The predominant formal models for blockchain systems, particularly smart contracts, have largely been drawn from the classical theory of computation, with the finite state machine (FSM) or labeled transition system serving as the primary conceptual tool. However, the FSM relegates the most difficult and novel aspect of a blockchain -- the achievement of consensus in a decentralized environment -- to a complex, often messy, implementation detail that lies outside the formal model itself. But the process of consensus is not an ancillary feature; it is the very essence of the computational phenomenon. To model it faithfully, a new mathematical language is required. The central thesis of this work is that topos theory, the theory of categories of sheaves, provides the native mathematical language for systems defined by local consistency and the construction of global truth.
The digitization of financial markets has produced two classes of platforms that price, in principle, the same state - contingent payoffs: centralized crypto-option exchanges and blockchain-based prediction markets. This paper provides the first option-implied benchmark test of prediction-market pricing for cryptocurrency threshold contracts. For each hour in a matched sample, we compare the Polymarket Yes price with the discounted risk-neutral binary value implied by a listed Binance call option on the same underlying, strike, and maturity, and study the gap between them. In the main September 2023 Bitcoin contract, the mean pricing gap equals 5.6 percentage points across 214 hourly observations (t = 6.46, p < 10^{-9}). Pooling three Binance-compatible Bitcoin threshold markets yields a mean gap of 6.3 percentage points across 287 observations, robust to HAC and block-bootstrap inference. The gap is persistent - with an AR(1) half-life of roughly four hours - yet mean-reverting, consistent with slow information transmission between segmented venues rather than mechanical noise. Cross-sectional regressions reveal that the wedge is largest at low option-implied probabilities and long maturities, a pattern consistent with speculative demand for prediction-market contracts rather than measurement error. A delta-hedged arbitrage proxy remains profitable after conservative transaction costs, though with marginal statistical precision. A Deribit extension on the same three Bitcoin contracts produces a larger pooled gap of 11 percentage points, while a smaller Ethereum exercise yields mixed evidence. The results demonstrate that digital fragmentation of financial markets generates systematic, persistent pricing wedges even for economically identical payoffs.
Contemporary human science is trapped in an extreme state of "involutionary stagnation": academic disciplines are increasingly hyper-fragmented, mathematical equations grow exponentially convoluted, and experimental precision pushes toward physical limits, yet the foundational core paradoxes remain fundamentally unresolved. Quantum mechanics and general relativity stand irreconcilable, the origin of fundamental physical constants remains unexplained, the essence of life and consciousness persists as a black box, and the development of both carbon-based life and silicon-based intelligence has hit a theoretical bottleneck. Based on the comprehensive theoretical ecosystem constructed across 118 core literature milestones of Yuanxian Theory (YXT / YD-T64), this paper systematically demonstrates that the root cause of all scientific involution is not a lack of empirical depth, but rather a failure to elevate the foundational paradigm's dimensionality. Grounded upon four absolutely self-consistent core axioms—True-Circle Self-Consistency (TCSC), Spacetime Uniqueness (STM), Self-Referential Mind-Field Generation (SRM), and Fine-Structure Conservation (FSC)—and validated by machine proof environments via Lean 4, Coq, and ZFC logical systems, Yuanxian Theory achieves a bottom-up reconstruction of all academic fields. This manifesto delineates the revolutionary breakthroughs achieved across mathematics (constructive proofs of the Millennium Prize Problems), physics and cosmology (first-principles constant derivations, dark energy suppression, and gravity-quantum unification), consciousness and life sciences (the topological origin of the 64 genetic codons and mind-field condensation theory), and silicon-based life applications (formal consciousness criteria, cellular hardware architectures, and controllable zero-point energy extraction). Building upon prior research that closed the paradigm loop from the four foundational cornerstones to the Monistic Unified Field, this paper declares that Yuanxian Theory is not a mere patchwork optimization of existing knowledge, but a definitive, wholesale replacement of the three-hundred-year-old scientific paradigm—offering humanity its singular pathway to move beyond disciplinary involution and actualize high-dimensional cognitive elevation. 当下人类科学正陷入一场极致的“内卷式停滞”:学科越分越细,公式越来越复杂,实验精度越来越高,但底层核心矛盾始终无解——量子力学与广义相对论无法统一,物理常数来源不明,生命与意识的本质始终是黑箱,碳基生命与硅基智能的发展陷入理论瓶颈。本文基于元宪理论(YXT / YD-T64)118篇核心文献构成的完整理论体系,系统阐述:所有内卷的根源,不是研究不够深入,而是底层范式没有升维。 元宪理论以四大核心公理——真圆自洽律(TCSC)、时空唯一性律(STM)、自指心场生成律(SRM)与宇宙因子守恒律(FSC)为绝对自洽的根基,依托Lean 4、Coq、ZFC逻辑体系完成机器验证,构建了从64维环面拓扑(YD-T64)到全学科的底层重构。本宣言梳理了元宪理论在数学(千禧年难题的构造性证明)、物理与宇宙学(常数推导、引力-量子统一、暗能量压制)、意识与生命科学(64密码子起源、心场凝聚论)、硅基生命与工程应用(意识判据、元胞架构、零点能可控开发)等领域的革命性突破。元宪理论的前序工作已完成了从四大基石到一元统一场的范式闭合,本文在此基础上宣告:元宪理论不是对现有科学的修补,而是对三百年科学范式的底层替换,是人类走出学科内卷、实现认知升维的唯一通道。
A rule-based logic solver resolves every instance in our benchmark in under 50 microseconds with 100% accuracy; the best frontier language model reaches 65% at best and drops to 23.5% under rendering-robust evaluation (worst case over four surface renderings). We introduce DeFAb (Defeasible Abduction Benchmark), a dataset and generation pipeline that converts four decades of publicly funded knowledge bases into formally grounded instances for defeasible abduction: constructing hypotheses that explain anomalies by overriding defaults while preserving unrelated expectations. Because every hypothesis must pass polynomial-time checks for valid derivation, conservativity, and minimality, DeFAb makes logical rigor the instrument for measuring creativity and theoretical reasoning, scoring the disciplined construction of theory revisions rather than fluent but theory-destroying prose. The pipeline pairs taxonomic hierarchies (OpenCyc, YAGO, Wikidata) with behavioral property graphs (ConceptNet, UMLS) to produce 372,648+ instances across 33.75M materialized rules from 18 sources, in three levels with polynomial-time verifiable gold standards. Four frontier models do not reliably internalize defeasible reasoning: rendering-robust Level 2 accuracy is 7.8-23.5%; chain-of-thought variance (~36 pp) exceeds any inter-model gap; and a matched contamination control isolates a +19.4 pp Level 3 gap. We further release DeFAb-Hard (a 235-instance Level 3 difficulty variant; best model 53.3% vs 100% symbolic) and CONJURE (a kernel-verified transformative-creativity variant of 560 Lean 4/Mathlib instances whose gold answers are definitions the proof kernel did not previously contain, judge-free verifier; a pilot finds zero novel concepts). The same verifier doubles as an exact reward for preference optimization (DPO, RLVR/GRPO). Released under MIT at https://huggingface.co/datasets/PatrickAllenCooper/DeFAb.
Daniel Pereira Alves de Abreu, Octávio Valente Campos, Aureliano Angel Bressan
Objective: This study aims to evaluate the performance of different ARMA-GARCH model specifications in the risk management of major cryptocurrencies, investigating whether the inclusion of exogenous variables improves the calibration of risk measures such as Value-at-Risk (VaR) and Expected Shortfall (ES). Methodology: To achieve this objective, 4,032 specifications of the ARMA-GARCH model applied to the ten main cryptocurrencies in trading were tested. The study incorporated the Fear and Greed Index and Bitcoin Trading Volume as exogenous variables in an ARMA-GARCH-X framework, comparing the performance of the different specifications against an ARMA(1,1)-GARCH(1,1) benchmark. Originality: Despite growing interest in crypto asset risk management, there are still gaps in the literature regarding the effectiveness of incorporating exogenous variables into forecasting models, as well as the increase in the quality of forecasts when using more complex models. Main results: The results indicate that the inclusion of external variables improves risk calibration in some assets, although the gains are marginal and heterogeneous. There is also no single optimal parameterization, requiring ARMA orders, GARCH specifications, and error distributions to be adjusted for each cryptocurrency. Theoretical/methodological contributions: From a methodological point of view, the study contributes by demonstrating the importance of specific calibration of ARMA-GARCH models for different cryptocurrencies in risk estimation. Furthermore, the results suggest that, although more complex models can improve tail risk estimation, the gains in predictive power over simpler models are limited. Keywords: Cryptocurrencies; Risk Management; ARMA-GARCH; Value-at-Risk; Expected Shortfall.
Deniz Erer, Tuna Can Güleç, Özge Korkmaz, Elif Erer
Rapid developments in blockchain, decentralized finance, and tokenization have raised the question of whether Sukuk can complement technology-based financial assets. This study compares the time-varying efficiency and multifractal dynamics of Sukuk indices, DeFi tokens, lending and borrowing tokens, and a FinTech index from May 25, 2020, to November 29, 2023. Using TGARCH, nonlinearity and long-memory tests, MF-DFA, and MF-DCCA, the study examines shock persistence, asymmetric volatility, market efficiency, and cross-market dependence. The findings show that negative shocks increase volatility more strongly than positive shocks and that all markets display nonlinear and multifractal behavior. Sukuk indices, particularly RMENA and RDJSUKUK, show lower market deficiency values than most technology-based assets. However, persistent cross-correlations indicate that Sukuk is not a direct substitute for these assets. Rather, Sukuk may serve as a relatively stable and efficient complementary asset in technology-exposed portfolios.Key Words: Sukuk, DeFi assets, Tokenization, Financial Economics, MF-DFA, MF-DCCAJEL Classification: F65, E44, G15, C58
This study presents the design and implementation of a blockchain-based decentralized portfolio management system that enables secure, immutable, and transparent storage of user records. The system is developed using Ethereum smart contracts and evaluated within a testing environment consisting of Remix IDE, Ganache, and MetaMask. The proposed architecture allows authorized actors to create records while enabling users to access and verify their data through blockchain-based identity mechanisms. Experimental results, based on gas consumption and insertion-time measurements, demonstrate that although smart contract deployment incurs relatively high initial costs, routine operations such as record insertion and retrieval remain efficient and predictable. The findings highlight the practical feasibility of the proposed system, while also revealing challenges related to scalability, cost variability, and system usability.
Ethereum and Hyperledger Fabric are architecturally heterogeneous‚ with Ethereum using the Ethereum Virtual Machine to execute Solidity contracts with order-execute transactions and pseudonymous ECDSA-based identity․ As Fabric runs Go chaincode under an execute-order-validate enforcement model‚ with MVCC conflict detection‚ X․509 certificate-based identity management‚ and an explicit key-value state API‚ smart contracts cannot be written to run on both Ethereum and Fabric without major duplication of effort‚ namely‚ maintaining two separate codebases‚ conducting two separate security audits‚ and manually re-implementing complex code․ This thesis aims to both design and test a Universal Intermediate Representation (UIR) for the migration of smart contract logic from Ethereum to Hyperledger Fabric in a structured, auditable and repeatable way. In the spirit of Design Science Research (DSR) (Peffers et al., 2007), this study investigates five portability barriers to the extent that they can be identified (PB-1 to PB-5) and relates them to six design requirements (R1 to R6). Based on this, two-stage prototype pipeline is created in Python, a front-end based on Solidity contracts and a back-end which generates Hyperledger Fabric Go chaincode. In three canonical case studies‚ SimpleStorage‚ Escrow and SimpleToken‚ we evaluated the translation with respect to four dimensions: feature translation rate‚ semantic approximation accuracy‚ barrier coverage and compilation success․ Out of the 14 categories of Solidity features‚ 6 (43%) are completely abstractable‚ 4 (29%) can be approximated with semantic gaps SG-1 to SG-2‚ and 4 (29%) are architecturally non-portable at the contract level․ All three Hyperledger Fabric Go chaincodes built using the UIR approach compiled successfully with go build‚ using Go version 1․22․5‚ showing the feasibility of the approach with Go․ The thesis is not about the fact that UIR is a production ready tool. The pipeline has no total automation; in the 3 case studies, the processing of function bodies was done manually in Stage 1. Also, the prototype currently only approximates 256-bit integers. The actual contribution is conceptual: It suggests a structured, auditable way to detect and overcome portability issues from Ethereum to Hyperledger Fabric. The master thesis consists of 94 pages; it contains 9 figures, 28 tables, 2 appendices, and 42 references.
In the digital revolution driven by blockchain technology, smart contracts emerge as a paradigm-shifting tool, poised to redefine traditional business practices across multiple domains. smart contracts stand as a cornerstone of innovation, promising to revolutionize the way we engage in business trustlessly. Driven by the pioneering spirit of exploration, this research delves into the expansive realm of smart contract use cases and applications, seeking to unveil the transformative potential they hold. Through meticulous analysis and case studies, this research illuminates the diverse array of scenarios where smart contracts can revolutionize processes, enhance accountability, and streamline operations in sectors such as finance, supply chain management, healthcare, and government services. By fostering collaboration and innovation, we seek to unlock the full potential of smart contracts, ushering in a new era of efficiency, integrity, and trust in the digital age while illuminating the path towards unlocking the untapped opportunities presented by smart contracts, reshaping the future of digital economies and organizational paradigms. Areas of application like decentralized Finance (DeFi), Non-Fungible Token (NFT), Regenerative Finance (ReFi) and many more where all discussed extensively.
Abstract Internal audit functions in U.S. manufacturing and retail face a growing disconnect between increasingly sophisticated fraud schemes and legacy detection methods that rely on static rules and manual sampling. The Association of Certified Fraud Examiners estimates that organizations lose approximately five percent of annual revenue to fraud, with manufacturing and retail sectors particularly vulnerable due to complex supply chains, high transaction volumes, and decentralized operations. Modern fraud schemes have evolved well beyond simple expense manipulation; they now involve multi-party collusion, cyber-enabled invoice fraud, supply chain manipulation through fictitious vendors, and coordinated point-of-sale skimming networks. Traditional audit approaches typically cover only three to five percent of transactions through periodic sampling, leaving the vast majority of activities unexamined and creating significant windows of exposure. Learningter demonstrates how applied AI (machine learning anomaly detection, natural language processing, and agentic AI) transforms fraud detection from reactive forensics into proactive, continuous assurance. Drawing on five anonymized case studies from active industry engagements, the presentation illustrates measurable outcomes: false-positive rates reduced by up to 70 percent, detection time compressed from months to minutes, and coverage expanded from sample-based testing to full-population analysis. Each case maps legacy controls against AI-augmented alternatives, providing a clear migration pathway. In particular, Agentic AI enables autonomous and continuous monitoring through self-correcting feedback loops that recalibrate detection models in real time without requiring manual intervention, adapting dynamically to emerging fraud patterns and shifting transaction behaviors. The presentation addresses practical adoption challenges (data quality, algorithmic bias, SOX/ICFR compliance, and change management) and offers a structured readiness framework for consulting engagements or dissertation research. Grounded in Boyer's Scholarship of Application, this work connects data science and auditing to real-world problems, demonstrating how cross-disciplinary collaboration produces actionable improvements in governance and risk management. The research is directly relevant to doctoral candidates seeking applied dissertation topics with measurable industry impact and to faculty developing curricula that bridge theoretical foundations with practitioner-oriented pedagogy.
Decentralized finance exposes supervisors to fast-moving, networked credit risks. General-purpose LLM agents fit this setting poorly: they over-read weak evidence and recommend high-stakes interventions, while existing evaluations offer no regulator-aligned way to measure the resulting false alarms. We introduce DeXposure-Claw, a forecast-grounded agentic supervision system that routes LLM decisions through structured evidence: (1) DeXposure-FM, a graph time-series foundation model, forecasts future exposure networks; (2) deterministic monitors and stress scenarios then turn those forecasts into typed alerts, attribution signals, and scenario evidence; and (3) data-health and confidence gates constrain escalation before DeXposure-Claw emits auditable supervisory tickets with rationales. We further develop DeXposure-Bench, a six-axis evaluation harness, whose decision axis scores tickets against a regulator-aligned absolute-loss ground truth and an explicit false-intervention rate. Experiments on five years of weekly real data fully support our system. Code is at https://github.com/EVIEHub/DeXposure-Claw.
Traditional philanthropic organizations often suffer from lim ited transparency, where donors have minimal visibility into how their contributions are utilized after donation [1,14]. To addressthisissue, this paper presents NGO-Chain, a hybrid Web3 platform designed to im prove accountability and transparency in charitable fund management. The proposed system utilizes a milestone-based conditional escrow mech anism in which donated funds are locked within blockchain smart con tracts and released incrementally only after administrative verification of uploaded proof documents stored on the InterPlanetary File System (IPFS) [4,5]. The architecture combines React-based frontend interfaces, Spring Boot middleware, decentralized IPFS storage, and Ethereum/Polygon smart contracts to create a scalable hybrid infrastructure capable of supporting real-time public transaction monitoring [14,12]. In addition, the platform integrates donor reputation tracking and blockchain-backed transaction auditing to strengthen trust between donors and NGOs [6,7]. By com bining decentralized financial management with milestone verification workflows, NGO-Chain provides a secure and transparent framework for milestone-driven charitable donations while reducing dependency on cen tralized trust mechanisms.
The rapidly expanding landscape of Web3 and the metaverse profoundly accentuates the escalating challenge of rigorously assessing and strategically selecting foundational Layer-1 digital blockchain platforms. Decision-makers frequently contend with the imperative of rational choice amidst a complex confluence of often conflicting technological attributes. This study directly addresses this critical exigency by utilizing robust benchmarking and validation for the comparative ranking of 10 prominent blockchain platforms. By applying a suite of five established multi-criteria decision-making (MCDM) methods, namely TOPSIS, ARAS, RAPS, RAMS, and RATMI, a comprehensive evaluation is undertaken, scrutinizing performance across three pivotal criteria categories: performance/scalability, security, and economic/activity. The weights for the entire criteria set were determined using the objective entropy method. Using the entropy approach to determine weights based on randomness, the criteria weights were determined as follows: Speed 12.9%, Market Cap 7.2%, Hash Rate 43.7%, Time to Finality 12.1%, Total Transactions 10.8%, and Number of Nodes 13.3%. The empirical analysis consistently identifies Bitcoin as the top-ranking platform, securing first position across all five MCDM methodologies. This finding validates its unparalleled robustness and security based on the defined criteria. Hyperliquid and Sui also emerged as exemplary performers, consistently exhibiting strong aggregate scores and securing second and third positions, respectively. Conversely, other blockchains, such as the BNB Chain and Tron, demonstrated significant ranking volatility across the different evaluation methods. This study provides a validated, data-driven benchmarking tool, offering stakeholders a transparent framework for strategic decision-making. This application contributes to the conceptual accuracy of evaluating sustainable digital infrastructure.
Seyed Salar Ghazi, Kaiwen Zhang, Mehdi feizi, Hans-Arno Jacobsen
Hierarchical Federated Learning (HFL) enables scalable collaborative model training across distributed devices while preserving data privacy. However, existing HFL client selection mechanisms suffer from a fundamental strategic inefficiency. By prioritizing stability over Pareto efficiency (PE), they produce suboptimal resource allocations, and without strategy proofness (SP), participants are incentivized to misrepresent their true preferences, both failures degrading system overall welfare in the Pareto sense in practice. To address it, we propose SCOPE-FL (Strategy-proof Chain-based Optimal pareto efficient Federated Learning), a synchronous HFL framework that formulates client selection as a two-sided school choice problem solved through the Top Trading Cycle (TTC) algorithm that simultaneously guarantees PE and SP. For reward distribution, SCOPE-FL employs a scalable Shapley value approximation based on One-Round Reconstruction (OR), ensuring compensation proportional to each client's contribution. The entire mechanism executes via blockchain smart contracts, providing the tamper-proof environment required for the SP guarantees to hold in practice. A comprehensive evaluation on MNIST, Fashion-MNIST, and CIFAR-10 demonstrates that SCOPE-FL outperforms state-of-the-art approaches, including DA, IAS, and other methods across model accuracy, convergence rate, and reward efficiency, while achieving communication latency comparable to DA and blockchain overhead significantly lower than DA at scale.
Consensus protocols form the core of blockchains and other replicated state machines, ensuring that all correct nodes process the same totally ordered log of input transactions. In fault-free executions, performance is driven by the good-case transaction latency -- the time between a transaction becoming known to all nodes and its confirmation by the consensus protocol -- which depends on both how frequently proposals are made and, once made, how quickly they are confirmed. While prior work has established tight lower bounds on confirmation latency that modern protocols already achieve, it remains open whether the inter-proposal time can be further reduced below the state-of-the-art of one network delay. We introduce Gatling, an atomic broadcast protocol that achieves arbitrarily small inter-proposal times under rotating leader schedules; in particular, smaller than the network delay. Gatling runs multiple parallel instances of a black-box atomic broadcast protocol and staggers their proposal schedules to generate proposals in faster succession than state-of-the-art protocols. A deterministic interleaving rule merges the outputs of these instances into a single global log. We analyze the effects of head-of-line blocking caused by crashed leaders, and derive Gatling's optimal number of parallel instances. We further study the impact of Gatling on predictable validity and present two variants that retain this property. Finally, our experiments confirm that Gatling can be used with off-the-shelf component protocols to achieve low latency without fine-tuning the component protocol for minimum latency.
Advances in Artificial Intelligence (AI) have led AI for Theorem Proving to become a promising means of formally verifying computer systems. Whilst formal verification is traditionally reserved for safety-critical systems due to the required amount of expertise and effort, AI can help to automate a large amount of this workload and make it far more accessible. Blockchain-based systems are becoming increasingly popular and are frequently targeted by malicious actors, often resulting in huge financial losses, highlighting the need to better verify these systems and mitigate vulnerabilities. Arguably the most important component of these systems is the consensus protocol, which allows nodes to agree on decisions in a potentially adversarial environment. In this paper, we improve upon IsabeLLM, the automated theorem proving tool in Isabelle. Namely, we implement a Retrieval-Augmented Generation framework, Error tracing and counterexample generation for improved context supplied to the Large Language Model. Compatibility with the latest version of Isabelle and Sledgehammer is also implemented for improved efficiency. We compare the performance of the two versions of IsabeLLM in their ability to complete the verification of Bitcoin's Proof of Work consensus.
Victor Gao, Wolfgang Grieskamp, Vineeth Kashyap, George Mitenkov · 8 authors
Move is a smart-contract language used to execute transactions on the Aptos blockchain. Move programs execute in a sandboxed VM as typed bytecode. The VM statically verifies foundational safety properties like type safety and reference safety at code loading time. In principle, this design gives strong guarantees for Move. However, the static verification logic is complex and continually evolving with the language; like any software, it is not immune to bugs. In a live blockchain setting, a missed rule violation can translate directly into loss of assets, forged authority, or unrecoverable corruption of on-chain state. For this reason, Aptos relies on defense-in-depth runtime safety checks that independently verify the critical invariants during execution, providing protection against latent verifier bugs and malicious bytecode. This paper motivates and describes the runtime safety checks for Move on Aptos.
Klaus M. Frahm, Leonardo Ermann, Dima L. Shepelyansky
According to the recent Wealth Thermalization Hypothesis (WTH) the wealth inequality in the world is described by the Rayleigh-Jeans (RJ) thermal distribution of interacting agents in a society with social stratification. In this concept, the wealth layers of society are associated with energy levels from a nonlinear dynamical system conserving two integrals of motion being total energy and probability norm. This leads to RJ condensation and the formation of a huge poverty phase of low wealth and a tiny oligarchic phase that captures a main part of total society wealth. This RJ phenomenon has similarities with self cleaning in multimode optical fibers and constraint driven condensation in various physical systems. We analyze real Lorenz and Pareto curves for wealth of households in countries and the world, Gross Domestic Product of countries, market capitalization of companies at stock exchange of Hong Kong, Shanghai, London, bitcoin transactions, world trade between countries and show that the WTH theory gives a good description of these curves. On the basis of this comparison we argue that the RJ thermal distribution provides a universal description of wealth inequality in the world.
<sec> <title>BACKGROUND</title> Home- and community-based services funded through the Medicaid program account for $125 billion in annual federal and state expenditure (Center for Medicare Services, 2023), serving millions of elderly and disabled individuals who receive care in private residences rather than institutional settings. The decentralized nature of home care delivery creates fundamental accountability challenges: services occur in private homes largely beyond direct supervisory oversight, making home care one of the highest-risk categories for Medicaid fraud. Nationwide investigations by the HHS Office of Inspector General from 2011 through 2015 recovered $975 million in fraudulent home health claims (OIG, 2016). A 2024 New York State Comptroller audit documented $14.5 billion in Medicaid personal care payments made without required electronic visit verification (Office of the New York State Comptroller, 2024). In Massachusetts, a 2024 federal conviction established that a home health agency co-owner defrauded MassHealth of at least $100 million over four years through billing for services never rendered (U.S Department of Justice, 2024). Electronic visit verification was mandated under the 21st Century Cures Act (Pub. L. No. 114-255, § 12006, 2016) to address these vulnerabilities by requiring real-time electronic capture of six data elements at each Medicaid-billable visit: service type, recipient identity, date, location, provider identity, and start and end times. MassHealth selected Sandata Technologies as the Commonwealth's designated EVV aggregator, with hard billing edits scheduled no earlier than July 2026 (MassHealth, 2025). Despite widespread EVV implementation nationally, no published peer-reviewed study has empirically characterized visit-level EVV anomaly patterns from operational agency data or documented the industry-wide pre-submission exception management infrastructure through which GPS verification failures are converted into billing-ready records before aggregator transmission. Direct telephone communication with Axxess customer support on June 16, 2026 confirmed that most agencies use the EVV Exception Center and that through this workflow an agency can achieve 100% compliance (Axxess, personal communication, June 16, 2026). WellSky customer support confirmed on the same date that flagged visits can be changed to verified visits prior to state aggregator transmission (WellSky, personal communication, June 16, 2026). </sec> <sec> <title>OBJECTIVE</title> This study had two primary objectives. First, to characterize the prevalence, typology, and distribution of EVV anomalies through quantitative analysis of 15,172 de-identified visit records from an operational Massachusetts Medicaid home care agency during the pre-enforcement window preceding MassHealth hard billing edits. Second, to document the industry-wide pre-submission exception management infrastructure across six major documentation platforms through direct vendor communication and systematic platform review, and to characterize the response pattern of Massachusetts home care agencies to voluntary research participation requests. </sec> <sec> <title>METHODS</title> This study employed a five-agency mixed-methods comparative design. Agency A: cross-sectional observational analysis of 15,172 de-identified Sandata EVV visit records from January 1 through May 20, 2026 (140 days; 120 unique patients; 52 caregivers; 11 procedure codes). Written data use authorization was obtained from Agency A leadership. Six anomaly categories were analyzed: GPS location exceptions (GPS_EXCEPTION field); non-verified visit status (VISIT_STATUS field); systematic minimum-time patterns (ACTUAL_TIME = 8.0 minutes exactly); manual time adjustments (both ADJUSTED_IN_TIME and ADJUSTED_OUT_TIME populated); batch backdating (entry creation timestamps versus visit dates); and geographic impossibility (Haversine formula applied to sequential GPS coordinates). Financial exposure was calculated by applying verified 2026 MassHealth fee schedule rates from 101 CMR 350.00 to actual billing units in non-verified visit records. Agency B: operational observation of Axxess Exception Center pre-submission workflows. Agencies C, D, and E: structured professional interviews and research participation solicitation. Twenty additional Massachusetts Medicaid-enrolled agencies were contacted by telephone for voluntary participation between June 15 and 16, 2026. Direct primary source telephone communication was conducted with Axxess and WellSky customer support on June 16, 2026, including step by step exception center workflow on how to correct a mismatched visit. Systematic review of published technical documentation was conducted for six major documentation platforms: Axxess, WellSky/Kinnser, HHAeXchange, AlayaCare, AxisCare, and Alora Health. All analyses were conducted in Microsoft Excel using raw Sandata export data. </sec> <sec> <title>RESULTS</title> Agency A: GPS exception flags were present in 12,683 of 15,172 visits (83.6%). The GPS_CALL_IN_DISTANCE field, available for 4,983 records, revealed a mean clock-in distance of 12,922 meters from the patient address, a median of 391 meters, and a maximum of 156,956 meters (97.5 miles). A total of 1,410 visits (9.3%) recorded distances exceeding 10 kilometers and 516 visits (3.4%) exceeded 50 kilometers. Non-verified visits totaled 3,333 (22.0%), with estimated potential financial exposure of $246,610 for the five-month period applying verified 2026 MassHealth rates (101 CMR 350.00), annualizing to approximately $642,948 at this single agency. A total of 1,992 visits (13.1%) were documented at exactly eight minutes duration, appearing across four procedure codes including G0299 registered nurse and G0300 licensed practical nurse. Employee E18 recorded 1,304 of 1,441 visits (90.5%) at exactly eight minutes — 6.9 times the agency-wide rate — across four service types, sustained over five months without attenuation. Manual time adjustments affected 601 records (4.0%), with five employees accounting for 299 of 601 adjusted visits (49.8%). Sequential visit records required implied travel speeds of 87 to 230 miles per hour between Massachusetts communities, constituting mathematical proof of fabricated location entries. A weekly batch backdating pattern was identified in which no real-time EVV entries were generated Monday through Thursday, followed by retroactive bulk entry on Friday. Agency B demonstrated systematic use of the Axxess Exception Center to normalize GPS exceptions before Sandata submission, self-reporting 96% compliance — illustrating the EVV Compliance Paradox. Agency C quality assurance professionals identified Drive-By Clock-In Fraud, in which caregivers clock in from within GPS geofence range of a patient's address without entering the premises. Agency D identified a theoretical Complicit Patient vulnerability through dual-device registration. Agency E declined research participation, stating their EVV data was problematic and they did not wish attention called to their records. Of 20 additional agencies approached, zero agreed to participate; responses included -direct refusals, non-responses, and one representative who stated no staff member had any knowledge of EVV. Vendor communication confirmed that most agencies use pre-submission exception management and that flagged visits can be reclassified as verified prior to aggregator transmission (Axxess, personal communication, June 16, 2026; WellSky, personal communication, June 16, 2026). Further documented photographic evidence from Axxess help system showing: The Exception Center workflow step by step, their own template example with a geographically mismatched visit, including a four- day visit error and correction steps: “select a reason code, type clinician signature, click update visit.” Upon completion, the visit is a verified record regardless of the original GPS mismatch or duration anomaly. </sec> <sec> <title>CONCLUSIONS</title> EVV data contains substantially more actionable fraud intelligence than current practice extracts. Six anomaly categories affecting thousands of visits in a single Massachusetts agency over five months reflect systemic rather than isolated non-compliance. Geographic impossibility requiring 87 to 230 mph implied travel speeds constitutes mathematical proof of GPS location fabrication. Employee E18's sustained eight-minute visit pattern across 1,441 visits and four procedure codes including licensed skilled nursing is statistically impossible as a naturally occurring clinical pattern. The estimated $246,610 in potential financial exposure over five months illustrates the scale of program integrity risk operating within apparently compliant EVV systems. The EVV Compliance Paradox - confirmed by direct vendor communication - demonstrates that compliance rates in GPS-based systems may reflect exception management sophistication rather than care delivery integrity, including the step by step exception center correction workflow that verifies a patient visit with clear original GPS mismatch. The 0% research participation rate across 21 Massachusetts agencies approached, including one that explicitly cited concern about its own EVV data, suggests widespread institutional awareness of compliance vulnerabilities. GPS-based EVV is necessary but structurally insufficient. Hardware-anchored verification requiring physical presence inside the patient's home, supervised biometric enrollment, and cryptographic visit records are the architectural requirements that GPS-based systems cannot meet. Six f
Alexandra Conda, Ștefan Găman, Raul Cristian Bag, Miruna Mazurencu-Marinescu-Pele · 6 authors
Abstract This study investigates the relationship between Facebook sentiment and Bitcoin market dynamics using AI-based emotion detection. We analyze 120,000 Facebook posts collected via CrowdTangle alongside Bitcoin financial data from the Blockchain Research Center, covering 2015–2023. Employing FinBERT for sentiment classification, we develop novel compound sentiment scores that integrate text-based sentiment with Facebook’s multi-reaction engagement system, then apply four analytical components: sentiment analysis, Dynamic Topic Modeling, sentiment-based trading strategies, and machine learning volume prediction. Results demonstrate that Facebook sentiment has substantial predictive power for Bitcoin trading volume. Sentiment-based trading strategies significantly outperform buy-and-hold, achieving superior cumulative returns and risk-adjusted performance. For volume prediction, Linear Regression and Bidirectional LSTM achieve comparable test performance, indicating that model complexity does not guarantee superior prediction. Topic modeling reveals that cryptocurrency investment and trading discussions dominate Bitcoin discourse on Facebook, with themes evolving over time in response to market conditions. This research contributes by being the first to apply post-level NLP sentiment analysis of Facebook data to cryptocurrency markets, extending beyond the Twitter and Reddit focus of prior research. The findings provide practical tools for traders and analysts navigating volatile digital asset markets while demonstrating that Facebook’s demographically diverse user base and rich reaction system offer unique advantages for sentiment quantification.
Overview This research introduces a production-ready agentic AI system designed to mitigate catastrophic forgetting in Large Language Models (LLMs). By anchoring six prime-indexed embedding rows $\{2, 3, 5, 7, 11, 13\}$ as fixed reference points, the system maintains historical knowledge with near-zero forgetting while requiring minimal memory overhead. Key Technical Contributions The Core Innovation: Prime Anchoring Topological Invariant: Utilizes the first six primes to create stable reference points. Mechanism: Anchor rows are snapshotted after initial training; gradient updates are blocked for these specific rows during subsequent tasks. Sparsity & Memory: Only 6 out of ~50,000 rows (0.01% of parameters) are used, resulting in an O(1) memory overhead of only 48–96 KB. Mathematical Foundation Euler Attenuation Product: These six primes account for 97.85% of total spectral weight, defined by: $$\Lambda = 1 - \prod_{p\in \{2,3,5,7,11,13\}}(1 - p^{-0.5}) \approx 0.9785$$ Spectral Trap: The anchors create a spectral peak at $\sigma = 0.5$, aligning with the critical line of the Riemann Hypothesis. Green-Tao Quantification: Establishes a decay law for coherence: $$\text{coherence}(k) = 2.1546\times k^{-0.8186} + 0.1218$$ Performance Metrics (Selected Models) Model Task C Accuracy Forgetting Std Dev Zero Forgetting Runs GPT-OSS-20B 92.3% ±1.28% 0/5 Sarvam-30B FP8 95.9% ±2.82% 0/5 Mixtral-8x7B FP8 89.7% ±2.53% 0/5 DeepSeek-V2-Lite FP8 95.4% ±0.21% 3/5 Multi-Agent System Architecture The system employs four specialized agents to manage task routing and classification: Classifier Agent: Routes documents based on keywords. Topic Agent: Performs unsupervised domain topic extraction. Sentiment Agent: Conducts autonomous tone analysis. Decision Agent: Acts as the final arbiter for task approval and routing. Efficiency: Achieves 96–100% classification accuracy with inference times between 252–446ms. Comparative Analysis The topological approach outperforms traditional methods by balancing plasticity and stability: Method Memory Cost Performance/Issue EWC 4.4 GB/task Memory intensive; fragments GPU Experience Replay O(k) Buffer growth issues; lower accuracy HOPE-like 2.3 GB High forgetting resistance but lower accuracy (88.1%) Topological AI 48 KB 99.5% accuracy; highly efficient Biological and Theoretical Insights Biological Analogy: The system treats 0% forgetting as a pathology. By allowing 99.99% of embedding rows to remain plastic, the model mimics biological brains that prioritize selective forgetting to facilitate adaptation. Riemann Hypothesis Connection: The research posits that the specific selection of the first six primes creates a unique "spectral trap" at $\sigma = 0.5$. Including any prime $\geq 17$ disrupts this trap and destroys the stability condition. Production Readiness and Certification TOPO-2026 Track II: The system passed all rigorous benchmarks, including Task C accuracy ($\geq 80\%$), Combined Forgetting ($\leq 10\%$), and O(1) memory overhead. Deployment: Fully compatible with commodity hardware, specifically tested on NVIDIA RTX PRO 6000 Blackwell GPUs. Resources: Implementation code, technical reports, and proof documents are available via the project's GitHub and Zenodo repositories.
Sovereign Personal Evidence is a defensively disclosed local-first architecture for preserving externally issued, high-assurance signed assertions and their verification transactions as durable, user-controlled evidence artifacts. The architecture extends the deterministic provenance engine first disclosed in Sovereign v1.0 (DOI 10.5281/zenodo.19056811) to a new evidence class: externally issued personal assertions such as verifiable credentials, selective-disclosure presentations, zero-knowledge identity proof results, and passport- or NFC-derived verification artifacts. The disclosed system ingests an external assertion, validates it according to its native trust model, cryptographically binds it to the specific request context and a local holder anchor, records it as a typed event in an append-only hash-chained personal provenance ledger, and exports a portable proof bundle for later independent verification — without requiring continued access to the original verification platform. This document constitutes a public defensive disclosure establishing prior art for the disclosed combination of elements, including composite assertion-to-context binding, a two-mode verification-engine fork, timestamped status and revocation evidence preservation, minimal-disclosure evidence packaging, and a personal evidence threat model. Publication is intended to prevent future patent claims covering the same or substantially similar system design.