diffusion models

PhysiCelldFBA: Linking single-cell genome-scale metabolism to spatially explicit multicellular dynamics

Genome-scale metabolic models can predict how individual cells allocate resources and respond to their environment, yet few frameworks link single-cell metabolism to the spatial organisation of multicellular systems. Here we introduce PhysiCelldFBA, an extension of the PhysiCell agent-based framework that couples genome-scale dynamic flux balance analysis to off-lattice multicellular simulations. Each simulated cell carries its own metabolic model, allowing local environmental conditions to shape metabolism while metabolic activity feeds back on the surrounding environment, cellular behaviour, and spatial organisation. We first validate this coupling by showing that glucose consumption, CO2 production, and biomass accumulation remain mass-balanced in a closed E. coli system, with simulated biomass agreeing with analytical predictions to within 1%. We then demonstrate how metabolic phenotypes emerge from this coupling across microbial and mammalian systems. Spatial nutrient gradients generate metabolic stratification and acetate cross-feeding in growing E. coli colonies; diffusion-limited metabolism produces proliferative, hypoxic, and necrotic zones across a broad panel of metabolites in a tumour-like tissue; distinct, organism-specific metabolic networks give rise to syntrophic cross-feeding and spatial niche formation in a two-species consortium; and metabolic state couples energy availability to transitions between cellular motility and growth. Across these examples, metabolic stratification, cross-feeding, and phenotypic adaptation emerge from local metabolic optimisation and environmental feedback rather than being explicitly prescribed. PhysiCelldFBA therefore provides a general framework for simulating genome-scale metabolism at single-cell resolution and linking intracellular metabolic state to cellular behaviour and emergent organisation across scales.

AI-Assisted Procedural Game Level Generation Using Diffusion Models and Reinforcement Learning Techniques

Robust Image Processing Techniques for Complex Real-World Environments

Image processing systems that are designed to work in real environments are subject to factors such as haze, rain, noise in sensors, motion blur, lack of proper illumination, and compression. Unlike the controlled environment, the aforementioned factors require an approach that can withstand various levels of degradations and still produce reliable results. In this chapter, a comprehensive discussion of how image processing systems have evolved, starting from basic signal-processing techniques to current advanced AI-based ones like CNNs, vision transformers, GANs, Retinex Enhancement Models, and Diffusion models is presented. Applications of these methods in autonomous vehicles, medicine, remote sensing, industry, and security surveillance are considered. Key issues such as computational complexity, dataset size, domain adaptation, and adversarial attacks are explored.

Deep-seated hydrogen outgassing as the cause of the 1908 Tunguska explosion.

Fluid-Dynamic and Thermodynamic Modeling of Sample No. 3 (Tunguska) 1. Initial Instrumental Parameters (Sample No. 3) According to the mass-spectrometric measurement protocol executed via the specialized hydrogen analyzer AB-1 in compliance with GOST 21132.1-98, the quantitative parameters for Sample No. 3 (Tunguska, mass m = 0.42 g) are established as follows: Diffusion-Mobile Hydrogen (DMH) fraction (Extraction temperature = 400 degrees C, Binding energy approx. 0.3-0.4 eV): 0.935 ppm Strongly Bound Hydrogen (SBH) fraction (Extraction temperature = 700 degrees C, Binding energy approx. 1.0 eV): 1.254 ppm Cumulative Protium Concentration (Q): 2.189 +- 0.348 ppm The absolute dominance of the high-temperature, strongly bound interstitial phase (SBH) over the mobile phase proves that the hydrogen is rigidly locked inside the deep nanotraps of the crystalline B.C.C. lattice rather than residing as superficial atmospheric sorbents or soil contamination. 2. Thermodynamic Evaluation via the Sieverts–Plitz–Altschuler Matrix To calculate the primordial fluid partial pressure required to chemically saturate the crystalline lattice of Sample No. 3 up to the instrumentally verified threshold of C = 2.189 ppm, we utilize Sieverts’ square-root law adapted for low-temperature conditions (150-200 degrees C). According to this framework, the concentration of hydrogen in the lattice is directly proportional to the square root of its partial pressure, where the Sieverts solubility constant (K_s) for the B.C.C. iron matrix drops exponentially at low temperatures and amounts to exactly 0.012 ppm * bar^(-0.5). Extrapolating these parameters yields the required thermodynamic chemical fugacity (f_H2) of the juvenile fluid stream: f_H2 = (C / K_s)^2 = (2.189 / 0.012)^2 = (182.416)^2 = 33,276 bar Utilizing the real-gas equation of state at the hyperbaric frontier, where the fugacity coefficient for highly compressed supercritical hydrogen diverges significantly from unity (amounting to approx. 1.18 within the target thermal window), we translate this chemical activity into actual mechanical fluid partial pressure (P): P = f_H2 / 1.18 = 33,276 / 1.18 = 28,200 bar = 2.82 GPa Conclusion of the Baric Model:The solid-state synthesis of Sample No. 3 occurred inside a closed natural autoclave under a monumental partial fluid pressure of 2.82 GPa (approximately 28,200 atmospheres). This extreme baric environment successfully substituted for high-temperature thermal activation, forcing the solid-state trapping of atomic protium. 3. Gas-Dynamic Inversion via Joule–Thomson Adiabatic Cooling The rapid migration of the supercritical fluid through the intersecting tectonic grid of the Siberian Platform triggered an avalanche-like decompression. In the gigapascal regime (2.82 GPa), the inversion parameters for real hydrogen shift radically, resulting in a strictly positive Joule–Thomson effect (mu_JT > 0). As the high-velocity reactive jet breached the local sedimentary strata, the intense expansion forced the fluid to perform massive internal work against intermolecular attractive forces. This adiabatic decompression triggered a localized "cooling flash" (instant cold crystallization), dropping the internal kinetic temperature precisely to the 150-200 degrees C boundary. This kinetic trap instantly locked the 2.189 ppm of protium inside the nascent lattice, simultaneously suppressing the Fischer–Tropsch methanation constant via the extreme volumetric work term in the Gibbs free energy equations. The rapid gas-to-solid transition during this flash expansion is visually fossilized as the highly scoriaceous, cavernous macro-porosity (vesicles) observed in the specimen's morphology. Discussion. The instrumentally verified cumulative protium concentration of 2.189 ppm and the calculated fluid partial pressure of 2.82 GPa inside the subsurface natural autoclave do not merely describe the localized solid-state genesis of Sample No. 3, but completely reshape the understanding of the macro-phenomenology of the 1908 Tunguska event. The calculated hyperbaric outgassing paradigm coupled with the positive Joule-Thomson adiabatic cooling mechanism provides a comprehensive physical explanation that fully resolves the long-standing anomalies recorded in the historical archives of the Kulik and Krinov expeditions, which documented verified testimonies from hundreds of observers within a radius of up to 800 km from the epicentral zone. The primary anomaly traditionally facing classical meteoritics is the acoustic-optical asynchrony, or the acoustic paradox, where eyewitnesses along the Angara River explicitly stated that they initially registered intense acoustic detonations and low-frequency seismic hums that shook the earth and shattered glassware, and only several minutes later observed the luminous plasma trail in the upper atmosphere. In a standard supersonic cosmic body entry, the acoustic shockwave invariably arrives after the object passes the observer's zenith, as sound cannot outrun a supersonic mass. Within the framework of the presented fluid-dynamic model, this chronological order is natural because the initial brittle failure of the crust's seams under a fluid pressure of 2.82 GPa generated immediate low-frequency seismic waves traveling through the lithosphere at 5–6 km/s, producing the pre-optical rumble and tremors across the region. Only minutes later did the high-pressure supercritical fluid successfully escape the lithosphere and auto-ignite in the troposphere, meaning that the observed flight was not the kinetic displacement of a solid asteroid, but the progression of a reactive gas jet’s combustion front This dynamic progression of the combustion front along the opening fractures also resolves the trajectory paradox, where observers from different Siberian villages recorded completely divergent and contradictory vectors of the object's progression at the exact same moment. Classical meteorites follow fixed linear trajectories governed strictly by the laws of ballistics and cannot alter their course in mid-air due to the total absence of aerodynamic control surfaces, yet empirical observations clearly document that the luminous zone executed a monumental non-ballistic zigzag maneuver, abruptly shifting its vector from south-north to east-west. This phenomenon is fully explained by the structural geology of the Siberian Platform, where the underlying tectonic faults do not follow isolated linear paths but intersect at varying angles, forming a dense orthogonal and diagonal structural grid. As the wave of tectonic stress propagated through this grid, the outgassing occurred sequentially, shifting from one structural seam to an adjacent one, which caused the atmospheric combustion front to abruptly jump from one reactive gas plume to the next. To a distant observer at a range of 500 km, this discrete transition between active plumes created the perfect visual illusion of a sharp zigzag and a 90-degree vector inversion. Furthermore, this planetary-scale fuel-air explosive mechanism provides a precise explanation for the severe thermal radiation pulse recorded at the Vanavara trading post, situated 65 km from the epicenter, where eyewitnesses stated that the sky fractured open, a broad band of fire swept across it, and the radiant thermal flux was so intense it felt as though their clothing would ignite. During a standard meteoroid entry, generating a thermal radiation pulse of such magnitude at a distance of 65 km is thermodynamically impossible because a solid space rock undergoes rapid surface ablation while its internal core remains frozen, operating as a localized point source of radiation where energy density decays rapidly according to the inverse-square law. Conversely, the outgassing of a hydrogen cloud saturates a massive atmospheric corridor within seconds, creating a prolonged, planar source of radiation where energy density decays much slower, inversely proportional to the distance to the first power. The ignition happened at a single point, and the fiery wall rushed at supersonic speed along this pre-made gas conduit, creating the volumetric detonation that generated the radial forest blowout. The corresponding visual phenomenon of the sky splitting in two was the direct result of the immense fluid-dynamic pressure of the outgassing plume, which physically displaced atmospheric air masses and opened a low-density vacuum corridor, exposing the high-temperature reaction zone within and perfectly linking the micro-level hydrogen saturation of Sample No. 3 with the global scale of the 1908 catastrophe. Conclusion, Fluid-Dynamic Mechanism of the Volumetric Explosion In summary, the physically verified scenario of the 1908 phenomenon unfolds as follows: ultra-high-pressure hydrogen breaches the lithosphere, undergoes an instantaneous pressure drop upon entering the atmosphere, and detonates through immediate chemical contact with oxygen. Initially, the supercritical fluid, oversaturated with protium, escapes the structural faults of the Siberian Platform under a monumental geostatic pressure of 2.82 GPa. Upon entering the troposphere, the gas experiences catastrophic decompression, triggering a positive Joule-Thomson effect that causes rapid adiabatic cooling along the extraction pathway. This highly diffusive expanding hydrogen saturates a multi-kilometer atmospheric corridor within seconds, instantly mixing with atmospheric oxygen to form a stoichiometric runaway fuel-air mixture (oxyhydrogen gas). Auto-ignition occurs due to high-velocity electrostatic friction, causing a supersonic combustion front to rush along the pre-made gas conduit, which eyewitnesses visually registered as a parallel-flying "glowing log." Finally, at the tectonic fault intersection directly above the ancient paleovolcano, the deflagration sharply transitions into a brittle volumetric detonation. This 40-megaton aerial explosion genera

The Ontology of Electric Charge: Finite Gradient Flux of the Spatial Deviation Field — With a Discussion on the Geometric Origin of Electron Charge, the Topological Root of Charge Quantization, Standing Waves in Potential Wells, and the Ultimate Energy Density of the Universe

AbstractThe Standard Model treats electric charge as an unaskable intrinsic quantumnumber of elementary particles. This paper argues that charge is not an intrinsicproperty of particles, but rather the finite gradient flux of the spatial deviation field.Starting from the ontological premise that space has a minimal primitive, primitivesmust fluctuate dynamically. Constrained refreshment clusters cause primitives todeviate from the vacuum ground state. The deviation field generates a gradientfield, and the divergence theorem—as a mathematical identity—enforces gradientflux conservation. Since space has no singularities, the conserved flux must be finite.The finite flux, multiplied by a geometric constant, is electric charge. This paperfurther argues that charge quantization originates from the topological invariantof constrained refreshment clusters; the deviation ratio of a single primitive hasa theoretical upper bound of approximately a few percent, so macroscopic chargemust be distributed across many carriers; the limiting electric field gives an ultimateenergy density of approximately the fine-structure constant times the Planck energydensity; the universe therefore could not have begun from infinite density, andthe Big Bang singularity is naturally dissolved. This paper also argues that theelectron is self-sustaining because the nearest-neighbor propagation chains insidethe constrained refreshment cluster form standing waves, and the conservation lawforces the real diffusion coefficient to zero. This paper introduces no adjustable parameters and proposes no new experimentalassumptions. The paper is positioned as a conceptual framework paper,not a complete mathematical derivation: it demonstrates that charge can be understoodas a geometric quantity of space, and provides testable order-of-magnitudepredictions. Precise numerical derivations are left as future work.Keywords: ontology of charge; spatial deviation field; finite gradient flux; constrainedrefreshment cluster; charge quantization; standing waves; ultimate energydensity; singularity dissolution

Synthetic Data Supporting "Elastic Velocity-Model Inversion Using Neural Operators Regularized by a Diffusion Model"

This dataset contains the synthetic velocity models, PP and PS migration images, and inversion results supporting the manuscript “Elastic Velocity-Model Inversion Using Neural Operators Regularized by a Diffusion Model.” The files are provided to support the reproducibility of the synthetic experiments and quantitative analyses presented in the manuscript. The Volve field data are not included and should be obtained separately from the official Equinor Volve Data Village.

Aging-aware residential load scenario generation for future energy systems: a foundation-model-augmented conditional diffusion method

Deep learning of position-dependent diffusivity from umbrella sampling molecular dynamics simulations

A widely used computational framework to calculate the membrane permeability coefficient of small molecules is the inhomogeneous solubility-diffusion (ISD) model. It requires two ingredients that can be calculated using molecular dynamics simulations: the potential of mean force, which is well-defined, and the position-dependent diffusivity, which is often problematic and challenging. Two methods (Woolf-Roux and Hummer) have been proposed to determine the position-dependent diffusivity profiles using biased umbrella sampling simulations. While both are constructed from similar time-correlation functions, yet, they can disagree quantitatively. Here, a reconciliation of these methods is achieved through deep learning memory functions in the time domain. The diffusivity extracted through this analysis is shown to be in best agreement with the equilibrium counting permeability for the same membrane system compared to both the Woolf-Roux and Hummer diffusivity. The effect of memory on the rate of barrier crossing is assessed through numerical simulations of the generalized Langevin equation (GLE). The GLE is efficiently simulated via Markovian embedding, which relies on the positive, decaying exponential form of the memory functions extracted by deep learning. Our results based on the ISD permeability, the known permeability from equilibrium MD, and the numerical GLE simulations indicate that memory effects most likely do not have a significant effect on the permeation of water.

Detection-Guided ROI-Constrained Diffusion for Weakly Supervised White Blood Cell Segmentation: A Retrospective Internal and External Dataset Evaluation

Background: Accurate white blood cell (WBC) segmentation is important for quantitative microscopic image analysis, but conventional supervised approaches depend on labor-intensive pixel-level annotations and may exhibit reduced robustness across datasets. This study proposes a detector-guided, region of interest (ROI)-constrained framework that combines automatic localization, pseudo-mask-based weak supervision, and diffusion-assisted segmentation. Methods: You Only Look Once version 13 Nano (YOLOv13-N) was used to localize WBCs and define ROIs, within which segmentation was performed using the proposed diffusion-assisted model. Automatically generated pseudo-masks served as segmentation-training targets, while expert masks were retained for reference evaluation. Experiments used a verified Dicle cohort of 14,721 records, partitioned into 10,305 training, 2208 validation, and 2208 held-out test records. Performance was evaluated separately at ROI-conditional and end-to-end levels. Generalization was assessed on 11,200 Raabin-WBC records using the frozen framework without external tuning. Results: The automatically generated pseudo-masks achieved a mean Dice score of approximately 0.702, demonstrating usable but imperfect weak supervision. On the held-out Dicle test set, the proposed framework achieved a Dice score of 0.7188, intersection over union (IoU) of 0.5796, precision of 0.8966, and recall of 0.6422 under ROI-conditional evaluation. End-to-end performance was 0.6705 Dice, 0.5408 IoU, 0.8378 precision, and 0.5986 recall, demonstrating the influence of localization on overall performance. Comparison with reference segmentation architectures showed that the proposed framework did not maximize internal Dice, while achieving the highest reported ROI-conditional precision. Frozen external evaluation on Raabin-WBC revealed further degradation under dataset shift, with failure analysis identifying detector/ROI transfer as an important end-to-end bottleneck. Conclusions: The proposed framework demonstrates the feasibility of WBC segmentation using detector-guided ROI processing and pseudo-mask-based weak supervision, reducing reliance on expert pixel-level segmentation targets. The findings further show that robust cross-dataset localization is critical to end-to-end generalization and provide a clear direction for improving weakly supervised WBC segmentation across heterogeneous microscopy datasets.

Poloxamer/HPMC/Carbopol-Based Thermosensitive Hydrogel Loaded with Ibuprofen for Potential Vaginal Drug Release

Vaginal drug delivery offers a critical route for local treatments but is limited by short formulation residence times. This study describes a thermosensitive in situ gel prepared by the cold-dissolution method from a ternary blend of Pluronic F127, Carbopol 940, and HPMC for localized vaginal therapy. We used ibuprofen as a model drug selected for its reported anti-inflammatory and antiproliferative activity. The hydrogels exhibited a constant gelation temperature of 28 °C and high viscosity under simulated physiological conditions; ibuprofen incorporation further reduced susceptibility to gravitational leakage. FTIR, XRD, and DSC analyses confirmed stable physical cross-linking of the polymer network and amorphous molecular dispersion of ibuprofen. Peppas–Sahlin modelling revealed a controlled, sustained release profile (>50% over 24 h) predominantly governed by Fickian diffusion (69%). The blank hydrogel exhibited high biocompatibility (>75% viability). In contrast, the ibuprofen-loaded matrix exhibited a concentration-dependent cytotoxic effect on HeLa cervical cancer cells, reducing cell viability to ~12% at the full extract concentration. Overall, this ternary hydrogel platform represents a stable, promising vehicle for sustained local administration of ibuprofen in the vaginal microenvironment.

Présentation de la DJE-Gueu

Abstract — DJE-Gueu Model The DJE-Gueu Model (Gueu’s Youth and Employment Dynamics Model) is a nonlinear differential model designed to mathematically represent the dynamics of the economically active youth population and the potential effects of economic activation policies on youth employment and inclusion. The model is formulated as : where Y(t) represents the economically active youth population, λ\lambda the demographic growth parameter, α\alpha the intensity of economic activation policies, κ\kappa the diffusion or effectiveness rate of these policies, and μ\mu the market saturation parameter. The model combines demographic dynamics, policy activation and a nonlinear saturation mechanism within a single differential framework. Its mathematical analysis makes it possible to study the existence and positivity of solutions, equilibrium points, local stability, critical thresholds, parameter sensitivity and the long-term evolution of the youth economically active population. Numerical simulations can further illustrate the effects of changes in policy intensity and market constraints. The DJE-Gueu Model aims to provide a mathematical decision-support framework for analysing youth employment dynamics, particularly in territories where statistical, institutional and economic resources are limited. It can therefore serve as a basis for research on public policies, territorial development, economic inclusion and the achievement of Sustainable Development Goal 8 concerning decent work and economic growth. Keywords: DJE-Gueu Model; youth employment; economic activation; differential equation; nonlinear dynamics; equilibrium; stability; parameter sensitivity; economic inclusion; public policy.

Robust Image Enhancement and Restoration for Real-World Computer Vision

This chapter presents a comprehensive study of image enhancement, restoration and preprocessing techniques for real-world computer vision applications. It begins by examining the physical sources of image degradation, including noise, blur, illumination variation, compression artifacts and adverse environmental conditions such as low light and weather effects. The limitations of vision systems trained under ideal conditions are highlighted through their degraded performance in unconstrained environments. The chapter reviews classical approaches, emphasizing their interpretability and efficiency. It then explores modern deep learning techniques such as convolutional neural networks, generative adversarial networks, transformers and diffusion models, which achieve superior performance through data-driven learning. Hybrid frameworks that integrate classical and learned methods are discussed as a practical solution for balancing robustness, interpretability and performance.