AI Paper Atlas

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

ALIM: Adjusting Label Importance Mechanism for Noisy Partial Label Learning

Mingyu Xu, Zheng Lian, Lei Feng, Bin Liu, Jianhua Tao

NeurIPS 2023 · Label-efficient and robust learning › Noisy labels · approach: Instance-dependent noise modeling · 1 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
Add ALIM label-importance weighting to PiCO with Onehot normalizationPiCO · CIFAR-10 (q=0.3, η=0.3)accuracy: 92.94±0.24 -> 94.67±0.21Improved
Add ALIM label-importance weighting to PiCO with Scale normalizationPiCO · CIFAR-100 (q=0.03, η=0.3)accuracy: 69.97±0.01 -> 75.92±0.14Improved
Add ALIM-Scale to PiCO+ with mixupPiCO · CUB-200 (q=0.05, η=0.2)accuracy: 60.65±0.79 -> 68.38±0.47Improved
Add ALIM-Scale to PiCO+ with mixupPiCO · CIFAR-100 (q=0.05, η=0.5)accuracy: 60.50±0.99 -> 72.26±0.25Improved
Add mixup training to ALIM-Onehot (ablation)ALIM-Onehot on PiCO · CIFAR-10 (q=0.3, η=0.3)accuracy: 92.42±0.17 -> 94.67±0.21Improved
Add mixup training to ALIM-Scale (ablation)ALIM-Scale on PiCO · CIFAR-100 (q=0.03, η=0.3)accuracy: 71.69±0.39 -> 75.92±0.14Improved
Use Onehot instead of Scale normalization (ablation)ALIM on PiCO · CIFAR-10 (q=0.3, η=0.3)accuracy: 92.01±0.19 -> 92.42±0.17Improved
Use Onehot instead of Scale normalization (ablation)ALIM on PiCO · CIFAR-100 (q=0.03, η=0.3)accuracy: 71.69±0.39 -> 70.14±0.25Hurt
Use the adaptively adjusted λ instead of manually selected λ (ablation)ALIM · CIFAR-10 and CIFAR-100accuracyNo clear effect
Add ALIM to RCRC · CIFAR-10 and CIFAR-100 under noisy conditionsaccuracy: 80.87±0.30 -> 88.81±0.17Improved
Add ALIM to CRDPLLCRDPLL · CIFAR-10 under noisy conditionsaccuracy: 93.48±0.17 -> 96.03±0.23Improved
Increase warm-up duration (ablation)ALIM · CIFAR-10 and CIFAR-100accuracyMixed

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