20.26 -> 2.03Hurt
add central differential privacy, fine labels, ImageNet initResNet-18 · FLAIR fineoverall AP: 27.31 -> 12.67Hurt
ImageNet pretrained init instead of random init, federated fineResNet-18 · FLAIR fineoverall AP: 22.68 -> 27.31Improved
federated training instead of centralized on binary structure labelResNet-18 · FLAIR binary structureAP: 87.72 -> 84.22Hurt
add differential privacy on binary structure label, ImageNet initResNet-18 · FLAIR binary structureAP: 90.49 -> 83.41Hurt

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