A Survey of Deep Learning-Based Methods for Detecting Anomalous Transactions in Blockchain
DOI:
https://doi.org/10.61173/ne7c8839Keywords:
Blockchain, Anomalous transaction detec-tion, Deep learning, Graph neural networkAbstract
The decentralized and anonymous nature of blockchain technology has driven the growth of the digital economy while introducing new challenges in detecting anomalous transactions. This paper systematically reviews recent advances in deep learning-based methods for identifying anomalous transactions in blockchain. It analyzes the main types of blockchain-based anomalous transactions and their risks, highlighting the limitations of traditional detection methods in handling high-dimensional, nonlinear transaction data. Furthermore, this survey provides a detailed introduction to six major categories of deep learning approaches: graph neural network-based methods, which effectively model transaction network topologies; autoencoder-based methods that identify anomalies through reconstruction errors; generative adversarial networks that mitigate data imbalance issues; attention mechanisms capable of capturing critical transaction features; temporal models suited for analyzing transaction timing patterns; and multi-feature fusion methods that enhance detection comprehensiveness. For each type of method, this paper summarizes the representative research achievements and their performance，and discusses the key challenges faced by current research and proposes future research directions.
References
rections to address these challenges: [2] Patel, V., Pan, L., Rajasegarar, S. (2020). Graph Deep · Developing novel deep learning architectures, such Learning Based Anomaly Detection in Ethereum Blockchain as dynamic graph neural networks and spatio-temporal Network. In: Kutyłowski, M., Zhang, J., Chen, C. (Eds.),
Transformers, to better adapt to blockchain data character- Network and System Security. NSS 2020. Lecture Notes in istics; Computer Science(), vol 12570. Springer, Cham. pp. 132–148. · Exploring self-supervised and incremental learning tech- [3] Chang, Z., Cai, Y., Liu, X. F., et al. (2025). Anomalous Node niques to reduce dependency on labeled data and enable Detection in Blockchain Networks Based on Graph Neural continuous learning capabilities; Networks. Sensors, 25(1), 1. · Investigating multimodal data fusion methods that inte- [4] Lin, Y., Jiang, P., & Zhu, L. (2025). Cross-chain Abnormal grate on-chain and off-chain data for more comprehensive Transaction Detection via Graph-based Multi-model Fusion. detection systems; In: Proceedings of the 6th ACM International Symposium on · Enhancing model interpretability through attention visu- Blockchain and Secure Critical Infrastructure, BSCI 2024. New alization and other techniques to improve detection credi- York. pp. 1-9.
bility; [5] Wu, Z., Liu, J., Wu, J., et al. (2023) TRacer: Scalable Graph- · Building cross-chain collaborative detection frameworks Based Transaction Tracing for Account-Based Blockchain to counter increasingly sophisticated cross-platform at- Trading Systems. IEEE Transactions on Information Forensics tacks. and Security, 18: 2609-2621.
With the rapid advancement of blockchain technology, [6] Scicchitano, F., Liguori, A., Guarascio, M., et al. (2020) Deep anomaly detection methods must continuously innovate. autoencoder ensembles for anomaly detection on blockchain. In the future, research should balance detection accuracy In: International Symposium on Methodologies for Intelligent and computational efficiency, take into account privacy Systems. Springer, Berlin. pp. 448-456.
protection and compliance requirements, and strengthen [7] Huang, G., Li, Y., Pleiss, G., Liu, Z., et al. (2017). Snapshot the cooperation between academia and industry to jointly ensembles: train 1, get m for free. arXiv e-prints.
promote the development of this field in a more intelligent [8] Xiong, A., Qiao, C., Li, W., et al. (2025). Blockchain and efficient direction. abnormal transaction detection method based on generative adversarial network and autoencoder. High-Confidence Computing, in press. 5. Conclusion [9] Li, D., Chen, D., Goh, J., et al. (2018) Anomaly detection Deep anomaly detection on blockchain is becoming a with generative adversarial networks for multivariate time key field, which is very important for protecting and series. In: the 7th International Workshop on Big Data, Streams Dean&Francis Qiujing Fan
and Heterogeneous Source Mining: Algorithms, Systems, [13] Zhu, H.J., Chen, J.F., Li, Z.Y., et al. (2021) Blockchain Programming Models and Applications on the ACM Knowledge abnormal transaction detection method based on multi-feature Discovery and Data Mining conference. London. adaptive fusion. Journal of Communications, 42: 41-50. [10] Liang, F., Wang, Y.J., Wang, Q., et al. (2025) Malicious [14] Lin, W. (2022) Detection of Abnormal Transactions in address detection technology in Ethereum based on Transformer Blockchain Based on Multi Feature Fusion. Netinfo Security, 22: and transaction information fusion. Network Security 24-30. Technology and Applications, 3: 55-59. [15] Perales Gómez, Á.L., Fernández Maimó, L., Huertas [11] Sun, X., Yang, T. & Hu, B. (2022) LSTM-TC: Bitcoin coin Celdrán, A., et al. (2024) A Review of SUSAN: A Deep Learning mixing detection method with a high recall. Applied Intelligence, based anomaly detection framework for sustainable industry. 52: 780–793. In: 9th National Conference on Cybersecurity Research (JNIC). [12] Tang, H., Jiao, Y., Huang, B., et al. (2018) Learning Sevilla. pp. 454-455. to Classify Blockchain Peers According to Their Behavior Sequences. IEEE Access, 6: 71208-71215.
Downloads
Published
Issue
Section
License
Copyright (c) 2025 by the authors.
This work is licensed under a Creative Commons Attribution 4.0 International License.
