AI Paper Atlas

Research map / Graph learning / GNN architectures

Learn to Propagate Reliably on Noisy Affinity Graphs

Lei Yang, Qingqiu Huang, Huaiyi Huang, Linning Xu, Dahua Lin

ECCV 2020 · Graph learning › GNN architectures · approach: GNN generalization theory · 13 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
increase maximum graph patch size from 100 to 3000 (ablation)reliable propagation · ImageNet 10K vertices, 50% noiseaccuracy: 64.63 -> 65.83Improved
increase maximum patch size from 3000 to 10000 (ablation)reliable propagation · ImageNet 10K vertices, 50% noiseaccuracy: 65.83 -> 65.92No clear effect
increase maximum patch size from 3000 to 10000 (cost) (ablation)reliable propagation · ImageNet 10K vertices, 50% noiseruntime: 316s -> 1077sHurt
increase maximum patch size from 100 to 3000 (cost) (ablation)reliable propagation · ImageNet 10K vertices, 50% noiseruntime: 34s -> 316sHurt
use the proposed confidence inclusion instead of design (1) (ablation)reliable propagation · ImageNet noisy graphaccuracy: 61.5 -> 65.8Improved
concatenate confidence to vertex features (design 2) instead of the proposed design (ablation)reliable propagation · ImageNet noisy graphaccuracy: 65.8 -> 65.9No clear effect
concatenate confidence to vertex features (design 2) instead of the proposed design (cost) (ablation)reliable propagation · ImageNet noisy graphmemory: 168M -> 205MHurt
use estimated confidence to choose samples for annotation (active learning) versus random or GCN-based selection (ablation)reliable propagation · ImageNet, 1% labels, 50% noiseaccuracyImproved
train face model on pseudo labels from the proposed method versus GCN or CDP pseudo labels (ablation)reliable propagation · MegaFace, Ms-Celeb-1M 1K identitiesaccuracyImproved
compare proposed method against SGC under 50% noisereliable propagation · ImageNet noisy graph, noise 50%accuracy: 65.63 -> 68.25Improved
compare proposed method against SGC under 0% noisereliable propagation · ImageNet noisy graph, noise 0%accuracy: 84.78 -> 85.16No clear effect
compare proposed method against GCN on Ms-Celeb-1M at 10% noisereliable propagation · Ms-Celeb-1M noisy graph, noise 10%accuracy: 99.6 -> 97.59Mixed

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