A.A. Ishmukhamedov1, I.A. Tarkhanov2

1 National University of Science and Technology «MISiS» (Moscow, Russia)
2 State Academic University for Humanities (Moscow, Russia)
2 Federal Research Center «Computer Science and Control» of the Russian Academy of Sciences (Moscow, Russia)
1 fesevu@gmail.com, 2 itarhanov@frccsc.ru

The aim of this research is identifying detecting anomalies in smart contracts of Ethereum network based on an analysis of their transaction history. The schemes used by attackers exploit not only known vulnerabilities that are determined by static methods, but also more complex distributed patterns of interaction that need to be able to detect in an automated way. The result of this work is a new multimodal machine learning-based model for determining the abnormal behavior of smart contracts, using not only well-studied metrics of static code analysis, but also metrics based on the transaction history of smart contract execution. An important part of the research is the preparation of an open dataset to verify the predictability of the results obtained and the application of other models.

Ishmukhamedov A.A., Tarkhanov I.A. A machine learning model for detecting anomalies in smart contracts based on graph neural networks and LSTM // Highly Available Systems. 2026. V. 22. № 2. P. 71−82. DOI: https://doi.org/10.18127/j20729472-202602-06

1. Glazkov A.A. Perspektivy ispol'zovaniya smart kontraktov v razvitii biznes-ekosistem. 2022. URL: https://cyberleninka.ru/article/n/per­spektivy-ispolzovaniya-smart-kontraktov-v-razvitii-biznes-ekosistem
2. Fedotova V.V., Emel'yanov B.G., Tipner L.M. Ponyatie blokchejn i vozmozhnosti ego ispol'zovaniya. European science. 2018. № 1 (33). S. 40–48. [Elektronnyj resurs]. URL: https://cyberleninka.ru/article/n/ponyatie-blokcheyn-i-vozmozhnosti-ego-ispolzovaniya (data obra­shcheniya: 18.02.2026).
3. Osmolovskaya A.S. Smart-kontrakty: funkcii i primenenie [Elektronnyj resurs]. URL: https://cyberleninka.ru/article/n/smart-kontrakty-funktsii-i-primenenie/viewer (data obrashcheniya: 08.02.2026).
4. Crypto Hacks 2025: Full List of Scams, Exchange Exploits & DeFi Vulnerabilities [Updated October] URL: https://www.ccn.com/educa­tion/crypto/crypto-hacks-exploits-full-list-scams-vulnerabilities/
5. Belous V.S., Tarhanov I.A. Poisk uyazvimostej v smart-kontraktah na osnove mashinnogo obucheniya. Trudy Instituta Sistemnogo Analiza Rossijskoj Akademii Nauk. 2024. T. 74. № 3. DOI 10.14357/20790279240310
6. Kamps J., Kleinberg B. To the moon: defining and detecting cryptocurrency pump-and-dumps. Crime Science. 2018 (data obrashche­niya: 14.06.2024).
7. Chen W., Zheng Z., Cui J., Ngai E., Zheng P., Zhou Y. Detecting ponzi schemes on ethereum: Towards healthier blockchain technology. Proceedings of the 2018 World Wide Web Conference. 2018. S. 1409–1418.
8. Bartoletti M., Pes B., Serusi S. Data mining for detecting bitcoin ponzi schemes. CVBC. 2018. S. 75–84.
9. Monamo P., Marivate V., Twala B Unsupervised learning for detection of abnormal credit card spending behavior. IEEE Access. 2016. № 4. S. 7106–7114.
10. Kalra S., Goel S., Dhawan M., & Sharma S. Analyzing safety of smart contracts. NDSS. 2018.
11. Security analysis tool for Ethereum smart contracts [Elektronnyj resurs]. URL: https://github.com/ConsenSys/mythril (data obra­shcheniya: 14.06.2024).
12. Wang L., Cheng H., Zheng Z. Temporal Transaction Information-aware Ponzi Scheme Detection for Ethereum Smart Contracts. Engineering Applications of Artificial Intelligence. 2023. S. 107022. DOI: 10.1016/j.engappai.2023.107022.
13. Cruzpero Reniel & Dajang El. LSTM-Based Approach: Enhancing Fraud Detection in Ethereum Transactions. 2024. 10.13140/RG.2.2.33297.20325
14. Tan Runnan & Tan Qingfeng & Zhang Peng & Li Zhao. Graph Neural Network for Ethereum Fraud Detection. 2021. 78–85. 10.1109/ICKG52313.2021.00020
15. Wang L. et al. PSPL: A Ponzi scheme smart contracts detection approach via compressed sensing oversampling-based peephole LSTM. Future Generation Computer Systems. 2025. T. 166. S. 107655.
16. Liang R., Chen J., Wu C. i dr. Towards Effective Detection of Ponzi schemes on Ethereum with Contract Runtime Behavior Graph [Elektronnyj resurs] // ACM. 2024. Rezhim dostupa: https://dl.acm.org/doi/10.1145, svobodnyj (data obrashcheniya: 21.12.2024).
17. Vagifa, Ethereum Fraud Detection Dataset [Elektronnyj resurs]. URL: https://www.kaggle.com/datasets/vagifa/ethereum-frauddetection-dataset (data obrashcheniya: 27.04.2025).
18. Suraj Jain, Ethereum Fraud Datasets [Elektronnyj resurs]. URL: https://github.com/surajsjain/ethereum-fraud-datasets/tree/main (data obrashcheniya: 27.04.2025).
19. Detecting Fraudulent Accounts on Blockchain: A Supervised Approach [Elektronnyj resurs]. arXiv. 2019. № 1908.07886. URL: https://arxiv.org/abs/1908.07886 (data obrashcheniya: 27.04.2025).
20. XBlock. Fishingovye adresa Ethereum (dataset №13) [Elektronnyj resurs]. URL: https://xblock.pro/#/dataset/13 (data obrashcheniya: 27.04.2025).
21. XBlock. Smart-kontrakty, uchastvuyushchie v Ponzi-skhemah (dataset №25) [Elektronnyj resurs]. URL: https://xblock.pro/#/da­taset/25 (data obrashcheniya: 27.04.2025).
22. XBlock. Smart-kontrakty, zamechennye v rugpull-moshennichestve (dataset №50) [Elektronnyj resurs]. URL: https://xblock.pro/#/da­taset/50 (data obrashcheniya: 27.04.2025).
23. Ishmukhamedov A.A. Ethereum Fraud Dataset by Activity. 10.34740/KAGGLE/DSV/12889568. URL: https://www.kaggle.com/data­sets/fesevu/ethereum-fraud-dataset-by-activity (data obrashcheniya: 28.02.2026).
24. Liang R. et al. Ponziguard: Detecting ponzi schemes on ethereum with contract runtime behavior graph (crbg). Proceedings of the 46th IEEE/ACM International Conference on Software Engineering. 2024. S. 1–12.