Ziqiao Kong, Wanxu Xia, Chong Wang, Yi LU · 9 authors
Smart contracts govern billions of dollars in decentralized finance (DeFi), yet automated vulnerability detection remains challenging because many vulnerabilities are tightly coupled with project-specific business logic. We observe that recurring vulnerabilities across diverse DeFi business models often share the same underlying economic mechanisms, which we term DeFi semantics, and that capturing these shared abstractions can enable more systematic auditing. Building on this insight, we propose Knowdit, a knowledge-driven, agentic workflow for smart contract vulnerability detection. Knowdit first constructs an auditing knowledge graph from historical human audit reports, linking fine-grained DeFi semantics with recurring vulnerability patterns. Given a new project, a multi-agent pipeline leverages this knowledge through an iterative loop of specification generation, Proof-of-Concept (PoC) synthesis, PoC execution, and finding reflection, driven by a shared repository index. We evaluate Knowdit on 11 recent Code4rena projects with 84 ground-truth vulnerabilities. Knowdit detects all 21 high-severity and 90% of medium-severity vulnerabilities without false positives, fully covering eight projects, significantly outperforming all baselines. Applied to seven real-world projects, Knowdit further discovers 9 high- and 36 medium-severity previously unknown vulnerabilities, securing millions in liquidity and proving its outstanding performance.
Bu çalışma, Bitcoin fiyat tahmininde mevsimsel ARIMA (SARIMA) modelinin öngörü performansını, basit bir Naive kıyas modeliyle açık biçimde karşılaştırarak yeniden değerlendirmeyi amaçlamaktadır. Analiz, 12 Mart 2021 ile 12 Mart 2026 dönemini kapsayan günlük Bitcoin kapanış fiyatlarına dayanmaktadır. Seri logaritmik forma dönüştürülmüş ve durağanlık özellikleri fark alma işlemleriyle incelenmiştir. İlk aşamada mevsimsel olmayan ARIMA modelleri tahmin edilmiş, ardından mevsimsel dinamikleri içeren alternatif SARIMA modelleri değerlendirilmiştir. Model seçiminde parametre anlamlılığı ile Ljung-Box tanı istatistikleri dikkate alınmış ve mevsimsel hareketli ortalama bileşeninin kısmen anlamlı olduğu görülmüştür. Bu çerçevede SARIMA(0,1,1)(0,1,1)[30] nihai mevsimsel aday model olarak belirlenmiştir. Tahmin performansı değerlendirmesi, yalnızca model uyumuna değil, örneklem dışı tahmin doğruluğuna odaklanmaktadır. Bu amaçla SARIMA modelinin performansı, ortalama mutlak hata (MAE) ve hata kareler ortalamasının karekökü (RMSE) ölçütleri kullanılarak Naive model ile karşılaştırılmıştır. Bulgular, örneklem dışı dönemde Naive modelin SARIMA modeline göre belirgin biçimde daha düşük tahmin hataları ürettiğini göstermektedir. Naive model için MAE 0.016011 ve RMSE 0.023173 iken, SARIMA modeli için bu değerler sırasıyla 0.23172 ve 0.28931’dir. Sonuçlar, Bitcoin gibi yüksek oynaklığa sahip finansal zaman serilerinde daha karmaşık mevsimsel yapıların her zaman daha üstün tahmin performansı sağlamadığını göstermektedir.
The landscape of e-commerce has witnessed a transformative shift in consumer behavior, driven by the rise of digital technologies and online platforms. As online purchases increase at an alarming rate, fraudulent activity has become a major concern for retailers and consumers alike. The objective of this research is to investigate methods for detecting fraudulent online transactions using machine learning algorithms. This paper proposes a Hybrid Agentic AI Architecture (HSAA) for edge-enabled e-commerce that incorporates intelligent agents and cryptographic security to enable real-time, trustworthy transaction processing. The architecture uses world-model distillation to enable efficient inference on edge devices. HSAA was tested on several large data sets such as a balanced credit card fraud set containing 2,952 transactions. The system scored 96.6% in detecting fraud, indicating very low false positives and high specificity. Negotiation exercises on 400 independent interactions were successful in 59%, with an average discount of 14.2%, using 1,142 zero-knowledge proofs that were verified with 100% validity. Some of the operational performance highlights include a throughput of 585 transactions per second, an average latency of 1.56 milliseconds, and a 81.9% reduction in bandwidth through selective state transfer. The findings support the argument that HSAA is a strong, secure, and high-performance edge-based e-commerce architecture, combining accuracy, efficiency, and reliability. Within HSAA, fraud detection functions as one of the core decision agents, while negotiation and secure execution mechanisms provide the broader operational context for trustworthy edge commerce. The architecture provides a solid basis for future studies in adaptive and autonomous AI-driven commercial systems.
Decentralised Finance (DeFi) applications involve a large volume of funds and exhibit diverse user behaviours, including malicious activities such as smart contract exploits and financial scams. Existing approaches struggle to capture complex behaviours. To address this gap, we propose a general Blockchain User Behaviour Analysis (BUBA) pipeline for DeFi security. The pipeline presents an automated action formation process that takes blockchain transactions as inputs and outputs user actions. In addition, BUBA introduces a dual Graph Neural Network (GNN) model that jointly captures user action features, contract and token interactions, and heterogeneous graph structure information to produce rich behavioural embeddings, enabling effective clustering of semantically meaningful user behaviours. We evaluate the proposed pipeline on Uniswap V3, where it outperforms baseline methods in identifying and differentiating suspicious behaviours. A further case study on Sushiswap V2 demonstrates the generalisability of the pipeline across DeFi applications.
Modern financial security issues have arisen as a result of the rapid proliferation of decentralized financial systems and cryptocurrencies. One such issue is the identification of individuals who are laundering money in blockchain networks. Blockchain technology's distributed ledgers clarify matters; however, the anonymity of wallet addresses facilitates illicit financial transactions by criminals. It is crucial to have effective methods to identify these crimes, as over $82 billion in cryptocurrencies were associated with money laundering in 2025. This study demonstrates a method for detecting indications of money laundering in blockchain transaction networks through the use of machine learning. The proposed method for identifying unusual patterns in transactions involves the combination of supervised machine learning, graph-based feature extraction, and data cleansing. The system examines transaction graphs to identify unusual patterns that are associated with illicit financial activities by employing techniques such as Random Forest, Gradient Boosting, and Graph Neural Networks.
The accelerating digitization of financial services has transformed global economic ecosystems while simultaneously amplifying the scale, speed, and structural complexity of financial fraud. Real-time payments, open banking infrastructures, fintech platforms, and decentralized finance environments have expanded transactional connectivity, creating highly dynamic and interconnected risk landscapes. Conventional rule-based and standalone machine learning systems demonstrate limited effectiveness against adaptive adversaries, coordinated fraud rings, synthetic identity schemes, and cross-platform laundering networks. Advanced detection strategies require intelligent architectures capable of modeling temporal behavior, relational dependencies, and large-scale streaming data within production-grade environments. This chapter presents a comprehensive framework for next-generation financial fraud detection integrating Artificial Intelligence, Deep Learning, and Graph Analytics. The discussion synthesizes supervised, unsupervised, and semi-supervised learning approaches with sequential deep learning architectures, transformer-based models, and graph neural networks for network-aware inference. Emphasis is placed on hierarchical multi-stage detection systems, cloud-native deployment strategies, adversarial robustness, privacy-preserving computation, and real-world validation methodologies. Critical challenges such as extreme class imbalance, concept drift, scalability of graph processing, explainability under regulatory constraints, and cross-institution collaboration are systematically examined. A unified hybrid AI–graph intelligence architecture is articulated to address both transactional anomalies and coordinated fraud ecosystems. The chapter contributes a structured taxonomy of modern financial fraud, an integrated modeling perspective combining temporal and structural intelligence, and a deployment-oriented evaluation framework aligned with real-world financial operations. By bridging theoretical advancements with production-grade implementation considerations, this work establishes a rigorous foundation for scalable, interpretable, and resilient fraud detection systems within evolving digital financial infrastructures.
The rapid expansion of digital banking, mobile payments, decentralized finance, and cross-border electronic transactions has fundamentally transformed global financial ecosystems while intensifying exposure to sophisticated cyber threats, fraud networks, synthetic identity schemes, and money laundering operations. Conventional rule-based security infrastructures lack the adaptability required to counter dynamic and large-scale financial crimes. Artificial Intelligence (AI) and Machine Learning (ML) have emerged as transformative enablers of intelligent financial security, supporting real-time fraud detection, behavioral authentication, transaction risk scoring, and regulatory compliance automation. This chapter presents a comprehensive examination of advanced machine learning techniques—including deep learning, graph neural networks, anomaly detection models, and reinforcement learning—for securing digital transactions and identifying coordinated fraud rings within complex financial networks. Integration of AI with blockchain consensus mechanisms, cryptographic infrastructures, and Regulatory Technology (RegTech) platforms is analyzed to demonstrate how adaptive intelligence enhances network resilience, transparency, and operational efficiency. Emphasis is placed on explainable and fairness-aware AI frameworks to ensure ethical accountability, regulatory alignment, and bias mitigation in automated financial decision systems. Privacy-preserving approaches such as federated learning and secure multi-party computation are also explored to address data governance constraints in cross-institutional collaboration. The chapter consolidates emerging research directions, identifies persistent technical and ethical challenges, and proposes an integrated AI-driven security architecture for scalable and trustworthy digital financial ecosystems.
The rapid digitalization of financial ecosystems has transformed online payments, transaction processing, and investment management into highly interconnected, data-intensive infrastructures. This transformation has simultaneously expanded exposure to cyber fraud, money laundering, identity theft, and market volatility, necessitating intelligent and adaptive security mechanisms. Advanced artificial intelligence techniques, including machine learning, deep learning, reinforcement learning, and graph-based analytics, have emerged as critical enablers of secure payment processing, real-time fraud detection, and predictive financial forecasting. Intelligent architectures embedded within online payment systems facilitate dynamic risk scoring, anomaly detection, behavioral profiling, and automated decision-making under strict latency constraints. This chapter presents a comprehensive examination of intelligent system frameworks for digital finance, integrating scalable cloud-based deployment, blockchain-enabled transaction integrity, explainable AI for regulatory compliance, and synthetic data generation for fraud simulation. Reinforcement learning approaches for portfolio optimization and risk-aware forecasting are analyzed to highlight adaptive investment strategies in volatile markets. Emphasis is placed on addressing class imbalance, adversarial threats, model interpretability, privacy preservation, and governance challenges within automated financial infrastructures. Emerging research directions such as federated learning, decentralized finance intelligence, and AI-driven anti-money laundering systems are also discussed to outline future technological trajectories. The presented synthesis establishes a structured foundation for developing secure, transparent, and scalable intelligent financial ecosystems aligned with regulatory and operational requirements of modern digital economies.
Artificial Intelligence (AI) has become a critical driver of firm survival in the banking industry, particularly for deposit money banks (DMBs) facing increasing challenges such as economic volatility, regulatory compliance, cybersecurity threats, and rising customer expectations. This study explores the role of AI in enhancing operational efficiency, risk management, fraud detection, customer experience, and financial resilience in the banking sector. AI-powered technologies, including machine learning, predictive analytics, robotic process automation (RPA), and natural language processing (NLP), are transforming how banks analyze financial risks, detect fraudulent transactions, automate operations, and provide personalized banking services. Research findings indicate that AI adoption has led to a 35% reduction in loan defaults, a 40% improvement in operational efficiency, and a 60% decline in financial fraud cases, highlighting its transformative potential in ensuring the survival and competitiveness of DMBs. Despite these advancements, AI adoption in the banking sector is hindered by high implementation costs, cybersecurity vulnerabilities, workforce resistance, and regulatory uncertainties. Many banks, particularly in developing economies like Nigeria, struggle with legacy banking systems, lack of AI governance frameworks, and concerns over algorithmic bias in lending decisions. Additionally, AI-driven financial innovations, such as blockchain integration, decentralized finance (DeFi), and AI-powered ESG compliance solutions, are reshaping the banking industry, yet require strategic policy alignment and investment to maximize their benefits. The study identifies gaps in existing literature, including the need for empirical research on AI’s long-term impact on firm survival, its role in financial inclusion, and the ethical challenges of AI governance in banking. To bridge these gaps, future research should focus on developing AI implementation models suited to the challenges of emerging economies, exploring AI’s potential in expanding financial access to underserved populations, and strengthening AI-driven sustainability and ESG compliance frameworks in banking. As AI continues to evolve, deposit money banks must embrace a balanced approach that integrates AI innovation with regulatory oversight, cybersecurity safeguards, and workforce upskilling to ensure long-term survival and competitiveness in the digital financial landscape
Federated learning enables financial institutions to collaboratively develop credit risk models while maintaining data privacy, yet existing implementations prioritize accuracy and confidentiality over transparency and regulatory compliance requirements. Current federated approaches treat explainability as a secondary concern addressed through separate post-processing workflows, creating significant gaps in auditability and stakeholder trust that limit adoption in regulated environments. This article introduces the Explainable Update Auditing framework, which embeds transparency mechanisms directly into federated training protocols through local explanation bundles and privacy-preserving audit trails. The framework generates standardized, model-agnostic explanations that characterize how institutional updates influence global model behavior without exposing proprietary data or competitive information. Cryptographic attestation mechanisms verify compliance with fairness, stability, and governance constraints throughout training processes using zero-knowledge proof systems that maintain institutional confidentiality while providing mathematical assurance of appropriate collaborative behavior. The dual-layer trust mechanism addresses distinct information needs across multiple stakeholder groups, including participating institutions, regulatory authorities, internal governance bodies, and affected borrowers. Implementation considerations reveal computational overhead challenges, privacy-utility trade-offs, and cryptographic protocol efficiency requirements that must be addressed for practical deployment. The framework transforms federated learning from an opaque collaboration protocol into a transparent, auditable ecosystem that satisfies regulatory requirements while preserving privacy guarantees essential for cross-institutional partnerships in credit risk modeling applications.
Smart contracts (SCs) implemented on blockchain represent a breakthrough in decentralized applications, enabling a range of functions such as managing supply chains and handling elections. As the adoption of SCs increases, the need to detect flaws and vulnerabilities in their execution grows. To address this challenge, we present Branch Reinforcement Learning Fuzzer (BRLF), a deep reinforcement learning-based solution for the detection of vulnerabilities in SCs. The novelty of our method is threefold: first, our deep model uses text-based embeddings of conditional branches to enhance its adaptability and flexibility. Secondly, we propose a reward function that considers multiple aspects of fuzzing, such as opcode analysis and gas usage. Finally, we incorporate evolutionary algorithms into our approach, which significantly bolsters its ability to produce varied outputs. Extensive evaluation on three datasets of Ethereum-based SCs shows that BRLF outperforms state-of-the-art methods, detecting more vulnerabilities and achieving higher code coverage than existing solutions. Our code and data are available at: https://zenodo.org/records/15022152
Fraud detection within the Ethereum network remains a major research challenge due to the strong statistical resemblance between legitimate and fraudulent transaction patterns, severe class imbalance, and the multiscale complexity of temporal-interaction dependencies. Proposing and evaluating a multi-branch attention-based system with automated architecture optimization, which can detect fraudulent Ethereum accounts with high accuracy, is the aim of this study. The experimental evaluation was performed on a dataset with 9,841 samples and 17 extracted features. The proposed system employed a hybrid multi-branch architecture combining CNN, Bi-LSTM, and LSTM with a Gated Fusion mechanism along with multiscale attention layers. The Artificial Bee Colony (ABC) algorithm was applied to automatically optimize sixteen key structural and learning parameters. The results indicate that the proposed system achieved an accuracy of 99.84 %, F1 score of 98.94 %, sensitivity of 98.76 percent, and precision of 99.12 percent. These results notably outperform eight algorithms, such as Random Forest, XGBoost, LGBM, and GADL. According to the confusion matrix analysis, there is a reduction in false negatives, confirming that the system produced only five such cases in the sample set. These findings show that the proposed system is an effective and efficient approach for detecting fraud in blockchain systems and enables deployment in exchanges, DeFi platforms, and regulatory institutions.
Amid the institutionalization wave of Decentralized Finance (DeFi), U.S. institutional Liquidity Providers (LPs) have emerged as the core incremental capital for leading Decentralized Exchanges (DEXs). However, the adaptation gap between Uniswap V4's concentrated liquidity mechanism and institutional risk preferences, as well as regulatory compliance requirements, has hindered their market entry. This study focuses on the integration of "technical characteristics - institutional constraints - precise pricing" and constructs a machine learning pricing model optimized across three dimensions: return, risk, and compliance. By integrating Uniswap V4 on-chain data, institutional risk preference data, and market data, a Stacking ensemble architecture combining LightGBM and CNN-LSTM is designed, incorporating 22 core features to achieve precise pricing. Empirical results show that the model's Mean Absolute Error (MAE) on the test set was reduced by 37% compared to the benchmark, and the Root Mean Square Error (RMSE) is reduced by 42%. The Sharpe ratio reaches 1.87 (an increase of 62% compared to the benchmark), with a volatility of 15.3% and a compliance adaptability score of 91. In the case study, a $150 million liquidity supply achieved a 19.7% annualized return and an 8.3% maximum drawdown, successfully passing SEC compliance review. This research fills the gap in institution-oriented pricing models for V4, improves the institutional extension of Automated Market Maker (AMM) pricing theory, and provides a risk-controllable and compliance-adaptable pricing tool for U.S. institutions participating in DeFi, promoting the transformation of the DeFi ecosystem towards standardization and institutionalization. By aligning the V4 Hook mechanism with U.S. regulatory frameworks, this research provides a scalable technical standard for institutional DeFi adoption, reinforcing the competitive advantage of the U.S. Web3 financial ecosystem.
Aijie Shu, Wenbin Wu, Gbenga Ibikunle, Fengxiang He
Credit exposure in Decentralized Finance (DeFi) is often implicit and token-mediated, creating a dense web of inter-protocol dependencies. Thus, a shock to one token may result in significant and uncontrolled contagion effects. As the DeFi ecosystem becomes increasingly linked with traditional financial infrastructure through instruments, such as stablecoins, the risk posed by this dynamic demands more powerful quantification tools. We introduce DeXposure-FM, the first time-series, graph foundation model for measuring and forecasting inter-protocol credit exposure on DeFi networks, to the best of our knowledge. Employing a graph-tabular encoder, with pre-trained weight initialization, and multiple task-specific heads, DeXposure-FM is trained on the DeXposure dataset that has 43.7 million data entries, across 4,300+ protocols on 602 blockchains, covering 24,300+ unique tokens. The training is operationalized for credit-exposure forecasting, predicting the joint dynamics of (1) protocol-level flows, and (2) the topology and weights of credit-exposure links. The DeXposure-FM is empirically validated on two machine learning benchmarks; it consistently outperforms the state-of-the-art approaches, including a graph foundation model and temporal graph neural networks. DeXposure-FM further produces financial economics tools that support macroprudential monitoring and scenario-based DeFi stress testing, by enabling protocol-level systemic-importance scores, sector-level spillover and concentration measures via a forecast-then-measure pipeline. Empirical verification fully supports our financial economics tools. The model and code have been publicly available. Model: https://huggingface.co/EVIEHub/DeXposure-FM. Code: https://github.com/EVIEHub/DeXposure-FM.
The financial sector is experiencing rapid transformation due to emerging technologies. Blockchain offers a decentralized, transparent, and immutable framework for secure transactions, while Artificial Intelligence (AI) enables advanced data analytics, predictive modeling, and intelligent automation. When combined, these technologies create a powerful synergy that is reshaping finance by enhancing fraud detection, improving credit evaluation, optimizing decentralized finance (DeFi) platforms, and automating compliance processes. This paper explores the combined benefits of blockchain and AI, highlighting practical applications such as AI-enabled fraud detection within blockchain networks, adaptive smart contracts, and blockchain-secured digital identity verification. It also addresses challenges in merging these technologies, including scalability limitations, regulatory ambiguity, interoperability concerns, and ethical considerations. The study underscores the potential future of autonomous financial systems, decentralized autonomous organizations (DAOs), and AI-driven sustainable finance solutions. Ultimately, the integration of blockchain and AI is seen as a transformative force capable of significantly improving transparency, efficiency, and inclusiveness in global financial systems.
Walter Kurz, Michel Malara, Wojtek Stricker, Eva Albrecht
The objective of this study is to define a compliance-first, conceptually generalisable architecture for a multi-agent artificial intelligence platform integrated with distributed ledger technology, designed to be domain-, deployment-, and vendor-agnostic. It addresses a persistent shortcoming in current AI deployments, where compliance is often treated as a secondary concern, applied retroactively through prompt engineering rather than embedded within the foundational design. The proposed model encodes regulatory, governance, and ESG requirements into an objective-under-constraints framework, ensuring that all specialised agents operate within legally admissible and verifiably auditable parameters prior to any domain-specific implementation. A DAG-based verification layer is incorporated to enable scalable, low-latency, and cost-efficient operation while preserving evidentiary integrity. The analysis evaluates the feasibility of this conceptual model to support sustainable, rapid-deployment vertical applications without inducing vendor lock-in, preserving operational neutrality, and ensuring environmental accountability. The findings suggest that integrating compliance, ESG metrics, and agent specialisation at the architectural level provides a transferable foundation for cross-domain AI–DLT infrastructures.
Cross-chain bridges constitute the single largest vector of systemic risk in Decentralized Finance (DeFi), accounting for over \$2.8 billion in losses since 2021. The fundamental vulnerability lies in the binary nature of existing bridge security models: a bridge is either fully operational or catastrophically compromised, with no intermediate state to contain partial failures. We present ASAS-BridgeAMM, a bridge-coupled automated market maker that introduces Contained Degradation: a formally specified operational state where the system gracefully degrades functionality in response to adversarial signals. By treating cross-chain message latency as a quantifiable execution risk, the protocol dynamically adjusts collateral haircuts, slippage bounds, and withdrawal limits. Across 18 months of historical replay on Ethereum and two auxiliary chains, ASAS-BridgeAMM reduces worst-case bridge-induced insolvency by 73% relative to baseline mint-and-burn architectures, while preserving 104.5% of transaction volume during stress periods. In rigorous adversarial simulations involving delayed finality, oracle manipulation, and liquidity griefing, the protocol maintains solvency with probability $>0.9999$ and bounds per-epoch bad debt to $<0.2%$ of total collateral. We provide a reference implementation in Solidity and formally prove safety (bounded debt), liveness (settlement completion), and manipulation resistance under a Byzantine relayer model.
Decentralized Finance (DeFi) has become a major component of digital asset markets, yet accurately valuing protocol performance remains difficult due to high volatility, nonlinear pricing dynamics, and persistent disclosure gaps that amplify valuation risk. This study develops an Optuna-tuned Super Learner stacked ensemble to improve risk-aware DeFi valuation, combining Extremely Randomized Trees (ETs), Support Vector Regression (SVR), and Categorical Boosting (CAT) as heterogeneous base learners, with a K-Nearest Neighbors (KNNs) meta-learner integrating their forecasts. Using an expanding-window panel time-series cross-validation design, the framework achieves significantly higher predictive accuracy than individual models, benchmark ensembles, and econometric baselines, obtaining RMSE = 0.085, MAE = 0.065, and R2 = 0.97—representing a 25–36% reduction in valuation error. Wilcoxon tests confirm that these gains are statistically significant (p < 0.01). SHAP-based interpretability analysis identifies Gross Merchandise Volume (GMV) as the primary valuation determinant, followed by Total Value Locked (TVL) and key protocol design features such as Decentralized Exchange (DEX) classification, while revenue variables and inflation contribute secondary effects. The findings demonstrate how explainable ensemble learning can strengthen valuation accuracy, reduce information-driven uncertainty, and support risk-informed decision-making for investors, analysts, developers, and policymakers operating within rapidly evolving blockchain-based digital asset environments.
Anti-money laundering (AML) remains a critical challenge in cryptocurrency ecosystems, where blockchain’s transparency paradoxically coexists with pseudonymity. Traditional methods often fall short in modeling the temporal and structural complexity of transaction networks. This paper introduces ChronoWave-GNN, a graph neural framework designed from the theoretical perspective of time-frequency representation learning. By combining wavelet-based frequency decomposition with temporal encoding, our model captures nonstationary and multi-scale patterns inherent in illicit financial activity. This dual-domain perspective enhances the expressive capacity of graph representations without relying on modular patching. We validate our approach on the Elliptic dataset, where ChronoWave-GNN achieves a test accuracy of 0.9802 and F1-score of 0.9799, surpassing prior state-of-the-art results. These findings suggest that unifying temporal dynamics and spectral compression offers a principled and effective pathway for robust AML in decentralized financial systems.
Ahmed Abbas Jasim Al‐Hchaimi, M. A. Khalifa, Walid El‐Shafai
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, real‐world 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 models‐XGBoost, LightGBM, and Decision Tree‐on the Ethereum‐based fraud detection data, in which 58% of transactions are identified as fraud. The methodology combines vast feature engineering, k‐fold cross‐validation, 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, F1‐Score, 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 precision‐recall balance, and Decision Trees demonstrate significant gains after resampling, despite their simplicity. SHAP analysis reveals that log‐transformed transaction amount, merchant‐based 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 cross‐chain 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