AI Paper Atlas

Research map / Graph learning / Graph self-supervised learning

Preference-Modulated Structural Attention for Multi-Objective Combinatorial Optimization

Rongsheng Jia, Jun Zhang, Yifan Zhang, Jian Cheng

ICML 2026 · Graph learning › Graph self-supervised learning · approach: Multi-view graph clustering

Read the paper

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 preference modulation to edge structural bias (ablation)PMSA · Bi-TSP100hypervolume: 0.7057 -> 0.7070Improved
add dynamic edge aggregation (ablation)PMSA · Bi-TSP100hypervolume: 0.7053 -> 0.7070Improved
add preference modulation (ablation)PMSA · Bi-TSP200hypervolume: 0.7384 -> 0.7408Improved
add dynamic edge aggregation (ablation)PMSA · Bi-TSP200hypervolume: 0.7371 -> 0.7408Improved
add edge-feature structural bias vs WE-CA node-only modelPMSA · Bi-TSP100hypervolume: 0.7034 -> 0.7070Improved
use PMSA at small scalePMSA · Bi-TSP20hypervolume: 0.6270 -> 0.6272No clear effect
use PMSA on larger unseen sizesPMSA · Bi-TSP200hypervolume: 0.7346 -> 0.7408Improved
add instance augmentationPMSA · Bi-TSP100hypervolume: 0.7070 -> 0.7083Improved
PMSA vs WE-CA on Bi-KPPMSA · Bi-KP200hypervolume: 0.3602 vs 0.3602No clear effect
use PMSA on TSPLIB OOD instancesPMSA · KroAB100hypervolume: 0.6948 -> 0.6996Improved
PMSA-Aug vs POMO-Aug at balanced preferencePMSA-Aug · Bi-TSP100 lambda=(0.5,0.5)gap: 0.77% -> 0.24%Improved

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