Ali M. Emran, Md Kamrul Islam -, Md Ashraful Islam Nayem -, Md Rubel · 5 authors
Abstract: Exploring GNNs as a cutting-edge approach to real-time detection of online money transfer fraud is the focus of this work. P2P payment systems, mobile money platforms, and decentralized financial infrastructures (DeFi) have all experienced explosive growth over the past decade due to their simplicity, speed, and affordability. Identity fraud, synthetic account misuse, coordinated fraud rings that exploit systemic vulnerabilities, and transaction laundering are some of the new types of fraud that can occur in these platforms, despite their desirability. In situations where fraud is predictable, isolated, and statistically distinct, logistic regression, rule-based algorithms, and standard ML models like Random Forests and SVMs have all proved effective in detecting it. Modern, hyper-connected, real-time financial ecosystems are seeing an uptick in non-linear, relational, and temporal fraud patterns, which these tactics struggle to combat. Because of their inherent bias, they fail to recognize the interconnected structural and relational processes that may point to coordinated fraud. The graph-like qualities of monetary exchanges, where elements (like IP addresses, users, and devices) are organically linked through edges that stand for transactions or relationships, are utilized by Graph Neural Networks to give a paradigm shift, on the other hand. Generalized neural networks (GNNs) are crucial for uncovering intricate fraud schemes because they represent these interactions as a graph structure that permits data to travel and accumulate across nodes. Because of this, the model may take global and regional effects into consideration. Relational learning excels when other methods fail, such as when trying to detect suspicious clusters of transactions, multi-hop collusions, or fraudulent subnetworks using separate features. In order to implement this method, we constructed an entirely new fraud detection system utilizing GNNs. Node feature engineering, graph generation, classification heads, message-passing layers, and a real-time processing optimized pipeline are all parts of it. We were able to empirically evaluate our technique using a real-world transactional dataset that was acquired from a leading financial services provider. As is typical in fraud detection tasks, the dataset had a highly skewed class distribution, which impacted both memory and accuracy. With an F1-score of 0.78, accuracy of 98.7 percent, precision of 0.81%, and recall of 0.76%, the model nevertheless performed admirably. The model's ability to detect fraudulent behaviors while maintaining dependable operations in the real world is demonstrated by these measures. Beyond its implications for technological performance, this study will help achieve broader aims in regulation, ethics, and national security. A number of federal agencies have issued advisories highlighting the need for strong, intelligent, and real-time fraud monitoring systems to safeguard national financial systems from fraudulent exploitation. These agencies include the DOJ, FinCEN, and DHS. Compliance with the USA PATRIOT Act and the Bank Secrecy Act (BSA) is of the utmost importance to financial institutions and fintech enterprises. As stated in the National Strategy to Combat Terrorist and Other Illicit Financing, they also want AI-driven surveillance systems to be resilient and explainable. This national goal is helped by our study, which provides a scalable, interpretable, and performance-driven GNN-based system. Along with helping with auditability, model explainability, and compliance reporting, all of which are crucial for regulated businesses, this strategy also helps with effective fraud detection. Integrating our suggested architecture for decentralized, privacy-preserving fraud detection into online learning extensions can further improve their functionality. Over time, these extensions can be integrated with federated learning systems and streaming data platforms. This work puts GNNs in a position to become a new weapon in the fight against digital payment fraud by combining cutting-edge graph representation learning with cybersecurity regulations and goals for financial integrity. Thanks to our research's careful analysis, innovative architecture, and adherence to statutory criteria, future financial systems will be reliable, safe, and robust. Additionally, it resolves a significant technical matter.
Moosa Uday Kumar, P. R. Pooja, Abhishek Dixit, M. A. Jabbar
The security of smart contracts is critical to the integrity of decentralized systems. Unlike traditional software, their immutability makes them particularly susceptible to zero-day vulnerabilities unseen flaws that can lead to catastrophic financial losses once exploited. Traditional detection methods, which rely on predefined attack patterns, are fundamentally incapable of addressing such unknown threats. This paper introduces a novel deep learning framework designed to proactively detect both known and previously unobserved zero-day vulnerabilities in Ethereum smart contracts. The approach employs a dual-path architecture that combines CodeBERT for deep semantic feature extraction with two parallel detection modules: a Graph Neural Network (GNN) for classifying known threats and a dedicated Autoencoder for unsupervised anomaly detection. This dual-path system leverages the strengths of both supervised and unsupervised learning. The GNN effectively classifies known attack vectors, while the Autoencoder identifies latent anomalies by flagging contracts with high reconstruction errors, a key indicator of unseen vulnerabilities. The framework was trained and validated on a balanced subset of the Malicious Smart Contract Detection dataset. The GNN demonstrated a high classification accuracy for known vulnerabilities, and the Autoencoder successfully identified anomalous contracts that deviated from learned patterns. This dual-pronged methodology represents a significant step forward in bolstering blockchain security by providing a robust, data-driven defense against the evolving landscape of smart contract vulnerabilities.
Humans have the ability to incrementally learn, accumulate, update, and apply knowledge from dynamic environments. This capability, known as continual learning or lifelong learning, is also a long-term goal in the development of artificial intelligence. However, neural network-based continual learning suffers from catastrophic forgetting: the acquisition of new knowledge typically disrupts previously learned knowledge, leading to partial forgetting and a decline in the model’s overall performance. Most current continual learning methods can only mitigate catastrophic forgetting and fail to incrementally improve the overall performance. In this work, we aim to incrementally improve performance within sample incremental context by utilizing inter-stage edges as a pathway for explicit knowledge transfer in continual graph learning. Building on this pathway, we propose a knowledge-augmented replay method by leveraging evolving subgraphs of important nodes. This method enhances the distinction between patterns associated with different node classes and consolidates previously learned knowledge. Experiments on phishing detection in Ethereum transaction networks validate the effectiveness of the proposed method, demonstrating effective knowledge retention and augmentation while overcoming catastrophic forgetting and incrementally improving performance. The results also reveal the relationship between average accuracy and average forgetting. Lastly, we identify the key factor to incremental performance improvement, which lays a foundation for convergence of continual graph learning.
V.Priyanga, A.Duraimurugan, P. Santhi, M.Sathya Sundaram · 6 authors
The explosive increase in decentralized data on blockchain systems has opened up new possibilities and problems of mining complex relational patterns in trustless, distributed systems. Conventional data mining models fall short of describing the complex relationships and heterogeneity of data stored in a blockchain, and more so when the data is distributed in graph form. This paper suggests GraphBlock-Mine, a new system that combines Graph Neural Networks (GNNs) with blockchain to support secure, scalable, and intelligent mining of patterns on decentralized graph-based data. The framework also can exploit the representational capacity of GNNs to capture dynamic relationship among nodes, temporal relationships, and structural anomalies with immutability and provenance via smart contracts and consensus mechanisms. GraphBlock-Mine uses an off-chain computation approach, which is reviewed using verifiable proofs on-chain to assure data integrity and privacy. We test the offered model on simulated blockchain networks and decentralized databases in real conditions, which proves the high level of pattern recognition accuracy, fault tolerance, and security with regard to traditional mining methods. The paper lays the foundations of the next-generation blockchain intelligence systems that can discover knowledge decentralized and in real-time across different fields, such as finance, supply chain, and social networks.
To our knowledge, this paper is the first to introduce an iterative semantic-clustering framework driven by Large Language Models (LLMs) for refined user profiling on Uniswap V4. Traditional numerical techniques such as K-Means falter in high-dimensional feature spaces and identify only three coarse user groups. In contrast, our LLM-based pipeline repeatedly refines clustering criteria through natural-language reasoning, producing 21 semantically interpretable sub-clusters. These fine-grained clusters can be hierarchically aggregated into three macro categories that align with the K-Means result, thereby preserving global consistency while revealing nuanced behavioral motivations and patterns. The proposed framework offers both a theoretical perspective and practical toolkit for decentralized-finance user analysis, opening new avenues for understanding on-chain behavior.
The rapid evolution of internet technologies has triggered exponential growth in multimodal data, intensifying security and efficiency challenges in digital copyright protection. This paper proposes a Multimodal Semantic Sharding Graph Convolutional Blockchain System (MSSGC) that innovatively integrates multimodal semantic fragmentation with graph convolutional neural networks (GCN). The system enables load-balanced blockchain dynamic sharding through GCN-based joint clustering of semantic features and transaction networks, effectively reducing cross-shard communication overhead. We develop a Trust-enhanced Proof of Stake (T-PoS) protocol to optimize account sharding via incentive mechanisms while maintaining decentralization and network equilibrium, complemented by queuing theory-based analysis of transactional performance ceilings. Experiments on blockchain simulators demonstrate that MSSGC significantly outperforms baseline systems across diverse sharding configurations: throughput improves by approximately$\mathbf{2 0 \%}$, latency is reduced by$\mathbf{3 0 \%}$, and load-balancing efficiency increases by roughly 2.3. Notably, the system maintains about 70% transaction integrity even under malicious attacks. This work bridges theoretical gaps in semantic-aware copyright protection while advancing practical blockchain implementations for digital rights management.
Weizhao Jin, Erik Kline, T. K. Satish Kumar, Lincoln Thurlow · 5 authors
Path validation in computer networks is used to enforce and verify data forwarding rules across network slices and administrative domains to satisfy specific service level requirements. Deviating from pre-established paths has the potential to downgrade network service quality, increase attack surface area, and disrupt network orchestration capabilities. Network operators regard the network infrastructure and topology as sensitive. This necessitates the need for privacy-preserving path validation techniques that leak minimal information about the overall network path to individual infrastructure owners. We present the design of a decentralized privacy-preserving path validation protocol using Non-Interactive Zero-Knowledge (NIZK) proofs to provide provable path privacy guarantees. The NIZK-based pairwise validation design identifies individual slice nodes that deviate from the prescribed path. Deploying this lightweight protocol periodically enables individual nodes to enforce and validate the network control path. We have implemented and evaluated our system on a testbed simulating a multi-authority network. Our results demonstrate the feasibility of preserving path privacy as well as the practicality of our proposed protocols for next-generation multi-authority sliced networks.
Bing Xue, Jun Zhang, Zhongwei An, Zhaoxiong Song · 6 authors
In recent years, graph neural networks have demonstrated strong capabilities in processing graph-structured data and have made significant progress in the field of smart contract vulnerability detection. This paper introduces HF-Sec, a novel framework for smart contract vulnerability detection. The framework first automatically generates heterogeneous contract graphs from the source code of Ethereum smart contracts to represent the control flow and function call relationships of the code. Then, by using a multi-source attention mechanism, the framework is able to synthesize features from different sources to capture key information from multiple perspectives. In addition, HF-Sec utilizes Fast Graph Transformer Networks and Kolmogorov-Arnold Networks to automatically extract mission-critical meta-paths and enhance the interpretability of the model. We performed experimental validation on a mixed dataset containing 423 contracts with vulnerabilities and 2742 contracts without vulnerabilities. The experimental results show that HF-Sec can significantly improve the accuracy of smart contract vulnerability detection, which is better than the methods based on machine learning or traditional analysis techniques. Through a series of ablation experiments, we further verified the importance of various key components in HF-Sec to improve the detection accuracy.
With the widespread application of blockchain technology, the security issues of smart contracts have become increasingly prominent. Traditional contract vulnerability detection methods mostly depend on source code, making it hard to adapt to the actual situation where only bytecode is available on-chain. This paper presents a novel bytecode-based vulnerability detection method for smart contracts (BC-GNN). By parsing the contract bytecode, it extracts opcode sequences, control flow graphs (CFGs), and data flow graphs (DFGs), integrates them into a unified heterogeneous graph structure, and uses graph neural networks (GNNs) to automatically detect multiple types of vulnerabilities. Experimental results show that BC-GNN achieves excellent performance on multiple actual and enhanced datasets. In particular, it demonstrates strong applicability and generalization ability in scenarios where source code is unavailable. This approach offers new ideas and technical pathways for blockchain smart contract security detection.
The rapid development of Web3 has generated massive amounts of on-chain data, making it crucial to effectively analyze and understand the complex relationships within blockchain ecosystems. Although standard RAG techniques augment LLMs through external data retrieval, it falls short in capturing the intricate network of relationships in Web3 data. In this paper, this work introduces an innovative method that combines GraphRAG with community detection algorithms to analyze Web3 textual data. By constructing knowledge graphs from Web3-related documents and leveraging community structures, our system can better understand the semantic relationships and contextual connections in Web3 content, delivering higher-precision answers to domain-specific questions. Our experiments on real-world Web3 textual data show that our method achieves superior response accuracy and contextual understanding compared to traditional RAG approaches, especially for complex Web3 concepts and community-driven insights.
Yi-Jing Liu, Long Zhang, Xiaoqian Li, Hongyang Du · 7 authors
Federated learning (FL) is integral to advancing edge intelligence by enabling collaborative machine learning. In FL-empowered edge networks, computing nodes first train local models and then send them to an or multiple aggregation node(s) for global model collaboration. However, the trustworthiness of both local and global models in conventional FL frameworks is compromised due to inadequate model security and transparency. Distributed ledger technique (DLT) can address this issue by leveraging multi-nodes trust capabilities to support distributed consensus. However, model training and consensus performance of DLT may significantly degrade due to instability and resource constraints of edge networks. Sharding technique provides an effective approach by dividing the ledger into smaller and manageable shards. In this paper, to improve model training and consensus performance, we propose a trusted FL framework by incorporating sharding DLT into FL frameworks. We construct a theoretical model to investigate the relationship between model training performance, consensus efficiency, and capacity of edge nodes regarding storage, computing and communications. Based on the theoretical model, we propose a trusted clustering scheme to aggregate local models. Numerical results show that our proposed scheme significantly improves network throughput for transmitting models while guaranteeing model learning performance in comparison with some classical baselines.
Decentralised applications (dApps) that run on public blockchains have the benefit of trustworthiness and transparency as every activity that happens on the blockchain can be publicly traced through the transaction data. However, this introduces a potential privacy problem as this data can be tracked and analysed, which can reveal user-behaviour information. A user behaviour analysis pipeline was proposed to present how this type of information can be extracted and analysed to identify separate behavioural clusters that can describe how users behave in the game. The pipeline starts with the collection of transaction data, involving smart contracts, that is collected from a blockchain-based game called Planet IX. Both the raw transaction information and the transaction events are considered in the data collection. From this data, separate game actions can be formed and those are leveraged to present how and when the users conducted their in-game activities in the form of user flows. An extended version of these user flows also presents how the Non-Fungible Tokens (NFTs) are being leveraged in the user actions. The latter is given as input for a Graph Neural Network (GNN) model to provide graph embeddings for these flows which then can be leveraged by clustering algorithms to cluster user behaviours into separate behavioural clusters. We benchmark and compare well-known clustering algorithms as a part of the proposed method. The user behaviour clusters were analysed and visualised in a graph format. It was found that behavioural information can be extracted regarding the users that belong to these clusters. Such information can be exploited by malicious users to their advantage. To demonstrate this, a privacy threat model was also presented based on the results that correspond to multiple potentially affected areas.
The rapid advancement of Large Language Models (LLMs) has catalyzed the development of multi-agent systems, where multiple LLM-based agents collaborate to solve complex tasks. However, existing systems predominantly rely on centralized coordination, which introduces scalability bottlenecks, limits adaptability, and creates single points of failure. Additionally, concerns over privacy and proprietary knowledge sharing hinder cross-organizational collaboration, leading to siloed expertise. To address these challenges, we propose AgentNet, a decentralized, Retrieval-Augmented Generation (RAG)-based framework that enables LLM-based agents to autonomously evolve their capabilities and collaborate efficiently in a Directed Acyclic Graph (DAG)-structured network. Unlike traditional multi-agent systems that depend on static role assignments or centralized control, AgentNet allows agents to specialize dynamically, adjust their connectivity, and route tasks without relying on predefined workflows. AgentNet’s core design is built upon several key innovations: (1) Fully Decentralized Paradigm: Removing the central orchestrator, allowing agents to coordinate and specialize autonomously, fostering fault tolerance and emergent collective intelligence. (2) Dynamically Evolving Graph Topology: Real-time adaptation of agent connections based on task demands, ensuring scalability and resilience. (3) Adaptive Learning for Expertise Refinement: A retrieval-based memory system that enables agents to continuously update and refine their specialized skills. By eliminating centralized control, AgentNet enhances fault tolerance, promotes scalable specialization, and enables privacy-preserving collaboration across organizations. Through decentralized coordination and minimal data exchange, agents can leverage diverse knowledge sources while safeguarding sensitive information. Experimental results demonstrate that AgentNet outperforms traditional centralized multi-agent systems, significantly improving efficiency, adaptability, and scalability in dynamic environments, making it a promising foundation for next-generation autonomous, privacy-respecting multi-agent ecosystems.
Junhao Wu, Yixin Yang, Chengxiang Jin, Silu Mu · 8 authors
With the widespread adoption of Ethereum, financial frauds such as Ponzi schemes have become increasingly rampant in the blockchain ecosystem, posing significant threats to the security of account assets. Existing Ethereum fraud detection methods typically model account transactions as graphs, but this approach primarily focuses on binary transactional relationships between accounts, failing to adequately capture the complex multi-party interaction patterns inherent in Ethereum. To address this, we propose a hypergraph modeling method for the Ponzi scheme detection method in Ethereum, called HyperDet. Specifically, we treat transaction hashes as hyperedges that connect all the relevant accounts involved in a transaction. Additionally, we design a two-step hypergraph sampling strategy to significantly reduce computational complexity. Furthermore, we introduce a dual-channel detection module, including the hypergraph detection channel and the hyper-homo graph detection channel, to be compatible with existing detection methods. Experimental results show that, compared to traditional homogeneous graph-based methods, the hyper-homo graph detection channel achieves significant performance improvements, demonstrating the superiority of hypergraph in Ponzi scheme detection. This research offers innovations for modeling complex relationships in blockchain data.
Hao-Tse Chung, Shao‐Hung Cheng, Yu‐Jia Chen, Li‐Chun Wang
Emerging Blockchain-empowered Federated Learning (BCFL) technology combines the decentralized security of blockchain with the privacy protection of federated learning. BCFL addresses the issue of single points of failure in centralized systems, making it an increasingly popular solution. However, current consensus mechanisms, such as Proof of Work (PoW), Proof of Stake (PoS), and Practical Byzantine Fault Tolerance (PBFT), lead to challenges such as high computational costs and limited scalability. This paper proposes a Batch Zero-Knowledge Proof-based practical Byzantine fault-tolerant (BZ-BFT) consensus mechanism for BCFL to enhance efficiency and reliability. By integrating Zero-Knowledge Proof (ZKP), our approach enables the verification of the primary node's proposal without revealing information from other network nodes, thereby ensuring the credibility of the aggregated results. To address the high computational overhead associated with ZKP, we present a batch quantization preprocessing technique called BatchZKP. Our proposed BZ-BFT reduces initialization, proof generation, and verification time by$97.81 \%, 70.0 \%$, and 47.64 %, respectively, significantly boosting BCFL system efficiency and reliability. Additionally, our approach reduces communication complexity from$O\left(n^{2}\right)$to$O(n)$and enhances Byzantine fault tolerance to${1/2}$.
The orchestration of agents to optimize a collective objective without centralized control is challenging yet crucial for applications such as controlling autonomous fleets, and surveillance and reconnaissance using sensor networks. Decentralized controller design has been inspired by self-organization found in nature, with a prominent source of inspiration being flocking; however, decentralized controllers struggle to maintain flock cohesion. The graph neural network (GNN) architecture has emerged as an indispensable machine learning tool for developing decentralized controllers capable of maintaining flock cohesion, but they fail to exploit the symmetries present in flocking dynamics, hindering their generalizability. We enforce rotation equivariance and translation invariance symmetries in decentralized flocking GNN controllers and achieve comparable flocking control with 70% less training data and 75% fewer trainable weights than existing GNN controllers without these symmetries enforced. We also show that our symmetry-aware controller generalizes better than existing GNN controllers. Code and animations are available at http://github.com/Utah-Math-Data-Science/Equivariant-Decentralized-Controllers.
Shiyang Chen, Yang Liu, Qun Zhang, Z. Shao · 5 authors
This article presents MDST‐GNN, a multi‐distance spatial‐temporal graph neural network for blockchain anomaly detection. To address challenges in detecting fraudulent cryptocurrency transactions, MDST‐GNN integrates a multi‐distance graph convolutional architecture with adaptive temporal modeling, enabling capture of both local and global spatial dependencies while inferring patterns from anonymized temporal data. The model incorporates self‐supervised learning to enhance generalization ability. Experiments on the Elliptic dataset demonstrate MDST‐GNN's superior performance over state‐of‐the‐art methods, achieving improvements of 1.5% in AUC‐ROC and 2.9% in AUC‐PR. The model's robustness to temporal granularity and effectiveness in identifying suspicious transactions underscore its practical value for blockchain forensics.
Dong-Seong Kim, Ikechi Saviour Igboanusi, Love Allen Chijioke Ahakonye, Goodness Oluchi Anyanwu
Blockchain networks must ensure secure and efficient consensus algorithms in consumer electronics and Internet of Things (IoT) devices. This paper proposes a novel consensus algorithm, “Proof-of-Authority-and-Association (PoA2)”, designed specifically for IoT blockchain networks in consumer applications. PoA2leverages redundancy-based mechanisms to enhance consensus efficiency and reliability, which is critical for real-time data processing and validation in IoT environments. The algorithm integrates proof of authority and association verification, ensuring network participants are authorized and associated with the transactions they validate. Combining these elements, PoA2mitigates common challenges such as latency, scalability, and energy consumption often encountered in traditional consensus algorithms. Through simulations and performance evaluations, we demonstrate PoA2's effectiveness in achieving consensus while maintaining high levels of security and reducing computational overhead, making it suitable for resource-constrained IoT devices in consumer electronics applications.
Public blockchain has outstanding performance in transaction privacy protection because of its anonymity. The data openness brings feasibility to transaction behavior analysis. At present, the transaction data of the public chain are huge, including complex trading objects and relationships. It is difficult to extract attributes and predict transaction behavior by traditional methods. To solve the problems, we extract the transaction features to construct the Ethereum transaction heterogeneous information network (HIN), and propose graph-neural-network-based transaction prediction method for public blockchain in HINs, which can divide the network into subgraphs according to connectivity and make the prediction results of transaction behavior more accurate. Experiments show that the execution time consumption of the proposed transaction subgraph division method is reduced by 70.61% on average compared with the search method. The accuracy of the proposed behavior prediction method also improve compared with the traditional random walk method, with an average accuracy of 83.82%.
Smart contract security constitutes the foundational cornerstone for ensuring the trusted operational integrity of blockchain ecosystems. In recent years, multi-task learning MTL architectures have been widely adopted in smart contract vulnerability detection, owing to their context-aware optimization and superior generalization capabilities compared to single-task learning STL frameworks. However, MTL-based approaches for smart contract vulnerability detection face two persistent challenges: 1 The negative transfer phenomenon, the mitigation of negative transfer via adaptive loss weighting in smart contract vulnerability detection remains underexplored in existing research. 2 Performance degradation caused by the homogeneous contribution assumption where undifferentiated contract representations impair expert layer learning efficacy. To overcome these limitations, we propose a novel detection framework incorporating adaptive loss weight and entropy-based feature enhancement. Our dual-weighting mechanism introduces: 1 dynamic loss coefficients that automatically balance task-specific optimization objectives based on evolving learning complexity and task significance, and 2 entropy-aware attention weights that prioritize high-information contract features during expert network training. Comprehensive evaluations on real-world smart contract datasets demonstrate the framework's superior detection performance compared to three state-of-the-art adaptive weighting baselines. Experimental results reveal significant improvements in F1-score across multiple vulnerability types, validating the effectiveness of our approach in mitigating negative transfer while maintaining robust concurrent detection capabilities. The experimental code will be systematically organized and made publicly available on GitHub shortly.
As the core component of blockchain applications, smart contracts are increasingly scrutinized for their security. Among various vulnerabilities, infinite loop flaws pose significant threats due to their hidden nature and potential for exhausting system resources. This paper proposes a static detection method based on Graph Convolutional Networks (GCNs), which transforms smart contracts into control flow and data flow graphs. Through graph-based modeling and vectorized encoding, semantic features such as loop structures and function dependencies are effectively captured. An improved GCN architecture is employed to identify potential infinite execution patterns through neighborhood aggregation and graph-level representation learning. Experimental results demonstrate that the proposed method achieves high accuracy and F1-score across real-world contract datasets, offering an effective and scalable solution for smart contract vulnerability analysis.
Smart contracts are self-executing programs on blockchains that manage complex business logic with transparency and integrity. However, their immutability after deployment makes programming errors particularly critical, as su... | Find, read and cite all the research you need on Tech Science Press