AI Paper Atlas

Research map / Graph learning / GNN architectures

Adaptive Recurrent Message Passing for Test Time Computing on Graphs

Junshu Sun, Wanxing Chang, Qingming Huang, Shuhui Wang

ICML 2026 · Graph learning › GNN architectures · approach: Expressive equivariant message passing · 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 step information to the recurrent update (ablation)AdaR · zero-shot transfer, averaged over 6 datasetsaverage accuracyImproved
encode step with fixed total time T=1 instead of fixed unit step dt=1 (ablation)AdaR · zero-shot transfer, unseen iteration budgetsaverage accuracyImproved
add representation-target relations R to the velocity function (ablation)AdaR · zero-shot transfer, averaged over 6 datasetsaverage accuracyImproved
add the full-trajectory gradient loss L_full (ablation)AdaR · zero-shot transfer, averaged over 6 datasetsaverage accuracyImproved
add the per-step gradient loss L_step (ablation)AdaR · zero-shot transfer, averaged over 6 datasetsaverage accuracyImproved
train with small budget (S=8) and infer with a much larger budget (up to 300 steps)AdaR · zero-shot transfer, averaged over 6 datasetsaverage accuracyImproved
use AdaR instead of N2 (adds step info and relations)N2 · zero-shot inductive, 6 datasetsaccuracy: average absolute improvement of 13.52%Improved
use AdaR instead of AdaR-style global-attention models (N2, graph transformers) in supervised settingglobal message passing models · supervised transductiveROC-AUC / accuracy: average absolute improvement of 1.33%Improved
choose number of inference steps by average node-target inner product peakAdaR · arxiv-year, CiteSeeraccuracyImproved
run very many recurrent steps without early exitAdaR · zero-shot transferaccuracyHurt

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