diffusion models

THE IMPACT OF FINTECH COMPANIES ON FINANCIAL INCLUSION IN NIGERIA.

Financial inclusion has become a major policy objective in developing economies due to its potential to reduce poverty, promote economic growth, and improve social welfare. In Nigeria, a significant proportion of the population has historically remained outside the formal financial system due to factors such as inadequ...

Pointwise Monotonicity of the Allen--Cahn Flow and Dynamical Limitations of Energy-Stable Schemes

Abstract Energy stability is fundamental in the numerical approximation of phase-field models, but it does not by itself guarantee faithful reproduction of local dynamics. Recent ODE-level studies have shown that energy dissipation alone does not ensure dynamical fidelity for spatially homogeneous phase-field models (X...

The existence and instability of blowing-up steady states for the Shigesada-Kawasaki-Teramoto competition model with cross-diffusion

Diffusion-Based Nanostructure and Self-Healing Characterization of Asphalt Binders Using Molecular Dynamics Simulations

Conventional macroscale and microscale experiments have not fully elucidated the thermodynamical behaviors of asphalt binders’ self-healing, particularly under varying thermal conditions and fracture dimensions, limiting understanding of nanoscale molecular processes. This research quantifies density changes and he...

A colour image encryption method based on a coupled map lattice–elementary cellular automaton neurodynamic chaotic system and dynamic dual-affine substitution-box architecture

Abstract Secure colour image transmission requires encryption schemes in which randomness generation, nonlinear substitution and diffusion are designed as a coherent whole. Conventional coupled map lattice-based methods are often constrained by fixed neighbourhood evolution and limited coupling diversity, while many ch...

Modeling long-range dependencies in knee joint contact mechanics: topology diffusion, global routing, and their hybridization.

Findings indicate that topology diffusion provides a robust foundation, while global routing offers complementary gains in reconstructing clinically relevant high-stress patterns.

Deterministic Topological Retrieval-Augmented Generation via Discrete Dynamical Basins

Retrieval-Augmented Generation (RAG) systems navigate a fundamental trade-off between the computational simplicity of flat vector indices and the structural richness of graph representations. Standard dense retrieval collapses document discourse into isolated embeddings, causing semantic fragmentation and context loss....

DeepFake-BMA: Photorealistic AI-Generated Faces for Binary Detection and Multiclass Attribution

This dataset contains photorealistic, high-resolution AI-generated facial portraits created by the authors to support research in synthetic media forensics, deepfake detection, and generative source attribution. Existing synthetic face benchmarks focus heavily on legacy Generative Adversarial Networks (GANs) or early l...

Rational AI Dependence Through Capability Augmentation: A Reference-Dependent Model Without Requiring Skill Degradation

Generative AI can augment human productivity and cognitive capability. Existing discussions of AI dependence have addressed cognitive offloading, overreliance, skill degradation, delegation of cognitive agency, and the social diffusion of AI use. This paper proposes a complementary pathway: rational AI dependence throu...

EWF 2.0: exact sampling of allele trajectories using the Wright–Fisher diffusion with time-varying demography

Abstract Background Accurate modelling of allele frequency trajectories requires incorporation of both genetic mechanisms such as selection and mutation, as well as realistic population demography. Accounting for a non-constant demography within a Wright–Fisher diffusion framework induces a time-inhomogenous drift coef...

Representation learning in diffusion and flow-based model: an application aspect

This work systematically categorizes representative methods across a wide range of downstream tasks, including image classification, dense visual prediction, instance-level perception, and annotation-scarce scenarios, and proposes a three-tier progressive framework that organizes existing works from three perspectives.

Comment on egusphere-2026-4368

Abstract. We propose a reduced axisymmetric tornado model in which the vertical transport of momentum is nonlocal, governed by a fractional derivative whose order is not a free parameter but is fixed by the observed coherent substructure of the vortex through the hierarchical cascade relation α = 2−2/(N+1). Building on...