AI Paper Atlas

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

Label-similarity Curriculum Learning

Ürün Dogan, Aniket Anand Deshmukh, Marcin Bronislaw Machura, Christian Igel

ECCV 2020 · Label-efficient and robust learning › Recognition and long-tail · approach: Partial and multi-label learning · 21 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
replace one-hot labels with label-similarity curriculum (LCL)ResNet-50 · ImageNet 5% datatop-1 accuracy: 38.21 -> 42.21Improved
replace one-hot labels with LCLResNet-50 · ImageNet 100% datatop-1 accuracy: 76.25 -> 76.89Improved
replace one-hot labels with LCLResNet-101 · AWA2 5% datatop-1 accuracy: 23.09 -> 45.78Improved
replace one-hot labels with LCLResNet-101 · AWA2 5% datatop-5 accuracy: 53.31 -> 76.47Improved
replace one-hot labels with LCLResNeXt-101 · ImageNet 100% datatop-5 accuracy: 93.95 -> 96.15Improved
replace one-hot labels with LCLWRN · CIFAR100 100% datatop-1 accuracy: 79.52 -> 81.25Improved
replace static label smoothing with LCLDenseNet-BC · CIFAR100 100% datatop-1 accuracy: 83.01 -> 83.2Improved
replace knowledge distillation with LCLResNet-50 · ImageNet 5% datatop-1 accuracy: 39.81 -> 42.21Improved
increase epsilon from 0.9 to 0.999 (slower decay to one-hot) (ablation)LCL ResNet-50 · ImageNet 5% datatop-1 accuracy: 41.24 -> 42.21Improved
increase epsilon from 0.9 to 0.999 (ablation)LCL ResNet-101 · AWA2 10% datatop-1 accuracy: 47.21 -> 53.39Improved
replace one-hot labels with LCL on SENet-154 top-5 at 10% dataSENet-154 · ImageNet 10% datatop-5 accuracy: 83.71 -> 83.6Hurt
use curriculum with limited data (5% vs 100%) gain size (ablation)LCL ResNet-50 · ImageNettop-1 accuracy: 76.25 -> 76.89Improved

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