Research map / Label-efficient and robust learning / Noisy labels
Learning From Noisy Data With Robust Representation Learning
Junnan Li, Caiming Xiong, Steven C.H. Hoi
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 changed | Where | Measured | Effect |
|---|---|---|---|
| add the mixup prototypical contrastive loss (ablation) | RRL · CIFAR-10 Sym 50% | accuracy (classifier): 85.9 -> 94.3 | Improved |
| add the mixup prototypical contrastive loss (ablation) | RRL · CIFAR-100 Sym 50% | accuracy (classifier): 65.6 -> 74.8 | Improved |
| add the consistency contrastive loss (ablation) | RRL · CIFAR-10 Sym 50% | accuracy (classifier): 93.7 -> 94.3 | Improved |
| add the consistency contrastive loss (ablation) | RRL · CIFAR-100 Sym 50% | accuracy (classifier): 71.9 -> 74.8 | Improved |
| add the consistency contrastive loss (ablation) | RRL · CIFAR-10 Sym 50% + image corruption | accuracy (classifier): 89.4 -> 91.4 | Improved |
| add the reconstruction loss (ablation) | RRL · CIFAR-10 Sym 50% | accuracy (classifier): 93.3 -> 94.3 | Improved |
| add the reconstruction loss (ablation) | RRL · CIFAR-100 Sym 50% | accuracy (classifier): 73.2 -> 74.8 | Improved |
| add mixup augmentation (ablation) | RRL · CIFAR-10 Sym 50% | accuracy (classifier): 89.5 -> 94.3 | Improved |
| add mixup augmentation (ablation) | RRL · CIFAR-100 Sym 50% | accuracy (classifier): 69.3 -> 74.8 | Improved |
| use AugMix strong augmentation instead of standard augmentation (ablation) | RRL · CIFAR-100 Sym 50% | accuracy (classifier): 74.5 -> 74.8 | Improved |
| classify with knn on embeddings instead of the softmax classifier | RRL · CIFAR-100 Sym 20% | accuracy: 79.1 -> 79.4 | Improved |
| add noise cleaning with smooth neighbors (ablation) | RRL · WebVision (mini) val | top-1 accuracy: 75.5 -> 76.3 | Improved |
Earlier methods this paper compares with
Named baselines from earlier accepted papers in this paper's results tables: better than 1 of 1 on most shared results.
- DivideMix — DivideMix: Learning with Noisy Labels as Semi-supervised Learning (ICLR 2020) · better on 2 of 2 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
- Ditch the Denoiser: Emergence of Noise Robustness in Self-Supervised Learning from Data Curriculum (NeurIPS 2025 · similarity 0.88)
- Partially View-Aligned Representation Learning With Noise-Robust Contrastive Loss (CVPR 2021 · similarity 0.88)
- FINE Samples for Learning with Noisy Labels (NeurIPS 2021 · similarity 0.88)
- Embedding Contrastive Unsupervised Features to Cluster in- and Out-of-Distribution Noise in Corrupted Image Datasets (ECCV 2022 · similarity 0.88)
- Robust Contrastive Learning Against Noisy Views (CVPR 2022 · similarity 0.88)
- Learning with Open-world Noisy Data via Class-independent Margin in Dual Representation Space (AAAI 2025 · similarity 0.86)
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.
