AI Paper Atlas

Research map / Graph learning / Graph self-supervised learning

Graph is a Substrate Across Data Modalities

Ziming Li, Xiao-Ming Wu, Zehong Wang, Jiazheng Li, Yijun Tian, Jinhe Bi, Yunpu Ma, Yanfang Ye, Chuxu Zhang

ICML 2026 · Graph learning › Graph self-supervised learning · approach: Heterophily and imbalance on graphs · 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 interleaved role-based training on top of unified multi-task schema (ablation)Unified multi-task (Qwen3-VL-2B) · Shortest pathaccuracy: 45.02 -> 48.59Improved
add interleaved role-based training on top of unified multi-task schema (ablation)Unified multi-task · HiEventF1: 19.37 -> 25.15Improved
use unified schema instead of native formats in task-isolated training (ablation)single-task training · Shortest path / bipartite matchingaccuracy: SP 38.27 -> 37.14; BM 92.05 -> 85.98Hurt
use unified schema instead of native formats in multi-task training (ablation)multi-task training · Shortest pathaccuracy: 41.27 -> 45.02Improved
add interleaving without unified schema (ablation)Naive multi-task · MAVEN-Causal / Cycle detectionF1 / accuracy: SP 41.27 -> 43.83; CD 94.72 -> 93.86Mixed
use XML-style serialization instead of the proposed schema realization (ablation)G-Substrate · MAVEN-Temporal etc.F1: 42.68 -> 36.40Hurt
use natural-language serialization instead of proposed schema (ablation)G-Substrate · Shortest pathaccuracy: 48.59 -> 44.80Hurt
pretrain on event graphs only (unified schema) before scene graph generation (ablation)Qwen3-VL-2B · Scene graph generation PCIsR@50: 19.10 -> 21.47Improved
replace reused graphs with structurally incorrect variants (ablation)interleaved training · GAR/MGD/SGG/EREdelta vs unified multi-task: gains of +0.62 to +2.23 become -1.20 to -2.03 in Figure 5Hurt
raise interleaved data ratio from 50% to 100% (ablation)G-Substrate · GARaccuracy: 94.5 -> 93.4Hurt
add 50% interleaved data instead of none (ablation)G-Substrate · EREaveraged F1: 38.0 -> 40.2Improved
corrupt 20% of reused graph edges (ablation)G-Substrate · MGDBLEU-4: 50.74 vs clean NMT 48.11Improved

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