AI Paper Atlas

Research map / Label-efficient and robust learning / Noisy labels

Drainage: A Unifying Framework for Addressing Class Uncertainty

Yasser Taha, Grégoire Montavon, Nils Körber

CVPR 2026 · Label-efficient and robust learning › Noisy labels · approach: Contrastive learning on noisy data · 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 drainage loss instead of cross-entropyCE baseline · CIFAR-100, asymmetric noise 0.4accuracy: 41.63 -> 61.55Improved
use drainage loss instead of cross-entropyCE baseline · CIFAR-10, asymmetric noise 0.45accuracy: 77.92 -> 82.23Improved
use drainage loss instead of GCEGCE · CIFAR-10, instance-dependent noise 0.5accuracy: 59.27 -> 64.22Improved
use drainage loss instead of GCEGCE · CIFAR-100, asymmetric noise 0.45accuracy: 42.70 -> 52.69Improved
use drainage loss on clean data instead of cross-entropyCE baseline · CIFAR-10, no noiseaccuracy: 92.16 -> 91.30Hurt
use drainage loss instead of CE on clean CIFAR-100CE baseline · CIFAR-100, no noiseaccuracy: 70.05 -> 73.31Improved
use drainage loss instead of GCEGCE · CIFAR-10N (worst labels)accuracy: 81.70 -> 79.85Hurt
use drainage loss instead of ANL CEANL CE · CIFAR-100Naccuracy: 56.62 -> 57.77Improved
use drainage loss instead of ANL CEANL CE · Clothing1Maccuracy: 68.44 -> 70.42Improved
use drainage loss instead of ANL CEANL CE · ILSVRC12 (mini, 50 classes)accuracy: 65.77 -> 64.90Hurt
use L2 instead of L1 regularization for CE (ablation)CE · CIFAR-10, asymmetric noise 0.4accuracy: 82.50 -> 73.38Hurt
use drainage node probability p_d as OSR score instead of CE max softmax probabilityDrainage OSR model · Open set recognition, CIFAR-10ROC AUC: 67.3 -> 72.4Improved

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