Ahmed Abbas Jasim AlâHchaimi, M. A. Khalifa, Walid ElâShafai
ABSTRACT Blockchain networks now support billions of dollars in daily transactions, making reliable and transparent fraud detection essential for maintaining user trust and financial stability. Yet, realâworld blockchain datasets are extremely imbalanced, with fraudulent activity representing less than 1% of all transactions. This imbalance causes conventional machine learning models to achieve deceptively high accuracy while still failing to detect a substantial portion of fraudulent events. To address this challenge, this study evaluates the performance and explainability of three modelsâXGBoost, LightGBM, and Decision Treeâon the Ethereumâbased fraud detection data, in which 58% of transactions are identified as fraud. The methodology combines vast feature engineering, kâfold crossâvalidation, and assorted resampling approaches, such as Synthetic Minority Oversampling Technique (SMOTE) and Adaptive Synthetic Sampling Nearest Neighbor (ADASYN), to revise the effect of class mismatch. Accuracy, AUC, recall, precision, F1âScore, and Matthews Correlation Coefficient(MCC) are used to measure model performance, and SHapley Additive exPlanations (SHAP) is utilized to give global and local interpretability. Experimental results show that XGBoost combined with SMOTE or ADASYN yields the strongest performance, achieving a recall over 99%, an AUC of 1.000, and a substantially improved MCC compared to training on the raw imbalanced data. LightGBM presents a favourable precisionârecall balance, and Decision Trees demonstrate significant gains after resampling, despite their simplicity. SHAP analysis reveals that logâtransformed transaction amount, merchantâbased encoding, geographic encoding, and temporal features are the primary contributors to fraud risk. These results are important in highlighting two implications: (i) the importance of dealing with extreme class imbalance, rather than choosing increasingly sophisticated approaches, and (ii) the ability to be trusted to be explained is a requirement of responsible working in both financial and blockchain settings. The research offers a pragmatic, interpretable framework on blockchain fraud detection and future directions, including sophisticated hybrid sampling, collective learning, as well as crossâchain generalization to enhance fraud detection in distributed systems.
Smart contracts have become the backbone of decentralized ecosystems, managing billions of dollars in assets across applications ranging from Decentralized Finance (DeFi) to digital governance. Given the immutable and autonomous nature of blockchains, the security of these contracts is paramount. A single vulnerability can lead to catastrophic and irreversible financial losses. However, despite these high stakes, a significant gap exists in understanding how developers utilize exception-handling mechanisms to enforce correctness and the specific types of logic flaws that arise from their misuse. This thesis aims to enhance smart contract security through comprehensive studies, beginning with an empirical characterization of defensive programming practices, followed by a systematic analysis of associated faults, and finally, the proposal of a novel vulnerability detection framework. It consists of the following three studies. The first study focuses on the fundamental safeguards of contract logic: state-reverting state-ments (i.e., require, revert, and throw). While these statements serve as the principal mechanisms for exception handling in Solidity, there is a lack of empirical understanding regarding their prevalence and usage patterns in the wild. To address this, the study conducts the first empirical study across thousands of real-world contracts. The results reveal that these statements are pervasive, appearing even more frequently than general-purpose if statements. The analysis further demonstrates that developers primarily use these statements to perform seven types of authority verification and input validity checks. This study establishes an understanding of how developers intend to secure contract logic. The second study investigates the landscape of faults arising from the improper use of these state-reverting statements. Although developers rely on these statements for security, incorrect implementation results in subtle bugs that traditional testing often misses. To understand these failures and benchmark detection capabilities, this study constructs the first comprehensive dataset of 320 real-world faults, curated from open-source project histories and security audit reports Through manual analysis, the study derives a taxonomy of 17 distinct fault types and distills 12 common fixing strategies. A subsequent evaluation of 12 state-of-the-art security tools against this benchmark reveals an average detection rate of only 14.4%, highlighting that existing tools are ineffective at identifying these critical logic flaws. The third study addresses the limitations of existing approaches in identifying high-level semantic vulnerabilities, specifically Price Manipulation. As indicated by the second study, traditional tools struggle with logic flaws because they often lack the ability to interpret complex economic context. To bridge this gap, this study proposes PMDETECTOR, a hybrid framework designed to proactively detect price manipulation. The framework employs a three-stage pipeline to model economic semantics: (1) static taint analysis to identify potentially vulnerable paths, (2) a two-stage Large Language Model (LLM) analysis to filter effective defenses and simulate exploitation, and (3) a final static checker to validate findings. Evaluated on 73 vulnerable and 288 benign contracts, PMDETECTOR achieves up to 100% precision and 88% recall, with GPT-4o achieving a state-of-the-art F1-score of 0.91. Furthermore, in a large-scale scan of over 8,000 recently deployed contracts, it identified 4 previously unknown vulnerabilities, confirming its practical utility in securing the DeFi ecosystem. In summary, this thesis advances the field of smart contract security by bridging the gap be-tween empirical study and automated tool development. By characterizing defensive practices and investigating the limitations of existing security tools, this work paves the way for more effective detection methods. The proposed hybrid framework demonstrates that integrating static analysis with the semantic reasoning of LLMs can effectively identify complex semantic smart contract vulnerabilities, providing the community with insights and tools to safeguard decentralized applications.
Context. Machine learning approaches for smart contract vulnerability detection are typically evaluated on synthetic benchmarks of programmatically generated code snippets. Practitioner reports and recent independent evaluations indicate that automated tools continue to miss critical vulnerabilities in production audits, yet the contribution of benchmark selection to this gap remains under-examined.Objectives. This study investigates whether surface-level Solidity features that correlate with vulnerability severity in synthetic benchmarks retain their predictive validity on professionally audited contracts, and proposes a quantitative metric for assessing benchmark suitability for severity prediction research.Methods. Fifteen features were extracted identically from a 10,448-sample synthetic Solidity benchmark and DAppSCAN, a corpus of 1,646 findings from 1,199 audit reports authored by 29 firms. Feature-severity correlations were compared using Fisher r-to-z, Kolmogorov-Smirnov, and Levene tests. Logistic Regression, Random Forest, and Gradient Boosted classifiers were trained in three conditions: in-distribution synthetic, in-distribution real, and cross-distribution.Results. Mean absolute correlation was 0.228 on synthetic data versus 0.057 on real data, a fourfold gap (all p
In this paper, we developed anomaly detection based on machine learning-based with the automated signing of the blockchain transaction system to effectively detect the anomalies to prevent the leakage of information from the bitcoin system. Initially, the anomalies data is collected from online resources. The automated signing of the transaction system is performed using machine learning. A blockchain transaction is used for the personalised identification of anomalies transactions. It secures the transactions from fraudulent blockchain transactions. Then, the anomaly detection is done by an optimised recurrent neural network with attention mechanism (ORNN-AM). Here, the parameters are optimised using fitness of firefly and driving training-based optimisation (FFDTO). Anomaly detection with the automated signing of blockchain transactions using machine learning techniques helps to detect anomalies effectively. The performance of anomaly detection with the automated signing of the blockchain transactions system is compared to other conventional anomaly detection models.
Smart contract vulnerabilities have caused documented financial losses exceeding $6 billion across decentralized finance (DeFi) ecosystems between 2020 and 2024. Existing automated security toolsincluding Slither, Mythril, and Manticoreemploy rule-based static analysis that systematically fails to detect contextual, indirect, and semantically complex vulnerability patterns that are commonly exploited in production attacks. This paper presents CyberTrust AI, a production-deployed security analysis framework that applies large language model (LLM) inference via Anthropic's Claude Sonnet to perform contextual, semantic vulnerability analysis of Solidity smart contracts. The system identifies critical vulnerability classes including reentrancy attacks, integer overflow and underflow, unprotected self-destruct, unchecked external call return values, tx.origin authentication abuse, timestamp manipulation, and access control logic flawsgenerating severity-classified structured findings, natural language attack vector explanations, automated Solidity remediation code, and cryptographically verifiable onchain NFT trust scores. The complete system has been deployed at cybersheild-sooty.vercel.app and implements seven production capabilities: multi-mode audit analysis (security audit, threat simulation, gas optimization), realtime streaming analysis output, batch multi-file contract auditing, side-by-side contract diff comparison, conversational AI chat assistance for vulnerability Q&A, and a public trust score leaderboard. We conducted preliminary evaluation on 35 annotated Solidity contracts covering five canonical vulnerability types, demonstrating that LLM-based contextual analysis successfully detects all vulnerability instances while conventional static analysis tools miss approximately 13% of contextual casesparticularly indirect reentrancy patterns and access control logic errors requiring cross-function semantic reasoning. A full quantitative comparative evaluation is in progress.
The rapid growth of decentralized finance (DeFi) has spurred innovation but also exposed blockchain systems to severe security threats. As of November 2025, cumulative losses from blockchain security incidents have exceeded${\$}$36.89 billion. Flash loan attacks account for 135 reported cases and rank fourth among all attack methods. Existing detection approaches either analyze contract source code, which is unavailable for many deployed contracts, or use transaction pattern matching tailored to specific scenarios, and therefore generalize poorly to diverse flash loan attacks. In this paper, we presentFlashShield, a general flash loan attack detection framework based on Hypergraph Neural Networks (HGNNs). We construct comprehensive datasets containing attack and benign transactions across multiple chains, and systematically analyze flash loan attack mechanisms along four DeFi protocol layers: code implementation, business logic, economic mechanisms, and cross protocol interactions.FlashShieldrepresents each transaction as a hypergraph of transfer actions and semantic relations, and employs a hybrid architecture that integrates spectral, spatial, and original features together with both node level and graph level representations. Experiments show thatFlashShieldimproves recall by 29% over leading methods and identifies 43 previously unknown malicious or suspicious activities (18 confirmed flash loan-related exploits and 25 suspected address poisoning incidents), demonstrating its effectiveness and scalability for automated DeFi security monitoring.
Dr. Pankaj Malik, Mohit Kapoor, Akshat Gupta, Aman Singhai · 5 authors
The rapid expansion of decentralized finance (DeFi) platforms has been accompanied by a surge in rug pull scams, where malicious actors exploit liquidity pools and abandon projects, causing substantial investor losses. Existing detection approaches are largely centralized and platform-specific, limiting their effectiveness due to privacy constraints, fragmented data sources, and the dynamic behavior of blockchain ecosystems. This paper proposes a novel Federated Time-Series Learning (FTSL) framework for cross-platform rug pull detection that enables collaborative model training without sharing raw transaction data. The proposed system integrates federated learning with advanced time-series modeling to capture temporal patterns in token price volatility, liquidity changes, transaction frequency, and smart contract activities. A hybrid deep learning architecture combining Long Short-Term Memory (LSTM) networks with an attention mechanism is employed to effectively learn sequential dependencies and identify early indicators of fraudulent behavior. The federated setup ensures privacy preservation while enabling knowledge sharing across multiple decentralized platforms. Experimental results on multi-chain DeFi datasets demonstrate that the proposed FTSL model achieves 96.3% detection accuracy, outperforming traditional centralized models (91.2%) and single-platform approaches (88.7%). The model also improves precision (95.1%), recall (94.6%), and F1-score (94.8%), indicating robust and balanced performance. Furthermore, the system is capable of detecting rug pull events 6â12 hours earlier than baseline methods, providing critical early warning signals. Communication overhead is reduced by approximately 28% through optimized federated aggregation, while maintaining scalability across heterogeneous platforms. These findings highlight that Federated Time-Series Learning offers a scalable, privacy-preserving, and highly effective solution for real-time rug pull detection, contributing to enhanced security, transparency, and trust in decentralized financial ecosystems.
Decentralized finance (DeFi) vaults are smart-contract-based asset management systems that pool deposits, execute programmable strategies, and mint tokenized shares representing claims on underlying assets and strategy performance. As vault designs have evolved from early yield aggregators to modular, actively managed systems, a new control layer, curation, has emerged to select strategies, configure risk parameters, and coordinate operational execution, introducing principal-agent dynamics and new failure modes. This paper systematizes DeFi vault architectures and curator-mediated control planes through (i) a unified system model and formal definitions for share accounting, roles, and operational dependencies, and (ii) three complementary taxonomies covering vault exposures and objectives, curator governance and accountability mechanisms, and strategy execution patterns together with their failure modes. We further map a representative set of production protocols to the proposed dimensions. The frameworks in this work aim to support rigorous analysis and safer design of blockchain-based financial applications.
Lending protocols in decentralized finance have traditionally relied on over-collateralization mechanisms, where investor protection is primarily ensured through the automatic liquidation of collateral. While effective from an operational perspective, this approach limits the economic role of credit when compared with under-collateralized structures. In such settings, the prudential management of credit risk becomes a central element for protocol sustainability. This paper proposes a prudential framework for decentralized lending protocols by introducing an additional protection layer based on the distinction between tokenized loss-absorbing capital, an operational buffer, and a prudential reserve. The model defines three classes of subordinated instruments-First Loss Token, Contingent Capital Token, and Subordinated Backstop Token-arranged according to a progressive loss waterfall. The model is first applied to public data from Goldfinch and then extended to a TrueFi dataset, with the aim of assessing the ability of the policy to reduce losses borne by senior liquidity providers. The model shows a net reduction in losses. The sensitivity analysis confirms that the mechanism maintains a positive net benefit across variations in instrument costs, risk weights, and loss severity. The results suggest that an explicit prudential layer may contribute to strengthening the resilience of DAO-based credit protocols by making the prudential cost of risk-taking more transparent and by distinguishing between available liquidity, loss-absorbing capital, and protective reserves.
This paper introduces AURORA, a predictive institutional risk intelligence architecture designed to detect market manipulation and systemic fragility within decentralized finance ecosystems. The framework integrates multi-chain blockchain monitoring, behavioral graph analytics, liquidity stress modeling, and causal verification to identify engineered market abuse and cross-protocol contagion risks. AURORA further translates technical detection outputs into structured compliance classifications aligned with emerging regulatory regimes including the UK Market Abuse Regime for Cryptoassets (MARC) and the EU Markets in Crypto-Assets Regulation (MiCA). The study contributes to financial regulation, market microstructure analysis, and systemic risk modeling by proposing a unified institutional surveillance architecture for decentralized financial markets.
Effective risk management has grown more and more crucial in the complex world of international trade finance, bolstered by security, trust, and openness. By creating an integrated system that blends Hyperledger Fabric blockchain technology, Supply Chain Finance (SCF) protocols, and Generative Adversarial Networks (GANs), this study seeks to improve the intelligence and dependability of financial risk assessment. Four interrelated steps make up the suggested approach: (1) preprocessing and encoding SCF datasets; (2) creating synthetic risk data with GANs to mimic uncommon or dishonest trade behaviors; (3) using Hyperledger Fabric to execute smart contracts and log transactions decentralized; and (4) using real-time SCF compliance modeling for dynamic risk assessment. While blockchain guarantees the transparency, immutability, and auditability of financial records, GAN integration improves the prediction model by adding value to the training corpus. Comparative studies show that the suggested system considerably lowers the likelihood of data tampering and improves risk prediction accuracy by 12% when compared to traditional machine learning models. The results demonstrate that integrating generative modeling with blockchain technology can significantly improve financial risk management, transparency, and adaptability in global trade settings.
Decentralized finance liquidity providers (LPs) who use their Uniswap v3 positions as collateral on lending platforms such as Aave often face liquidations because these platforms rely on fixed Loan-to-Value (LTV) rules. These rules do not account for Impermanent Loss which can increase rapidly when asset prices move outside an LP's chosen price range. To address this, we simulated Uniswap v3 LP positions using historical ETH/USD price data and compared the standard fixed-LTV lending model with a hybrid risk-management framework that incorporates stress testing and dynamic exposure reduction. Under the conventional 65% LTV model, liquidations were frequent, with 39,649 liquidation events observed over roughly a decade of daily price data (2015-2025). The proposed framework reduced liquidations by 97.66% while maintaining healthier collateral positions, achieving this through adaptive reductions in effective leverage rather than full liquidation. These findings suggest that incorporating impermanent loss-aware risk controls into DeFi lending protocols could significantly reduce liquidations while keeping leveraged positions safer through periods of volatility. By combining the accessibility of DeFi with the risk management techniques used in traditional finance, lending platforms can become more stable and efficient, benefiting liquidity providers.
Blockchain technology has revolutionized numerous industries by providing decentralized, transparent, and immutable ledgers. However, its adoption is hindered by persistent security challenges, including arbitrage attacks, liquidity exploits, and noncompliance with antimoney laundering (AML) regulations. This paper proposes an enhanced framework to address these issues, combining dynamic pricing mechanisms, AIâbased anomaly detection, and regulatory compliance checks within a multilayered architecture. The framework is composed of five interconnected layers: the input layer for data collection and validation, the data warehouse layer for structured data classification, the processing layer for anomaly detection and pricing adjustments, and the decision layer for transaction validation, execution, and reporting. The integration of these layers ensures robust security and compliance mechanisms, reducing system vulnerabilities while optimizing efficiency. To validate the proposed framework, we conducted simulations using realâworld blockchain scenarios, including decentralized finance (DeFi) platforms and cryptocurrency exchanges. Results demonstrate significant reductions in arbitrage opportunities and liquidity risks, with improved accuracy in anomaly detection and compliance adherence. For instance, the dynamic pricing mechanism mitigated 87% of arbitrage attack attempts, while the AIâbased anomaly detection achieved an 89% accuracy rate in identifying highârisk transactions. This study provides actionable insights and a scalable solution for enhancing blockchain security and trust. Future work will focus on integrating crossâchain interoperability, realâtime threat intelligence, and privacyâpreserving techniques to further expand the frameworkâs applicability. By addressing critical vulnerabilities, this research contributes to the development of secure, transparent, and compliant blockchain ecosystems, paving the way for wider adoption across industries. Unlike previous blockchain security models, our framework introduces a realâtime, AIâenhanced risk assessment mechanism that dynamically updates transaction risk scores, mitigating financial threats in decentralized environments. This holistic approach provides a scalable, explainable, and adaptive security system that not only protects decentralized financial infrastructures but also aligns with emerging regulatory requirements, ensuring longâterm applicability.
Sai Srikanth Madugula, jose Luis de la Rosa Esteva, Daya Shankar
This paper presents an integrated framework for decentralized invoice-backed loan underwriting combining interpretable machine learning, dynamic pricing algorithms, and on-chain trust infrastructure. We develop and validate SHAP-explainable ML models for real-time default probability assessment, design a Reverse Kelly AMM smart contract for optimal risk-adjusted loan pricing, integrate ERC-725 identity and on-chain reputation scoring with an automated insurance reserve, and deploy the system on Ethereum testnet with end-to-end functional and security testing. Stress testing across simulated default and fraud scenarios demonstrates the model achieves AUC-ROC of 0.89 on validation data, maintains LP yields of 12â18% under normal conditions while containing non-performing loan ratios below 3% under adverse scenarios, and sustains reserve solvency across 95th percentile stress events. The framework addresses critical gaps in DeFi lending by bridging regulatory interpretability requirements with decentralized credit assessment, demonstrating both technical feasibility and economic viability for permissionless SME financing at scale.
Ye Tian, Liangliang Song, Yuanyuan Ma, Yanbin Wang · 7 authors
Detecting fraudulent accounts on Ethereum is critical for securing decentralized finance (DeFi) ecosystems. Although temporal transaction dynamics offer richer behavioral signals than static graphs, existing methods struggle to jointly model the continuous, directional, and community-level nature of fraudulent fund flows. Specifically, they inadequately capture two key cues: directional-temporal transaction patterns (integrating directionality and temporal dynamics) and higher-level behavioral communities formed by accounts with similar transactional and temporal traits.To address these, we propose TimeTrace, an unsupervised graph representation learning framework for fraud detection in Ethereum transaction networks. TimeTrace models temporal behaviors via a Directed Temporal Aggregation mechanism that explicitly distinguishes incoming and outgoing flows while adaptively emphasizing recent interactions to capture pattern evolution. Building on these directional representations, TimeTrace further incorporates a differentiable clustering module to identify latent behavioral communities and encode cluster-level relational semantics. To enhance representation coherence and structural consistency, we introduce Cohesive Embedding Regularization (CER)âa graph refinement objective combining a global Laplacian term for structural consistency, an intra-cluster Laplacian term for cluster compactness, and a fidelity term to retain initial clustering information. Additionally, we construct a new phishing account benchmark with up-to-date Ethereum transaction records for realistic evaluation. Extensive experiments on three datasets demonstrate that TimeTrace outperforms state-of-the-art methods in both accuracy and efficiency, achieving F1-score improvements of 3.62\%-10.83\% and inference speeds at least as fast as the quickest baseline.
DharmaCoin is an artificial intelligence-based cyber fraud detection system that is supposed to establish a safe and moral online financial system. The suggested model enables combining blockchain security, fraud detection with the use of artificial intelligence, biometric multi-factor authentication, and secure document verification to enhance the safeguarding of digital financial crimes. It has a framework built on the values of Indian philosophies based on logic, transparency, and ethical governance and in the name of responsible and trust-worthy digital finance. DharmaCoin manages to identify transaction fraud with a 97% F1-score and forged/altering documents with 99% accuracy in using state-of-the-art technologies SatyaAI, a real-time deepfake and identity-checking tool, and RishiGuard, an anomaly financial activity detecting tool, in 200 milliseconds. The system uses a Proof-of-Stake blockchain registry to ensure unalterable records of transactions, and to increase the level of transparency within financial deals. Performance checks show that the blockchain infrastructure has the capability of handling a rate of transactions of over 1000 transactions in less than five minutes as well as be able to verify transactions in less than five seconds hence scaling to high financial volume environments. A Pilot project involving a banking partner found that a reduction of up to 40 in time to process KYC manually led to a faster approach to operations and better fraud detection. In general, DharmaCoin offers a safe, transparent, sustainable digital finance system, which is a moral and tech-centered solution to the prevention of cyber fraud and promotes the trustful financial ecosystem in accordance with the global sustainability development objectives.
The rapid growth of Ethereum blockchain transactions has led to increased vulnerability to fraudulent activities. Traditional rule-based detection systems fail to capture complex and evolving fraud patterns. This paper presents an unsupervised deep learning approach using an Autoencoder to model normal transaction behavior and detect anomalies based on reconstruction error. The model is trained on normal transactions and evaluated on a dataset of 9,841 Ethereum accounts with 45 features. The system achieves a recall of 78.7%, precision of 69%, F1-score of 0.74, and ROC-AUC of 0.70. The proposed approach demonstrates the effectiveness of unsupervised learning for fraud detection without requiring labeled datasets.
With the rapid development of financial technology and the digital economy, fraud detection in financial transactions faces increasing challenges due to complex transaction networks, temporal dependencies, and nonlinear interactions.This study proposes an RL-LGNN framework that integrates long short-term memory (LSTM) networks, graph neural networks (GNN), and reinforcement learning (RL) for fraud detection in the financial transaction settlement process.LSTM is used to encode historical transactions as temporal sequences and extract time-dependent behavioural features.GNN then models inter-node transaction relationships and captures structural information from the transaction graph.On this basis, RL is introduced to dynamically optimise the detection strategy, thereby improving model adaptability and robustness.Experimental results on both public and real-world datasets show that the proposed framework outperforms conventional methods and achieves fraud detection accuracy above 90%.The proposed framework provides an effective solution for fraud detection in financial transaction settlement.
The book Solutions and Technologies for Modern Business stands as a relevant contribution to understanding the technological and strategic transformations impacting the contemporary business environment.With a broad approach, the work offers reflections on innovative solutions, technological tools, and management practices aimed at strengthening and adapting organizations in the face of constant market changes.By emphasizing the integration of theory and practice, the book contributes to the development of critical analyses regarding the use of technology in organizational processes, highlighting its importance for competitiveness, innovation, and decision-making.The work brings together diverse perspectives that enrich academic debate and encourage the development of more efficient and sustainable strategies in the business context.This book is recommended for professors, students, and professionals in the fields of administration, management, technology, and business, as well as for anyone interested in expanding their knowledge of solutions and technologies applied to the corporate environment.It is a work that fosters learning, reflection, and the improvement of organizational practices in the digital era.