122 papers
This paper proposes a simple numerical method based on inverse transform sampling to efficiently load relativistic Maxwellian-type distributions by approximating the cumulative distribution of a shifted-Maxwellian energy distribution with an invertible function to generate random energy and momentum variates.
This paper proposes an improved objective function for the optAPM code, featuring a simplified hotspot cost formulation and pre-interpolation of trail data, to reduce modeling errors and enhance the accuracy of absolute plate motion reconstructions over geological timescales.
This paper introduces SMARL, a reinforcement learning framework that uses enstrophy spectrum-based rewards to develop stable, data-efficient subgrid-scale closures for geophysical turbulence, enabling accurate prediction of extreme events with significantly reduced computational costs.
The paper introduces **peapods**, a high-performance, open-source Python package leveraging a Rust core to efficiently simulate Ising spin systems on periodic Bravais lattices using a comprehensive suite of single-spin-flip, cluster, and replica-based Monte Carlo algorithms validated against exact critical temperatures.
This paper introduces the Open Polymers 2026 (OPoly26) dataset, a publicly released collection of over 6.57 million density functional theory calculations on polymeric systems designed to overcome previous computational limitations and enhance machine learning models for predicting polymer properties.
Using extensive parallel-tempering Monte Carlo simulations, this study reveals that in the three-dimensional random-bond Ising model, a secondary percolation transition involving two equal-density clusters occurs above the thermodynamic ordering points, with the subsequent divergence of these cluster densities serving as a distinct percolation signature for the ferromagnetic and spin-glass phase transitions.
This paper presents a Hybrid High-Order (HHO) formulation for variable density incompressible flows that ensures exact volume conservation and pure density advection through a combination of hybrid spatial discretization and ESDIRK time stepping, demonstrating robustness, pressure-independence, and high-order accuracy in simulating immiscible fluid mixtures and Rayleigh-Taylor instabilities.
This paper proposes an accurate and memory-efficient sum-of-Gaussians tensor neural network (SOG-TNN) algorithm that overcomes the curse of dimensionality and handles Coulomb singularities in high-dimensional Schrödinger equations through a low-rank tensor representation and a novel range-splitting scheme for electron-electron interactions.
This paper proposes a scalable physics-informed deep generative model (sPI-GeM) that overcomes the limitations of existing methods by effectively solving forward and inverse stochastic differential equations in high-dimensional stochastic and spatial spaces through a combination of physics-informed basis networks and a deep generative model.
This paper proposes a general framework for adaptive sensing of physical dynamical systems, where a trainable attention module learns optimal spatial probing and measurement combination to significantly improve prediction accuracy, effectively reframing neural networks as trainable devices for extracting information from physical systems.
This paper identifies that measurement-feedback Ising machines suffer from a significantly reduced range of effective hyperparameters compared to their time-continuous analog counterparts due to discrete operation, and proposes an experimentally verified method to mitigate this sensitivity.
This paper introduces a novel modified-gradient (MG) method that employs an implicit projection to reformulate magnetic field gradients, thereby achieving exact divergence-free results with round-off precision in meshless magnetohydrodynamics and demonstrating superior performance over constrained-gradient techniques and the GIZMO code across various test cases.
This paper presents a robust, machine-learned interatomic potential for TiC MXenes that enables efficient molecular dynamics simulations of ion irradiation, providing critical insights into defect statistics and guidelines for defect engineering.
This paper introduces a comprehensive database and two efficient, general-purpose machine-learned interatomic potentials (tabGAP and NEP) for nine refractory elements, enabling accurate large-scale simulations of diverse multiphase alloys and complex phenomena like phase transitions and radiation damage.
This paper proposes a novel symmetry-driven generative framework that combines large language models for chemical semantics with a linear-complexity heuristic beam search to rigorously enforce algebraic consistency in Wyckoff patterns, thereby overcoming the combinatorial bottleneck in crystal structure prediction to achieve state-of-the-art performance in discovering new materials without relying on existing databases.
This study utilizes atomistic and multiscale simulations to demonstrate that misfit and threading dislocations significantly reduce the surface energies of PbTe-PbSe interfaces, with heteroepitaxial processes yielding up to 50% lower energy compared to coherent interfaces.
This paper introduces a two-tier extension for the GPU-accelerated micromagnetic framework mumax+ that enables efficient, spatially resolved simulation of multimode cavity magnonics in both ferromagnets and antiferromagnets, successfully validating the tool through eight benchmarks covering phenomena ranging from coherent coupling and mode-selective addressing to dissipative interactions and level attraction.
This paper introduces a discontinuous Galerkin time-domain method that computes Casimir forces by recasting the Maxwell stress tensor into classical scattering problems driven by dipolar excitations, enabling accurate evaluation of interactions across diverse geometries and material properties at finite temperatures.
This study demonstrates that incorporating biologically motivated, state-dependent synaptic feedback into a Lipkin-Meshkov-Glick quantum brain model significantly reshapes its phase diagram by expanding the paramagnetic phase and displacing critical boundaries, a phenomenon rigorously characterized through ground-state Husimi distributions, Wehrl entropy, and mean-field dynamical analysis.
This paper introduces GET-SEI, a general framework combining graph contrastive learning, extended dynamic mode decomposition, and transition path theory to automatically characterize local atomic environments and quantify lithium transport kinetics and pathways across diverse solid-state electrolyte/lithium metal interfaces for targeted SEI engineering.