Research map / Label-efficient and robust learning / Noisy labels
Class-Dependent Label-Noise Learning with Cycle-Consistency Regularization
De Cheng, Yixiong Ning, Nannan Wang, Xinbo Gao, Heng Yang, Yuxuan Du, Bo Han, Tongliang Liu
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 backward transition matrix T' with L2 loss to forward loss L1 (ablation) | Ours · CIFAR-100 Sym-60% | test accuracy (%): 40.89 -> 41.94 | Improved |
| add cycle-consistency regularization on top of T + T' (ablation) | Ours · CIFAR-100 Sym-60% | test accuracy (%): 41.94 -> 49.30 | Improved |
| add cycle-consistency regularization on top of T + T' (ablation) | Ours · CIFAR-10 Pair-45% | test accuracy (%): 78.62 -> 91.08 | Improved |
| add cycle-consistency regularization on top of T + T' (ablation) | Ours · CIFAR-10 Sym-60% | test accuracy (%): 73.39 -> 81.01 | Improved |
| add cycle-consistency regularization on top of T + T' (ablation) | Ours · CIFAR-100 Asym-60% | test accuracy (%): 42.72 -> 50.29 | Improved |
| use only the backward loss L2(w,T') instead of only forward loss (ablation) | Ours · CIFAR-10 Asym-20% | test accuracy (%): 89.46 -> 89.97 | No clear effect |
| plug the proposed module into DivideMix | DivideMix · Clothing1M | accuracy (%): 74.67 -> 75.12 | Improved |
| plug the proposed module into DivideMix | DivideMix · Food-101N | accuracy (%): 84.39 -> 86.11 | Improved |
| replace VolMinNet by Ours when combined with DivideMix | DivideMix · Clothing1M | accuracy (%): 74.83 -> 75.12 | Improved |
| replace CE training by Ours | ResNet-50 · Clothing1M | accuracy (%): 68.94 -> 70.73 | Improved |
| replace CE training by Ours | ResNet-50 · Food-101N | accuracy (%): 81.44 -> 83.71 | Improved |
| set lambda to 0.3 (ablation) | Ours · CIFAR-10/100 hyper-parameter study | test accuracy (%) | 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
- A Parametrical Model for Instance-Dependent Label Noise (TPAMI 2023 · similarity 0.92)
- Learning Noise Transition Matrix from Only Noisy Labels via Total Variation Regularization (ICML 2021 · similarity 0.92)
- Dual T: Reducing Estimation Error for Transition Matrix in Label-noise Learning (NeurIPS 2020 · similarity 0.91)
- Extended : Learning with Mixed Closed-set and Open-set Noisy Labels (TPAMI 2022 · similarity 0.91)
- Estimation of the Label-Noise Transition Matrix with Performance Guarantees via Selective Classification (NeurIPS 2026 · similarity 0.90)
- Estimating Noisy Class Posterior with Part-level Labels for Noisy Label Learning (CVPR 2024 · similarity 0.89)
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.
