Blockchain Papers

Xubin Luo

A decentralized ecosystem can capture value and still fail to fund the actors who keep it running. Users may pay fees, tokens may appreciate, issuers may earn revenue, and protocols may burn value, but none of these facts by itself shows that authors, miners, validators, suppliers, storage providers, or other critical participants are actually compensated. This paper argues that traditional value-capture analysis often assumes a centralized pool: once value is captured, it can be reallocated through budgets, contracts, payroll, or managerial discretion. Decentralized ecosystems do not have this default pool. They require routed closure: captured value must pass through a verifiable route to a specified critical incentive recipient, and it must be sufficient relative to that recipient's reward requirement. We formalize this distinction through Route-Admissible Value and operationalize it with the External Value Routing Closure protocol. A contrast set including YouTube, Steem/Steemit, Bitcoin, Ethereum, Aave, Filecoin, USDC, and XRP shows why revenue, fees, burns, token prices, or market capitalization should not be mistaken for sustainable incentive funding.

José Ignacio Peinador Sala

Conditional Refutation of Erdős Problem #463 via Arithmetic Quantum Chaos Author: José Ignacio Peinador Sala Overview This repository contains the full manuscript, companion computational notebooks, and formal Lean 4 verification for the paper "Conditional Refutation of Erdős Problem #463 in Hyper‑Slow Growth Regimes via Arithmetic Quantum Chaos". We demonstrate that, under the hypothesis that the survival variance of rough numbers around primorials is controlled by the fractal dimension D2≈0.24338 of the Riemann‑GUE Hamiltonian (Bridge Conjecture), no function f(n)≤log⁡(log⁡n) satisfies Erdős' condition for all sufficiently large n. The proof is constructed by bridging Galois projection operators, power‑law random banded matrices (PRBM), the Altshuler‑Shklovskii effect, and optimal transport (Kantorovich–Rubinstein duality). The ultimate goal of this program is to elevate this conditional result to an unconditional proof by integrating the supersymmetric Non-Linear Sigma Model (NLσM) limit with the most recent 2025 sieve bounds on rough numbers in short intervals. Contents Article: Open pdf One‑Click Reproducibility This project is designed for frictionless, one‑click reproducibility. No compiler installation, no supercomputing cluster. All experiments run on Google Colab with zero local setup — you can audit the physics of the arithmetic vacuum from a browser on your laptop or even your phone. What the notebooks validate You can run the experiments directly in your browser: Notebook Contents What it certifies Main experiments: Open in Colab Experiments 1–4 + Chirikov map Collapse of Nₖ, monotonic decrease of D₂, massive suppression of Σ²(L), sub‑diffusive SFF ramp, classical chaos suppression Lean 4 verification: Open in Colab Lean 4 formal proofs Idempotence of the Galois projector, discrete variance floor lemma, modular classification of primes Experiments (Main Notebook) Collapse of the survival variable Nk – deterministic emptiness of the critical interval for primorials k≥10 (M=5,000 samples). Fractal dimension D2 of pruned Hamiltonians – monotonic decrease under Galois projection (Numba‑accelerated up to N=10,000). Number variance Σ2(L) and Thouless energy – massive spectral suppression (up to 96% below GUE) with the Thouless scale plunging below L=0.5 (M=10,000 realizations). Spectral Form Factor and Finite‑Size Scaling – robust sub‑diffusive ramp (γ→0.61) and convergent D2≈0.106 in the thermodynamic limit (M=100 realizations, N up to 6,000). Chirikov Map (Classical) – Galois projection completely strangulates chaotic transport (D≈0.00 vs D≈11.05), proving universal ergodicity suppression. Formal Verification in Lean 4 Erdős Problem #463 is actively tracked by the mathematical community, including Google DeepMind's formal-conjectures repository. Laying the formal groundwork to resolve this, the notebook Notebooks/erdos_refutation.ipynb compiles and mechanically verifies three foundational lemmas in Lean 4 (v4.29.1, Mathlib4): Discrete Variance Floor Lemma — ∀ x ∈ ℕ, x ≤ x² Galois Projector Idempotence — χ² = χ for the coprimality indicator Modular Classification of Primes — ∀ p > 3 prime, p ≡ 1 ∨ p ≡ 5 (mod 6) These lemmas form the unshakeable logical bedrock of the conditional refutation. 🔭 Philosophical Context "Mathematics is not about numbers, equations, computations, or algorithms: it is about understanding." — William Thurston For decades, the distribution of prime numbers and the behaviour of chaotic quantum systems were studied as separate continents of knowledge, occasionally glimpsing each other across a narrow strait —the Hilbert–Pólya conjecture, the Montgomery–Odlyzko law— but never truly merging. This work builds a bridge across that strait. The key insight is that the ring ℤ/6ℤ is not merely a convenient sieve for eliminating multiples of 2 and 3. It is a topological substrate —a discrete analogue of the KO‑dimension in noncommutative geometry— that partitions the integers into resonant channels (𝒞₁ and 𝒞₅) and sterile channels (𝒞₀, 𝒞₂, 𝒞₃, 𝒞₄). When this partition is imposed as a superselection rule on a quantum Hamiltonian, the system does not thermalise. It enters a Non‑Ergodic Extended (NEE) phase where fluctuations are systematically suppressed, variance collapses, and the arithmetic vacuum swallows the survivors. The philosophical lesson is profound: randomness is not the default state of complex systems. The apparent chaos of prime numbers, long regarded as the quintessence of unpredictability, harbours a rigid geometric order. That order can be harnessed —through Galois projection, through PRBM Hamiltonians, through the Altshuler–Shklovskii effect— to prove theorems that have resisted classical sieve methods for half a century. This project also embodies a conviction about how science should be done in the age of artificial intelligence. Every line of code, every formally verified lemma, and every numerical experiment was developed using freely accessible tools. The massive simulations of quantum chaos, which traditionally would demand exclusive access to institutional supercomputers, were executed entirely on Google Colab, democratizing high-performance computing. The formal verification of the mathematical bedrock was achieved using the open-source proof assistant Lean 4. Furthermore, the theoretical framework was built in a genuine symbiosis with DeepSeek, an open-weight AI freely provided to the world. No proprietary models, no paywalled platforms, no computational aristocracy. This work demonstrates that the absolute frontier of mathematical research is now accessible to anyone with a good idea, a standard laptop, and the willingness to engage in dialogue with tools that amplify, rather than replace, human creativity. "The universe is written in the language of mathematics." — Galileo Galilei Perhaps it is written, more precisely, in the language of modular arithmetic. Last Update: May 2026 | Status: Under Peer Review in IOP/LMS Nonlinearity (Ref: NON-110856) | Built with ❤️, 🐍 & 🤖

Paolo Antonelli, Pierangelo Marcati, Laura V. Spinolo

We study the zero-dispersion limit for a class of Korteweg--de Vries (KdV)-type initial-boundary value problems on the half-line, with Dirichlet boundary conditions assigned at \(x=0\). We focus on the outflow regime, where the solution of the limiting scalar conservation law does not attain the boundary condition imposed on the dispersive problem. We construct a boundary layer profile, depending on the fast variable, which is uniquely determined, through the associated stationary third-order boundary layer equation, by the mismatch between the boundary conditions, and by the exponential decay at infinity in the fast variable. Our main result shows that, under suitable regularity and compatibility assumptions on the data, the dispersive solution is well approximated by a WKB expansion given by the sum of the smooth solution of the conservation law and the boundary layer profile. In particular, we establish stability of the boundary layer profile by proving quantitative estimates for the remainder term in a weighted energy norm, and show that it converges to $0$ in $H^1$, uniformly in time and up to the lifespan of the smooth solution of the conservation law. The proof is based on the analysis of a linearized energy functional and does not rely on complete integrability or inverse scattering techniques. It applies to general fluxes and requires no smallness assumption on the amplitude of the boundary layer. To the best of our knowledge, this is the first stability result for boundary layers of KdV-type equation on the half line.

Kunt

No abstract is available for this record.

Dr. Pankaj Malik, Soham Bundela, Mohd. Ivaid, Mohnish Dhurve · 5 authors

The rapid expansion of Decentralized Finance (DeFi) has enabled open and permissionless token trading, but it has also led to a surge in fraudulent activities such as rug pulls, wash trading, and pump-and-dump schemes. This paper presents a novel fraud detection approach based on correlation analysis between token price and liquidity, leveraging the inherent relationship between these two market variables. In legitimate markets, price movements are typically supported by corresponding changes in liquidity, whereas fraudulent tokens often exhibit abnormal or decoupled behavior due to artificial price manipulation. To investigate this, we analyze time-series data of token price and liquidity across multiple decentralized exchanges and compute statistical correlation metrics alongside liquidity variation patterns. Experimental results show that legitimate tokens maintain strong positive correlations (r > 0.7) between price and liquidity, while fraudulent tokens exhibit weak or unstable correlations (r < 0.3), often accompanied by sudden liquidity withdrawals or artificial volume spikes. The proposed framework achieves high detection performance with an accuracy of 92.4%, precision of 90.1%, recall of 93.6%, and F1-score of 91.8%, demonstrating its effectiveness in identifying suspicious tokens at early stages. The findings confirm that deviations in price–liquidity correlation serve as a reliable and computationally efficient indicator for fraud detection in DeFi ecosystems. This approach can be integrated with existing blockchain analytics tools to enhance real-time monitoring and improve investor protection.

Ide Prima

Pesatnya perkembangan teknologi finansial telah melahirkan aset digital baru berupa Cryptocurrency dan Non-Fungible Token (NFT) yang memiliki nilai ekonomi signifikan. Namun, regulasi di Indonesia saat ini lebih menitikberatkan pada aspek perdagangan (komoditas) melalui aturan Bappebti, sementara pengaturannya sebagai objek hukum dalam ranah keperdataan, khususnya hukum kewarisan, masih belum spesifik. Penelitian ini bertujuan untuk menganalisis kedudukan hukum aset digital sebagai objek waris menurut Kitab Undang-Undang Hukum Perdata (KUHPerdata) dan mekanisme pemindahannya kepada ahli waris. Metode penelitian yang digunakan adalah yuridis normatif dengan pendekatan perundang-undangan (statute approach) dan pendekatan konseptual (conceptual approach). Data yang digunakan adalah data sekunder yang terdiri dari bahan hukum primer, sekunder, dan tersier. Meskipun bersifat imateriel, Cryptocurrency dan NFT memenuhi kualifikasi sebagai "Benda" (Zaak) bergerak yang tidak berwujud sebagaimana diatur dalam Pasal 499 dan Pasal 503 KUHPerdata, karena memiliki nilai ekonomi dan dapat dimiliki secara hukum. Oleh karena itu, aset digital secara yuridis sah untuk dikategorikan sebagai bagian dari harta warisan (boedel waris). Dalam pewarisan aset digital terletak pada sifat anonimitas dan desentralisasi teknologi blockchain. Tanpa penyerahan private key atau akses dompet digital dari pewaris kepada ahli waris, aset tersebut terancam menjadi "aset beku" yang tidak dapat dieksekusi meskipun secara hukum hak kepemilikannya telah berpindah demi hukum (Le Mort Saisit Le Vif). Diperlukan pembaharuan hukum atau pedoman teknis mengenai tata cara pembuktian kepemilikan dan prosedur eksekusi aset digital dalam penetapan waris agar memberikan kepastian hukum dan perlindungan hak bagi ahli waris.

Pedro Henrique F. S. Oliveira, Heder S. Bernardino, Saulo Moraes Villela, Edelberto Franco Silva · 6 authors

O Ethereum é uma plataforma de criptomoedas que permite a execução de contratos inteligentes, programas autônomos que operam em uma rede descentralizada. As vulnerabilidades nesses contratos representam grandes riscos financeiros e de segurança nos ecossistemas blockchain, motivando a automatização do processo de detectá-las. Este trabalho estuda a detecção de vulnerabilidades em contratos inteligentes Ethereum usando embeddings derivados de bytecode. Embeddings são representações vetoriais geradas por modelos de linguagem, que capturam as características estruturais de texto. Essas representações foram usadas como entrada para os algoritmos de regressão logística, árvore de decisão e floresta aleatória, com o fim de detectar quais contratos possuem vulnerabilidades. Os resultados mostram que os embeddings contêm informações úteis para distinguir contratos vulneráveis de não vulneráveis. O estudo também constata que a alteração da distribuição original dos dados durante o treinamento afeta significativamente o desempenho, destacando a sensibilidade das abordagens baseadas em embeddings às estratégias de amostragem.

Jieran Cui, Zhengkai Wen, Haowen Fang, Yinan Zhu · 9 authors

Zero-knowledge virtual machines (zkVMs) are a key technology for driving the large-scale adoption of zero-knowledge proofs (ZKP), but their performance bottlenecks severely limit their practicality. While current hardware acceleration research has exclusively focused on backend proving, we identify that the frontend execution and trace generation phase is rapidly emerging as the new system bottleneck. To address this challenge, we propose ZK-Tracer, the first hardware accelerator architecture specifically designed for the zkVM frontend. ZK-Tracer features a novel heterogeneous design comprising a Main Trace Unit and parallel Permutation Trace Units. It exposes a fine-grained interface to the host software through a lightweight instruction set extension, enabling efficient task offloading. Our ASIC implementation results demonstrate that ZK-Tracer achieves up to 1829x speedup in trace generation over a high-performance multi-core CPU. When integrated with existing backend proving accelerators, it delivers a remarkable 963x end-to-end performance improvement for the entire ZKP system.

Vladimir Kovšca, Zrinka Lacković Vincek, Suzana Keglević Kozjak

Prior research on cryptoasset valuation has largely adapted discounted cash flow (DCF) models by treating staking rewards and transaction fees as productive cash flows, while insufficiently accounting for monetary characteristics and strategic flexibility inherent to decentralized platforms. This study investigates whether such cashflow based approaches systematically undervalue Ethereum. The central hypothesis is that Ethereum’s intrinsic value cannot be adequately explained by DCF valuation alone, and that monetary premium and technological optionality constitute economically significant components of value. To examine this hypothesis, a multi-layer valuation framework is applied using network and market data from 2022–2025, combining a DCF model, a monetary premium benchmarked against gold based on relative scarcity and adoption, and a real option uplift reflecting future expansion potential. Monte Carlo simulation is employed to test the robustness of the results. The findings indicate that while DCF-based valuations remain relatively stable, total intrinsic value is highly sensitive to assumptions regarding monetary adoption and strategic optionality. These results underline the importance of layered valuation frameworks for decentralized platforms.

DANIEL GUEDES, Douglas Wegner

Objetivo: propor a Decentralized Autonomous Franchise (DAF) como uma arquitetura organizacional alternativa para redes de franquias, baseada em blockchain, contratos inteligentes e governança tokenizada. Estado da arte: embora o franchising seja amplamente reconhecido como modelo eficiente de expansão, enfrenta limitações estruturais relacionadas à centralização de poder, à incompletude contratual e às assimetrias informacionais. Paralelamente, a literatura científica sobre Decentralized Autonomous Organizations (DAOs) tem avançado na discussão de governança descentralizada, ainda com pouca articulação com o campo de franchising. Originalidade: o artigo aproxima os campos de franchising e DAOs, propondo a DAF como modelo alternativo que reconfigura mecanismos de coordenação, participação e controle em redes de franquias. Impactos: o artigo oferece um referencial inovador para redes de franquias interessadas em atualizar seus mecanismos de governança, ampliar a participação dos franqueados e incorporar princípios de transparência e descentralização apoiados por tecnologias digitais descentralizadas. ODS: 8 – Trabalho decente e crescimento econômico, 9 – Indústria, inovação e infraestrutura, 17 – Parcerias e meios de implementação.

Rodrigo Reis Lastra Cid

Editorial: Governance as a Problem of Collective DecisionBy Prof. Dr. Rodrigo Cid, The PhilosopherGovernance, in its essence, is the art of answering an unavoidable question: how do we decide together? From tribal councils to modern corporations, from cooperatives to DAOs and artistic communities, the challenge repeats itself. And at the heart of this question lies a mechanism that is deceptively simple yet philosophically treacherous: the vote. Which voting system is the fairest? Arrow’s impossibility theorem taught us that there is no universal answer. Every context — its size, its urgency, its level of trust among participants — demands a specific institutional design. This dossier does not offer ready‑made formulas. Instead, it maps how different agents and sectors are confronting this problem. VAN Ameneyro, in their conversation with One Love DAO, shows us how governance becomes a “necessary evil” in the world of digital art. The promise of horizontality runs up against concrete questions: who gets a seat at the table? Who is still left out? The answer, for VAN, lies in transparency, real participation, and the construction of institutional memory — something fragile in the volatile environment of Web3 platforms. Fer Caggiano delivers an incisive diagnosis: power has not disappeared with decentralisation. It has merely moved — from institutions and curators to wallets and tokens. Community curation, in practice, often reproduces plutocracy. Their article forces us to ask: before voting, who defines what can be voted on? Governance begins with visibility. Steve Coulter (aka 45renegade) offers an unexpected interpretative key: punk as an operating system. Before lean startup, before bootstrapping, before the creator economy, punk already practised independent production, direct distribution, and community building as survival strategies. His text reminds us that governance is not only about formal rules and votes. It is also about ethics, refusal, and the courage to build without asking for permission. Vessy Mink appears twice in this issue, each time with a different governance experiment. In Governing the Sound, she presents Optimus Goddess, a licensing model where the artist retains full publishing rights and curation is replaced by equitable partnerships — an attempt to rewrite the rules of a historically extractive industry. In Music Train S9 E1‑6, co‑created with BK Han, she documents a live, collaborative song‑minting project. The audience does not merely listen; it participates in real‑time creation, turning musical production into a participatory governance performance. Here, the very act of making music becomes a collective decision process. Vitor Emanuel Gripp, writing from inside Token Nation, shows how a technology event can become a living laboratory for governance. By bringing together academics such as Maria Goreti (Fiocruz), Carlos Frederico (UFOP/KryptoLab) and Rodrigo Cid (UFOP/GIFLABS), Token Nation does not merely discuss decentralisation — it practises it, in the curation of its stages, the selection of its projects, and the constant negotiation between efficiency and participation. His account, grounded in his own journey from exhibiting artist to community manager, reminds us that governance is not an abstract protocol but a daily, messy, collective achievement. Rodrigo Cid, in Voting or Governing, returns to the philosophical bedrock of the problem. He reminds us that Arrow’s theorem is not a mathematical curiosity but a structural warning: no voting system is neutral. By applying this lesson to blockchain, his article demonstrates that digital governance does not escape the aporias of collective choice — it merely translates them into code. He also develops the problem of many hands and the fragility of institutional memory in decentralised systems, connecting directly with the questions raised by VAN Ameneyro and Fer Caggiano. Felipe Farinha, from the University of Saint Joseph in Macau, brings a comparative perspective. In Direct Democracy in the Age of the Extended Mind, he examines how different jurisdictions and cultural contexts shape the possibility of legitimate digital governance. His reflection on personal exocortices — AI assistants that extend a citizen’s cognitive capacity — asks whether direct democracy might finally become feasible at scale, provided we solve the problems of authenticity, manipulation, privacy, and civic deskilling. The exocortex, for Farinha, is not a substitute for democratic agency but a prosthesis for it. Daniel Gomides, from the Federal University of Ouro Preto, revisits Rousseau’s First Discourse in Scientific Progress and Moral Progress. He argues that technical sophistication does not guarantee ethical advancement. His analysis of AI, hunger, and climate agreements shows that the gap between what we can do and what we should do remains as wide as ever — a sobering reminder that governance cannot be reduced to algorithmic efficiency. Rousseau’s warning, written in 1750, still echoes today: science may teach us how to build better tools, but it does not teach us how to be better humans. Rafaela Ferrari Kley, from Degenerados Club, takes a complementary path in The Logic of Irrationality. Diagnosing a civilisation that has never been so technologically advanced yet remains emotionally manipulable, she recalls that Aristotle’s logic was a civilising attempt to contain collective hysteria. Today’s algorithms, however, are optimised for engagement, not truth. Her reflection forces us to ask whether governance can ever be purely rational, or whether it must always wrestle with the irrational architectures of attention, fear, and belonging. The question she leaves us with is disarmingly simple and profoundly uncomfortable: are we seeking to understand reality, or merely constructing emotionally bearable versions of it? Finally, this issue closes with a special launch. Rodrigo Cid’s The Philosophy of Artificial Intelligence (co‑authored with Pedro Luiz Caetano Filho) is not merely reviewed here; it is treated as a governance artifact. Its systematic collaboration with AI systems, its transparent use of the CRediT‑IA framework, and its rigorous discussion of Arrow’s theorem, opacity, bias, and responsibility provide a conceptual toolkit for exactly the questions this magazine raises. The book is available for free download (DOI: 10.5281/zenodo.20143966), and we invite our readers to read it alongside the articles in this issue. At the end of this volume, we hope that the reader will not find definitive answers. Instead, find a richer repertoire of questions. About who votes, about who decides what is put to a vote, about how we move from paper to action. About whether code can ever fully replace trust, and whether we would want it to. About what the classroom, the punk venue, the music studio, the DAO, the tokenised event, and the blockchain have in common: all are arenas where the same question echoes, again and again — how shall we decide together? Enjoy the reading. Prof. Dr. Rodrigo CidThe Philosopher(GIFLABS / Universidade Federal de Ouro Preto)

Stefan Stefanović

Validators on generic Proof of Stake chains earn the same fees whether they handle attestation work correctly or selectively censor it. For chains whose main activity is moving tokens around, that indifference is fine. For chains whose primary economic activity is recording attestations (content provenance, AI-output attribution, threshold-signed credentials, supply-chain receipts), the indifference becomes a problem. Proof of Useful Attestation (PoUA) makes attestation handling first-class in the consensus weighting itself. Validator vote weight is the product of bonded stake and a reputation scalar in [r_min, r_max] that accumulates from valid attestation work. The reputation update is additive, fee-weighted, non-transferable, and capped per epoch. We prove a cost-to-grind floor (Lemma 1): under chain-wide adaptive burn fraction tau_burn, the non-recoverable cost an adversary pays to inflate reputation by Delta_r is bounded below by tau_burn * Delta_r / (eta * alpha_eff). Under the recommended v0 calibration (r_max/r_min in [4, 10]), the cost premium against a capital adversary is 4x to 10x over equivalent pure-stake PoS at steady state. The paper specifies the mechanism, six layered Sybil and grinding defenses, empirical Monte Carlo strategy-search across the full layered defense, and grinding detectors with explicit threshold derivations. It is a mechanism-design proposal with a formal economic floor and inherited BFT safety and liveness, not a complete cryptographic security proof. This release incorporates feedback from Jiangshan Yu (University of Sydney) and Marko Vukolić (Bitcoin Scaling Labs).

Guilherme Martins Soares, João L. D. S. Filho, Nicholas P. Fontanini, Bruno Evaristo

Contratos inteligentes gerenciam ativos digitais de alto valor, mas falhas de segurança frequentemente causam perdas financeiras irreversíveis. Embora existam diversas ferramentas de auditoria automatizada, seu uso isolado gera altas taxas de falsos positivos e falsos negativos. Este trabalho propõe e avalia um framework unificado para auditoria de contratos inteligentes em Solidity, orquestrando análise estática (Slither), execução simbólica (Mythril) e testes dinâmicos (Foundry). A arquitetura unifica os resultados heterogêneos utilizando o padrão SARIF e aplica um Modelo de Linguagem de Grande Escala (LLM) para traduzir logs brutos em relatórios contextuais explicáveis. Avaliado em um dataset curado de 53 contratos do repositório SmartBugs, o framework alcançou um F1-Score de 92,93%, superando substancialmente o desempenho isolado do Slither (72,28%) e do Mythril (88,42%). Os resultados demonstram que a orquestração híbrida mitiga as limitações estruturais de cada motor, reduz a carga cognitiva do auditor e consolida-se como uma plataforma robusta e eficaz para o desenvolvimento seguro no ecossistema Web3.

Mycelia Research Team

Abstract Web3 governance systems have expanded the visibility of institutional processes through on-chain records of votes, proposals, and token balances. Yet the availability of governance data does not necessarily ensure that governance structure or decision dynamics remain interpretable. This paper introduces governance observability as a distinct analytical layer within Web3 systems — concerned not with data availability, but with the capacity to relate observable governance activity to the institutional structures and processes that give it meaning. Building upon previous work on VCS governance geometry, the paper defines governance observability through two complementary dimensions: structural observability, which concerns the visibility of governance configuration and role relationships; and flow observability, which concerns the traceability of decision processes as they move through governance structures over time. Together, these dimensions provide a framework through which governance systems may be examined as both structural and dynamic phenomena. The paper further introduces structural invariants as continuity conditions supporting coherent interpretation of governance structure across institutional change, and flow concentration detection as a means of examining how operational patterns may gradually shape the exercise of authority differently from formal governance design. Web3 infrastructure is examined as an enabling environment for such observability, while remaining distinct from the interpretive frameworks through which governance conditions become institutionally meaningful. Governance observability is throughout distinguished from governance automation. The framework does not prescribe intervention or establish universal governance standards. Rather, it proposes that governance systems may benefit from maintaining the capacity to sense and interpret their own structural and operational conditions as governance evolves — contributing an interpretive layer that supports institutional discernment without replacing institutional judgment. Keywords: Web3 governance, governance observability, structural audit, flow observability, VCS governance geometry, institutional design, decentralized governance

Junjie Peng, Sheng Cao, Ying Zhou, Xiaosong Zhang

No abstract is available for this record.

AKATERINH XENOPOULOU-TYROKOMOU, Epameinondas Xenopoulos

Xenopoulos’ Historical Genetic Logic: A New Framework and the XEPTQLRI Theorem DOI:10.5281/zenodo.20367121Date: May 2026 Aikaterini Xenopoulou TyrokomouIndependent ResearcherORCID: 0009 0004 9057 7432Email: katerinaxenopoulou@gmail.com Theoretical Foundation: Epameinondas Xenopoulos †Based on the Historical Genetic Logic of Epameinondas Xenopoulos, Epistemology of Logic: Logic Dialectic or Theory of Knowledge (posthumous 2nd ed., 2024) [1, 2]ORCID: 0009 0000 1736 8555† In memoriam (1920–1994) Methodological NoteThe present work simplifies and mathematizes central ideas of the formal-dialectical logic of E. Xenopoulos in order to create an applicable computational tool. It does not constitute a faithful rendering of his philosophical theory in its full depth, but a focused operationalization for the purpose of computational application. Statement of AuthorshipThe present work is founded on the logical system of Epameinondas Xenopoulos (1920–1994). The XEPTQLRI index does not constitute an independent theory, nor does it introduce a new autonomous logical framework. The theoretical background, the basic categories, the logical relations, the fundamental principles, and the dialectical operators belong to the work of Epameinondas Xenopoulos. The contribution of the present work consists in the formal mathematical operationalization of specific principles of this logical system through a computable index, capable of being applied to dynamic and historically evolving systems. Consequently, the theoretical authorship belongs entirely to Epameinondas Xenopoulos, while the present work belongs to the level of systematic formalization, proof, application, and methodological development of his framework. The XEPTQLRI index expresses in quantitative form the logic of Being, Non-Being, Becoming, historical memory, and dialectical sublation, while adapting these concepts for computational use. In this sense, the present work constitutes a continuation, clarification, and applicative deepening of the Xenopoulos system, not a displacement or replacement of it. ABSTRACT We present the Xenopoulos Pre-Transitional Qualitative Leap Risk Index (XEPTQLRI), a novel mathematical index grounded in the Historical-Genetic Logic of the Greek philosopher Epameinondas Xenopoulos [1, 2]. Unlike conventional statistical summaries, XEPTQLRI captures the dialectical interplay between Being (B), Non‑Being (N), historical memory (τ), and a historical paradox factor (Π). The index is defined as Ξ = [T · τ · (1 + Π)] / Θ₀ with Θ₀ = 0.85, where T = 2BN/(B+N) is the dialectical tension expressed through the harmonic mean. Its construction respects strict causality, min‑max or logistic normalization, and a negative feedback mechanism (∂σ/∂Ξ < 0) in its dynamical extensions, though the index itself remains exogenous and purely diagnostic. We prove five theorems establishing constructive computability, scale homogeneity, non‑preservation of dynamical structure, representation dependence, and linear‑time computability. Two additional theorems (non‑self‑inversion and logical phase transition) are proved within the extended framework of the 34 Principles. Numerical experiments with the Ferrari–Xenopoulos v4.0 stochastic model show reproducible and persistent exceedance of the Aufhebung threshold, with endogenous volatility remaining low (σ ≈ 0.058). An extreme parameter run (α₅ = 1.6, σ₁ = 1.0, Θ₀ = 0.0867) reaches Ξ = 16.1, demonstrating that the critical value is not a universal constant but a local, parameter‑dependent realization. A “Dialectical War” experiment (LSTM vs. Xenopoulos system under noise = 1.0) reveals a striking dissociation: technical performance (MAE = 0.1039, 67.1% improvement) coexists with universal dialectical risk (20/20 high‑risk steps, Ξ_max = 2.99, zero paradoxality and false stability). This dissociation is mathematically consistent, as MAE and Ξ are distinct functions measuring different aspects of system behavior (MAE ⇏ Ξ). A null model comparison confirms that this risk is structurally generated (AUC 0.949 vs. 0.501, p < 0.001), with ground truth defined by the condition Ξ(t) ≥ Θ₀ for at least three consecutive time steps and binary classification threshold optimized via the Youden index. A strictly endogenous application of the canonical XEPTQLRI index to 13 distinct COVID‑19 waves in Greece (JHU CSSE) yields early warnings 48–90 days in advance (mean 84.0 days) with a mean EWS Score of 0.785, successfully detecting 10 of 13 waves (76.9%). The system substantially outperforms a simple cases‑threshold baseline (mean EWS 0.42, 23.1% success) without any reliance on AUC or external classifiers. Beyond its diagnostic function, the XEPTQLRI framework demonstrates a transformative capacity: non‑dialectical codes exposed to the Xenopoulos environment undergo systematic improvement, with documented gains ranging from 52.3% to 95.65% across multiple independent experiments. A banking crisis application correctly identified Lehman Brothers (z=3.2, p<0.001) and Bear Stearns (z=2.9, p<0.01) two years before their collapse using only pre‑2006 data. A financial early warning application achieved statistically significant predictive correlations (r=0.29–0.44, p<0.001) with lead times of 10–77 days across S&P 500, VIX, Treasury yields, and Bitcoin. Two complete experimental protocols (XENO‑EXP‑2026‑002 and XENO‑EXP‑2026‑003) provide systematic, statistically significant evidence that the Xenopoulos System, when fully embedded in machine learning architectures, functions as an improvement catalyst with measurable economic value (ROI 63:1, break‑even 6 days). Thus, XEPTQLRI bridges formal dialectics with practical early warning systems, establishing a universal law of qualitative transition while keeping its numerical expression local and context‑dependent. The present system constitutes a proto‑formalized theoretical framework — a structured mathematical–dynamical system with axiomatic foundation (34 Principles), provable theorems (7 Theorems), and computational implementation (Ferrari–Xenopoulos v4.0, COVID‑19 application), whose applicative and transformative value has been verified on real data. The system is internally consistent under its stated principles, though its full formalization in the sense of a Hilbert‑style formal system remains a subject for future work. Keywords: XEPTQLRI, Historical‑Genetic Logic, dialectical logic, qualitative leap, Aufhebung, early warning systems, stochastic differential equations, LSTM, COVID‑19, proto‑formalized framework, non‑classical negation, harmonic mean, paradox factor, historical memory, dialectical transformation, financial crisis prediction, code optimization. Lead paragraph Complex dynamical systems often undergo sudden, qualitative transformations—critical transitions that are difficult to anticipate with conventional statistical tools. This paper introduces a new mathematical framework for detecting such transformations, grounded in the Historical‑Genetic Logic of the Greek philosopher Epameinondas Xenopoulos (1920–1994). The central contribution is the Xenopoulos Pre‑Transitional Qualitative Leap Risk Index (XEPTQLRI), defined as Ξ(t) = T(t) · τ(t) · (1 + Π(t)) / Θ₀, where T is the dialectical tension between Being and Non‑Being, τ captures historical memory, and Π encodes the accumulated paradox of extreme past states. The index is fully endogenous, requires no external training or classifiers, and is accompanied by a typology of ten dialectical stages (τ₀–τ₉). We prove five constructive theorems, validate the framework through stochastic simulations, and apply it to real COVID‑19 data from Greece. Across 13 epidemic waves, XEPTQLRI issued early warnings with an average lead time of 84.0 days and a mean Early Warning Score of 0.785, substantially outperforming a simple cases‑threshold baseline. The framework thus bridges formal dialectics with operational early warning capability, offering a new lens for the study of critical phenomena. Part I — Definition and Foundation of XEPTQLRI 1. Theoretical Foundation This section presents the fundamental principles underlying the Xenopoulos Pre-Transitional Qualitative Leap Risk Index (XEPTQLRI), as formulated in the Historical-Genetic Logic of the Greek philosopher Epameinondas Xenopoulos (1920–1994) [1, 2]. These principles constitute the axiomatic framework of the index and determine both its mathematical form and its interpretive function. XEPTQLRI is neither a simple numerical magnitude nor a mere statistical summary. Instead, it is defined as a complex historical-dialectical index that captures the relationship between Being, Non-Being, their dialectical tension, historical tendency, and the probability of transcending a critical threshold of transformation. The index is embedded within the broader system of 34 Principles as the 23rd Principle, expressed through the general dialectical operator: Ξ(t) = N[F₂₃(G₂₃)]. 1.1 Principle 5: Complementarity According to the theory [1, 2], Non-Being is not an independent quantity but the complement of Being. This relationship is expressed by Principle 5: N(t)=1−B(t)N(t)=1−B(t) This equation implies that: B(t)+N(t)=1B(t)+N(t)=1 Therefore, the two quantities B(t) and N(t) are complementary aspects of the same dynamic state. If B(t) expresses the degree of presence of Being, then N(t) expresses the degree of presence of Non-Being. From the same principle it immediately follows that it is impossible for both of the following to hold simultaneously: B(t)>0.8andN(t)>0.8B(t)>0.8andN(t)>0.8 because then we would have B(t) + N(t) > 1.6, in contradiction with B(t) + N(t) = 1. Important clarification: In Theorem 2 (Paradoxical Transcendence), the condition B > 0.8 ∧ N > 0.8 refers to a special paradoxical state where the usual complementarity is suspended due to the historical accumulation of contradictions. In this state, B and N are not understood as instantaneous values at

Anthony Coslett

The numerical receipt that allows independent verification of which AI model is serving a frontier API endpoint — the top-*K* log-probability vector computed on every forward pass — is being withdrawn across every major frontier lab, without announcement. xAI silently ignores the parameter on Grok 4.20 and newer. Google Vertex began returning errors on Gemini 3 Pro without notice. OpenAI excludes the entire reasoning-model class and the GPT-5 line. Anthropic has never exposed the field. The withdrawal is not universal: legacy and non-reasoning models at the same providers continue to return logprobs on the same infrastructure. The capability is not technically infeasible. It is a decision. This note documents the current state of logprob access across four frontier providers, establishes what the access enables and what it does not, and provides six operational contract clauses that preserve the enterprise's right to verify model identity at the API layer. The mathematics of establishing model identity from top-*K* logprob output is documented in the companion research [1, 2]; this note concerns whether the numbers will continue to be available at all. The Neural Network Identity Series — Mathematical foundations, empirical validation, and governance frameworks for verifying which model is running Paper 1: The δ-Gene: Inference-Time Physical Unclonable Functions from Architecture-Invariant Output Geometry (DOI: 10.5281/zenodo.18704275) Paper 2: Template-Based Endpoint Verification via Logprob Order-Statistic Geometry (DOI: 10.5281/zenodo.18776711) Paper 3: The Geometry of Model Theft: Distillation Forensics, Adversarial Erasure, and the Illusion of Spoofing (DOI: 10.5281/zenodo.18818608) Paper 4: Provenance Generalization and Verification Scaling for Neural Network Forensics (DOI: 10.5281/zenodo.18872071) Paper 5: Beneath the Character: The Structural Identity of Neural Networks — Mathematical Evidence for a Non-Narrative Layer of AI Identity (DOI: 10.5281/zenodo.18907292) Paper 6: Which Model Is Running?: Structural Identity as a Prerequisite for Trustworthy Zero-Knowledge Machine Learning (DOI: 10.5281/zenodo.19008116) Paper 7: The Deformation Laws of Neural Identity (DOI: 10.5281/zenodo.19055966) Paper 8: What Counts as Proof? — Admissible Evidence for Neural Network Identity Claims (DOI: 10.5281/zenodo.19058540) Paper 9: Composable Model Identity — Formal Hardening of Structural Attestations in the Enterprise Identity Stack (DOI: 10.5281/zenodo.19099911) Paper 10:Where Identity Comes From: Path Sensitivity and Endpoint Underdetermination in Neural Network Training (DOI: 10.5281/zenodo.19118807) Paper 11: Post-Hoc Disclosure Is Not Runtime Proof: Model Identity at Frontier Scale (DOI: 10.5281/zenodo.19216634) Paper 12: Family-Dependent Response to Reasoning Distillation Across Structural and Functional Identity Layers (DOI: 10.5281/zenodo.19298857) Paper 13: Safety-Alignment Removal as a Model-Identity Failure — Structural Evidence from Published Weight-Level Mutation Checkpoints (DOI: 10.5281/zenodo.19383019) Technical Note: Agent Identity Is Not Model Identity (DOI: 10.5281/zenodo.19240883) Technical Note: Gap Invariance: Why PPP Measurements Are Domain-Independent by Construction (DOI: 10.5281/zenodo.19275524) Technical Note: Measured Model Substitution Under Valid Agent Credentials (DOI: 10.5281/zenodo.19342848) Technical Note: Artifact Identity Is Not Runtime Identity — Trustfall Lite and the Boundary of File-Level Model Verification (DOI: 10.5281/zenodo.20019127) Technical Note: Artifact Identity Is Not Runtime Identity — Trustfall Lite and the Boundary of File-Level Model Verification (DOI: 10.5281/zenodo.20019127) Technical Note:: The Disappearing Window — AI Logprob Access Withdrawal and the Structural Verifiability of Frontier Model Contracts (DOI: 10.5281/zenodo.20362098) Formal Verification Stack for Neural Network Structural Identity (IT-PUF Coq Proofs) (DOI: 10.5281/zenodo.18930621) Copyright (c) 2026 Anthony Ray Coslett / Fall Risk AI, LLC. All Rights Reserved. Confidential and Proprietary. Patent Pending (Applications 63/982,893, 63/990,487, 63/996,680, 64/003,244).

Ilir Mehmetaj

Abstract This concept documents a complete physics-first authorisation architecture for the quantum-permanent era — applicable across AI cluster security, interbank settlement, space and interplanetary infrastructure, critical infrastructure protection, digital identity, supply chain integrity, and optional democratic participation tools. The architecture rests on a single physical principle: a cryptographic credential that no longer exists cannot be recovered by any computation, quantum or classical, regardless of future advances in hardware or algorithms. The concept extends the Temporal Rotation Security Protocol (TRSP v3, DOI: 10.5281/zenodo.20324081) and the TRSP Digital Coin (TDC v2, DOI: 10.5281/zenodo.20332811) with a unified Layered Temporal-Quantum Security (LTQS) framework. LTQS combines NIST FIPS 203/204-standardised Post-Quantum Cryptography (ML-KEM, ML-DSA) as Layer 0 — mathematical transit security — with TRSP temporal rotation as Layer 1 — physical credential elimination through hardware-enforced destructive readout within a configurable rotation window (10–500 ms). Layers 2 and 3 add geographically distributed hybrid dynamic quorum validation and LEO satellite orbital entropy anchoring with relativistic timestamp verification. An integrated adaptive AI management layer selects security profiles dynamically across High-Assurance, Standard, Degraded, and Emergency modes — guaranteeing graceful degradation to pure PQC fallback when physical infrastructure is unavailable. The hardware commitment module previously documented as CRATON is architecturally designated URDHR, after the Norse Norn of the irrecoverable past. The two complementary quorum layers are designated VERÐANDI (present-moment ground quorum) and SKULD (future-anchoring orbital quorum) — the three Norns mapped to the three temporal dimensions of cryptographic security. Prior art established under the CRATON designation in all previously published documents extends fully to the URDHR designation. Fifteen novel contributions are placed on the public record as defensive prior art: NC-TDC-21 (AI-to-AI Micropayment Architecture), NC-TDC-22 (Macroscopic Environmental Entropy as Optical Physical Unclonable Function), NC-TDC-23 and NC-TDC-23a (Macroscopic Polymorphic Cipher with Dynamic Dimensional Entropy — exploratory), NC-TDC-24 (TRSP Democratic Coercion Shield — exploratory, extending Juels-Catalano-Jakobsson coercion-resistant voting literature), NC-TDC-25 (Continuous Anonymous Democratic Pulse — exploratory), NC-TDC-26 (Physical Proof of Presence consensus mechanism operating at the Landauer thermodynamic minimum), NC-TDC-27 (Temporal Scarcity Value Architecture anchored in thermodynamic time-arrow irreversibility), NC-TDC-28 (AI Exchange Consortium Architecture), NC-TDC-29 (Biometric Supply Architecture), NC-TDC-30 (CRATON Chain Coin Identity Architecture without persistent private key), NC-TDC-31 (Three-Phase Value Architecture), and NC-TDC-32 (Cooperative Multi-Anchor Currency Architecture with Founder-Operator Equity-Plus-Operating-Margin Compensation Structure). NC-URDHR-1 and NC-TRSP-Hybrid-1 formalise the Three-Norn naming framework and the four-layer hybrid post-quantum/temporal architecture respectively. NC-TDC-32 is the central economic contribution of this version. It formalises a digital currency architecture in which multiple stakeholder classes — AI infrastructure operators, financial institutions, sovereign states, and individual participants — coexist as independent issuing classes within a single cooperative cryptographic framework. Each class mints its own coin contingent backed by its own economic activity rather than by shared monetary authority. Phase transitions admit new classes through supply expansion, not through re-pricing of existing coins. Coin denomination is calibrated from inception across micropayment to reserve-asset volume regimes via the monetary identity M·V = P·Q. The infrastructure operator class — the AI companies that build and continuously operate the adaptive security layer — is compensated through a two-component structure: bounded equity recognition at phase transitions (capped, independently audited) plus formula-bound operating margin on continuing services. This two-component compensation model is economically required to keep operating margins moderate and the architecture competitive against established settlement infrastructures. Monetary sovereignty remains exclusively with the issuing class for each contingent; the operator class operates the cryptographic issuance infrastructure but does not exercise monetary authority over any contingent. The architecture is the first formalised digital implementation of the cooperative multi-stakeholder economic model previously demonstrated at continental scale only by the Hanseatic League (twelfth to seventeenth century). All fifteen contributions are documented as conceptual frameworks. Production Concepts (NC-TDC-21, NC-TDC-22, NC-TDC-26 through NC-TDC-32, NC-URDHR-1, NC-TRSP-Hybrid-1) represent architecturally sound design patterns ready for implementation evaluation. Exploratory Concepts (NC-TDC-23, NC-TDC-23a, NC-TDC-24, NC-TDC-25) document underlying architectural ideas requiring further formal research. All specific implementation parameters — quantities, ranges, governance percentages, consortium composition — are illustrative starting points belonging to the institutions that choose to implement the architecture. A dedicated Part 9 — Engineering Considerations and Open Challenges — documents five anticipated technical reviewer questions with referenced solution pathways from current research literature: global consensus latency under M-of-N geographically distributed validation (Sliding Window Key Rotation with Dual-Key Buffers, TLS 1.3 RFC 8446); fuzzy extractor Helper Data leakage in optical entropy capture (Controlled PUF Finite State Machine architectures eliminating Helper Data transmission, addressing Becker 2015); orbital quorum availability under atmospheric and orbital dynamics constraints (Multi-Path Delivery with configurable Grace Periods and Layer 2 graceful degradation); post-quantum zero-knowledge proof latency for autonomous AI agent commerce (Off-Critical-Path ZKP architecture separating HMAC authorisation from asynchronous identity verification); and multi-anchor synchronisation between independent issuance classes (Key-ID and class-identification headers preserving structural separation between technical operation and monetary sovereignty). Part 9 introduces no additional Novel Contributions — it documents that the engineering challenges anticipated by reviewers have established research-backed pathways, demonstrating readiness for Proof-of-Concept implementation phases without modifying or weakening any architectural element documented in Parts 1 through 8. The concept is published as defensive prior art under CC BY-NC-ND 4.0 , preventing future patent claims on the documented conceptual architectures while preserving open non-commercial use for evaluation, research, citation, and standards consideration by IETF, ISO/IEC JTC 1/SC 27, NIST Post-Quantum Cryptography programme, or any institution choosing to adopt all or any independent component of the architecture.

Paul Mincher

This paper establishes the SAFE-Matter™ Offline Evidential Continuity Doctrine as the constitutional governance architecture governing execution authority during periods of partial connectivity loss, ledger fragmentation, degraded synchronisation, isolated runtime operation, and delayed evidential reconciliation within consequence-bearing systems. The framework recognises that modern operational systems increasingly depend upon distributed evidential architectures involving telemetry exchange, runtime admissibility computation, dependency verification, provenance validation, and continuous constitutional execution governance across interconnected environments. Under such conditions, temporary disconnection, synchronisation degradation, network fragmentation, and partial ledger unavailability become operationally inevitable. SAFE-Matter™ rejects both unrestricted assumption-based continuation and rigid fail-stop absolutism during disconnected operation. Instead, the framework establishes bounded constitutional survivability governance under which offline execution authority may continue only where locally retained evidence remains independently verifiable, cryptographically attributable, time-bound, dependency-aware, provenance-consistent, tamper-resistant, and constitutionally reconcilable following restoration of wider evidential synchronisation. The doctrine further establishes that admissibility legitimacy decays more aggressively during fragmented operation than during fully synchronised runtime conditions. Offline authority therefore exists only as a constrained constitutional survivability condition subject to accelerated expiration, bounded execution scope, restricted operational permissibility, mandatory reconciliation, and deterministic collapse into UNKNOWN where constitutional continuity can no longer be sufficiently sustained. This paper forms part of the SAFE-Matter™ constitutional runtime governance architecture governing admissibility, evidential continuity, runtime legitimacy, UNKNOWN-state treatment, dependency integrity, operational survivability, deterministic enforcement, and execution authority within consequence-bearing systems.

Angelo Ferrando, Blondelle Kana Zanlefack, Vadim Malvone

The exponential growth of Decentralized Finance (DeFi) has underscored the critical need for formal verification methods that can reason about the financial properties of smart contracts. Traditional formal methods such as Alternating-time Temporal Logic (ATL) cannot express liquidity properties—guarantees about users' ability to access assets based on wallet balances. We introduce Wallet ATL (WATL), an extension of ATL with wallet predicates and financially constrained strategic operators. WATL ensures that actions are both strategically and economically feasible. We formalize the semantics of WATL, provide model checking algorithms within the VITAMIN framework, and address scalability through the Meta-Agent Abstraction, which collapses all non-coalition agents into a single meta-agent with a sum-aggregated wallet. This abstraction preserves liquidity properties while significantly reducing the verification space. Through case studies such as a crowdfunding smart contract, we demonstrate how WATL formally specifies and verifies liquidity guarantees. Our results show that WATL, implemented in the VITAMIN tool, bridges the gap between multi-agent strategic reasoning and financial correctness, providing a practical step towards the formal verification of smart contracts with liquidity-awareness.

Rohini Lakshané

This working paper is an output of the Community Privacy Residency held in Taipei in 2025. https://community-privacy.github.io/ Keywords: Image-based abuse; non-consensual intimate imagery; evidentiary privacy; protected identity; sexual autonomy; Global South; digital evidence; hash evidence; zero-knowledge proofs; privacy-enhancing cryptography; survivor-auditable governance.

tsenhsuan Chang, tsenhsuan Chang

Decentralized Federated Intelligence Protocol (DFIP) Engineering a Post-Cloud Autonomous Swarm Intelligence Architecture Integrating LEO Satellite Backbone, Edge Vectorized Compute,and Zero-Knowledge Proof Validation for Next-Generation Defense Applications

Amina Ladhari, Heni Boubaker

This study investigates the predictive performance of decomposition-based deep learning models through a focused case study on Ethereum price forecasting. Using hourly Ethereum price data from 5 September 2020 to 13 July 2025, we develop hybrid forecasting frameworks that integrate three signal decomposition techniques—Wavelet Decomposition (WD), Variational Mode Decomposition (VMD), and Empirical Mode Decomposition (EMD)—with a Long Short-Term Memory network enhanced by an attention mechanism (LSTM–Attention). The decomposition methods are first applied to extract multiple frequency components from the original time series, allowing the forecasting model to capture both short-term fluctuations and long-term dynamics inherent in this specific digital asset. Each decomposed component is then modeled using the LSTM–Attention architecture, and the forecasts are aggregated to produce the final prediction. The predictive performance of the proposed models is evaluated using MAE, MSE, RMSE, and MAPE, and the results are compared with benchmark models including ARIMA-GARCH and standard LSTM–Attention. Forecast accuracy is assessed through out-of-sample one-step-ahead predictions, and robustness is ensured by averaging results across 10 independent runs. The empirical results demonstrate that incorporating decomposition techniques substantially improves forecasting accuracy. Among the tested models, the EMD–LSTM–Attention framework achieves the best performance, producing the lowest forecasting errors. While focused on the Ethereum market, these findings highlight the effectiveness of combining signal decomposition and attention-based deep learning architectures to enhance predictive performance in high-volatility cryptocurrency environments.

Shashank Joshi, Wojciech Golab

Reentrancy attacks remain a persistent threat to decentralized applications (DApps), with malicious actors siphoning around 80M USD from the DApp ecosystem last year by exploiting EVM's inter-contract message-passing semantics. Existing research focuses primarily on detection, relying on known attack patterns, and fails to provide deployable solutions that eliminate the vulnerability. Traditional reentrancy guards are similarly limited, offering incomplete coverage across attack variations and lacking robustness against complex DApp interactions. In this paper, we introduce Sentinel, a novel proxy-based approach that mitigates reentrancy vulnerabilities in a type-agnostic way by integrating reentrancy logic directly into the proxy layer, intercepting all calls to the underlying implementation contract. Key features include a dual-mode operational system offering both a gas-optimized internal guard and a high-security external lock registry for cross-contract reentrancy prevention. The proxy also intelligently handles static calls, enabling safe view-function execution while protecting against Read-Only Reentrancy (ROR) attacks. Through rigorous evaluation on a dataset of 70 vulnerable smart contracts, Sentinel achieves 100% security coverage across four major reentrancy attack categories, outperforming existing solutions by over 40%