AI Paper Atlas

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

Enhancing Learning with Noisy Labels via Rockafellian Relaxation

Louis Chen, Bobbie Chern, Eric Eckstrand, Amogh Mahapatra, Johannes O. Royset

ICLR 2026 · Label-efficient and robust learning › Noisy labels · approach: Sample selection and co-training

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
wrap Divide-Mix with RRMDivide-Mix · CIFAR-10N (Worst)accuracy: 92.56 -> 94.75Improved
wrap Divide-Mix with RRMDivide-Mix · CIFAR-100N (Noisy Fine)accuracy: 71.13 -> 73.98Improved
wrap ProMix with RRMProMix · CIFAR-10N (Worst)accuracy: 96.34 -> 96.32No clear effect
wrap ProMix with RRMProMix · CIFAR-100N (Noisy Fine)accuracy: 73.79 -> 74.19Improved
wrap CCE with RRMCCE · Clothing1Maccuracy: 68.94 -> 71.48Improved
wrap CC with RRMCC · Clothing1Maccuracy: 75.4 -> 75.69Improved
wrap CCE with RRMCCE · Food-101Naccuracy: 81.67 -> 84.21Improved
wrap CCE with RRMCCE · CIFAR-10 symmetric noise 30%accuracy: 81.90 -> 88.49Improved
wrap MSE with RRMMSE · CIFAR-10 symmetric noise 30%accuracy: 86.34 -> 89.95Improved
wrap MAE with RRMMAE · CIFAR-10 symmetric noiseaccuracy: 92.04 -> 90.10 at 10%; 83.72 -> 87.82 at 20%; 88.28 -> 82.98 at 30%Mixed
add RRM reweighting to adversarial training (A-RRM vs AT) (ablation)adversarial training with label noise · MNIST-10 20% noise, test eps 0.00accuracy: 58 -> 96Improved
use overestimated contamination C' = C + 10% instead of C (ablation)RRM-wrapped MSE · CIFAR-10 symmetric noise 30%accuracy: 89.95 -> 88.47Hurt

Earlier methods this paper compares with

Named baselines from earlier accepted papers in this paper's results tables: better than 1 of 2 on most shared results.

Matched by method name in results tables read by AI. Authors' own tables tend to favour the newer method, and a comparison can be missed when a method is renamed.

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