DG-VDT: Dynamic Graph-Guided Reinforcement Learning for Low-Latency Vulnerability Detection and Attack Traceability in Ethereum Smart Contracts
Open Access
- 30 September 2026
- journal article
- research article
- Published by MDPI AG in Applied Sciences
- Vol. 16 (19) , 9698
Abstract
Ethereum hosts more than 50 million deployed contracts, and cumulative losses from contract exploits exceed three billion US dollars, making automated detection and attacker attribution a practical priority. Existing tool-augmented language models and reward-driven reinforcement learning reason over transaction text or a single graph view; they infer maliciousness from superficial topology and degrade when applied to attacks whose traces lack it. This article presents DG-VDT, a reinforcement-learning framework in which tool-centric rewards are replaced by graph-structured feedback derived directly from Ethereum Virtual Machine execution traces. Each trace is decomposed into three complementary graphs (fund flow, contract creation, and contract call), and a dual-stage reward first credits embedding similarity to a canonical attack signature, then enforces exact subgraph-isomorphism matching, with a smooth curriculum governing the transition. The framework is scoped to three structurally distinct vulnerability classes: reentrancy, short-address attack, and timestamp dependence. Zero-shot evaluation on two author-independent public corpora yields 87.7% macro- on SolidiFI-Bench and 84.0% (95% confidence interval 83.4–84.7) on SmartBugs Wild, exceeding fine-tuned GPT-4o by 6.8 and 6.3 points at 180 ms per trace on a single consumer graphics card. Ablation attributes 13.9 macro- points to the multi-graph representation. Extension beyond the three evaluated classes remains future work.
Keywords
This publication has 28 references indexed in Scilit:
- GPTScan: Detecting Logic Vulnerabilities in Smart Contracts by Combining GPT with Program Analysis2024
- DeFiRanger: Detecting DeFi Price Manipulation AttacksIEEE Transactions on Dependable and Secure Computing, 2023
- Combining Graph Neural Networks with Expert Knowledge for Smart Contract Vulnerability DetectionIEEE Transactions on Knowledge and Data Engineering, 2021
- Understanding Ethereum via Graph AnalysisACM Transactions on Internet Technology, 2020
- Towards Automated Reentrancy Detection for Smart Contracts Based on Sequential ModelsIEEE Access, 2020
- Graph neural networks: A review of methods and applicationsAI Open, 2020
- A (sub)graph isomorphism algorithm for matching large graphsIEEE Transactions on Pattern Analysis and Machine Intelligence, 2004
- Least Median of Squares RegressionJournal of the American Statistical Association, 1984
- A Coefficient of Agreement for Nominal ScalesEducational and Psychological Measurement, 1960
- Note on the Sampling Error of the Difference Between Correlated Proportions or PercentagesPsychometrika, 1947