AI Paper Atlas

Research map / Label-efficient and robust learning / Recognition and long-tail

DLME: Deep Local-Flatness Manifold Embedding

Zelin Zang, Siyuan Li, Di Wu, Ge Wang, Kai Wang, Lei Shang, Baigui Sun, Hao Li, Stan Z. Li

ECCV 2022 · Label-efficient and robust learning › Recognition and long-tail · approach: Correspondence and causal representation learning · 26 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 DLME instead of UMAP for 2D embeddinglinear SVM · EMnistclassification accuracy: 0.588 -> 0.657Improved
use DLME instead of best CL baselineSimCLR-style linear test · TinyImageNetlinear-test accuracy: 0.388 -> 0.449Improved
use DLME instead of CC for deep clusteringCC · TinyImageNetclustering accuracy: 0.140 -> 0.182Improved
use DLME instead of CRLC for deep clusteringCRLC · CIFAR10clustering accuracy: 0.799 -> 0.822Improved
add LD gradient on structure network f (versus CL loss on f) (ablation)DLME · CIFAR10 clusteringclustering accuracy: 0.792 -> 0.822Improved
add t-distribution kernel (versus standard kernel) (ablation)DLME-A1 · TinyImageNet clusteringclustering accuracy: 0.145 -> 0.133Mixed
add two-network structure (versus single-network CL) (ablation)DLME-A2 · CIFAR10 linear testlinear-test accuracy: 0.888 -> 0.902Improved
add data augmentation-based CL framework with LD loss (ablation)DLME-A3 · CIFAR100 linear testlinear-test accuracy: 0.624 -> 0.661Improved
use DLME vs tSNE on KMnist classificationlinear SVM · KMnistclassification accuracy: 0.738 -> 0.782Improved
use DLME on HCL vs tSNElinear SVM · HCLclassification accuracy: 0.874 -> 0.874No 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