Our Work

Unrefereed mathematical candidate presenting an explicitly admissible mass-action realization of the established Phillipson–Rojas extremal polynomial chain. For every integer n≥2 the displayed quadrat…

Unrefereed mathematical candidate presenting an explicitly admissible mass-action realization of the established Phillipson–Rojas extremal polynomial chain. For every integer n≥2 the displayed quadratic network has n species, n+2 reactions and exactly n+1 positive equilibria, all nondegenerate. Two exact three-species examples include an inflow-free network with maximum product molecularity ten, an exact one-rate interval, and a reciprocal example with two local sinks and two saddles. The PDF and TeX are authoritative; exact programs and receipts provide computational corroboration.The scholarly creator is Anonymous; Ian Pitchford maintains the package for Evidence Press. The supplied referee-style material has unverified reviewer identity and is not journal acceptance or an official research assessment.

Reproducibility release accompanying the manuscript on a measurement-state-aware, monotonicity-constrained DC/quasi-static compact model for the GaN Systems GS65030. The archive contains the custom Py…

Reproducibility release accompanying the manuscript on a measurement-state-aware, monotonicity-constrained DC/quasi-static compact model for the GaN Systems GS65030. The archive contains the custom Python identification and validation code, frozen model coefficients, the validated LTspice behavioral subcircuit, a candidate Verilog-A portability implementation, support-envelope data, and derived validation/ablation summaries. The primary measurement data are not redistributed; they are publicly available from the source dataset at DOI 10.57745/IZ6LY2. Accuracy claims are limited to the measured-data support of one device at one temperature. The Verilog-A file is provided as a candidate implementation and was not compiled or numerically validated; the LTspice subcircuit is the validated simulator realization. v1.0.1 patch release. Adds the A0–A4 verification and clean-room re-execution harness, its frozen inputs and reference artifacts, release-portable export/validation launchers and documentation, and source-data licensing/attribution records. The complete v1.0.0 package snapshot is retained unchanged inside this release. No model coefficients, fits, numerical results, figures, or conclusions have changed.

This replication package contains the data and analytical materials supporting the study “The Regulatory Gap in African AI Governance: Institutional Coordination in Tanzania, Uganda, and Zambia.…

This replication package contains the data and analytical materials supporting the study “The Regulatory Gap in African AI Governance: Institutional Coordination in Tanzania, Uganda, and Zambia.” The study examines how institutional functions associated with artificial intelligence governance and development are organized across Tanzania, Uganda, and Zambia between 2020 and 2026. The package includes the document inventory, event-level coded dataset, institutional-function coding framework, policy commitment mapping, intercoder reliability materials, and analytical files used to produce the descriptive statistics and comparative analyses reported in the article. The analysis covers twelve institutional functions: funding, implementation, partnership, advising, regulation, training, infrastructure provision, hosting, co-development, consultation, research collaboration, and convening. The original corpus contained 103 records, of which 91 met the study’s inclusion criteria. These records were consolidated into 79 distinct institutional events, including 71 national events used for comparative analysis: 25 in Tanzania, 29 in Uganda, and 17 in Zambia. Eight regional or continental events were retained for contextual analysis but excluded from national comparisons. The repository is intended to support transparency, verification, and reproducibility of the analyses reported in the accompanying article. Third-party source documents are not redistributed unless permitted; where applicable, the dataset provides bibliographic information, source identifiers, and links to the original publicly accessible materials.

Fine-tuned MatterGen diffusion-model checkpoints accompanying the paper An Environment-Conditioned Generative Model Designs Inert Anodes across Opposite Molten-Salt Chemistries.Three fifteen-property …

Fine-tuned MatterGen diffusion-model checkpoints accompanying the paper An Environment-Conditioned Generative Model Designs Inert Anodes across Opposite Molten-Salt Chemistries.Three fifteen-property fine-tunes of the public MatterGen base checkpoint, trained on 30,161 transition-metal oxides from Alex-MP-20 with identical architecture and hyperparameters. They differ only in the per-cation anodic-stability table used to label the bath-aware and weakest-link corrosion properties:Fluoride model, first generation (LiF–NdF3–Nd2O3; 200,000-candidate main campaign)Fluoride model, second generation (same bath, Mn and Al scores recalibrated; Ni2+ high-entropy campaign)Chloride model (CaCl2–CaF2–CaO, calcium electro-reduction)Each model is provided as the final PyTorch Lightning checkpoint (*_last.ckpt) with its training configuration (*_config.yaml). See README.md for the directory layout MatterGen expects and SHA256SUMS for checksums. Code: https://github.com/jonglee69/OxideMatterGen

This repository contains the analytic dataset and reproducible research materials supporting the study “Dengue y neumonía infantil en el Perú: perfiles territoriales y factores soc…

This repository contains the analytic dataset and reproducible research materials supporting the study “Dengue y neumonía infantil en el Perú: perfiles territoriales y factores socioambientales asociados, 2000–2024” (“Dengue and childhood pneumonia in Peru: territorial profiles and associated socio-environmental factors, 2000–2024”). The study is a longitudinal ecological analysis at the department-year level covering 25 Peruvian departments, including Callao, from 2000 through 2024. The repository contains the aggregated analytic dataset, variable dictionary, reproducible analysis code, supplementary tables, methodological documentation, reference outputs, and supporting files required to understand and reproduce the analytical workflow. The analytic dataset contains department-year indicators derived from publicly available epidemiological, demographic, socioeconomic, environmental, climatic, and health-system sources. The deposited dataset is aggregated at the department-year level and does not contain individually identifiable personal information. The repository supports the analyses of dengue and childhood pneumonia, territorial profiles, persistence over time, socio-environmental associations, sensitivity analyses, and external validation described in the accompanying manuscript. Primary source data from CDC-MINSA, INEI, MINAM, and other institutions retain their respective ownership and terms of use. Original individual-level surveillance records and third-party source documents are not redistributed in this repository. Derived aggregate data and reproducible research materials are provided for scientific transparency and reproducibility. Repository version: 1.0Corresponding author: Rodrigo Guerrero LópezInstitution: Universidad Nacional San Luis Gonzaga, Peru.

Preprint, version 0.22b (29 September 2026). Grid cells exist before an animal explores. We ask how little has to be learned for a fixed hexagonal phase template, advanced by self-motion, to localise …

Preprint, version 0.22b (29 September 2026). Grid cells exist before an animal explores. We ask how little has to be learned for a fixed hexagonal phase template, advanced by self-motion, to localise sensory content that is arbitrary in what it is, provided it is unique to its site and stable across visits. The only learned quantity is an association between a site's sensory pattern and the template's phase, written in one shot at the first visit and thereafter rewritten at a small gain, with the site's identity supplied by the simulator to tell a first visit and to address the rewrite, a shortcut we declare; at every step the phase is nudged towards what is stored. Without binding the template drifts across the room; with it, the decoding error falls within the first tenth of the walk to about one arena cell and stays there under velocity noise of up to ten percent per step, and it remains bounded when the noise is tripled. Shuffling the velocity or the associations destroys the alignment. Since the sensory input alone identifies sites, what the bound template adds is measured directly: while the input is off the template path-integrates like any integrator, and when the input returns the bound template re-locks on the stored position within about ten steps, where an unbound template does not. None of this depends on the spatial smoothness of the input: with content that is white noise per site the binding anchors and re-locks the same way. What is stored is a memory of the room and not a re-fit to it: when the room is stretched after the map is written, decoded positions keep the old room's frame, compressed by the stretch, while a map bound from scratch in the stretched room follows the new room exactly. Kept anchors are fields that stretch with the room: in the metric of the deformation experiments the model's field spacing follows 93% of the room change without the rewrite and about two thirds with it, where rats follow about half (a Hebbian write at every step, which does not use the site identity, anchors as well with smooth content but, at the correction gain used throughout, keeps the old frame whole, so in the model as run the partial rescaling rests on the shortcut), and two rooms bound separately merge by a local reorganisation at the old wall that is an order of magnitude smaller than the animal's. A place stage of stored exemplars, never trained, aligns the template as well as a trained one, and the same binding on a continuous-attractor network reproduces the anchoring and the stretch. The regime is the easy one by design, unique content and a sensor that identifies every site on its own, and the result is reported as the measured floor of the design, with its controls and with what it costs. Data and code: per-seed result files, figures, the model code and the scripts that reproduce every figure are in the evidence pack, doi:10.5281/zenodo.22267940 (version 3; all versions: doi:10.5281/zenodo.22239562).Use of AI tools: Generative AI (Claude and Claude Code, Anthropic) assisted the author in writing the text, in implementing and running the simulation code, and in exploring and critically evaluating hypotheses and experiments. All scientific content, decisions and claims are the author's, who takes full responsibility for them; every number reported is an output of the deposited simulation code and is anchored to a versioned result file.

Exact Gershon successor-set weights; capacity-feasible min-width smoothing; the destination rung (per-destination earliest period, lift of a CPIT plan, exact OPBSP-[D] local search, PCPSP LP bound thr…

Exact Gershon successor-set weights; capacity-feasible min-width smoothing; the destination rung (per-destination earliest period, lift of a CPIT plan, exact OPBSP-[D] local search, PCPSP LP bound through HiGHS, which reproduces the published newman1 PCPSP LP bound 24,486,549); an LP-guided, predecessor-closed candidate set for the sliding time window. Full notes in CHANGELOG.md. Closes #24.

THINKIT: workflows for AI-assisted scientific research THINKIT is a set of coordinated workflows for using AI in scientific research: understanding physical problems, examining literature and evidence…

THINKIT: workflows for AI-assisted scientific research THINKIT is a set of coordinated workflows for using AI in scientific research: understanding physical problems, examining literature and evidence, developing research projects and proposals, and preparing scientific manuscripts. It is intended for researchers, doctoral students, and research teams, particularly in the physical sciences and engineering. Its purpose is to make scientific interaction with AI more disciplined, transparent, and controllable by the researcher. THINKIT provides a shared framework for clarification, reasoning, evidence assessment, and scientific decisions, reducing the need to repeatedly explain how the AI should approach the work. Status: Validated experimental release Feedback: If you use THINKIT, I would be grateful for comments on usability, scientific workflow, and problems you encounter: Feedback link   Guiding principles Researcher-led discussion. The researcher controls the direction, depth, and pace of the work. THINKIT starts from the available observations, evidence, and ideas, asks questions when consequential ambiguities remain, and addresses the current question without automatically expanding the task. Physical understanding before calculation. Mechanisms, forcing, constraints, and relevant regimes guide the choice of equations and models. Numerical inputs and computational work serve a defined scientific purpose. Iterative hypothesis development and testing. Hypotheses are formulated, examined, refined, or rejected through checks of internal consistency and comparison with relevant literature, experiments, or numerical results. The researcher’s hypothesis is examined before alternatives are developed. Broader idea generation follows the researcher’s request or choice; evaluating an idea does not authorize replacing it. Literature integrated into reasoning. Targeted literature checks support focused research questions and consequential claims. Supporting, contradictory, and missing evidence are made explicit, together with the limits of source access. Clear evidence status. Observations, source claims, assumptions, derived results, interpretations, and provisional ideas remain distinguishable. Adopting a working hypothesis does not establish its validity, and polished wording must not imply stronger evidence than is available. Four coordinated methods PHYSIT — physical understanding and modelling PHYSIT supports the interpretation of physical phenomena, examination of mechanisms, hypothesis testing, and development of theoretical, reduced, scaling, semi-empirical, or empirical models. Equations express and test the physical picture. Once a new physical research question is sufficiently focused, PHYSIT initiates a bounded LITIT literature check without requiring a separate search workflow. Numerical calculation, fitting, CFD, and code development become primary tasks when explicitly requested. LITIT — literature and evidence LITIT supports literature search, source screening, claim-specific evidence assessment, identification of existing models, source cards, and bibliographies. It extracts relevant assumptions, mechanisms, validity ranges, and limitations for use in modelling, proposals, or manuscripts. Source content remains separate from analytical interpretation, and the actual access level—metadata, abstract, partial text, or full text—is stated. PROPOSIT — research projects and proposals PROPOSIT supports research-topic development, feasibility assessment, objectives, methods, work packages, consortium design, and proposal writing. Project development and review are guided by the target call’s official requirements and assessment criteria. Scientific gaps, novelty claims, feasibility, and risks are examined before they are turned into polished proposal text. KETAVIT — scientific writing and revision KETAVIT supports manuscript drafting, revision, polishing, and reviewer responses. It improves clarity, paragraph coherence, and precision while preserving scientific meaning, claim strength, uncertainty, author voice, notation, citations, and LaTeX structure. Polishing, a requested post-polishing audit, and an optional final prose pass are distinct stages. Editing must not invent scientific content or conceal unresolved problems. Connected workflow and implementation The four methods share context, assumptions, terminology, and the evidence status of ideas as work moves between physical reasoning, literature assessment, proposal development, and writing. Focused questions are handled directly; broader work proceeds in bounded steps with substantive decisions left to the researcher. The current implementation is a ChatGPT plugin containing PHYSIT, LITIT, PROPOSIT, and KETAVIT as coordinated skills. The workflow instructions and supporting reference materials are included in the plugin. ChatGPT Projects can keep relevant papers, manuscripts, data, code, terminology, and project decisions together. Separate Project Instructions are not required for normal THINKIT operation. Where an accessible project terminology record is maintained, LITIT can contribute source-grounded terminology and KETAVIT can use it consistently in writing. Adaptation to other AI assistants is a possible direction for further development; compatibility must be assessed separately for each platform. Scope and limitations THINKIT provides methodological instructions to the underlying AI model. Their execution depends on the host model, available tools, source access, and access to the required instructions and resources. Its rules require explicit uncertainty and prohibit fabricated sources, data, and validation, but they do not guarantee error-free responses. THINKIT supports scientific reasoning and writing; it does not replace scientific verification or the researcher’s responsibility for conclusions.   Developed by Ilia V. Roisman.

LREM — Low-Flow Reservoir Evolution Model Description for Zenodo LREM — Low-Flow Reservoir Evolution Model — is a parsimonious empirical framework developed to characterize reservoir…

LREM — Low-Flow Reservoir Evolution Model Description for Zenodo LREM — Low-Flow Reservoir Evolution Model — is a parsimonious empirical framework developed to characterize reservoir behavior during low-flow seasons and to connect reservoir elevation dynamics with potential operational impacts on hydropower systems. The method is structured through the FGET–VRR–NFF methodological sequence: FGET — Formula Genomics and Evolution Theory explores candidate mathematical relationships among observable variables. VRR — Variable Reduction and Ranking identifies the dominant variables, coefficients, and relative scales. NFF — Nested Formula Framework reorganizes the reduced relationships into a compact chain of linked equations. In its application to Mazar Reservoir in Ecuador, LREM uses the September reservoir-elevation loss to define the dimensionless Invernal Type factor, \(I_T\). This factor characterizes the relative severity of the low-flow season and is then incorporated into a time-dependent quadratic formulation for reservoir elevation, \(Z(d)\). A second empirical relationship links reservoir elevation and \(I_T\) to estimated daily power-rationing hours, \(H\). The complete operational chain is: \[ \Delta Z_{sep}\rightarrow I_T\rightarrow Z(d)\rightarrow H \] LREM provides a compact way to compare low-flow seasons, classify relative severity, estimate the date and elevation of the reservoir minimum, identify medium-risk and maximum-risk periods, and construct daily scenarios of potential power-rationing hours. The method has been developed from historical observations of the Mazar Reservoir and Ecuadorian power-system conditions, using the 2017–2025 period as a historical reference and 2023–2024 as key verification cases. THEK-LREM-MAZAR is the specific application of the LREM framework to the Mazar Reservoir. Keywords LREM; Low-Flow Reservoir Evolution Model; Mazar Reservoir; hydropower; reservoir modeling; drought; low-flow season; water resources; energy security; power rationing; blackout risk; energy modeling; hydrological risk; reservoir elevation; FGET; VRR; NFF; Invernal Type; Ecuador; Paute Integral Hydroelectric Complex; renewable energy; climate resilience; energy resilience; water-energy nexus; hydropower risk assessment; mathematical modeling; empirical modeling; predictive modeling; infrastructure resilience; power system risk; THEK Research Institute. Hashtags #LREM #LowFlow #ReservoirModeling #Hydropower #Mazar #Ecuador #EnergySecurity #EnergyResilience #WaterResources #WaterEnergyNexus #Drought #ClimateResilience #PowerSystems #BlackoutRisk #PowerRationing #HydrologicalRisk #RenewableEnergy #EnergyModeling #MathematicalModeling #PredictiveModeling #InfrastructureResilience #FGET #VRR #NFF #THEK #THEKResearchInstitute #Science #Engineering #Research #OpenScience Collaboration THEK Research Institute welcomes international collaboration in reservoir modeling, hydropower systems, water–energy risk, mathematical modeling, infrastructure resilience, and applications of the LREM, FGET, VRR, and NFF methodologies. We are available for research collaboration, scientific conferences, keynote presentations, professional courses, technical advisory services, consulting, institutional cooperation, and licensing opportunities. Contact: research@hithek.com THEK Research Institute — Transforming Knowledge into Solutions.

This article analyzes the impact of developing Uzbekistan's economy and tourism sectors based on international standards on sustainable economic growth. The research highlights how adherence…

This article analyzes the impact of developing Uzbekistan's economy and tourism sectors based on international standards on sustainable economic growth. The research highlights how adherence to global norms is crucial for attracting foreign direct investment, enhancing competitiveness, and fostering job creation. It further explores how modernizing tourism infrastructure and improving service quality contribute to strengthening the country's position in the global economic landscape. The study provides practical recommendations for implementing effective strategies aimed at achieving long-term sustainable development.