AI Paper Atlas

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

Self-Supervised Learning Disentangled Group Representation as Feature

Tan Wang, Zhongqi Yue, Jianqiang Huang, Qianru Sun, Hanwang Zhang

NeurIPS 2021 · Label-efficient and robust learning › Self-supervised and contrastive · approach: Invariance-based self-supervised learning · 4 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 IP-IRM to HCLHCL · STL10 400 epochslinear accuracy: 87.60 -> 87.81Improved
add IP-IRM to HCLHCL · STL10 400 epochsk-NN accuracy: 80.06 -> 84.29Improved
add IP-IRM to SimCLRSimCLR · STL10 400 epochslinear accuracy: 78.89 -> 84.44Improved
add IP-IRM to SimCLRSimCLR · Cifar100 400 epochslinear accuracy: 66.63 -> 69.55Improved
add IP-IRM to DCLDCL · Cifar100 400 epochslinear accuracy: 68.59 -> 68.76Improved
add IP-IRM to SimCLR with MixUp, 1000 epochsSimCLR · STL10k-NN accuracy: 79.80 -> 85.08Improved
add IP-IRM to SimCLR with MixUp, 1000 epochsSimCLR · Cifar100linear accuracy: 72.18 -> 73.99Improved
add IP-IRM to MoCo-v2MoCo-v2 · few-shot 5-way-5-shot averageaccuracy: 67.22 -> 69.44Improved
IP-IRM vs SSL baselines on NICOSSL baselines · NICOaccuracy: 1.5-4.2%Improved
IP-IRM vs SimCLR on disentanglementSimCLR · Shapes3Daverage score: 0.645 -> 0.672Improved
IP-IRM vs SimCLR on disentanglementSimCLR · CMNISTaverage score: 0.873 -> 0.906Improved
set lambda1 and lambda2 to 1.0 instead of 0.2-0.5 (ablation)IP-IRM · STL10 and Cifar100, 200 epochslinear accuracyHurt

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