AI Paper Atlas

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

Learning From Web Data With Self-Organizing Memory Module

Yi Tu, Li Niu, Junjie Chen, Dawei Cheng, Liqing Zhang

CVPR 2020 · Label-efficient and robust learning › Noisy labels · approach: Instance-dependent noise modeling · 20 citations (OpenAlex)

Check 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 the discriminative score to ROI weighting (ablation)SOMNet · Clothing1Maccuracy: 76.8 -> 82.1Improved
Add the representative score to ROI weighting (ablation)SOMNet · Clothing1Maccuracy: 73.5 -> 82.1Improved
Add the area score to ROI weighting (ablation)SOMNet · Clothing1Maccuracy: 74.8 -> 82.1Improved
Gradually increase p from 10% to 40% rather than starting at 40% (ablation)SOMNet · Clothing1Maccuracy: 81.3 -> 82.1Improved
Add region proposals to handle background noise (ablation)SOMNet · Clothing1Maccuracy: 74.1 -> 82.1Improved
Use the end-to-end self-organizing memory module instead of K-means (ablation)SOMNet · Clothing1Maccuracy: 79.5 -> 82.1Improved
Add the SOM neighborhood constraint (ablation)SOMNet · Clothing1Maccuracy: 78.2 -> 82.1Improved
Use two images per bag instead of one (ablation)SOMNet · Clothing1Maccuracy: 72.9 -> 82.1Improved
Increase bag size beyond two images (ablation)SOMNet · Clothing1Maccuracy: 82.1 -> 81.1; 82.1 -> 78.5; 82.1 -> 75.9Hurt
Use too few memory key slots (L=4^2) (ablation)SOMNet · Clothing1Maccuracy: 82.1 -> 78.7Hurt
Use L=8^2, L=16^2, or L=20^2 key slots instead of L=12^2 (ablation)SOMNet · Clothing1Maccuracy: 82.1 -> 81.7; 82.1 -> 81.9; 82.1 -> 81.6Mixed

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.

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