AI Paper Atlas

Research map / Graph learning / Graph self-supervised learning

BugSweeper: Function-Level Detection of Smart Contract Vulnerabilities Using Graph Neural Networks

Uisang Lee, Changhoon Chung, Junmo Lee, Soo-Mook Moon

AAAI 2026 · Graph learning › Graph self-supervised learning · approach: Graph anomaly detection · 4 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 second-stage GNN after CGNN (SAGE + GAT vs single-stage GAT) (ablation)BugSweeper · SmartBugs Wild reentrancyF1 (%): 84.77 -> 91.61Improved
use GraphSAGE first and GAT second instead of GAT then SAGE (ablation)BugSweeper · SmartBugs Wild reentrancyF1 (%): 82.46 -> 91.61Improved
use GAT instead of GraphSAGE in second stage (SAGE+SAGE to SAGE+GAT) (ablation)BugSweeper · SmartBugs Wild reentrancyF1 (%): 88.89 -> 91.61Improved
use GAT in first stage instead of GraphSAGE (ablation)BugSweeper two-stage · SmartBugs Wild reentrancyF1 (%): 88.89 -> 85.66Hurt
two-stage GAT + SAGE vs single-stage SAGE on time manipulation (ablation)BugSweeper · SmartBugs Wild time manipulationF1 (%): 74.77 -> 69.33Hurt
use CGPool instead of ASAPool (ablation)BugSweeper (SAGE + GAT) · SmartBugs Wild multi-classF1 (%): 82.41 -> 87.32Improved
use CGPool instead of TopKPool (ablation)BugSweeper (SAGE + GAT) · SmartBugs Wild multi-classF1 (%): 75.29 -> 87.32Improved
add second-stage GNN with TopKPool (ablation)BugSweeper with TopKPool · SmartBugs Wild multi-classF1 (%): 78.71 -> 75.29Hurt
add second-stage GNN with ASAPool (ablation)BugSweeper with ASAPool · SmartBugs Wild multi-classF1 (%): 80.32 -> 82.41Improved
function-level FLAG with learned GNN instead of rule-based/other deep modelsBugSweeper · AME dataset (reentrancy)F1 (%): 95.42 -> 98.57Improved

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