AI Paper Atlas

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

A Gift from Label Smoothing: Robust Training with Adaptive Label Smoothing via Auxiliary Classifier under Label Noise

Jongwoo Ko, Bongsoo Yi, Se-Young Yun

AAAI 2023 · Label-efficient and robust learning › Noisy labels · approach: Instance-dependent noise modeling

Read the paper

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 ALASCA to Standard trainingStandard · CIFAR-10, symmetric 20% noiseaccuracy: 87.0 $\pm$ 0.1 -> 92.2 $\pm$ 0.2Improved
Add ALASCA to Co-teachingCo-teaching · CIFAR-100, symmetric 80% noiseaccuracy: 20.5 $\pm$ 1.3 -> 42.2 $\pm$ 1.2Improved
Add ALASCA to ELR+ELR+ · Clothing1Maccuracy: 74.81 -> 74.92Improved
Add auxiliary classifiers (ablation)Uniform label smoothing · CIFAR-10, symmetric 20% noiseaccuracy: 87.3 $\pm$ 0.5 -> 91.1 $\pm$ 0.1Improved
Add adaptive rather than uniform label smoothing (ablation)Label smoothing with auxiliary classifiers · CIFAR-10, symmetric 20% noiseaccuracy: 91.1 $\pm$ 0.1 -> 91.8 $\pm$ 0.1Improved
Add EMA confidence to adaptive label smoothing without auxiliary classifiers (ablation)ALS · CIFAR-10, asymmetric 40% noiseaccuracy: 85.1 $\pm$ 0.2 -> 86.9 $\pm$ 1.1Mixed
Add EMA confidence to ALS + AC (ablation)ALS + AC · CIFAR-10, symmetric 20% noiseaccuracy: 91.8 $\pm$ 0.1 -> 92.2 $\pm$ 0.2Improved
Ensemble predictions from the main and auxiliary classifiers at inference (ablation)ALASCA · CIFAR-10, symmetric 20% noiseaccuracy: 92.2 $\pm$ 0.2 -> 92.8 $\pm$ 0.1Improved
Use bottleneck rather than MLP auxiliary classifiers at all blocks (ablation)ALASCA · CIFAR-10, instance-dependent 40% noiseaccuracy: 80.8 $\pm$ 0.3 -> 77.9 $\pm$ 0.4Hurt

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