AI Paper Atlas

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

RankMatch: Fostering Confidence and Consistency in Learning with Noisy Labels

Ziyi Zhang, Weikai Chen, Chaowei Fang, Zhen Li, Lechao Chen, Liang Lin, Guanbin Li

ICCV 2023 · Label-efficient and robust learning › Noisy labels · approach: Noise-aware loss and pairwise methods · 25 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 sample selection via confidence voting (SCV) instead of GMM selection (ablation)RankMatch · CIFAR-100 (sym. 90% noise)test accuracy (best): 45.1 -> 50.6Improved
add SCV instead of GMM selection (ablation)RankMatch · CIFAR-10 (sym. 90% noise)test accuracy (best): 89.6 -> 92.6Improved
add K-means prototypes in SCV (ablation)RankMatch · CIFAR-100 (sym. 90% noise)test accuracy (best): 47.9 -> 50.6Improved
add rank contrastive loss (RCL) (ablation)RankMatch · CIFAR-100 (sym. 90% noise)test accuracy (best): 45.6 -> 50.6Improved
add rank contrastive loss (RCL) (ablation)RankMatch · CIFAR-10 (sym. 90% noise)test accuracy (best): 90.8 -> 92.6Improved
add SCV and rank contrastive loss in low-noise setting (ablation)RankMatch · CIFAR-10 (sym. 50% noise)test accuracy (best): 95.1 -> 95.6No clear effect
RankMatch vs RRL at low symmetric noiseRankMatch · CIFAR-10 (sym. 20% noise)test accuracy (last 10 epochs): 96.4 vs 96.4 (RRL); CIFAR-100 20% 79.3 vs 80.3 (RRL)No clear effect
use RankMatch instead of DivideMix under extreme noiseLNL sample selection · CIFAR-100 (sym. 90% noise)test accuracy (best): 31.5 -> 50.6Improved
vary confidence threshold tau 0.90-0.99 and prototype count K 5-30 (ablation)SCV · CIFAR-10 (sym. 50% and 90% noise)test accuracy (best)No clear effect

Later papers that compared with this method

4 later accepted papers report RankMatch as a baseline in their results tables; 3 of them report a better number on most of the results they share with it.

Earlier methods this paper compares with

Named baselines from earlier accepted papers in this paper's results tables: better than 2 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