AI Paper Atlas

Research map / Label-efficient and robust learning / Knowledge distillation

Generalist Graph Anomaly Detection via Prototype-Based Distillation

Yiming Xu, Zihan Chen, Zhen Peng, Song Wang, Bin Shi, Bo Dong, Chao Shen

ICML 2026 · Label-efficient and robust learning › Knowledge distillation · approach: Patient consistent distillation · 0 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 prototype-driven soft-label distillation (ablation)ProMoS · 6 graphs avgAUROC: 56.32 -> 75.46Improved
add discrepancy-aware commitment and refinement loss (ablation)ProMoS · 6 graphs avgAUROC: 70.20 -> 75.46Improved
use a pre-trained SSL teacher instead of random GCN (ablation)ProMoS · 6 graphs avgAUROC: 69.93 -> 75.46Improved
alternate prototype and teacher updates instead of joint optimization (ablation)ProMoS · 6 graphs avgAUROC: 75.46 vs 75.32 / 75.03No clear effect
swap GCA teacher for GraphCL (ablation)ProMoS · 6 graphs avgAUROC: 80.92 -> 79.45Mixed
swap GCA teacher for GraphMAE (ablation)ProMoS · 6 graphs avgAUROC: 80.92 -> 77.81Hurt
swap GCA teacher for DGI (ablation)ProMoS · 6 graphs avgAUROC: 80.92 -> 80.01Hurt
use unsupervised distillation from SSL teacher vs supervised generalist methodsProMoS · CoraAUROC: 66.53 -> 84.56Improved
ProMoS vs DOMINANT on WeiboProMoS · WeiboAUROC: 92.88 -> 91.74Hurt
ProMoS efficiency vs GCNProMoS · training/inference timespeedup: 4.8x faster training, 1.4x faster inferenceImproved

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