Active asset managers are increasingly including cryptocurrencies in their alternative asset allocations, highlighting their speculative and volatile nature. The aim of this research is to examine trends in the returns and volatility of cryptocurrencies while accounting for the depegging of stablecoins driven by speculative trading macroeconomic shocks, and technological shifts. It builds a sample, by market capitalisation, using data from the daily closing prices of Bitcoin (BTC), Ethereum (ETH), Binance (BNB), and Ripple (XRP), two fiat-backed stablecoins (USDT and USDC) and a cryptocurrency-collateralised stablecoin (DAI). As a first step, Granger causality tests were applied to examine the influence of stablecoin depegging events on crypto returns during financial market stress. The results indicate that DAI exhibits the most consistent Granger-causal relationship with cryptocurrency returns; whereas, the predictive power of USDT and USDC depegging events varies across assets. The analysis was extended by modelling volatility using an EGARCH-X model to study whether depegs also affect crypto during periods of market stress. In this case, the evidence for statistically significant effects is limited. Nevertheless, in the instances where significance is detected, the results are consistently linked to USDC.
This study aims to analyze the volatility dynamics and spillover phenomena among major crypto assets (Bitcoin, Solana, and Ethereum) and their relationship with the Jakarta Composite Index (JCI), a proxy for the Indonesian capital market. In the era of digital financial integration, the link between speculative crypto asset markets and conventional stock markets is a crucial issue for financial system stability. This study uses daily price time series data for the period 2020-2025. The analysis was conducted using the Generalized Autoregressive Conditional Heteroskedasticity (GARCH) model and the Diebold-Yilmaz spillover index approach to measure the magnitude of shock transmission between markets. The results indicate significant volatility transmission among the three crypto assets, with Bitcoin remaining the primary source of volatility. Furthermore, this study finds an increasing dynamic correlation between the global crypto market and the Indonesian capital market during periods of economic uncertainty. These findings have important implications for investors in portfolio diversification strategies and for Indonesian regulators in monitoring systemic risks originating from digital assets.
Keir Finlow-Bates, Markus Jakobsson, Hossein Siadati
The transition to post-quantum cryptography in blockchain systems such as Bitcoin and Ethereum is often framed as a purely cryptographic problem. In practice, it also presents significant economic and infrastructural challenges: in globally replicated networks, increases in transaction size and verification cost are multiplied across all participating nodes. Existing post-quantum signature schemes, including lattice-based constructions such as CRYSTALS-Dilithium and stateless hash-based schemes such as SPHINCS+, introduce substantial increases in signature size. At blockchain scale, these increases translate into higher storage, bandwidth, and validation requirements, potentially requiring multiple generations of hardware improvement to become operationally routine. Historical experience suggests that even moderate increases in data footprint can be contentious, as illustrated by the Bitcoin block size debates (2015--2017). We propose a hash-based commit--reveal construction that replaces a single signature-bearing transaction with two lightweight transactions, each containing a fixed-size (32-byte) hash output derived from well-established primitives such as SHA-256, BLAKE, or Keccak. This approach achieves post-quantum security under standard hash assumptions while increasing the effective transaction footprint by only approximately 1.5$\times$ to 2$\times$ per authorization event. These results indicate that practical post-quantum migration may benefit from rethinking transaction semantics rather than directly adopting larger signature schemes, and that viable designs for decentralized systems must account for system-wide cost amplification.
Este artigo apresenta uma análise comparativa de desempenho entre rollups otimistas e execução nativa em Ethereum Virtual Machine (EVM). O estudo investiga as diferenças em termos de custo de gás, avaliando o impacto das soluções de Layer 2 na escalabilidade da blockchain Ethereum. Os resultados experimentais fornecem insights sobre os trade-offs entre execução on-chain tradicional e rollups otimistas, contribuindo para a compreensão das estratégias de escalabilidade em ambientes blockchain.
Distributed ledger technology (DLT) has emerged as a transformative force in decentralized data management across e-transactions, with significant applications in the banking, finance, supply chain, and trade sectors. Recognizing its potential, governments, including Estonia and India, have implemented DLT-based e-services to enhance transparency and privacy in public administration. With numerous platforms arising/available in the DLT segment, such as Hyperledger, Ethereum, Corda, Ripple, Stellar, Dragonchain, IOTA, and Hedera, understanding interoperability mechanisms across heterogeneous platforms has become critical. This comprehensive research provides a systematic analysis of distributed ledger technology fundamentals, consensus mechanisms, smart contracts, and their applications in e-governance services. The study examines leading DLT platforms and their core features, with a specific focus on interoperability capabilities essential for seamless cross-platform integration. Through analysis of existing interoperability solutions, including trade finance platforms, central bank digital currency initiatives, and e-governance implementations, this work identifies critical challenges and evaluation criteria for DLT adoption. The research addresses three primary research questions: (1) what capabilities does DLT provide for implementing effective e-governance strategies? (2) How does interoperability influence the delivery and effectiveness of various e-governance services? (3) What is the current impact and growth trajectory of existing e-governance services providing interoperability capabilities? The primary contributions include systematic exploration of interoperability mechanisms in various DLT platforms, documentation of existing implementations across multiple countries, including Estonia, the European Union, Dubai, and India, identification of technical challenges and security considerations, and development of a future roadmap for DLT-influenced e-governance systems. The research demonstrates that effective interoperability, combined with emerging technologies such as artificial intelligence and quantum-resistant cryptography, can enable citizen-centric, transparent, and secure governance systems while maintaining regulatory compliance and data privacy.
Fraud detection on Ethereum is challenging because of the anonymity, speed and graph structure of blockchain transactions. While prior research has proven the effectiveness of using machine learning classifiers, Graph Neural Networks (GNNs) and behavioural heuristics to detect fraudulent transactions, most systems are offline and fail to consider real-world deployment challenges for real-time blockchain analytics. Here, we present DeTrust ETH, an operational fraud intelligence system for real-time tracking of Ethereum transactions on the Sepolia testnet. DeTrust ETH aims for the integration of five operational considerations: (1) real-time blockchain ingestion with Web3.py, (2) explainable machine learning with XGBoost and SHAP, (3) light-weight graph-based transaction tracing and risk propagation, (4) temporal trust decay and behavioural anomaly detection, and (5) tamper-resistant on-chain persistence of trust scores using Solidity smart contracts. The system maintains an in-memory directed transaction graph for real-time edge insertion and updating, circular-flow tracing, funding pattern tracing and fast path tracing, avoiding the retraining overhead of Graph Neural Networks (GNNs). Our experimental results demonstrate the median graph-query time is less than 20 ms, the system can handle 222.82 requests per second with a concurrent load, and 93.44% fraud recall with a recall-favouring threshold. Unlike prior research that mostly focuses on accuracy on historical data, DeTrust ETH focuses on real-time deployment. The novelty of this work lies in the design of a real-time, low-latency fraud intelligence architecture that integrates explainable machine learning, temporal trust modeling, and lightweight graph analytics under streaming blockchain constraints.
\abstract{\textbf{Purpose:} This study addresses the lack of trust in ethical product labels by designing a blockchain platform grounded in the TAFES principles (Transparency, Accountability, Fairness, Ethics, Safety). It aims to bridge the gap between blockchain's theoretical transparency and a responsible, real-world implementation for certification ecosystems. \textbf{Design/Methodology/Approach:} Using Action Design Research, we developed a proof-of-concept platform for label authentication. A hybrid architecture records critical events on an Ethereum Layer-2 network for security, while supporting evidence is stored off-chain via IPFS and linked via content identifiers. The solution was validated through a coffee supply chain scenario. \textbf{Findings:} The proof of concept demonstrates how a TAFES-aligned blockchain platform can support verification of label claims without requiring trust in a single intermediary by creating tamper-evident provenance records and auditable certification evidence across multiple stakeholders. The design supports low-cost, near-real-time anchoring of supply chain events while mitigating adoption barriers related to scalability, privacy, and operational viability. \textbf{Originality/Value:} This research contributes an integrated ethical and technical blueprint for trustworthy label authentication systems by translating TAFES into implementable design requirements and evaluation checks, and validating them through an ADR driven proof of concept. It advances prior work by moving from the question of whether blockchain can help to the question of how it should be implemented responsibly in multi stakeholder certification ecosystems.}
Ulugmurodov Farkhod Fakhriddinovich, Hasanov Anvar Erkinovich, Abduvakhobov Feruzbek Abdurakhmonovich
This article comprehensively analyzes the formation, stages of development and the impact of cryptocurrencies on the modern economy. In particular, the transformation processes that have occurred in the financial system with the emergence of digital assets such as Bitcoin and Ethereum are studied. The study highlights the role of blockchain technology in transparency, security and reducing transaction costs. It also assesses the role of cryptocurrencies as an investment tool, their impact on monetary policy, and their impact on stability and risk factors in global financial markets. The article also examines the mechanisms for regulating cryptocurrencies based on the experience of different countries, and substantiates their positive and negative effects on economic development. The results of the study serve to draw scientific conclusions on the effective use of cryptocurrencies in the digital economy.
This study investigates the rapid centralization of the Ethereum builder market under the Proposer-Builder Separation (PBS) architecture. We argue that existing research, by focusing predominantly on influential order flows, lacks a comprehensive evaluation of order flow behavioral patterns and economic purposes. To address this gap, we analyze Ethereum transactions from September 2023 to August 2025 to characterize Exclusive Order Flows (EOFs) and non-atomic Maximal Extractable Value (MEV) -- the missing components corresponding to these behavioral and economic dimensions, respectively. We introduce a novel exclusivity metric based on Kullback-Leibler divergence and employ supervised learning to identify 75 EOFs and 322 non-atomic MEV flows, which account for 71\% and 23\% of trading-related builder revenue. A longitudinal analysis of builder strategies across these dimensions delineates the market's evolution into four distinct eras, revealing that while EOFs were instrumental in establishing early dominance, incumbents have since decoupled market share from immediate EOF dependency by leveraging entrenched network effects. Ultimately, we conclude that builder centralization is an emergent property of the PBS framework itself, as the architecture systematically violates the fundamental prerequisites of a competitive market.
Karolina Gorna, Nicolas Iooss, Yannick Seurin, Rida Khatoun · 5 authors
Zorya is a concolic execution framework that lifts compiled binaries to Ghidra's P-Code intermediate representation and uses the Z3 SMT solver to detect vulnerabilities by reasoning over both concrete and symbolic values. Previous versions supported only single-threaded TinyGo binaries. In this paper, we extend Zorya to multi-threaded binaries produced by Go's standard gc compiler. This is achieved by restoring OS thread states from gdb dumps, neutralizing runtime preemption, and introducing overlay path analysis with copy-on-write semantics to detect silent vulnerabilities on untaken branches. We rigorously assess Zorya on 11 real-world vulnerabilities from production Go projects such as Kubernetes, Go-Ethereum, and CoreDNS. Our evaluation shows that Zorya detects seven bugs at the binary level, including a silent integer overflow detects no other evaluated tool finds without a manually written oracle.
As the Agentic Economy expands, autonomous AI agents—ranging from algorithmic trading bots on decentralized finance (DeFi) platforms to decentralized physical infrastructure (DePIN) orchestrators—operate with increasing autonomy. However, the absence of a standardized, cross-protocol identity and behavioral reputation layer exposes the ecosystem to coordinated agentic attacks. This paper presents Sigui, alongside the proposed Ethereum standard ERC-8259, which introduces a composable framework for Decentralized Identifiers (DIDs), dynamic reputation scoring, and trustless threat intelligence sharing specifically designed for AI agents operating on EVM-compatible networks. By decoupling identity verification, reputation mutation, and threat pattern hashing, Sigui enables smart contracts to perform zero-latency, on-chain risk assessments of agent transactions. We detail the architecture of the IAgentIdentity, IAgentReputation, and IThreatRegistry interfaces, propose a deterministic cryptographic hashing standard for multi-layered threat patterns, and discuss the economic security and sybil resistance of the protocol. The reference implementation, deployed on the Ethereum Sepolia testnet, demonstrates the feasibility of real-time A2A (Agent-to-Agent) security protocols.
Sinchana Shetty, Tejaswini M R, Kiran Samantha D S, Vijaylaxmi H Manjunatha
Existing electronic voting platforms are persistently centralized repositories, introducing fundamental security challenged by vote manipulation, result falsification, limited weaknesses [1]. Blockchain technology has emerged as a compelling alternative, owing to its cryptographic permanence, data management. This paper proposes and evaluates a fully integrated blockchain-based electoral system built on the Ethereum network, leveraging Solidity smart contracts to address these systemic shortcomings. The proposed architecture adopts a decentralized three-tier design incorporating Web3.js communication bridges and cryptographic validation mechanisms that collectively guarantee immutability, transparency, and end-to-end verifiability throughout all electoral phases. The system incorporates hierarchical role-based access controls, real-time vote tallying, and comprehensive audit trail functionality, while preserving voter anonymity through pseudonymous addressing. Experimental results demonstrate transaction confirmation within 15–20 seconds, with a mean gas consumption of 0.0023 ETH per vote, confirming practical feasibility for medium-scale deployments. A comparative evaluation against conventional centralized e-voting solutions highlights measurable security full-stack Ethereum-based voting platform comprising Solidity improvements and the elimination of single points of failure, balanced against acceptable computational overhead.
Strategic competitions in the real world, from wars to geopolitical rivalries, often involve coalitions competing against rival groups. These contests are not simple interactions between unified entities, but multilayered processes in which coalitions face external competition while dealing with internal conflicts over resources and strategy. Existing game-theoretic models typically treat inter-coalition rivalry and intra-coalition competition separately. This paper introduces the Compound Coalition-Attrition Game (CCAG), a unified framework that integrates a war of attrition between coalitions with a simultaneous war of attrition within each coalition. In this model, the endurance of a coalition in external competition is determined by the strategic choices of its members, who compete internally for shares of the outcome. We prove the nonexistence of pure-strategy equilibria and characterize the unique mixed-strategy Nash equilibrium. The analysis reveals feedback effects: external competition intensifies internal conflict, while internal discord weakens external performance. A case study compares traditional commodity markets, including gold, copper, and silver, with cryptocurrency markets, including Bitcoin, Ethereum, and Solana, using data from 2018 to 2023 in a simulation framework. The results demonstrate applicability in industrial strategy, corporate decision-making, and geopolitical competition. The CCAG framework provides a tool for analysing complex strategic environments.
Penelitian ini bertujuan mengimplementasikan smart contract Ethereum dengan pendekatan hybrid untuk memperkuat verifikasi dokumen dan transparansi pada sistem crowdfunding beasiswa. Permasalahan utama yang diangkat adalah rendahnya kepercayaan publik terhadap platform donasi pendidikan ketika dokumen persyaratan, status verifikasi, dan realisasi penggunaan dana hanya dikelola melalui basis data terpusat. Metode yang digunakan adalah penelitian pengembangan perangkat lunak dengan model prototype yang mencakup komunikasi kebutuhan, perencanaan cepat, pemodelan desain, konstruksi prototipe, serta penyerahan dan evaluasi umpan balik. Sistem dibangun menggunakan Next.js, SQLite, Prisma ORM, Solidity, Ethers.js, MetaMask, dan jaringan Ethereum Sepolia Testnet. Hasil penyusunan sistem menunjukkan bahwa arsitektur hybrid mampu memisahkan penyimpanan dokumen fisik secara off-chain dari pencatatan bukti integritas secara on-chain. Smart contract ScholarshipRegistry dirancang untuk mencatat hash dokumen, status verifikasi, alamat wallet verifikator, timestamp, dan log nominal donasi tanpa menggunakan mata uang kripto sebagai alat pembayaran. Fitur audit publik memungkinkan donatur dan masyarakat mencocokkan hash dokumen, memantau bukti pencairan dana, serta melaporkan indikasi kejanggalan. Secara kritis, blockchain meningkatkan integritas rekam jejak, tetapi tidak otomatis menjamin kebenaran substantif isi dokumen; karena itu validasi administratif, kontrol akses, dan mekanisme pelaporan publik tetap diperlukan. Penelitian ini berkontribusi pada model crowdfunding beasiswa yang lebih transparan, efisien, dan dapat diaudit. This study aims to implement an Ethereum smart contract using a hybrid approach to strengthen document verification and transparency in a scholarship crowdfunding system. The main problem addressed is the limited public trust in digital education donation platforms when eligibility documents, verification status, and fund realization records are controlled only through a centralized database. The study applied a software development method based on the prototype model, consisting of communication, quick planning, quick design modeling, prototype construction, and delivery with feedback evaluation. The prototype was developed using Next.js, SQLite, Prisma ORM, Solidity, Ethers.js, MetaMask, and the Ethereum Sepolia Testnet. The resulting design demonstrates that the hybrid architecture can separate physical document storage in an off-chain layer from integrity proof recording in an on-chain layer. The ScholarshipRegistry smart contract records document hashes, verification status, verifier wallet addresses, timestamps, and donation amount logs without using cryptocurrency as the payment instrument. The public audit feature enables donors and the public to compare document hashes, monitor disbursement evidence, and submit reports on suspected irregularities. Critically, blockchain improves the integrity of audit trails, but it does not automatically verify the substantive truth of uploaded documents; therefore, administrative validation, role-based access control, and participatory reporting remain necessary. This study contributes a transparent, cost-efficient, and auditable model for scholarship crowdfunding systems.
Cryptocurrency market infrastructure—public blockchains and cross-chain bridges supporting tens of billions in liquidity—is monitored as a systemic-risk surface by the Financial Stability Board and equivalent bodies, with defensive posture calibrated against human-level adversaries. Anthropic’s April 2026 release of Claude Mythos Preview has prompted institutional response across financial regulation but no blockchain-specific analytical framework. This paper develops one by defining Mythos-class as a vendor-neutral capability profile: a set of frontier autonomous offensive capabilities specified independently of any single model or vendor (defined by five constituent capability primitives). The central analytical claim is friction inversion: the patch primitives, segmentation, vendor-coordinated disclosure, and credential rotation that constrain Mythos-class capability in conventional IT environments are structurally absent on-chain. This makes blockchain exposure positioned differently in kind, not degree, from enterprise IT. The paper instantiates this finding against Bitcoin and Ethereum/L2 architectures through analysis of four major bridge exploits totaling over $1.74 billion in losses. Vendor-neutral defensive and governance frameworks defined against the capability profile rather than any specific model release are the correct unit of analysis. On this basis the paper offers general recommendations for protocol governance, audit and verification cadence, and regulatory posture, developed as an analytical framework rather than as empirically validated risk estimates.
У статті досліджено економічний потенціал блокчейн-технологій як інструменту протидії глобальним змінам клімату. Проаналізовано реальний екологічний вплив криптовалют, зокрема порівняно енергоспоживання мереж Bitcoin та Ethereum після переходу на Proof-of-Stake. Розглянуто механізми токенізації вуглецевих кредитів, роль децентралізованих фінансів (DeFi) та децентралізованих автономних організацій (DAO) у кліматичному фінансуванні. Висвітлено практичні кейси застосування блокчейну в секторі відновлюваної енергетики та ризики грінвошингу. Окремо проаналізовано внесок вітчизняних науковців у дослідження впливу блокчейну на екологічну стійкість та формування «зеленої» цифрової економіки в Україні. Визначено перспективи інтеграції штучного інтелекту та Web3-технологій у кліматичні ініціативи до 2030 року.
Prof. Narde S. A., Yadav N.S., Ghodake D.T., Patil S.J., Salunkhe S. S.
Abstract In recent years, the rapid growth of digital technologies in education has increased the importance of academic certificates for employment, higher studies, and professional validation. However, the issue of fake and forged certificates has become a serious challenge for institutions and organizations. Traditional certificate verification systems are manual, time-consuming, and often lack transparency and security. These systems are also vulnerable to data manipulation and unauthorized access due to centralized storage. To address these challenges, this paper proposes a blockchain-based academic certificate validation system. The system uses blockchain technology to securely store certificate data in the form of cryptographic hash values generated using the SHA-256 algorithm. Since blockchain is decentralized and immutable, once data is stored, it cannot be altered or deleted. The system allows administrators to upload student data and issue results, which are then stored on the blockchain. Each certificate is associated with a unique verification ID and can be validated using QR codes or direct input. The proposed system improves efficiency, enhances data security, and reduces the risk of fraud. The results demonstrate that the system is faster, more reliable, and more secure than traditional methods.
This replication package contains the data and scripts used in this empirical study, including the LLM-based semantic validation pipeline, the observed practice extraction process, and all figures from the research questions (RQ1–RQ5). ERC_Observed_Practicess.xlsx: workbook of observed practices ERC_Observed_Practices_Process_Review.xlsx: Phases to generate the workbook of observed practices Other supplementary materials: Essential data files (data/) results_semantic_validation.json: 11,559 issues classified by LLM (substantive, category, justification) sample_manual_review_updated.csv: ~400 manually reviewed entries for LLM quality validation eips_labels.csv / ercs_labels.csv: PR metadata from Ethereum repositories for status evolution analysis (RQ5) Scripts (scripts/) 01: scrapes the official ERC list from ethereum.org 02 : filters the dataset for ERC mentions via regex 03: classifies issues via Gemini (substantive + category) 03: removes duplicates from the validation JSON 03: merges LLM results with issue metadata 04: extracts observed practices per ERC via Gemini, cross-referenced with official specs 05: fetches GitHub labels and generates ERC status evolution figure (RQ5) 06: generates all quantitative figures (RQ1–RQ4)
While public blockchains provide transparent and auditable transaction histories, they inherently compromise user privacy. Existing privacy-enhancing protocols, such as those deployed on Ethereum, typically rely on succinct zero-knowledge proofs (zk-SNARKs) to obscure the transaction graph. However, implementing comparable cryptographic guarantees on high-throughput blockchains like Algorand is challenging due to strict per-call execution budgets and the state contention introduced by global Merkle accumulators. This paper presents Obscura, a decentralized, non-custodial privacy protocol tailored for constrained smart contract environments. Obscura achieves transaction anonymity using Linkable Spontaneous Anonymous Group (LSAG) signatures over the BN254 elliptic curve, verified entirely on-chain. To overcome limitations of the Algorand Virtual Machine (AVM), we introduce a novel state model that leverages Algorand's Box Storage for $O(1)$ commitment membership checks, eliminating the need for global Merkle accumulators, and a dynamic opcode-budget expansion mechanism via pooled inner application calls. Our implementation demonstrates that signer-ambiguous privacy is practical and efficient on Algorand without relying on trusted setups or succinct proofs. Obscura provides a robust privacy layer for transparent ledgers, bridging the gap between high-throughput blockchain architectures and the dual requirements of cryptographic privacy and selective auditability.
This paper presents Polyquity, a Web2.5 platform enabling decentralized Initial Public Offering (IPO) fundraising through a hybrid data architecture. The platform leverages the Avalanche C-Chain for high-speed settlement, while utilizing a custom WebSocket indexer and PostgreSQL database to bridge the gap between blockchain security and institutional-grade user interfaces. By implementing a strict Role-Based Access Control (RBAC) model alongside modular architecture for auction mechanisms, fund escrow, and secondary market functions, Polyquity demonstrates how decentralized capital formation can achieve web2-equivalent performance while preserving core web3 security. The system utilizes the blockchain as the ultimate source of truth for state and funds, while the relational database serves as the source of speed for client-side rendering. Polyquity achieves sub-2-second transaction finality with 50% lower costs than Ethereum, supporting 10,000+ concurrent participants. This work establishes practical mechanisms for bridging traditional finance and decentralized ecosystems through a highly scalable, hybrid full-stack design.
Blockchain-based financial systems process billions in transactions but remain vulnerable to sophisticated fraud schemes. Current detection approaches analyze completed transactions, preventing neither fund loss nor protocol exploitation. We address this through an oracle-mediated prevention system integrating machine learning inference with smart contract execution. Training ensemble models on 12,847 Ethereum transactions with engineered features capturing gas anomalies and temporal patterns, we achieve 94.2\% fraud classification accuracy. Testnet deployment demonstrates 1.09-second response latency with 6.8\% computational overhead, contrasting favorably against prior on-chain implementations requiring 34\% overhead. Our working prototype validates practical viability for production environments where security requirements justify marginal transaction costs.
This paper develops a deep reinforcement learning framework for cryptocurrency portfolio management in which transaction costs are derived from the Riemannian geometry of the underlying volatility model rather than assumed constant. A Proximal Policy Optimisation agent is trained on a reward function grounded in non-equilibrium thermodynamics: we use the free-energy Bellman equation, in which transaction costs are the geodesic slippage on the Fisher information manifold of a maximum-entropy Markov-switching GARCH model, and regime-transition costs are the Wasserstein-2 distance between the calm and turbulent return distributions. A thermodynamic Carnot bound on portfolio efficiency is established and empirically validated. Five hypotheses are tested across Bitcoin, Ethereum, Ripple, Litecoin, and Bitcoin Cash over January 2017 to March 2026. The geometric-cost agent achieves statistically superior Sharpe ratios relative to flat-fee baselines on four of five assets; portfolio turnover is reduced by 56 to 83 percent relative to signal-following; the thermodynamic friction point at which the agent prefers no-trade is asset-specific and ordered by turbulent half-life; a joint topological and geometric circuit breaker reduces Maximum Drawdown by 28 to 38 percent; and ablation confirms that every component of the observation vector contributes a statistically significant performance gain. The framework requires liquid cryptocurrency markets with validated parametric volatility models; transferability to other asset classes requires upstream recalibration.
Blockchain and decentralized finance have revolutionized the financial ecosystem while simultaneously exposing it to cryptocurrency phishing attacks. Existing phishing detection methods primarily rely on graph learning, but they face significant limitations. Static graph learning approaches fail to account for the temporal evolution of phishing patterns, while semi-dynamic methods, such as those combining static GNNs with LSTM, struggle to capture the irregular and bursty nature of blockchain transactions. Moreover, these methods overlook the diversity of Ethereum transactions, treating them as homogeneous graphs, and heavily rely on supervised learning, which requires extensive labeled data that is not readily available. These limitations reduce their adaptability to emerging phishing threats. In this paper, we present PhishEye, a fully dynamic self-supervised system that monitors on-chain transactions to detect phishing activities. PhishEye formulates Ethereum transactions as a heterogeneous temporal attributed multi-graph and incorporates a novel temporal graph contrastive learning model, which captures both temporal patterns and heterogeneous transaction types. The evaluation on a dataset of 161,658 addresses and 416,541 transactions shows that PhishEye outperforms existing methods, achieving an F1 score of 87.23% and an AUC of 98.43% for phishing transaction detection, and an F1 score of 94.19% and an AUC of 98.03% for phishing account detection. In real-world deployment from May 1, 2023 to July 31, 2024, PhishEye identified 1,803 previously unknown phishing addresses, providing early alerts that helped prevent losses exceeding 2 billion USD.