Mehmetaj, Ilir

9 files · 12 MB · jpeg, pdf, pngdeclared

ABSTRACT - Seraphim Skin v1.0 Can the Seraphim UV-Protection Layer Reshape the USD 2.26 Trillion Clothing Industry? The global apparel market - worth roughly USD 1.8 trillion in 2024/25 and projected to reach USD 2.26 trillion by 2030 - is not the target of a niche within it, but of a material layer that integrates into any garment across the whole of it. Offered two identical garments at one price, one plain and one with permanent UV protection and passive cooling woven into the material itself, the buyer needs no persuasion. That is the lever, and the trend. The stakes are health, not fashion alone: ultraviolet radiation drives 83 percent of melanoma, the World Health Organization projects a 50 percent rise in incidence by 2040, two billion people work outdoors under rising heat stress, and fifty million live with UV-sensitive medical conditions. For millennia the answer has been oils, fabrics, and chemicals - each partial, each reapplied. This concept proposes that the protection be the fabric itself: permanent, physical, effective from first wear to hundredth wash. Seraphim Skin is a 35-micrometre, 45-gram-per-square-metre multilayer laminate - the same material family documented across this series, tuned here to the everyday-climate design point rather than the fire or energy-harvest ones. A diamond-like-carbon outer layer reflects ultraviolet across UV-A and UV-B through bandgap engineering and carries a self-cleaning lotus surface; a roughly 20-micrometre gradient graphene layer manages the infrared; a passivation layer isolates the active layer; a bio-compatible inner layer carries comfort and sensing. The concept states its layer hierarchy explicitly to prevent a contradiction: the 20-micrometre graphene layer carries the thermal work through thickness-tolerant mechanisms - isotope (¹²C/¹³C) phonon scattering and emission in the 8-to-13-micrometre atmospheric window - while the electronic decoupling of the Wiedemann-Franz violation at the Dirac point (Nature Physics, 2025) belongs only to the tens-of-nanometre sensor zone that powers biosensors from body heat. The fabric manages the full electromagnetic spectrum in one architecture: permanent UV-B and UV-A reflection at the material level, and emission of the body's 9-micrometre heat through the atmospheric window to the cold sky - the passive daytime radiative cooling demonstrated in peer-reviewed metafabrics. The thermal behaviour is described as a bounded rectifier (an outward bias in the class of 1.3-to-2-to-1, not an absolute valve), and the cooler-than-skin result is stated with its mechanism so it cannot be read as a violation of thermodynamics. Layer count is a design variable - the N-Factor: stacking N laminates compounds protection and answers what happens if a layer is damaged, since the remaining skins keep functioning and performance degrades gradually. Twelve novel contributions (NC-SKN-1 through NC-SKN-12) span five garment markets - outdoor labour, children, medical photoprotection, defence signature management, and luxury performance - and every surface between people and the sun: building and glazing envelopes, closed UV-filtering radiatively cooled desert greenhouses, and vehicle surfaces that lower cabin heat load to extend electric-vehicle range. Those who bear the downstream cost of UV damage and heat stress - insurers, employers with outdoor workforces, public health systems - are the structural payers: prevention shifts value from treatment toward avoidance. Because sub-ambient radiative-cooling textiles exist as prior art, each contribution is framed by its integration rather than any single function. Every load-bearing element exists at high maturity in another industry; the integrated laminate stands at TRL 2 to 3. The decisive experiment is a fabric coupon under a solar simulator - measured UV transmission, 8-to-13-micrometre emissivity, and sub-ambient temperature difference against exposed skin - turning the central claims into data for a few thousand euros. All parameters are theoretical design estimates requiring independent validation. This concept consolidates and refines a disclosure of March 2026 and operates at the everyday-climate design point (around 50°C), standing independent of the fire branch (600 to 1,200°C) and the energy-harvest branch of the same laminate family. The twelve novel contributions are placed on the public record of prior art as of the Zenodo publication date under CC BY-NC-ND 4.0, preventing future patent claims on these specific architectures by any party. Ilir Mehmetaj | Independent Concept Developer | CC BY-NC-ND 4.0 | 2026

HUDA, SABIN

1 files · 729 MB · csvdeclared

AMLNet - Synthetic Anti-Money Laundering Benchmark Dataset, Version 2.0 Relation: Is new version of identifier: 10.5281/zenodo.16736515 AMLNet is a synthetic anti-money laundering benchmark dataset created for machine learning evaluation. This Version 2.0 release is associated with the paper: "AMLNet: A Knowledge-Guided Synthetic Benchmark for Machine Learning Evaluation in Anti-Money Laundering" The dataset contains a fixed synthetic benchmark instance generated using the AMLNet framework. It includes approximately 1.09 million transactions over a 195-day simulation period, from 13 October 2025 to 27 April 2026. The benchmark contains 1,411 suspicious transactions, corresponding to a suspicious rate of approximately 0.13%. The dataset is fully synthetic. The accounts, transactions, timestamps, locations, balances, metadata, labels, and customer activity patterns do not correspond to real customers, real institutions, or real banking activity. AMLNet was designed to support machine learning experiments under rare-event anti-money laundering conditions. The dataset includes ordinary transactions and suspicious transaction sequences representing structuring, layering, and integration patterns. Suspicious activity is embedded within ordinary account activity to make the detection task more realistic and non-trivial. CONTENTS This release includes: - the fixed AMLNet Version 2.0 synthetic transaction dataset; - transaction labels; - laundering typology labels; - transaction metadata; - dataset documentation and column descriptions. Evaluation scripts will be added to this Zenodo record after publication of the associated paper. The AMLNet generator source code is not included in this release due to security concerns. DATA FORMAT The main dataset is provided as a CSV file with the following columns: - step: Sequential simulation step or transaction index. - type: Transaction type, such as BPAY, CASH_OUT, DEBIT, EFTPOS, NPP, OSKO, PAYMENT, or TRANSFER. - amount: Transaction amount in Australian dollars. - category: Transaction category, such as housing, food, transport, recreation, healthcare, education, utilities, shell company, property investment, cryptocurrency, or other. - nameOrig: Originating account or customer identifier. - nameDest: Destination account, customer, or merchant identifier. - oldbalanceOrg: Originating account balance before the transaction. - newbalanceOrig: Originating account balance after the transaction. - isFraud: Binary label used for suspicious/fraudulent transaction detection. - isMoneyLaundering: Binary AML label, where 1 indicates suspicious money laundering activity and 0 indicates ordinary activity. - laundering_typology: Laundering typology label. Values include normal, structuring, layering, and integration. - metadata: JSON-style metadata containing timestamp, location, device information, payment method, risk indicators, and typology-specific details where applicable. - fraud_probability: Model-generated or risk-score field used in selected experiments. This field may be empty for some records. - hour: Hour of transaction. - day_of_week: Day of week. - day_of_month: Day of month. - month: Month number. LABELS The dataset includes two binary label fields: - isFraud: Binary suspicious/fraudulent transaction label. - isMoneyLaundering: Binary anti-money laundering label. The laundering_typology column provides the suspicious activity type for labeled suspicious transactions. Values include: - normal - structuring - layering - integration DATASET STATISTICS - Total transactions: approximately 1.09 million - Suspicious transactions: 1,411 - Suspicious transaction rate: approximately 0.13% - Simulation period: 195 days - Simulation dates: 13 October 2025 to 27 April 2026 - Generated customer accounts: 10,000 - Observed graph nodes: 11,000, including customer and merchant nodes - Transaction types: BPAY, CASH_OUT, DEBIT, EFTPOS, NPP, OSKO, PAYMENT, and TRANSFER - Payment identifiers: BSB_Account, CardNumber, and PayID - Laundering typologies: structuring, layering, and integration - Geographic setting: Australian synthetic banking context SPLIT AND EVALUATION PROTOCOL The associated manuscript describes the split protocol and evaluation protocol used in the reported experiments. For transaction-level experiments, transactions are ordered chronologically to reduce temporal leakage. For node-level graph experiments, the dataset is converted into an account/entity graph, and node features are computed from aggregated transaction statistics. The manuscript describes the chronological train, validation, and test protocol used for model evaluation. Evaluation scripts will be added to this Zenodo record after publication of the associated paper. SYNTHETIC DATA NOTICE This dataset is fully synthetic. It does not contain real customer records, real account information, real banking transactions, real IP addresses, or real financial institution data. All customer identifiers, merchant identifiers, balances, timestamps, locations, transaction patterns, labels, and metadata values were generated for research purposes. The dataset should not be interpreted as a sample of actual banking activity. GENERATOR SOURCE CODE The AMLNet generator source code is not included in this release due to security concerns. Requests for access to the generator source code may be considered for legitimate research purposes, subject to identity verification, institutional affiliation, and appropriate use conditions. The associated manuscript provides the generator pseudocode, main configuration details, demographic assumptions, regulatory constraints, laundering typologies, timing rules, routing rules, split protocol, and evaluation protocol. This public release provides the fixed generated dataset instance used in the associated paper. Evaluation scripts will be added to this Zenodo record after publication. USAGE This dataset can be used for: - anti-money laundering machine learning research; - rare-event classification experiments; - transaction-level suspicious activity detection; - account-level graph classification; - feature ablation studies; - explainability experiments; - synthetic-to-real transfer learning research; - educational and academic research on financial crime detection. LICENSE This dataset is released under the Creative Commons Attribution-NonCommercial 4.0 International License. You may share and adapt the dataset for non-commercial purposes, provided that appropriate credit is given. Commercial use is not permitted without prior permission. For commercial licensing enquiries, please contact: s.huda@griffith.edu.au, hudasabin@gmail.com CITATION If you use this dataset, please cite the Zenodo dataset record and the associated paper: Huda, S., Foo, E., Newton, M. A. H., Jadidi, Z., Huda, S., Morgan, G., and Sattar, A. AMLNet: A Knowledge-Guided Synthetic Benchmark for Machine Learning Evaluation in Anti-Money Laundering. Under review. Dataset: Huda, S. et al. AMLNet Synthetic Anti-Money Laundering Benchmark Dataset, Version 2.0. Zenodo. 10.5281/zenodo.21237971 CONTACT For questions about the dataset, please contact: Sabin Huda School of Information and Communication Technology Faculty Member, Academy of Excellence in Financial Crime Investigation & Compliance Griffith University Email: s.huda@griffith.edu.au, hudasabin@gmail.com

Mehmetaj, Ilir

6 files · 25 MB · jpeg, pdf, pngdeclared

ABSTRACT - Seraphim Fire Response Ecosystem v1.0 Fire rescue runs on two clocks that have never agreed: an average emergency response of 8 to 14 minutes, and a survivable window of 2 to 4 minutes once a person stands inside an active fire zone. The gap is a material gap - Nomex, the aramid standard of firefighting since the 1970s, chars beyond 260°C, while a wildfire front burns at 600 to 900°C and a firestorm exceeds 1,100°C. For fifty years that difference has been managed with tactics: retreat lines, safety zones, the discipline of staying away. This concept closes it with material, and documents the ground branch of the HAAP platform family - the systems that operate inside the zone every current doctrine writes off. The foundation is one laminate. Seraphim S-100 is a four-layer stack of 35 micrometres and 45 grams per square metre: a diamond-like-carbon outer armour (tetrahedral amorphous carbon class with an oxidation-barrier overcoat, rated for transient direct-flame contact beyond 1,200°C, with sustained-air oxidation above 600°C carried as a named validation parameter), a Gradient Graphene layer whose active thickness of roughly 35 nanometres is a functional requirement of its physics, a Parylene HT dielectric, and a refractory nano-felt backplane. Two mechanisms carry the thermal claim: phonon scattering by engineered ¹²C/¹³C isotope disorder as the load-bearing channel at fire temperature, and the experimentally confirmed violation of the Wiedemann-Franz law in ultraclean graphene at the Dirac point (Nature Physics, 2025) as the laboratory anchor of the electronic channel. A division of labour governs the family: v1 skin rejects, v2 patches harvest - 1.3 to 2 kW of modelled power from the fire's own gradient feeding the coolant pump, the ember-deflection field, and the beacon, stated as a separate and smaller quantity than the 22 kW/m² resistance rating of the stack. Fourteen novel contributions (NC-SFR-1 through NC-SFR-14) document five systems built on that laminate. The AEGIS SUIT states its envelope in the grammar of protective-equipment standards: 600 to 1,000°C operating environment, 300 to 600°C sustained outer surface, 1,200°C peak contact - with an electrostatic Micro-PAML field of 5 to 10 kV/m deflecting embers before contact and a self-powered microfluidic-PCM cooling chain behind the armour. The Seraphim Survival Tent shelters six to eight standing people behind an atmosphere-capable multilayer-insulation wall - ten laminate skins with aerogel interlayers suppressing the gas conduction that defeats spacecraft MLI at ground level - with onboard oxygen, HCN/CO filtration, and an automatic beacon, differentiated point by point from the single-person USFS fire shelter, and deployable four ways, including stationary mounting in fire-extinguisher status for homes and businesses. The Seraphim House Plane is a direct-contact-rated, reusable building envelope in the USD 25,000 class protecting structures of two hundred times that value. The MGU is a single-mission tracked robot delivering one shelter through conditions that exclude aircraft and humans. HAAP 10 is the aircraft that lands inside the burn perimeter on a steam-pre-cooled corridor under automatic limits, with a fixed 60/40 water split reserved for its own protection and the turbine intake-temperature constraint placed openly on the record. A WASP Zone Guardian doctrine and the Prometheus Link - laser power down, biometrics up - bind every suit, tent, robot, and aircraft into one AI-coordinated network. Every load-bearing component exists at TRL 7 to 9 in other industries; the integrated systems stand at TRL 1 to 3. The fastest way forward is a coupon flame test: a laminate sample under metered radiant and direct-flame load, rear-face temperature and harvested watts measured for a few thousand euros, setting the tent wall, the full-scale burn test, and everything above them. A dedicated insurance chapter carries the adoption logic from expected-loss arithmetic to the market re-entry argument, with an instrumented pilot fire season as the underwriting-data instrument. All parameters are theoretical design estimates requiring independent validation. These contributions consolidate and refine three disclosures of March 2026 (two LinkedIn articles, 22 and 24 March; one X publication, 29 March). The fourteen novel contributions are placed on the public record of prior art as of the Zenodo publication date under CC BY-NC-ND 4.0, preventing future patent claims on these specific architectures by any party. Ilir Mehmetaj | Independent Concept Developer | CC BY-NC-ND 4.0 | 2026

Ayala-Godoy, Jairo Arturo · Alvarado-Vargas, Lizbeth · Agosto-Rivera, Jose L.

5 files · 1.6 MB · pngdeclared

BeeChrono is an interactive R Shiny web application for processing, visualizing, cleaning, and analyzing honey bee time data collected across multiple field locations. The application aligns entry and exit records for each individual bee, computes time at the feeder and feeder-to-hive travel time, identifies potential outliers using Tukey's rule, and provides interactive visualizations and exportable datasets in wide and long formats. It also includes final statistical analyses using generalized linear mixed models, including diagnostic checks and tests for differences by location, trip, and temporal variability across locations.

Souza, Caio · Correia, João

54 files · 16 MB · csvdeclared

This repository contains the second version of the metamodel-based rescoring framework for protein-ligand binding affinity prediction. The first version introduced a linear metamodel combined with random train/test splits, but did not explicitly account for structural similarity or information leakage between training, validation, and test partitions. In this updated version, we included: Python scripts for data preprocessing, model training, and evaluation Pairwise ΔΔG evaluation code and associated data Visualization scripts for generating publication figures This dataset and codebase are intended to support reproducibility and further development of metamodel-based rescoring strategies in structural bioinformatics and computational drug design.

Costa, Carlos Fransley Scatambulo

3 files · 614 KB · csv, pdfdeclared

Dados anonimizados referentes ao estudo na tese: Formação Docente em Inteligência Artificial para o Ensino Superior: dos Fundamentos Computacionais ao Uso Crítico e Pedagógico da IA Generativa .

Kirjonen, Miikka · Kirjonen, Miikka

33 files · 5.5 MB · pngdeclared

The purpose of this manuscript is to formalise the point at which a seam-bearing support becomes a loop-bearing support. This distinction is essential. A divider, ridge, throat, caustic, or mirror trace can be visually compelling, but visibility is not yet transport. A visible loop can be drawn around nearly any sufficiently structured image, but a drawn loop is not yet a Wilson memory object. A shell can look cyclic, ternary, heptagonal, or recurrent, but visible sectoring is not yet a carrier law. The present paper therefore begins by refusing the shortcut from morphology to ontology. It asks instead for a declared sequence of admissibility: first the mirror-involution grammar, then the no-cancellation certificate, then binary seam admission, then post-seam loop support, then closed Wilson memory, then page continuation, then homebound closure, and only finally local ternary refinement. This order matters because RSM's seam programme has reached a point where several previously separate threads now meet. The first thread is the mirror-involution seam: a two-sheet comparison carrier, a quotient projection, an order-two branch exchange, and an even-channel public divider. The second thread is loop inheritance: once the seam gate has passed, the post-seam support admits closed reinjection loops whose holonomy is stable and whose memory is not attached to a local branch but to the loop as a whole. The third thread is the lattice continuum: once a closed loop carries Wilson memory, one must ask how a page of memory relates to the next, how the continuation coordinate is transported, and how local ternary public events may arise without altering the underlying binary seam law. The novelty of the present paper lies in connecting these three threads without changing their order. The mirror-involution grammar is binary. The post-seam loop is closed and Wilson-bearing. The lattice is a continuation law on that loop-capable support. The local ternary law is a refinement, not a replacement. This resolves a tension that has accompanied earlier lattice language: if a visible shell was introduced too early, the lattice risked sounding like a primitive geometry. If the ternary cycle was introduced too early, the public event law risked sounding like a universal carrier law. The present paper avoids both problems. It places the lattice after the seam gate and places the ternary event after homebound closure. The deepest algebraic reason for this order is the no-cancellation principle. In a mirror-involution setting, arbitrary sign cancellation can erase the very event that the seam certificate is meant to classify. A pre-seam carrier must therefore be treated as an oriented comparison support whose local samples remain in a coherent sign cell until a declared boundary is reached. This is where the Grassmannian picture becomes useful, but only as a certificate layer. The positive or totally non-negative Grassmannian does not define the vortex tube, the Wilson loop, or the ternary event. It supplies the finite-dimensional pre-seam audit: it tells us whether local orientation data can be compared without hidden cancellation, and it gives a precise language for the boundary at which a graph-like or sheet-like support ceases to be the correct chart. After that boundary, the correct observable changes. The support is no longer primarily a divider-bearing graph; it is a loop-capable neighbourhood of the certified seam. The mature object is then the closed reinjection loop with a Wilson holonomy. This is the conceptual pivot of the manuscript. Before the seam, the question is whether the local support can be certified. At the seam, the question is whether the mirror and residual gates pass. After the seam, the question is whether the inherited loop phase is stable. The loop is not a left branch and not a right branch. It is the quotient memory object created when the lifted return legs project to one closed public transport class.

Rockefeller Archive Center · Shubinski, Barbara · Krishna, Sanjaya · et al.

1 files · 738 MB · tardeclared

RAC-KG is a knowledge graph, stored in a Neo4j database, created from approximately 12,000 pages of typed, paper, analog records from the Rockefeller Foundation (RF) collection. The knowledge graph covers the activites of RF's Natural Sciences division between the years of 1932 and 1941 as documented in two primary types of source archival material - the daily activies of Natural Sciences division oficers, as documented in what are referred to as Officer Diaries and Minute Books from Board and Executive Committee meetings which document where RF actually granted money for work in the Natural Sciences. Some additonal source archival material is also contained, but the two primary sources drive the vast majority of entities and relationships contained in the graph.