1147 papers
Computational physics bridges the gap between abstract theory and real-world observation by using powerful computers to solve complex physical problems. This field allows scientists to simulate everything from the collision of subatomic particles to the swirling dynamics of galaxies, offering insights that traditional experiments alone cannot provide.
On Gist.Science, we continuously process every new preprint in this category from arXiv to make these breakthroughs accessible to everyone. Each entry is accompanied by both a clear, plain-language explanation and a detailed technical summary, ensuring that researchers and curious readers alike can grasp the significance of the latest findings without getting lost in dense equations.
Below are the latest papers in computational physics, curated to keep you at the forefront of this rapidly evolving discipline.
This paper establishes a thermodynamic theory of algorithmic catalysis within the watts-per-intelligence framework, proving that task-specific speed-ups are fundamentally limited by the algorithmic mutual information between the substrate and task descriptor, with a minimum thermodynamic cost for information installation that determines the energy-efficient deployment horizon for reusable computational structures.
This paper introduces a chaos-diagnostic framework combining multireference electronic structure theory, Adiabatic Gauge Potentials, and Random Matrix Theory to demonstrate how quantum chaos suppression at transition states enhances proton-transfer tunneling in ultracold astrophysical environments, thereby offering a new metric for refining ion-molecule reaction models in planetary atmospheres.
This study utilizes an observational dataset and Random Forest modeling to quantify how atmospheric conditions, particularly moisture convergence, instability, and water vapor, nonlinearly control the spatial and seasonal variability of mesoscale convective systems over tropical oceans, explaining up to 50% of their frequency and precipitation variance.
This study utilizes satellite-based indices and long-term MCS tracking data to demonstrate a robust two-way feedback mechanism where the Madden-Julian Oscillation organizes mesoscale convective systems through environmental modulation, while the collective upscale impacts of these systems actively reinforce the MJO's circulation and support its eastward propagation.
This paper reviews data-driven and machine learning approaches, particularly descriptor-based models and interatomic potentials trained on DFT data, that accelerate point defect simulations in solid-state materials by enabling rapid, quantum-mechanically accurate predictions of properties like formation energies and vibrational free energies for high-throughput screening and experimental integration.
This paper introduces DiffSRDA, a computationally efficient, uncertainty-aware data assimilation framework based on denoising diffusion models that generates high-resolution ensemble analyses from low-resolution forecasts and sparse observations, achieving performance comparable to high-cost methods while supporting training-free adaptation to changing sensor layouts.
This paper introduces a high-order nodal Galerkin formulation for the Müller boundary integral equation that bypasses the need for divergence-conforming basis functions by exploiting kernel cancellation to reduce hypersingularities, thereby enabling robust, superlinear convergence through a metric-weighted tangent frame and specialized preconditioning.
This Monte Carlo study verifies that clusters formed by two-dimensional simple random walks exhibit marginal logarithmic fractal behavior, possess a hull fractal dimension of consistent with SLE predictions, and display chemical distances scaling as , which aligns with the theoretical upper bound for Gaussian free field level-set percolation.
This paper introduces JAX-BEM, a differentiable Boundary Element Method solver built on the JAX framework that matches the accuracy of traditional codes while enabling efficient gradient-based acoustic shape optimization and inverse problem solving.
This paper introduces a flexible, integrated interface within GROMACS that enables hybrid machine learning/molecular mechanics simulations using PyTorch-trained neural network potentials, facilitating their seamless application to diverse biomolecular systems and advanced sampling workflows.