AI Paper Atlas

Research map / Label-efficient and robust learning / Self-supervised and contrastive

Understanding the Robustness of Distributed Self-Supervised Learning Frameworks Against Non-IID Data

Xuanyu Chen, NAN YANG, Shuai Wang, Dong Yuan

ICLR 2026 · Label-efficient and robust learning › Self-supervised and contrastive · approach: Data-efficient and robust self-supervision · 0 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 MAE (MIM) instead of SimSiam (CL) pre-trainingdistributed ViT pre-training · CIFAR-100 label non-IIDIID-to-non-IID accuracy gap: SimSiam 5.11 drop vs MAE 1.09 dropImproved
use MAE instead of SimSiamdistributed CNN pre-training · CIFAR-10 label non-IIDaccuracy gap: 1.70 drop -> 0.31 dropImproved
use MAE instead of SimSiamdistributed ViT pre-training · ImageNet feature non-IIDaccuracy gap: 2.84 drop -> 0.18 dropImproved
add MAR loss to MAEMAE + ViT · local-global weight distancel2-norm difference: 36.10 -> 35.75 (local 1 vs local 100)Improved
replace SimSiam by MAEViT · local-global weight distancel2-norm difference: 45.37 -> 36.10 (local 1 vs local 100)Improved
increase average connectivity from about 5 to about 10decentralized MIM · CIFAR-100accuracy: 46.62 -> 52.28 (data size weights)Improved
use federated instead of decentralized learning (connectivity about 5)distributed MIM · CIFAR-100accuracy: 45.22-46.62 -> 52.72Improved
change consensus matrix scheme (data size, degree normalized, doubly stochastic, push sum) (ablation)decentralized MIM · CIFAR-100accuracy: 52.28, 52.08, 52.18, 52.22 at connectivity about 10No clear effect
add A-MMD (mean sigma) instead of cosine similarity (ablation)MAR loss · CIFAR-100accuracy: 52.47 -> 54.39 (gamma 1e-2)Improved
use mean instead of median pairwise distance for A-MMD bandwidth (ablation)MAR loss · CIFAR-100accuracy: 54.13 -> 54.39 (gamma 1e-2)Improved
add alignment term (gamma 1e-2, A-MMD mean) to vanilla MIM (ablation)MAE · CIFAR-100accuracy: 51.45 -> 54.39Improved
set regularization weight gamma to 1 instead of 1e-2 (ablation)MAR loss · CIFAR-100accuracy: 54.39 -> 51.50Hurt

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