The web3 applications have recently been growing, especially on the Ethereum platform, starting to become the target of scammers. The web3 scams, imitating the services provided by legitimate platforms, mimic regular activity to deceive users. However, previous studies have primarily concentrated on de-anonymization and phishing nodes, neglecting the distinctive features of web3 scams. Moreover, the current phishing account detection tools utilize graph learning or sampling algorithms to obtain graph features. However, large-scale transaction networks with temporal attributes conform to a power-law distribution, posing challenges in detecting web3 scams. To overcome these challenges, we present ScamSweeper, a novel framework that emphasizes the dynamic evolution of transaction graphs, to identify web3 scams on Ethereum. ScamSweeper samples the network with a structure temporal random walk, which is an optimized sample walking method that considers both temporal attributes and structural information. Then, the directed graph encoder generates the features of each subgraph during different temporal intervals, sorting as a sequence. Moreover, a variational Transformer is utilized to extract the dynamic evolution in the subgraph sequence. Furthermore, we collect a large-scale transaction dataset consisting of web3 scams, phishing, and normal accounts, which are from the first 18 million block heights on Ethereum. Subsequently, we comprehensively analyze the distinctions in various attributes, including nodes, edges, and degree distribution. Our experiments indicate that ScamSweeper outperforms SIEGE, Ethident, and PDTGA in detecting web3 scams, achieving a weighted F1-score improvement of at least 17.29% with the base value of 0.59. In addition, ScamSweeper in phishing node detection achieves at least a 17.5% improvement over DGTSG and BERT4ETH in F1-score from 0.80.
We introduce the Overlap-Weighted Hierarchical Normalized Persistence Velocity (OW-HNPV), a novel topological data analysis method for detecting anomalies in time-varying networks. Unlike existing methods that measure cumulative topological presence, we introduce the first velocity-based perspective on persistence diagrams, measuring the rate at which features appear and disappear, automatically downweighting noise through overlap-based weighting. We also prove that OW-HNPV is mathematically stable. It behaves in a controlled, predictable way, even when comparing persistence diagrams from networks with different feature types. Applied to Ethereum transaction networks (May 2017-May 2018), OW-HNPV demonstrates superior performance for cryptocurrency anomaly detection, achieving up to 10.4% AUC gain over baseline models for 7-day price movement predictions. Compared with established methods, including Vector of Averaged Bettis (VAB), persistence landscapes, and persistence images, velocity-based summaries excel at medium- to long-range forecasting (4-7 days), with OW-HNPV providing the most consistent and stable performance across prediction horizons. Our results show that modeling topological velocity is crucial for detecting structural anomalies in dynamic networks.
The accurate and timely classification of toddlers' nutritional status is critical for early intervention, particularly in remote or underserved communities with limited access to healthcare professionals. However, data security, especially for children's health data, is equally essential to ensure safe storage and access. To address these challenges, this study proposes a hybrid AI-powered chatbot that integrates ensemble learning, blockchain, and decentralized storage to support both nutritional status classification and educational interaction. The system combines a random forest model for classification with GPT-3.5 Turbo for bilingual (Indonesian–English) stunting education deployed via Telegram. Preprocessing includes standardizing, normalizing, and encoding Indonesian-language nutrition data to ensure machine learning readiness. Six ensemble algorithms are evaluated using stratified five-fold cross-validation, with classification results hashed using SHA-256 and immutably stored on the Interplanetary File System (IPFS) and a local Ethereum blockchain. The chatbot effectively manages both structured inputs and natural language queries, ensuring secure, transparent, and real-time nutritional assessments. Results demonstrate high classification performance, with the random forest model achieving the highest mean F1-score (0.9987) and the lowest deviation. Its robustness was validated by a 20% hold-out test set and stratified five-fold cross-validation, which obtained excellent balanced performance across nutritional status categories (F1-macro, precision, recall, accuracy ≈ 0.99; ROC AUC = 1.00). External validation also yielded robust and consistent results (F1-macro = 0.97, precision = 0.97, recall = 0.96, ROC AUC = 0.98, and accuracy = 0.97), demonstrating the model's generalization ability and mitigating concerns regarding overfitting. Blockchain evaluation confirmed stable and linear CID transaction throughput (blocks 29–46) with no observed latency, ensuring reliable and continuous data recording. Furthermore, gas prices decreased by ~87.5%, highlighting significant improvements in cost efficiency and scalability, which reinforces blockchain's feasibility for decentralized, AI-driven health data management. Received: 9 June 2025 | Revised: 29 September 2025 | Accepted: 31 October 2025 Conflicts of Interest The authors declare that they have no conflicts of interest to this work. Data Availability Statement The data that support the findings of this study are openly available in Kaggle at https://www.kaggle.com/datasets/rendiputra/stunting-balita-detection-121k-rows and https://www.kaggle.com/datasets/jabirmuktabir/stunting-wasting-dataset. Author Contribution Statement Wa Ode Siti Nur Alam: Conceptualization, Methodology, Software, Validation, Formal analysis, Investigation, Resources, Data curation, Writing – original draft, Writing – review & editing, Visualization, Project administration. Riri Fitri Sari: Conceptualization, Writing – review & editing, Supervision, Funding acquisition.
Abstract Fair exchange is a cryptographic protocol that enables two parties to exchange their electronic data fairly, i.e., it ensures that no one can steal the other party’s item. FairSwap and OptiSwap are well-known fair exchange protocols for files and coins based on smart contracts. Although the design principles for the two protocols are similar, there is an extreme difference in round complexity and communication overhead between the two protocols. This gap can be a barrier to users since they must choose one of them. In order to remove the barrier, this work generalizes these two protocols. It proposes a fair exchange protocol that allows users to adjust the communication overhead and the number of rounds by a newly introduced parameter. Our protocol contains FairSwap, OptiSwap, and protocols with intermediate efficiency between them. Moreover, to evaluate the gas costs of our protocol, we present a performance evaluation of the proposed protocol by Solidity implementation over Ethereum.
Despite the growing emphasis on the nexus between growth and macroeconomic indicators, research on the influence of cryptocurrencies on economic performance remains limited. This study compares the impact of two leading cryptocurrencies, Bitcoin and Ethereum, on economic growth, alongside inflation, market uncertainty, and oil and gold prices, using panel data from 14 countries between Q3 2015 and Q3 2023. The results demonstrate robust cross-sectional dependence, indicating that economic shocks in one country affect the entire group. Therefore, second-generation tests are employed to confirm the presence of stationarity in the variables. Except for Bitcoin’s trading volume, panel fully modified ordinary least squares estimations reveal a significantly positive impact of cryptocurrencies on growth. Cointegration is present in the long run, while in the short run, strong bi- and unidirectional causality is found for all cryptocurrency proxies. The study provides insights that can help policymakers develop strategies to align economic growth with the crypto market, benefiting the broader economy.
Abstract This paper investigates the optimization of data sampling and target labeling techniques to enhance algorithmic trading strategies in cryptocurrency markets, focusing on Bitcoin (BTC) and Ethereum (ETH). Traditional data sampling methods, such as time bars, often fail to capture the nuances of the continuously active and highly volatile cryptocurrency market and force traders to wait for arbitrary points in time. To address this, we propose an alternative approach using information-driven sampling methods, including the CUSUM filter, range bars, volume bars, and dollar bars, and evaluate their performance using tick-level data from January 2018 to June 2023. Additionally, we introduce the Triple Barrier method for target labeling, which offers a solution tailored for algorithmic trading as opposed to the widely used next-bar prediction. We empirically assess the effectiveness of these data sampling and labeling methods to craft profitable trading strategies. The results demonstrate that the innovative combination of CUSUM-filtered data with Triple Barrier labeling outperforms traditional time bars and next-bar prediction, achieving consistently positive trading performance even after accounting for transaction costs. Moreover, our system enables making trading decisions at any point in time on the basis of market conditions, providing an advantage over traditional methods that rely on fixed time intervals. Furthermore, the paper contributes to the ongoing debate on the applicability of Transformer models to time series classification in the context of algorithmic trading by evaluating various Transformer architectures—including the vanilla Transformer encoder, FEDformer, and Autoformer—alongside other deep learning architectures and classical machine learning models, revealing insights into their relative performance.
Smart contracts rely on blockchain oracles to access off-chain data, yet existing oracle designs often face challenges such as untrustworthy data sources, weak temporal guarantees, and limited verifiability. This work presents Ivy Oracle, a robust and time-trustworthy data feed framework that enhances the reliability and auditability of off-chain information for smart contracts. Ivy Oracle integrates trusted execution environments (TEEs) for secure data acquisition, an external time server for authenticated timestamps, and a PageRank-based trust model to evaluate source credibility. We implement and evaluate Ivy Oracle on the Ethereum Sepolia testnet, demonstrating that it achieves up to 63.6% lower on-chain gas consumption than Chainlink for signature verification while maintaining only a slight increase in communication overhead due to its dual-attestation mechanism. These results confirm that Ivy Oracle provides strong time trustworthiness and data reliability with minimal performance cost, making it suitable for latency-sensitive blockchain applications.
This study asks whether Ethereum’s proof-of-stake (PoS) incentives not only make economic sense on paper but also feel attractive to real validators who may be loss-averse and sensitive to risk. We take a canonical Eth2 slot-level model of rewards, penalties, costs, and proposer-conditional maximal extractable value (MEV) and overlay a prospect-theoretic valuation that captures reference dependence, loss aversion, diminishing sensitivity, and probability weighting. This Prospect-Theoretic Incentive Mechanism (PT-IM) separates the “money edge” (expected accounting return) from the “felt edge” (behavioral value) by mapping monetary outcomes through a prospect value function and comparing the two across parameter ranges. The mechanism is parametric and modular, allowing different MEV, cost, and penalty profiles to plug in without altering the base PoS model. Using stylized numerical examples, we identify regions where cooperation that pays in expectation can remain unattractive under plausible loss-averse preferences, especially when penalties are salient or MEV is volatile. We discuss how these distortions may affect validator participation, economic security, and the tuning of rewards and penalties in Ethereum’s PoS. Integrating behavioral valuation into crypto-economic design thus provides a practical diagnostic for adjusting protocol parameters when economics and perception diverge.
Dan Alexandru Mitrea, Constantin Viorel Marian, Rareş Alexandru Manolescu
In many jurisdictions, property registration and transfers remain constrained by inefficient, paper-based processes that depend on multiple intermediaries and bureaucratic approvals. This paper proposes a decentralized, blockchain-based property platform designed to streamline these processes using Non-Fungible Tokens (NFTs) and artificial intelligence (AI) agents to modernize public-sector asset management. The work addresses the persistent inefficiencies of paper-based property registration and ownership transfer by embedding legal and administrative logic within smart contracts and automating compliance through an intelligent conversational interface. The system was implemented using Ethereum-based ERC-721 standards, React for the user interface, and Langfuse-powered AI integration for guided user interaction. The pilot implementation presents secure, transparent, and auditable property-transfer transactions executed entirely on-chain, while hybrid IPFS-based storage and decentralized identifiers preserve privacy and legal validity. Comparative analysis against existing national initiatives indicates that the proposed architecture delivers decentralization, citizen control, and interoperability without compromising regulatory requirements. The system reduces bureaucratic overhead, simplifies transaction workflows, and lowers user error risk, thereby strengthening accountability and public trust. Overall, the paper outlines a viable foundation for legally aligned, AI-assisted digital property registries and offers a policy-oriented roadmap for integrating blockchain-enabled systems into public-sector governance infrastructures.
Mutiullah Shaikh, Uffe Kock Wiil, Ali Ebrahimi, Yumna Memon
Blockchain technology has revolutionized digital systems by ensuring trust, transparency, decentralization, and security. However, in the democratic nature of blockchain networks, there is a huge underlying dependency on consensus mechanisms, but the challenges associated with these, such as energy costs, network attacks, preservation of privacy, centralization, and limited scalability, hinder miners and stakeholders from adopting appropriate consensus mechanisms. In this paper, we present a conceptual literature overview of most consensus mechanisms by highlighting potential areas of exploration and considerations before adopting blockchain technology for various applications. This exploration turned our focus toward analyzing three prominent underlying aspects of consensus mechanisms, i.e. energy consumption, security, and decentralization. A simulation-based comparative analysis of five prominent blockchain consensus mechanisms, such as Proof of Work (PoW), Proof of Stake (PoS), Delegated Proof of Stake (DPoS), Proof of Authority (PoA), and Proof of Capacity (PoC), is presented in various network load scenarios to further evaluate their performance metrics. The simulated metrics were cross-validated using empirical data from real blockchain networks (e.g., Ethereum, Bitcoin, VeChain, and Chia) collected between 2022 and 2025, ensuring alignment between theoretical performance models and observed on-chain behavior across diverse consensus mechanisms. Results overall indicate that PoW excels in decentralization and security while costing the highest energy, making it less scalable for high-throughput scenarios. PoS balances energy efficiency and moderate decentralization, while DPoS achieves scalability at the expense of decentralization. PoA and PoC are shown to be energy-efficient alternatives, but vary in their levels of centralization and security. Our findings constitute a comprehensive guide for researchers, miners, and practitioners aiming to optimize blockchain performance for diverse applications.
Decentralized lending protocols, exemplified by Aave V3, have transformed financial intermediation by enabling permissionless, multi-chain borrowing and lending without intermediaries. Despite managing over $10 billion in total value locked, empirical research remains severely constrained by the lack of standardized, cross-chain event-level datasets. This paper introduces the first comprehensive, event-driven data infrastructure for Aave V3 spanning six major EVM-compatible chains (Ethereum, Arbitrum, Optimism, Polygon, Avalanche, and Base) from respective deployment blocks through October 2025. We collect and fully decode eight core event types -- Supply, Borrow, Withdraw, Repay, LiquidationCall, FlashLoan, ReserveDataUpdated, and MintedToTreasury -- producing over 50 million structured records enriched with block metadata and USD valuations. Using an open-source Python pipeline with dynamic batch sizing and automatic sharding (each file less than or equal to 1 million rows), we ensure strict chronological ordering and full reproducibility. The resulting publicly available dataset enables granular analysis of capital flows, interest rate dynamics, liquidation cascades, and cross-chain user behavior, providing a foundational resource for future studies on decentralized lending markets and systemic risk.
The aim of this dissertation is to test the applicability of two strategies – Dollar Cost Average (DCA) and Lump-Sum (LS) – in the context of the crypto market. We tested these strategies on three assets, namely Bitcoin, Ethereum and Ripple. We developed a simulation using daily historical data recorded over a period of nine years. We then calculated performance ratios and created an AR-GARCH model to analyse their properties and predictive capacity more effectively. Our empirical results show that all assets are highly volatile and exhibit heavy tails and asymmetry. Additionally, they are moderately to highly correlated with each other. We also presented proof of higher Sharpe and Sortino ratios for DCA strategies, with Bitcoin performing better than the other two assets. The results also show that Bitcoin has low-to-moderate shock sensitivity and high persistence; Ethereum has low shock sensitivity and high persistence; and Ripple has both high shock sensitivity and persistence. Furthermore, we observed the impact of strategy choice on volatility. When compared to DCA, LS lowered shock sensitivity in Bitcoin and Ripple, enhancing persistence, while having an insignificant effect on Ethereum. Finally, we demonstrate that our model exhibits superior predictive capacity with regard to Ripple compared to Bitcoin and Ethereum, and that all three assets are inefficient. These findings contribute to previous literature by providing novel empirical data and attesting to the attributes of cryptocurrencies. Furthermore, this thesis improves financial awareness and provides investors with valuable information.
Love Allen Chijioke Ahakonye, Hamza Ibrahim, Jae-Min Lee, Dong‐Seong Kim
Smart contract environments are increasingly targeted by stealthy, adaptive attacks that evade conventional rule-based or static anomaly detection systems. Inspired by the anglerfish’s bioluminescent filament, which perceives and lures activity in dark, dynamic environments, this research introduces a Bioluminescent Filament-Inspired Artificial Intelligence Perception framework for smart contract intrusion detection. The proposed model emulates biological sensory adaptation through multi-modal attention layers that dynamically illuminate anomalous behaviors in contract execution flows. By integrating self-supervised temporal perception with context-driven feedback, the framework continuously refines its detection sensitivity while maintaining low computational overhead. We evaluate the framework using fuzz-tested smart contract vulnerability datasets that simulate diverse malicious execution behaviors observed in Ethereum environments, demonstrating over 98% detection accuracy with a 40% reduction in latency compared to traditional deep learning-based IDS models. This biologically inspired perception paradigm offers a scalable, energy-efficient solution for securing blockchain-based decentralized systems against evolving threat vectors.
Even in this day and age, when digital technologies are becoming more and more prevalent, it is still extremely important for democratic systems to maintain the honesty and openness of their voting procedures. This article introduces NextGenVote, a decentralised online voting platform developed to address the security, transparency, and confidence issues traditional electronic voting systems face. Automation of election operations, including voter registration, candidate administration, ballot casting, and result computation, is achieved through smart contracts written in the Solidity programming language. The system is built on the Ethereum blockchain. MetaMask is a React-based frontend that uses Web3.js to connect to the blockchain. MetaMask is responsible for ensuring that user authentication and transaction signatures are secure. Therefore, to prevent unauthorised manipulation, the platform utilises a role-based access control approach that clearly distinguishes between administrative capabilities and voter credentials. NextGenVote assures that election results are tamper-proof, traceable, and auditable. It was deployed and tested in a local blockchain environment powered by Ganache. The system provides a solid foundation for scalable, secure, and transparent digital elections by eliminating centralised intermediaries and relying solely on processes executed on the blockchain.
The rising demand for collaborative machine learning and data analytics calls for secure and decentralized data sharing frameworks that balance privacy, trust, and incentives. Existing approaches, including federated learning (FL) and blockchain-based data markets, fall short: FL often depends on trusted aggregators and lacks Byzantine robustness, while blockchain frameworks struggle with computation-intensive training and incentive integration. We present \prot, a decentralized data marketplace that unifies federated learning, blockchain arbitration, and economic incentives into a single framework for privacy-preserving data sharing. \prot\ enables data buyers to submit bid-based requests via blockchain smart contracts, which manage auctions, escrow, and dispute resolution. Computationally intensive training is delegated to \cone\ (\uline{Co}mpute \uline{N}etwork for \uline{E}xecution), an off-chain distributed execution layer. To safeguard against adversarial behavior, \prot\ integrates a modified YODA protocol with exponentially growing execution sets for resilient consensus, and introduces Corrected OSMD to mitigate malicious or low-quality contributions from sellers. All protocols are incentive-compatible, and our game-theoretic analysis establishes honesty as the dominant strategy. We implement \prot\ on Ethereum and evaluate it over benchmark datasets -- MNIST, Fashion-MNIST, and CIFAR-10 -- under varying adversarial settings. \prot\ achieves up to 99\% accuracy on MNIST and 90\% on Fashion-MNIST, with less than 3\% degradation up to 30\% Byzantine nodes, and 56\% accuracy on CIFAR-10 despite its complexity. Our results show that \prot\ ensures privacy, maintains robustness under adversarial conditions, and scales efficiently with the number of participants, making it a practical foundation for real-world decentralized data sharing.
The rapid proliferation of Internet of Things (IoT) devices across various industries, including healthcare, smart cities, and industrial automation, has introduced significant security, authenticity, and traceability challenges within increasingly complex supply chains. Although existing approaches have utilised blockchain-based digital identity solutions to address some of these concerns, persistent issues of counterfeit products and inadequate lifecycle transparency highlight the need for more robust, hardware-anchored identification mechanisms. Our work presents a novel architecture that integrates Physically Unclonable Functions (PUFs) and blockchain-based Soulbound Tokens (SBTs) to establish secure and verifiable digital identities directly tied to the physical hardware of IoT devices. By employing cryptographic tools such as fuzzy extractors, Merkle trees, and zero-knowledge proofs, the proposed architecture ensures accurate lifecycle tracking through key operational stages, including manufacturing, procurement, provisioning, maintenance, and eventual disposal or recycling. Performance evaluations conducted on the Ethereum Sepolia testnet demonstrate reasonable computational overhead in terms of gas usage and transaction confirmation times. The findings reveal that this approach aligns with NIST Special Publication 800-161 guidelines, as well as emerging regulatory standards, notably the European Union’s Digital Product Passport initiative, and has significant implications for enhancing transparency, sustainability, and security across global IoT supply chains.
Open access
Physical Unclonable Functions (PUFs) and Hardware Security
Ethereum has become a significant trading platform for financial activities such as Dapps, ICOs, and DeFi. However, it has also become a hub for criminal activities such as fraud, money laundering, and illicit fundraising. The construction of fraud detection models employing machine learning techniques is currently a mainstream research direction. Nevertheless, existing studies face significant challenges, including class imbalance in data samples and a lack of model interpretability. In this content, this work proposes a novel explainable model for Ethereum illicit account detection, ETHIAD (Ethereum Illicit Account Detection). Firstly, we pre-process the dataset by ADASYN oversampling and Lasso feature selection, etc., to more efficiently achieve feature modeling of transaction structures. Then, the ETHIAD model is trained using the XGboost algorithm, with an accuracy, precision, recall, F1 score, and AUC value of 99.70%, 99.51%, 99.02%, 99.26%, and 99.45%, respectively, the model outperforms the existing SOTA model by 0.05%-1.1%. Finally, we introduce SHAP framework to analyze the key influencing factors of illicit accounts from multiple perspectives, and the conclusions strongly enhance the explainability of the model.
High-frequency trading, in both traditional and decentralized markets, induces latency races and redundant order flow as traders spend resources to win time-sensitive opportunities. We show that auctioning artificial time priority can redirect resources away from wasteful speed races toward auction payments. While such waste is difficult to measure in traditional markets, blockchain transactions provide transparent records of these competitive costs through observable duplicate submissions. We study the introduction of Timeboost, a time-priority auction mechanism on Arbitrum, a blockchain that batches transactions before settlement on Ethereum, as a natural experiment. We find that redundant transactions decrease and platform revenue increases relative to comparable networks, consistent with our theoretical predictions.
The transition to electric vehicles (EVs) plays a critical role in reducing global carbon emissions. However, the end-of-life management of electric vehicle batteries (EVBs) presents significant sustainability and operational challenges. This study proposes a blockchain-based framework that enables full lifecycle tracking of EVBs, from production to disposal or reuse, while addressing issues of transparency, efficiency, and regulatory compliance. The framework incorporates a multi-criteria decision model to guide data-driven end-of-life routing—whether for second-life reuse or direct recycling—based on technical, environmental, and economic indicators. By integrating smart contracts with a hybrid web/mobile platform, the system ensures tamper-proof documentation, stakeholder accountability, and compliance with the EU battery passport regulation. A detailed cost analysis of deploying the framework on Ethereum is also presented. The proposed solution aims to enhance the sustainability of EVB management, reduce environmental impact, and promote circular economy practices within the EV industry.
Despite the recognition of Blockchain Technology’s disruptive potential, there is ongoing debate about its ontological and axiomatic foundations. This study develops a theoretical framework to explain the underline structural principles of blockchain technology through the lens of Arthur’s theory of technology, and the framework is developed through adopting Narrative Literature Review. By integrating conceptual analysis with a structural examination of Ethereum, this study reveals that blockchain technology is not a single invention but a composite technological system developed through recursive interactions among sub-technologies. The proposed framework identifies three interrelated structural patterns—the Combinatorial Pattern of Components elucidating blockchain technology’s structural ontology, the Capturing Pattern of Algorithms revealing the operational source of its innovation, and the Recursive Pattern of Technologies characterizing its inner logical structure of components—that together explain blockchain technology’s generative and evolving nature. The study extends Arthur’s theory by clarifying the “technology within technology” dynamic that underlies blockchain technology innovation. The Ethereum case confirms the framework’s applicability and generalizability, showing that blockchain systems, despite their diversity, share a consistent structural logic. Beyond its theoretical contribution, the framework offers practical guidance for sustainable technological innovation. It provides analytical support for designing blockchain-based applications’ architectures that enhance transparency, efficiency, and adaptability, contributing to the sustainable evolution of digital technologies.
Topicality. Fraudulent activities on the Ethereum blockchain pose a substantial risk to decentralized finance and require capable models not only to respond to already detected abuses but also to identify suspicious accounts proactively before losses escalate. The subject of study is the application of graph and temporal neural models to the task of classifying Ethereum accounts as benign or fraudulent, considering the structural relationships between addresses and the temporal dynamics of transactions. The purpose of this article is to develop and experimentally evaluate a neural architecture based on a multilayer perceptron as a baseline component for the subsequent integration of graph and temporal mechanisms, and to analyze its performance on the open Ethereum Fraud Detection dataset, which features a high-class imbalance. The following results were obtained. A baseline deep model for binary account classification was constructed using feature preprocessing, stratified data splitting, class weight balancing, L2 regularization, Dropout, and early stopping, which enabled the achievement of an ROC AUC value of approximately 0.98 under conditions of a pronounced dominance of the safe class. A detailed analysis of the confusion matrix and the precision, recall, and F1 metrics demonstrated an acceptable trade-off between reducing false positives and minimizing the proportion of missed fraudulent accounts, which is critical for real-world financial scenarios. Conclusion. The results indicate that a properly designed baseline neural model on tabular features can ensure high-quality proactive identification of fraudulent Ethereum accounts and serve as a starting point for further integration of graph and temporal architectures aimed at improving interpretability and robustness to the evolution of malicious behavior patterns.
Vladyslav Prosolov, Oleksandr Kushnerov, Vladyslav Sokol, Ruslan Trofymenko
Topicality. Fraudulent activities on the Ethereum blockchain pose a substantial risk to decentralized finance and require capable models not only to respond to already detected abuses but also to identify suspicious accounts proactively before losses escalate. The subject of study is the application of graph and temporal neural models to the task of classifying Ethereum accounts as benign or fraudulent, considering the structural relationships between addresses and the temporal dynamics of transactions. The purpose of this article is to develop and experimentally evaluate a neural architecture based on a multilayer perceptron as a baseline component for the subsequent integration of graph and temporal mechanisms, and to analyze its performance on the open Ethereum Fraud Detection dataset, which features a high-class imbalance. The following results were obtained. A baseline deep model for binary account classification was constructed using feature preprocessing, stratified data splitting, class weight balancing, L2 regularization, Dropout, and early stopping, which enabled the achievement of an ROC AUC value of approximately 0.98 under conditions of a pronounced dominance of the safe class. A detailed analysis of the confusion matrix and the precision, recall, and F1 metrics demonstrated an acceptable trade-off between reducing false positives and minimizing the proportion of missed fraudulent accounts, which is critical for real-world financial scenarios. Conclusion. The results indicate that a properly designed baseline neural model on tabular features can ensure high-quality proactive identification of fraudulent Ethereum accounts and serve as a starting point for further integration of graph and temporal architectures aimed at improving interpretability and robustness to the evolution of malicious behavior patterns.
Uisang Lee, Changhoon Chung, Junmo Lee, Sung Jun Moon
The rapid growth of Ethereum has made it more important to quickly and accurately detect smart contract vulnerabilities. While machine-learning-based methods have shown some promise, many still rely on rule-based preprocessing designed by domain experts. Rule-based preprocessing methods often discard crucial context from the source code, potentially causing certain vulnerabilities to be overlooked and limiting adaptability to newly emerging threats. We introduce BugSweeper, an end-to-end deep learning framework that detects vulnerabilities directly from the source code without manual engineering. BugSweeper represents each Solidity function as a Function-Level Abstract Syntax Graph (FLAG), a novel graph that combines its Abstract Syntax Tree (AST) with enriched control-flow and data-flow semantics. Then, our two-stage Graph Neural Network (GNN) analyzes these graphs. The first-stage GNN filters noise from the syntax graphs, while the second-stage GNN conducts high-level reasoning to detect diverse vulnerabilities. Extensive experiments on real-world contracts show that BugSweeper significantly outperforms all state-of-the-art detection methods. By removing the need for handcrafted rules, our approach offers a robust, automated, and scalable solution for securing smart contracts without any dependence on security experts.
Deepika Dash, Bipin Raj C., B Jnyanadeep, Anala M R
The proliferation of decentralized finance (DeFi) has highlighted critical challenges in cross- chain oracle reliability and performance assessment. Traditional blockchain networks remain isolated from external data sources, creating the fundamental Oracle Problem that hinders institutional adoption of DeFi protocols. This paper presents DeFiLens, a comprehensive benchmarking framework that provides standard- ized performance metrics and real-time analytics across multiple blockchain ecosystems including Ethereum, Binance Smart Chain, Polygon, and Avalanche. Our framework addresses the gap between traditional finance’s seamless market data access and blockchain’s data isolation through systematic oracle assessment. DeFiLens implements a six-layer security scoring system encompassing cryptographic verification, attack detection, and network health monitoring. Through extensive evaluation of major oracle providers including Chainlink, Band Protocol, and Tellor, we demonstrate significant performance variations across chains, with response times ranging from 2.1 seconds to 8.7 seconds and reliability scores varying between 72% and 95%. Our statistical analysis reveals critical arbitrage opportunities with price discrepancies up to 2.3% across chains. The framework serves as a ‘‘Bloomberg Terminal’’ for oracle data, enabling financial institutions, DeFi protocols, and researchers to make data-driven decisions for oracle integration and risk management.