Our Work

Abstract: A geometric construction of inverse-square gravity is presented in which solid angle accounts only for spatial dilution, while universal coupling to mass-energy supplies the physical strengt…

Abstract: A geometric construction of inverse-square gravity is presented in which solid angle accounts only for spatial dilution, while universal coupling to mass-energy supplies the physical strength of the interaction. This separation avoids treating a body-dependent geometric parameter as a substitute for gravitational mass. For a static spherical source, the construction recovers the exterior Newtonian potential per unit test mass, Φ(r) = -GM/r. Combining this static scalar potential with the special-relativistic energy of the escaping particle gives:   v_esc(r) = c * sqrt( 1 - 1 / (1 + GM / (r * c^2))^2 ) The result is subluminal for all finite r and reduces to the Newtonian escape velocity in the weak-field limit. The model is not presented as a replacement for general relativity, nor as a complete Lorentz-covariant field theory. Rather, it is a flat-spacetime scalar-potential construction that clarifies what follows from solid-angle dilution, universal mass coupling, and special-relativistic energy conservation before additional assumptions, such as wave propagation or optical-gravity extensions, are introduced.   Note on Versioning & Relationship to Prior Work: This paper presents a standalone, refined mathematical derivation focusing specifically on the static scalar formulation and relativistic escape condition. It serves as an independent baseline related to the foundational concepts outlined in the author's broader work (A Solid-Angle Postulate as a Geometric Foundation for Gravity, DOI: 10.5281/zenodo.19333670).

Replication package for the paper "Use-dependent depreciation of intangible capital: thresholds, history dependence, and generative artificial intelligence." The paper studies a firm that accumulates …

Replication package for the paper "Use-dependent depreciation of intangible capital: thresholds, history dependence, and generative artificial intelligence." The paper studies a firm that accumulates an intangible stock — interpreted as the quality of its data estate — while the intensity with which it deploys a generative AI system accelerates the depreciation of that same stock. Optimal deployment therefore carries an intertemporal contamination wedge: the marginal value of scale is reduced by the shadow price of the stock times the marginal contribution of use to its decay. Under an S-shaped effectiveness function the problem admits three steady states, and optimal policy is history-dependent, separated by a threshold in the initial state obtained by value matching rather than by any local property of the unstable equilibrium. This package reproduces every number, table and figure reported in the paper. It contains three Python modules and no data: all results are generated from the model primitives and the stated parameter values. Contents model.py — model primitives and the numerical solver: steady states, stable manifolds, value functions, and the threshold by value matching. run_all.py — driver that reproduces all reported quantities and writes results/all_results.json. Runtime is approximately 25 minutes on a laptop; the Monte Carlo exercises dominate. make_figures.py — regenerates Figures 1 to 5 as PDF from the solver output. README.md — numerical methods, seeds, validation, and the classification scheme for perturbed draws. RequirementsPython 3.10 or later with NumPy, SciPy and Matplotlib. No other dependencies. Tested on NumPy 2.4, SciPy 1.17 and Matplotlib 3.10. Numerical methodsSteady states are the roots of a scalar residual, bracketed on a grid and refined by Brent's method to a tolerance of 1e-13. Stable manifolds are obtained by eigen-decomposing the Jacobian at each saddle point, displacing along the stable eigenvector, and integrating the time-reversed canonical system with an adaptive Runge–Kutta scheme at relative tolerance 1e-11. Value functions exploit the identity between the costate and the derivative of the value function along the optimal path, with the constant of integration fixed by the perpetuity value at the steady state; this avoids forward integration, which is unstable off the saddle path. The threshold solves the value-matching condition on the overlap of the two manifold domains. ReproducibilityAll stochastic exercises use a fixed seed, numpy.random.default_rng(20260731), so re-running reproduces the reported frequencies exactly. The shape of the effectiveness function is a parameter of the model rather than a code branch, so the exercises that compare specifications require no modification of the solver. ValidationValue functions are cross-validated against direct time integration of discounted profit along the manifolds; the two computations agree to relative errors of 2.4e-6 and 6.2e-6 on the low and high branches. Single crossing of the difference of the two value functions is not established analytically, so the package records, for every perturbed draw admitting a threshold, whether that difference is strictly monotone on a 400-point grid. Draws on which the solver fails are reported as errors rather than discarded silently. CitationPlease cite both this software and the associated paper. A CITATION.cff file is included.

Abstract: While Africa has a disproportionately high share of the global disease burden, it is also plagued by poor connectivity, a lack of genomic reference data, and diverse digital and network infr…

Abstract: While Africa has a disproportionately high share of the global disease burden, it is also plagued by poor connectivity, a lack of genomic reference data, and diverse digital and network infrastructure that impede the continent's shift toward precision healthcare. To work around these constraints, three converging technologies offer a way forward: federated artificial intelligence (federated learning), which allows for model training to be done collaboratively across institutions without sharing sensitive patient data; health data interoperability standards, which make it possible to exchange health information from heterogeneous electronic health record (EHR) and mobile-health systems; and digital twins, dynamically updated virtual patients or virtual population models for simulation-based health-related decisionmaking. This narrative review collates literature published from 2020–2025 on these three technologies in Africa and other resource-constrained environments, including federated-learning pilots for tuberculosis and foetal-ultrasound screening, continentwide scoping of interoperability, and early digital-twin architectures proposed for low-resource African health systems. We propose an integrated, layered structure that connects local federated-learning nodes, an interoperable semantic data layer, and a regional digital-twin simulation layer, linked by privacypreserving mechanisms and Africa-specific data-governance safeguards. The paper discusses obstacles and limitations to cross-border data transfer, such as weak institutional trust in data sharing, algorithmic bias from non-representative training sets, and unreliable connections, as well as measures being taken to overcome them. In conclusion, federated AI, interoperability, and digital twins are all promising and essential for achieving privacypreserving precision healthcare at scale in Africa.

Batch upload of files from directory 'allium_unifolium_FSCR'.

Dataset used to reproduce the results from paper: Open-Set Power-System Event Classification with Angular Compactness and Rank-Based Score Fusion

This dataset supports the structure-guided analysis of the MFN2 C-terminal modulator surface in the review "Conformational allocation of mitofusin 2 in neuronal mitochondrial dynamics: mechanisms and …

This dataset supports the structure-guided analysis of the MFN2 C-terminal modulator surface in the review "Conformational allocation of mitofusin 2 in neuronal mitochondrial dynamics: mechanisms and therapeutic targeting". It contains receptor and ligand coordinates, docking configurations, raw poses and logs, solvent-accessibility and pocket-detection results, analysis tables, scripts, software records and checksums.Version 1.1.0 adds the complete methods and two computation tables relocated from the manuscript, the current integrated Figure 6 and its caption, source-data and coordinate crosswalks, a field dictionary and file inventories. Four historical rendering inputs omitted from version 1.0.0 are also supplied with their original filenames and restoration paths. The 784 files from version 1.0.0 are preserved byte for byte. No calculations or numerical results have changed.The complete ZIP is self-contained. The separate methods document and two CSV tables are identical convenience copies of files inside it. The former Figures 6 and 7 are represented by the integrated main Figure 6. Historical labels remain unchanged within the original archive.Docking tests local geometric compatibility and coordinate-derived exposure. It does not establish native-membrane accessibility, binding affinity, selectivity or therapeutic response. Fixed search seeds are technical searches, not biological replicates.

Batch upload of files from directory 'allium_stellatum_FSCR'.

Batch upload of files from directory 'allium_shevockii_FSCR'.

  Introducing children to the world of books is essential, and guiding the current generation toward reading is a vital task. Children's literature is created specifically for this purpose; ideal…

  Introducing children to the world of books is essential, and guiding the current generation toward reading is a vital task. Children's literature is created specifically for this purpose; ideally, it should be easy for them to comprehend while also fostering intellectual curiosity. In this context, author Vishnupuram Saravanan, in his book ‘Otrai Siragu Oviya’, explains concepts to children with remarkable simplicity and clarity. Organic farming is currently facing the threat of decline, and today's generation often remains confined indoors, immersed in the online world. The author addresses this by presenting the importance of agriculture and its precarious state in a manner that is both accessible and engaging for young readers. Furthermore, the issues associated with methane gas are well known; the author skillfully explains the resulting environmental problems to children. This article examines the environmental themes presented in Vishnupuram Saravanan's book, ‘Otrai Siragu Oviya’.