AI Paper Atlas

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

Learning From Noisy Labels With Distillation

Yuncheng Li, Jianchao Yang, Yale Song, Liangliang Cao, Jiebo Luo, Li-Jia Li

ICCV 2017 · Label-efficient and robust learning › Noisy labels · approach: Noise-aware loss and pairwise methods · 571 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 distillation from clean-data teacherBaseline-Noisy · SportsmAP: 50.7 -> 53.5Improved
add distillation from clean-data teacherBaseline-Noisy · Species-ImAP: 38.5 -> 41.6Improved
add distillation from clean-data teacherBaseline-Noisy · ArtifactsmAP: 22.0 -> 26.0Improved
add knowledge graph guidance to distillationDistillation · SportsmAP: 53.5 -> 53.7Improved
add knowledge graph guidance to distillationDistillation · Species-ImAP: 41.6 -> 42.3Improved
add knowledge graph guidance to distillationDistillation · Species-YmAP: 26.1 -> 25.2Hurt
add knowledge graph guidance to distillationDistillation · ArtifactsmAP: 26.0 -> 26.0No clear effect
use distillation instead of label smoothingnoisy label learning · ArtifactsmAP: 22.9 -> 26.0Improved
use distillation instead of bootstrapnoisy label learning · Species-YmAP: 23.6 -> 26.1Improved
use distillation instead of importance re-weightingnoisy label learning · ArtifactsmAP: 24.8 -> 26.0Improved
use finetuning from clean model instead of training on noisy data onlyBaseline-Noisy · Species-YmAP: 23.7 -> 22.2Hurt
use distillation instead of ensembling clean and noisy modelsBaseline-Ensemble · ArtifactsmAP: 26.9 -> 26.0Hurt

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