Browse Categories

Fluid / CFD - Airfoil RANS

Parametric airfoil RANS flow-field dataset - compact and ML-pipeline friendly, the go-to starting point for learning fluid surrogates.

CFD Benchmark / Multi-Scenario

Multiphysics benchmark for computational fluid dynamics covering internal/external flows and thermal convection - includes transient and steady-state data.

CFD/Computational Fluid Dynamics

A computational fluid dynamics dataset hosted on HuggingFace, focusing on wind-related fluid simulation data for CFD researchers and engineers.

Materials

A large-scale pretraining corpus for materials science, containing approximately 780M tokens of RefinedWeb data, designed for training AI models for materials discovery and molecular dynamics research.

CFD/Computational Fluid Dynamics

A comprehensive benchmark dataset for training and evaluating neural PDE solvers, containing six standard tasks on regular grids, structured grids, and irregular geometries, used for unified evaluation in the ICML 2024 paper Transolver.

ScientificML

MaCroScope physical macroscopic dataset with 10k samples for machine learning training and evaluation in physics-related scenarios.

CFD/Computational Fluid Dynamics

A large-scale benchmark dataset for machine learning in computational fluid dynamics, covering four classic CFD problems—lid-driven cavity, laminar pipe flow, dam-break, and flow around a cylinder—to evaluate neural operator generalization under unseen conditions.

Materials

A technical Q&A dataset covering fundamental and advanced topics in materials science, including material properties, synthesis methods, characterization techniques, and engineering applications, designed for training technical response models.

Materials

A comprehensive materials science dataset with ~500K peer-reviewed publications, 139K phase diagrams, 409K crystalline nanostructures, and 1M+ physical property sets, enabling computational materials design and ML-driven materials informatics research.

ScientificML

A HuggingFace-hosted machine learning dataset for physics domain, part of the MaCroScope series, with Physics-5k containing approximately 5000 samples.

ScientificML

A HuggingFace dataset repository related to PDE interpolants, supporting research at the intersection of physics modeling and machine learning.

CFD Surrogate

A HuggingFace-hosted dataset repository for computational fluid dynamics surrogate modeling, containing data resources for CFD simulation and surrogate model development.

Computational Fluid Dynamics

A large-scale CFD-generated wind field simulation dataset containing over 12,000 cases, designed for training and evaluating terrain-informed machine learning models to reconstruct fine-scale 3D wind fields

ScientificML

A cleaned version of the MaCroScope physics-related experimental dataset with approximately 20k entries, hosted on HuggingFace for research purposes.

CFD

A dataset of 2000 steady-state CFD simulations for a liquid-cooled dual-chip pin-fin cold plate, accompanied by a POD-NN surrogate model and code to reproduce all figures in the associated paper.

CFD

Large-scale CFD-generated wind field dataset with over 12,000 cases, designed for training and evaluating the terrain-informed machine learning model FuXi-CFD for fine-scale 3D wind field reconstruction.

ScientificML

A provenance-bearing benchmark for verifying symbolic candidate solutions to PDEs, containing 20 records with generator outputs, verification cases, and human annotations for assessing symbolic solver reliability.

Multibody/Robotics

A robot arm manipulation dataset collected using the LeRobot framework, containing 51 episodes and 29803 frames of video data for robot manipulation policy learning.

ScientificML

A benchmark dataset of complete numerical solutions for seven PDE families including Darcy, Poisson, and Helmholtz, generated by the benchmark's own solvers for scientific ML model evaluation.

Materials

A materials science demonstration dataset hosted on HuggingFace under the SciFy account, showcasing data organization and access patterns for materials science applications.

ScientificML

MaCroScope-Physics-1k is a physics-related scientific dataset published by the ai4collaboration team, designed to support research at the intersection of physical modeling and machine learning.

ScientificML

A PDE evaluation dataset containing 32 physical systems, each with four representations: source code, natural-language description, governing equation, and numerical trajectory, with both correct and corrupted variants.

CFD

A physics simulation dataset accompanying the GeoPT paper, covering 5 downstream tasks including aerodynamic simulations, with metadata on mesh sizes, variables, and train/test splits.

Computational Fluid Dynamics

A computational fluid dynamics surrogate model training dataset hosted on HuggingFace by user harryshah14.

Materials

A materials science dataset hosted on HuggingFace, supporting machine learning and materials science cross-disciplinary research, containing relevant domain training or evaluation samples.

CFD

A 36M-parameter UNet-based model for solving Navier-Stokes equations using a point-wise coupling strategy for high-accuracy flow field prediction.

ScientificML

A benchmark dataset for PDE surrogate tasks within the ai4sci-tasks evaluation framework, containing training trajectories and evaluation splits for assessing neural operator and surrogate modeling performance.

Materials

A collection of domain ontologies for materials science and engineering, providing semantic organization of core concepts including structure, properties, processing, and performance for knowledge reasoning applications.

ScientificML

MaCroScope-Physics-4k is a physics-related dataset hosted by ai4collaboration on HuggingFace, with the name suggesting approximately 4,000 samples, potentially for machine learning and physics modeling research.

ScientificML

Simulation training datasets for PDE control containing 3000 controlled trajectories generated by classical numerical solvers under randomly sampled control sequences, designed for training TI-DeepONet surrogate models.

Multibody

Proposes specification-conditioned human demonstration collection for embodied AI in construction, bridging the gap between design model configurations and physical environments.

Physics-Informed Neural Networks

A HuggingFace repository containing a potential flow PINN clustering model, providing physics-informed neural network resources for potential flow modeling.

ScientificML

A downstream physics simulation dataset accompanying the GeoPT paper, containing multi-scale mesh data across 5 simulation tasks to validate the generalization of geometric pre-trained neural simulation models across diverse physics domains.

Hydrology Data

A daily average discharge dataset from Ait Melloul hydrological station in Morocco spanning 1970 to 2024, sourced from Morocco Open Data, containing 34 rows and provided in ML-ready Parquet format.

ScientificML

A code perturbation dataset with 256 PDE solver implementations comprising 64 base programs and 4 lexical perturbation variants each, designed for evaluating LLM understanding of physics simulation code.

Automotive Aerodynamics / Large-Scale CFD

MIT's large-scale automotive aerodynamics dataset: 8,000 parametric car geometries with CFD flow fields, meshes and aerodynamic coefficients.

PDE Benchmark / Differentiable Solver

Microsoft Research's open-source PDE benchmark suite with differentiable solvers and spatiotemporal dynamics data for evaluating long-horizon rollout stability.

Industrial Heterogeneous-Mesh Simulation

ICML 2026 industrial physics-ML dataset with 6 heterogeneous-mesh simulation sets, CGNS-compatible, distributed via HF+Zarr - closest to real CAE workflows.

PDE Benchmark / Multi-Equation

Large-scale PDE benchmark covering heat, advection-diffusion and fluid dynamics - supports fair comparison of data-driven and physics-informed methods.

Multiphysics / PDE - 16 Domains

Unified benchmark spanning 16 physical domains, dubbed the 'ImageNet moment of physical ML', with PyTorch API, pretrained baselines and HF streaming.

3D Turbulence (DNS Validation)

2026 3D turbulent DNS benchmark with 5 tasks and controlled generalization splits, designed to test extrapolation of turbulence models.

Turbulence / Experimental & Simulation Validation

Classic experimental & simulation dataset of turbulent wakes/cavity flows from TU Delft - covers multiple bluff-body flows, a standard CFD validation benchmark.

Conjugate Heat Transfer / Parametric Simulation

10,000 parametric 90 degree bend fluid-thermal coupling simulations across curvature ratios and Reynolds numbers - for heat-exchanger design and surrogates.

PINN Starter Framework

The most beginner-friendly PINN framework, with excellent Chinese docs and TensorFlow/PyTorch backends - the top choice to start with physics-informed networks.

Finite Element Framework / Multiphysics

Open-source FEM-based multiphysics simulation framework; describes variational problems in UFL and solves large-scale PDEs - a standard academic tool.

Differentiable Fluid Solver

Google Research's differentiable JAX-based fluid solver enabling end-to-end gradients for embedding neural networks into CFD time stepping.

Data Analysis / Parallel Computing

Distributed DataFrame library, a drop-in pandas replacement that auto-parallelizes across cores/clusters (Dask/Ray/Intel) for large-scale simulation data analysis.

Differentiable CFD Solver

Fully auto-differentiable CFD solver on JAX (CPU/GPU/TPU) for end-to-end differentiable fluid simulation, coupling seamlessly with neural networks.

Neural Operator Library

Reference implementation library for FNO / DeepONet and other neural operators (MIT) - the de-facto standard for learning 'infinite-dimensional maps' as surrogates.

General-Purpose CFD Solver

The de facto open-source CFD solver covering incompressible/compressible flow, heat transfer, multiphase and combustion - the workhorse for generating physics-AI training data.

Numerical Solver Library / HPC

High-performance scientific linear/nonlinear solver library for large sparse systems - the underlying engine of top-tier CFD and multiphysics simulation.

Solver / Training Framework

NVIDIA's physics-ML training framework (formerly Modulus) with PINN, neural operators and physics-constrained training - commercial-friendly, industry-first choice.

PDE Foundation Model / Downstream Benchmark

ETH Zurich's PDE foundation model and downstream fine-tuning suite, using semigroup augmentation to mine abundant training pairs, covering 15 downstream PDE tasks.

Visualization / Post-Processing

Python 3D visualization and mesh-processing framework for VTK - concise API for field data, mesh rendering and animation; the go-to viz library for simulation results.

Aerodynamics / Design Optimization Solver

Open-source CFD and adjoint solver with built-in aerodynamic shape optimization, widely used for airfoil/aircraft analysis and gradient-driven design.

PINN

A heat conduction equation solver based on Physics-Informed Neural Networks, using neural networks embedded with heat conduction physics to predict temperature fields.

ScientificML

Warp-Geo provides a GPU-accelerated differentiable framework for reconstructing signed distance fields and evaluating surface normals from complex 3D point clouds, enabling gradient-based shape optimization.

ScientificML

Qwen3-8B base LLM fine-tuned on mixed biology and physics data via GRPO; not a specialized physical simulation or modeling tool.

PINN Model

A HuggingFace repository hosting a Physics-Informed Neural Network model for hydrostatic fluid mechanics computations, applying physics-based constraints to deep learning approaches.

PDE Solving

DeepXDE v1.15.0 released with L-LAAF activation function support for JAX backend and Apple MPS GPU acceleration for PyTorch backend, improving deep learning-based PDE solving performance.

Scientific Computing

A PINN-related model resource hosted on HuggingFace by OneScience-Group, designed for scientific computing and physics-based problem modeling.

Inference Optimization

TensorRT Edge-LLM achieves significant inference acceleration on Jetson AGX Thor edge platform, substantially improving AI agent deployment performance at the edge.

Scientific Computing/Finite Element

FEniCS/dolfinx releases a patch version that fixes a Mathjax rendering issue in the built documentation. This release is functionally identical to v0.11.0, so users do not need to upgrade.

Neural Operator

A neural operator model based on spherical Fourier transforms, designed for learning partial differential equations on spherical geometries.

Embodied AI / Robot Foundation Model

Unitree releases UnifoLM-WLA-1.0, a 6B parameter general-purpose embodied AI model trained with ~2500 hours of real-machine data, achieving 7 open-source SOTA results in embodied reasoning.

Neural Operator

A neural operator-based predictor model hosted on HuggingFace, designed for sampled measurement prediction tasks.

Physics Simulation

NVIDIA PhysicsNeMo v2.2.2 release fixes documentation alignment with the PyPi package name, improving installation accuracy.

Scientific Computing

NeuralOperator 2.0.0 major release introduces Codano model, improves documentation, and enhances usability for training Neural Operators in PyTorch.

Neural Network / AI for Science

A fundamental physics neural operator model project published by Delta Lab on HuggingFace. Refer to official repository documentation for specific functionality and performance details.

Quantum/Chemistry

A context-conditioned molecule transfer model based on InternS1-Mini, designed for drug-induced liver injury (DILI) related molecular transformation tasks in chemical reaction prediction and drug discovery.

PINN

A Physics-Informed Neural Network model for solving compressible Euler equations, hosted on HuggingFace.

Physics-Informed Neural Networks

A physics-informed neural network model hosted on HuggingFace, with name suggesting relevance to MRI imaging and SIREN architecture; detailed information requires consulting the official page.

Fourier-PINN

A HuggingFace model repository with a name suggesting Fourier-enhanced Physics-Informed Neural Networks implementation.

CFD

NVIDIA Earth-2 leverages AI atmospheric models integrated with observation data to provide high-resolution weather forecasts and decision support for weather-sensitive industries.

Physics-informed Neural Network / CFD

A HuggingFace repository hosting a physics-informed neural network model for computational fluid dynamics, implementing physics constraints embedded within neural network architectures.

CFD

A 36M-parameter supervised learning model based on U-Net architecture designed for solving Navier-Stokes fluid dynamics PDEs, hosted on HuggingFace.

Materials

Presents a general theory for rolling arbitrary 2D crystals into nanotubes, implemented as open-source NTBuilder with desktop and web interfaces, including a catalog of 20 million nanotube structures.

ScientificML

A 36M-parameter neural network solver for partial differential equations based on U-Net architecture, enabling data-driven solving and prediction of PDE systems

Scientific Computing

A scientific computing model on HuggingFace that integrates Dynamic Mode Decomposition with Neural Operator for scientific data modeling and prediction tasks.

PINN Model

A HuggingFace-hosted model repository for bubble dynamics using Physics-Informed Neural Networks (PINN), with the model name indicating focus on PINN-based bubble dynamics applications.

ScientificML

A deep learning-based context-conditioned molecule transfer model designed for carcinogen-related molecular structure migration and generation in cheminformatics.

Aerodynamic Design / Surrogate Model

Airfoil aerodynamic simulation and surrogate study: large parametric CFD datasets of NACA airfoils plus FNO/DeepONet surrogate comparison.

Automotive Aerodynamics / Drag Prediction

Train surrogates on high-fidelity LES data for millisecond drag-coefficient (Cd) prediction, enabling digitalization and aero optimization of vehicle external flow.

Thermal Management / Physics-Informed ML

A physics-informed convolutional neural network whose loss encodes the finite-difference heat conduction equation, used as a surrogate for the temperature field of a liquid-cooled battery pack and more accurate than a purely data-driven model trained on the same data.

Structural Health Monitoring / Digital Twin

Real-time SHM digital twin of a long-span suspension bridge via PINN, fusing strain measurements with physics constraints for damage detection and remaining-life prediction.

AI Simulation

This NVIDIA blog post describes how agentic AI workflows can be used to prepare and validate digital twins for physical AI systems, with agents capable of inspecting 3D scenes and authoring simulation-relevant data.

AI Inference Optimization

NVIDIA Groq 3 LPX leverages deterministic execution technology to enhance inference efficiency on the Vera Rubin platform, enabling high-interactivity and power-efficient computing for AI factory workloads.

Geophysics

NVIDIA Earth-2 enables weather-sensitive industries like energy to make timely decisions by integrating observational data with AI-powered atmospheric simulation.

Materials Science / First Principles

Google DeepMind's GNN system for materials discovery predicting 2.2M new crystal structures with 380k stable ones, expanding known stable materials by nearly an order of magnitude.

Weather / Large-Scale Physical Prediction

A graph neural network for medium-range weather prediction that outperforms conventional numerical weather prediction on many targets while running in minutes instead of hours.

Biomedical Flow / Reduced-Order Model

Neural networks predict 0D reduced-order model parameters from geometry, cutting cardiovascular flow error by over half and running complex pulmonary cases in under 2 seconds on a laptop.

Flow Control / Turbulent Drag Reduction

An open-source reinforcement learning platform for flow control that couples six CFD solver backends with dozens of pre-configured environments; trained policies transfer zero-shot to a wing and cut skin-friction drag substantially.

Molecular Dynamics / ML Interatomic Potential

A general-purpose machine-learned interatomic potential built on an equivariant graph neural network, trained on 150k public inorganic crystals, that runs stable molecular dynamics out of the box for solids, liquids, gases and reactions, and can be fine-tuned to ab initio accuracy with only a handful of configurations.

Geophysics / Inverse Problem

LANL's large-scale FWI benchmark - 12 datasets, 2.1 TB - reconstructing subsurface velocity from seismic data with data-driven methods, serving carbon storage, reservoir and earthquake applications.

Energy / Fusion Control

Deep reinforcement learning directly commands tokamak magnetic coils to hold diverse plasma configurations, a landmark public case of AI controlling a real physical device.

Structural Optimization / Generative Design

A conditional diffusion approach to structural topology optimization from MIT that uses surrogate-model guidance to optimise both mechanical compliance and manufacturability, outperforming conditional GANs by a wide margin.

Wind Energy / Load Prediction

Data-driven prediction of blade fatigue load and power curve, solving hard-to-evaluate turbulent loads under extreme winds - for farm O&M decisions.

Survey / Reading Material

A focused survey of PINN improvements via transfer learning and meta-learning, addressing PINN's notorious convergence difficulty and the need to re-optimize on every PDE parameter change - the engineering reference for productionizing PINNs.

ScientificML

A survey developing a unified, concept-driven view of major continuous-time machine learning branches, characterizing their mathematical relationships and design trade-offs.

Electromagnetics

A systematic review of physics-informed machine learning applications in electricity system design, monitoring, and control, focusing on hybrid architectures and physics constraint embedding methods.

Multibody

HuggingFace blog tutorial on using NVIDIA Warp physics framework and MjWarp bridge for robotics simulation, covering rigid body dynamics and contact mechanics in learning workflows.

Multibody

A survey on embedding physics priors in robot learning, covering how physical constraints like contact mechanics, rigid-flexible coupling, and multibody dynamics improve sample efficiency and reliability of learned policies.

ScientificML

PhD course lecture notes from University of Bozen/Bolzano covering Physics Informed Neural Networks and Neural Operators with PyTorch implementation tutorials.

Multibody

NVIDIA tutorial demonstrating GPU acceleration techniques for ROS 2 computational graphs using Isaac ROS and AI Agent integration to speed up robotics workloads.

Fourier Neural Operator / Beginner Tutorial

Systematic FNO from theory to code - for engineers and researchers wanting neural operators as PDE surrogates.

Finite Element / From Scratch

Build a 2D linear-elastic FEM solver from NumPy/SciPy - no FEniCS/Abaqus - to deeply understand every matrix derivation and assembly step.

Neural Operator / Beginner Tutorial

Systematic DeepONet theory and DeepXDE implementation - for PINN developers handling multiphysics-parameter scenarios.

Survey / Reading Material

A survey of data-driven and physics-informed neural operators, organizing the architecture landscape (DeepONet / FNO / GNO / PINO / Transformer-based / diffusion) along three axes: scalability, generalizability, and data sensitivity.

Tutorial / Reading Material

A practical focused review of neural operators covering DeepONet, FNO and PCANet with hands-on training strategies; benchmarked on Poisson and linear elasticity, and demonstrated as surrogates for Bayesian inversion - an engineering-ready primer.

Survey / Reading Material

A systematic survey of operator learning covering the theory, architectures and error analysis of DeepONet, FNO and their variants - the reference entry point to learning infinite-dimensional maps.

Textbook / Systematic Course

A free online textbook from TU Munich covering physics + deep learning: differentiable physics, PINNs, neural operators and turbulence modeling, with runnable code.

Survey / Reading Material

A systematic survey of physics-knowledge + ML methods for PDE prediction, splitting the field into architectural-level (loss, structured prediction, augmentation) and data-level (multi-task, meta, contextual) approaches, with industrial perspective and open-source ecosystem review.

Learning Roadmap / Systematic Overview

Systematic path from classical PDE numerics to data-driven physics AI - FEM, FVM, surrogates, PINN and their evolution.

Beginner Tutorial

A classic Chinese structured tutorial implementing PINN from minimal dependencies, paired with Raissi's original repo - the best start for Chinese-speaking learners.

Original Research Paper

ICML 2026 paper proposing APT, a geometry-/mesh-/physics-agnostic neural operator that fuses graph-based local feature extraction with global attention for long-range physical effects; achieves 4× saturation error reduction on 2D faulted CO2 storage vs prior SOTA, 19,873× speedup over GEOS and 6,800× over ECLIPSE.

Original Research Paper

ICLR 2026 paper proposing ATOM, a quasi-equivariant pretrained transformer neural operator that requires no explicit molecular graph and uses temporal attention for parallel decoding of multiple future states; curates TG80 (80 compounds, 2.5M fs trajectories) and demonstrates zero-shot generalization on MD17/RMD17/MD22.

Foundational Paper

Presents a branch-trunk architecture realizing a deep learning version of the universal operator approximation theorem - the first systematic architecture for learning nonlinear operators and still a mainstream baseline.

Foundational Paper

Introduces neural operators with integral kernels parameterized in Fourier space, enabling resolution-invariant PDE solving - a foundational and highly cited work in physics AI.

Original Research Paper

Reframes Transolver's Physics-Attention as a special case of linear attention and proposes LinearNO (Linear Attention Neural Operator); achieves SOTA on six PDE benchmarks with 40.0% fewer parameters and 36.2% less compute, plus superior performance on AirfRANS and Shape-Net Car industrial datasets.

Electromagnetics

Develops a droop-aware extension of GridFM foundation model for power systems, embedding droop gains and deadband parameters as node features for control-aware AC power flow analysis.

ScientificML

HIPNO leverages symmetry-aware physics-informed neural operators to infer hemodynamic states from non-invasive signals, addressing scale symmetry challenges in physics-informed hemodynamic inference and expanding access to advanced monitoring.

CFD

A comparative study benchmarking autoencoders against parsimonious diffusion maps (PDMs) grounded in classical numerical analysis for reduced-order modeling of Navier-Stokes turbulent flows.

Materials

Presents FRPN, a Full-Resolution Polymer Network that retains both atom-level details and monomer-level connectivity for multi-scale polymer property prediction.

Materials

Investigating nonequilibrium phase-ordering kinetics in binary mixtures of uniaxial ellipsoidal particles using large-scale molecular dynamics simulations with anisotropic Gay-Berne potential, revealing two-stage ordering processes across different time scales after temperature quench.

Electromagnetics

A novel computationally efficient surrogate model for linear mode coupling in graded-index multimode fibers, tunable via a single measurable parameter, benchmarked numerically and experimentally on a 90-polarization-mode fiber.

CFD

Proposes PITT methodology combining symbolic equation tokenization with Fourier neural operator encoder to solve nonlinear hyperbolic balance laws for relativistic Schwarzschild-Burgers fluid flows with shock waves.

Geophysics

FREESIA addresses multimodal posterior estimation under high-dimensional sparse observations by using covariance-aware posterior transport, improving inference accuracy for nonlinear or non-injective observation operators in dynamical systems.

ScientificML

Proposes a ladder of information limits to quantify predictability boundaries for reduced-order models under limited information, covering memoryless and memory-augmented trajectory forecasts, event prediction, stationary statistics, and generative laws.

Quantum/Chemistry

A spectral co-design framework that directly couples fabrication-constrained SPSA-Adam optimizer with full-wave finite-element eigenmode simulations for multifrequency Rydberg electrometry, operating without analytical or adjoint gradients.

ScientificML

LPINNs propose multiplying the first hidden activation of a dense PINN by input-dependent localization functions, endowing first-layer units with receptive fields to improve optimization on long domains and for high-order operators without domain partitioning.

Multibody/Robotics

PIVOT combines physically-informed reasoning with vision-language models for off-road traversability assessment, enhancing navigation safety in unstructured terrain for field robots.

CFD

Proposes an integrated CFD-ML framework for optimizing pod geometry in supersonic evacuated tube transport, leveraging Latin Hypercube Sampling and Bezier curve parameterization for aerodynamic design.

Electromagnetics

A GPU-accelerated implicit solver for electromagnetic field computations in full 3D tokamak simulations, featuring co-designed hierarchical domain decomposition and fast exact local solver for efficient large-scale parallel GPU computing.

Multibody

This paper proposes a data-driven policy construction method for dynamical systems using non-expert MPC demonstrations, introducing a memory-based nonparametric policy with Lipschitz-regularized upper envelope of the finite-horizon surrogate value function to approximate infinite-horizon optimal control.

ScientificML

This paper investigates using artificial neural networks as surrogate models for black-box optimization, enabling efficient optimization when analytical function descriptions are unavailable by obtaining objective values through numerical simulation.

Quantum/Chemical

This paper presents a privacy-preserving hybrid quantum-classical computing framework that integrates parameterized quantum circuits with classical components, targeting quantum computing proxy applications such as chemistry simulation.

ScientificML

Proposes the Fractional Fourier Neural Operator (FrFNO) that embeds the analytic Mittag-Leffler propagator to address the computational challenge of memory-dependent time derivatives and nonlocal fractional Laplacians in fractional PDEs, enabling efficient surrogate modeling.

Thermal

Therm-FM adapts a pretrained PDE foundation model to steady-state and transient 3D-IC thermal simulation, reducing data-generation costs for cross-design reuse.

ScientificML

Introduces a decision-epoch-based framework for dynamic electric autonomous dial-a-ride, jointly optimizing vehicle routing, scheduling, and charging decisions using combinatorial optimization and machine learning.

ScientificML

Proposes a graph hypernetwork approach using an operator graph that explicitly represents PDE fields, derivatives, terms, and residuals as nodes, enabling amortized neural solver meta-training across related partial differential equations.

ScientificML

Proposes a triply-scalable equivariant Gaussian process method that encodes symmetries while maintaining probabilistic predictions and uncertainty quantification for large-scale scientific computing.

CFD

Presents physics-informed hemodynamic modeling for coronary intervention, enabling data-free prediction and sparse-data assimilation to overcome limitations of traditional numerical analysis and purely data-driven approaches.

Quantum/Chemical

For quantum many-body physics and ab initio quantum chemistry, this work derives an unbiased direct gradient estimator and introduces the adaptive minimum-variance phase (AMVP) estimator to overcome training failures in weak-gradient regimes for neural-network variational optimization.

Electromagnetics

K-TRAIL is a simulator-guided generative framework combining diffusion-based layout generation with derivative-free ensemble Kalman guidance for automated EM/RF circuit inverse design.

Materials

Presents the first complete machine learning method for accelerating plane-wave DFT in materials under the PAW formalism, formalizing seven criteria that a Complete Neural Electronic Initializer must satisfy and addressing two missing structure-dependent components: augmentation occupancies and spin initialization.

Quantum/Chemistry

A physics-informed dual-state query learning framework for GPCR ligand bioactivity prediction, addressing the neglect of dynamic conformational equilibria in existing deep learning approaches to improve functional efficacy discrimination.

Materials

Proposes high-order variable-scaled energetic variational neural networks (VS-EVNN) for phase-field gradient flows, using interface-scale normalization of spatial inputs to improve discretization accuracy.

Geophysics

This paper presents a domain-decomposed random-feature evolutional deep neural network (DD RF-EDNN) for solving transient pressure diffusion in heterogeneous porous media with discontinuous and high-contrast permeability, combining spatial approximation with temporal evolution.

ScientificML

Proposes Graph Wavelet Compressed Sensing (GWCS) for offline compression of graph signals via sparse wavelet-domain representations, reducing data preparation costs for scientific machine learning training.

CFD

This paper presents an unconditionally stable second-order numerical method for fluid-poroelastic structure interaction, where free fluid governed by incompressible Stokes equations couples with poroelastic medium via the Biot system, reformulated using Robin-type interface conditions.

ScientificML

This paper presents a comparative analysis of two neural network models for Initial Orbit Determination from Very Short Arcs, addressing challenges in Near-Earth Object surveillance where classical methods fail due to insufficient observational arcs.

ScientificML

Proposes a variational latent Markov operator framework that encodes physical states as latent distributions and evolves them through probabilistic transitions, addressing error accumulation and distribution shift in neural PDE solvers for long-horizon prediction.

Biomechanics / Computational Medicine

This paper introduces a Bayesian surrogate-modeling framework for inferring regional Holzapfel-Ogden material parameters in a left ventricle partitioned into five physiological zones, using cardiac MRI data to overcome the limitation of treating the myocardium as mechanically homogeneous.

CFD

Proposes a physics-guided hybrid architecture combining UNet and FNO, integrating multi-modal Sentinel-1/2 remote sensing with shallow water equation constraints for end-to-end flood prediction.

ScientificML

Proposes a geometry-aware incremental neural operator to address error accumulation in long-horizon autoregressive PDE prediction, mitigating spectral inconsistency and unstable channel couplings.

ScientificML

A joint training approach that addresses neural operator out-of-distribution degradation and PINN multi-solution convergence errors by leveraging imperfect neural operators to guide PINN optimization.

CFD/Fluid Dynamics

A Transformer-based neural operator approach for 3D wind field prediction over complex mountainous terrain, addressing bottlenecks in traditional CFD methods that require expert-intensive mesh generation and computationally expensive iterative solvers.

ScientificML

A unified framework combining physics-informed neural networks with finite element methods via functional Gauss-Newton problems and Petrov-Galerkin discretization.

ScientificML

A deep-learning surrogate model based on residual CNN predicts Earth-induced transition probabilities for solar neutrinos across energy and zenith angle parameter space.

ScientificML

Investigates three variational objectives for VAE-based amortized Bayesian inference in inverse problems, revealing how posterior regularization impacts weakly identified parameter directions.

Electromagnetics

Proposes two ML-based calibration methods to correct electromagnetic shower feature discrepancies between Monte Carlo simulations and real collision data in the CMS calorimeter.

ScientificML

Proposes a kernel estimator for nonparametric tangent vector field regression on Riemannian manifolds without boundary, using parallel transport and volume correction, with theoretical derivation of uniform second-order bias and stochastic convergence rates.

Electromagnetics

The Anisotropic Maxwell Neural Operator enables rapid parametric modeling of ion cyclotron resonance heating full-wave responses for the EAST tokamak by learning the one-parameter dielectric-field family within a single operator.

ScientificML

Introduces backward logarithmic orthogonal functions constructed via terminal-endpoint logarithmic mapping of generalized Laguerre polynomials, designed for backward spectral approximation of problems with weak endpoint singularities.

Structural

Proposes a ROM-based preconditioner to accelerate topology optimization, reducing computational cost of high-fidelity finite element analysis in structural conceptual design.

ScientificML

Proposes Transformer-based learning of structured representations for symbolic ODE expressions, preserving commutative properties to enable mathematical reasoning in symbolic form, distinct from traditional methods relying on numerical fitting.

Computational Geometry

This paper presents smooth spherical splines combined with conformal factors to solve Poisson equations and eigenvalue problems for the Laplace-Beltrami operator on genus-zero Riemannian surfaces.

Quantum/Chemistry

Proposes finite-state automata encoding recursive lattice geometry for exact MPO construction of tight-binding Hamiltonians across fractals, quasicrystals, and hyperbolic lattices.

Geophysics

A continental-scale electrical resistivity model of Australia's shallow crust (up to ~650 m depth) derived from probabilistic deep learning inversion of 26.6+ million AEM soundings, with quantified uncertainty for geology, groundwater, and critical mineral exploration.

ScientificML

Evaluates six sleep staging foundation models under reduced PSG channel conditions, providing insights for lightweight sensing solutions in long-term home monitoring.

ScientificML

Proposes the Kuramoto Neural Operator, introducing coupled oscillator dynamics into operator learning to overcome limitations of fixed-basis architectures for physical phenomena governed by local interactions.

ScientificML

Proposes ADEx-FNO, a framework that extends Fourier neural operators to varying geometries and independent discretizations via ambient-domain embedding and signed distance functions, without modifying core Fourier-operator layers.

ScientificML

Proposes a nonlinear reduced-order model with convolutional operator compression to overcome the Kolmogorov N-width limitation of linear projection-based ROMs, achieving accurate approximations with fewer modes for parametrized PDEs.

Multibody/Robotics

This paper addresses kinematics estimation and closed-loop control for tensegrity-based continuum robots by parameterizing complex continuous bending configurations using strut inclination angles, overcoming limitations of elastic backbone models.

ScientificML

Proposes a unified scalable framework combining structured decomposition with SDP certificates and small-gain analysis for robust stability and performance analysis of feedback systems with high-dimensional NNs and unmodeled dynamics.

Geophysics

Proposes a three-step framework for ensemble smoother with multiple data assimilation (ES-MDA), unifying existing adaptive variants to balance accuracy and computational cost in reservoir engineering inverse problems.

CFD

Proposes a geometry-aware anisotropic boundary correction method using neural operators as an efficient alternative to traditional CFD solvers for aerodynamic simulation, improving physical modeling near solid boundaries.

ScientificML

Proposes a method to enforce hard physical constraints in generative models for semiconductor manufacturing by construction, ensuring that generated masks, layouts, and process recipes strictly comply with lithography, transport, reaction, and device-physics constraints.

Geophysics

A continually learning neural-operator surrogate model replaces the traditional forward solver for efficient Bayesian inversion of three-dimensional time-domain airborne electromagnetic data, addressing prohibitive computational costs.

ScientificML

Proposes a low-rank alternating direction implicit method for solving large-scale discrete-time non-symmetric algebraic Riccati equations arising in game-theoretic Nash equilibrium computations, significantly reducing storage and computational complexity.

ScientificML

Proposes a collaborative framework of AI virtual labs under a citation-based influence economy to automate neural operator architecture search and training evaluation.

CFD

This paper develops a structure-informed data-driven ROM for transport-dominated systems governed by scalar hyperbolic conservation laws, using the kinetic defect formulation to separate characteristic transport from entropy defects localized on shock manifolds.

ScientificML

Proposes a Bayesian decision-theoretic framework for multiobjective optimization under uncertainties, maximizing the expected hypervolume with respect to input points using gradient-based stochastic optimization.

ScientificML

Exploring vision-language models for jet analysis in high-energy physics, converting JetClass jets into physics-informed four-panel images for multi-task jet tagging classification.

CFD

Proposes WaveOp-LiteFM, a lightweight neural-operator flow matching approach that addresses the trade-off between computational cost and accuracy in satellite-to-radar precipitation retrieval for atmospheric remote sensing applications.

ScientificML

The Fractional Fourier Neural Operator (FrFNO) integrates the analytic Mittag-Leffler propagator into a Fourier-based neural operator to provide a resolution-robust surrogate for space-time fractional PDEs with memory-dependent derivatives.

Structural

Investigates the effectiveness of Physics-Informed Neural Networks in simulating transverse vibration of one-dimensional linear stiff strings by incorporating PDE constraints and boundary conditions into the loss function.

Electromagnetics

Proposes a 3D FDTD framework using Quantized Tensor Trains that reduces memory and computational costs from conventional volumetric scaling to logarithmic scaling, enabling high-resolution electromagnetic scattering simulations with limited resources.

Materials

A method to predict how force correction changes energy exchange and structural statistics in molecular dynamics by measuring leading response coefficients on shared reference trajectories.

Quantum / Chemistry

A probabilistic physics-informed framework using WaveNet and ParamNet to solve the inverse problem of recovering Woods-Saxon nuclear potential parameters from sparse spectral data.

Materials

Hybrid quantum neural network extracts quantitative parameters like phase weight fractions directly from 1D powder diffraction patterns without iterative refinement, suitable for complex multiphase materials and noisy measurements.

ScientificML

A 36M-parameter UNet-based PDE solver v2 by Los Alamos National Laboratory, designed for surrogate modeling of partial differential equations in scientific computing.

ScientificML

PRONE introduces a Petrov-Galerkin regression framework Ψ(X)K≈Φ(Y) that unifies DMD, SINDy, Koopman regression, and sparse regression as variants of a single construction for data-driven dynamics learning.

Structural

A modal integrator framework (MIISO) is proposed to stabilize non-physical outlier modes at the top of the discrete spectrum in high-order isogeometric discretizations, improving numerical solutions in wave propagation and structural dynamics.

CFD

Proposes an improved Fourier Neural Operator for modeling 2D Rayleigh-Bénard convection by predicting time increments, achieving a compact model (314k parameters, 7ms inference) while revealing accuracy limitations imposed by training data resolution.

Multibody

Proposes AffordanceWAM, an affordance-aware generative World Action Model that leverages both action-labeled robot videos and egocentric human videos for generalizable robot manipulation.

CFD

Diff-NekRS is an open-source framework that enables automatic differentiation through the spectral element CFD solver NekRS, enabling gradient backpropagation across multi-timestep rollouts to train neural correction networks for error mitigation.

Structural

A mesh-free KAN-PINN method for solving nonlinear strain-limiting partial differential equations, featuring exact hard boundary ansatz and pseudo-Huber residual loss for enhanced training stability.

ScientificML

Reveals compatibility conditions under which multi-symplectic Hamiltonian PDEs regularized by Brinkman-type penalization retain a multi-conformal symplectic structure, yielding exact local conservation laws for structure-preserving operator learning on complex geometries.

CFD

This paper presents an integrated CFD-machine learning framework for optimizing pod geometry in supersonic evacuated tube transport, using composite cubic Bezier curves with seven design variables and Latin Hypercube Sampling to evaluate 200 configurations across subsonic and supersonic regimes.

HEP

This high-energy physics work develops a Parnassus conditional flow matching surrogate for the CLD detector, emulating full detector simulation and event reconstruction to address computationally expensive collider studies.

Electromagnetics

DNNsolver proposes a deep learning approach for accurate and efficient modeling of polarization-sensitive diffractive neural networks, addressing limitations of thin element approximation in capturing intra-layer electromagnetic coupling.

Electromagnetics

DNNsolver introduces a deep learning-based approach for modeling polarization-sensitive diffractive neural networks, addressing the limitation of thin element approximation in capturing intra-layer electromagnetic coupling for improved accuracy and efficiency in high-density optical computing.

Electromagnetics

DNNsolver introduces a deep learning approach for modeling polarization-sensitive diffractive neural networks, addressing the challenge of accurate and efficient electromagnetic coupling simulation within cascaded diffractive layers.

Structural

DeepFEAv2 applies deep learning surrogate models to transient finite element analysis on unstructured meshes, enabling joint prediction of node-based and element-based outputs to reduce computational cost in structural mechanics.

Materials

Investigates the relationship between residual scale and learnability in Δ-learning, demonstrating through molecular graph neural network evaluation on total energy targets that smaller residual targets do not necessarily yield better learning outcomes.

Multibody

GINIO introduces a geometric SO(3)-equivariant interface ensuring consistent neural inertial odometry measurements under arbitrary IMU mounting orientations in filtering pipelines.

CFD

FAST-ML is a hybrid framework combining data-driven efficiency with physical constraints for tropical cyclone rapid intensification forecasting.

ScientificML

DissipNet enforces dissipativity constraints to model dissipative dynamical systems, balancing prediction accuracy with physical property preservation for control theory and complex dynamics analysis.

Electromagnetics

Gnomon introduces a physics-informed system that reconstructs LEO satellite serving geometry and signal attenuation to forecast user-perceived downlink/uplink throughput and RTT under varying trace availability.

Power Systems

A physics-informed GNN-based foundation model for AC-OPF, pretrained across 54 grid topologies (500-4,000 buses), achieving 2.45% zero-shot generation-cost error on 10,000-bus systems.

ScientificML

This paper investigates neural network solvers for infinity and p-Laplace problems in nonlinear analysis, applying PINNs and DeepONets to 2D and 3D domains with p values ranging from 2 to 1000.

Multibody

HOTICE introduces Humanoid-Object Decoupled Potential Fields for joint collision encoding, enabling whole-body coordination and collision avoidance during object transportation in cluttered environments.

Multibody

JAMB proposes jointly denoising bimanual actions and motion states, explicitly modeling geometric interactions between arms in shared 3D scenes to enable coordinated manipulation.

Structural

KATOsuper presents an objective-agnostic framework that couples neural-reparameterized topology optimization with sensitivity-consistent Fourier neural operators, addressing gradient inconsistency between predicted objectives and sensitivities for stable structural optimization.

Multibody

Proposes bilinear Koopman model-based diffusion (BK-MBD) to reduce computational cost in real-time robot control by decoupling state lifting from candidate propagation.

Biomedical Engineering / Hemodynamics

This study develops a patient-specific image-based computational modeling framework to evaluate the hemodynamic impact of mitral and aortic valve repair in five LVAD-supported patients, addressing valve dysfunction threats to long-term LVAD therapy success.

CFD

LeafNet algorithm reformulates hyperbolic conservation laws as a coupled system with smooth solutions and approximates smooth functions via physics-informed neural networks, enabling accurate computation of multidimensional hyperbolic PDEs without diffusion.

ScientificML

A high-performance PDE neural solver based on Manifold Function Encoder, featuring resolution-independent memory footprint and arbitrary PDE input support for large-scale physical simulations.

Electromagnetics

Proposes a geometric constellation shaping (GCS) framework to enhance sensing privacy in OFDM-ISAC systems, addressing the privacy risk where passive eavesdroppers exploit ISAC data signals for unauthorized target sensing, revealing receiver-specific ranging performance dependent on constellation design.

Geophysics

PISCES is a convolutional autoencoder trained without catalog labels on OMNI solar wind measurements under physics constraints, designed for detecting solar wind transients and providing space weather early warning at the Sun-Earth L1 point.

Electromagnetics

A physics-informed deep learning framework performs single-snapshot 3D phase-only positioning in narrowband phase-coherent distributed MIMO networks under two-ray electromagnetic propagation.

Plasma / Fusion

Presents PQLS, a quasilinear gyrokinetic transport solver using an eigenvalue solver in general magnetic geometry with Bayesian saturation-rule closure to improve turbulent transport prediction in tokamak plasmas.

Geophysics

PR-Smoother is an amortized smoother for prescribed-simulator data assimilation, supporting non-Gaussian posterior estimation over physical state variables in high-dimensional simulators with calibration compatibility.

Acoustics

RFlash uses differentiable acoustic simulation and radiance field decomposition to reduce bone-induced shadow artifacts in ultrasound imaging, improving symmetric assessment of fetal brain hemispheres.

Materials

Proposes SO(3) Equivariant Neural Kalman Networks (SENK) with response-state cascade to improve transferable vibrational spectral prediction across chemical spaces using equivariant transformer backbone.

ScientificML

SHRAV is an architecture-independent computational framework for physical modeling and inverse design, featuring a state-continuation core with declared reuse boundaries and explicit roles for learned evolution.

CFD

TM4FF introduces a physics-constrained operator learning framework combining Transformer attention with Mamba state-space models for accurate flow field prediction and denoising.

Multibody

A Transformer-based warm-starting approach for non-convex trajectory optimization in cislunar rendezvous, proximity operations, and docking scenarios.

Electromagnetics

This study assesses RF-EMF exposure across all floors of a 20-story building adjacent to a multi-operator shared base station, revealing peak levels of 31.86 V/m near the center of the antenna's vertical coverage zone.

Multibody

ForceTwin learns joint physical properties including inertia, friction, and spring mechanisms from human interaction data to build physics-informed digital twins that predict robotic manipulation actions.

Multibody

This paper investigates physical diagnosis for robots facing unfamiliar conditions, defining six ordered empirical conditions required for behavior improvement, and validates the chain in controlled and public environments.

Quantum

Using physics-informed neural networks to solve electron spin dynamics in few-cycle laser pulses, deriving analytical expressions for the polarization angle of electrons initially at rest in CEP-stabilized linearly polarized plane-wave pulses.

Electromagnetics

Proposes a hybrid CNN-Transformer architecture for solar flare prediction based on the physical process of magnetic energy release in the solar atmosphere, enabling early warning for C/M/X-class flares.

CFD

Proposes a neural operator-based surrogate model for CFD-level transient thermal-hydraulic simulation of helical coil steam generators in SMRs, enabling real-time digital twin applications.

CFD

Proposes periodic Fourier Neural Mapping (p-FNM), a neural-operator framework for rapid prediction of unsteady pressure distributions on turbine rotor blades, targeting aeroelastic analysis in turbomachinery.

ScientificML

NeuIDO proposes a neural intrinsic dynamics operator that internalizes physical constraints into world models for predicting 3D object interaction trajectories, bridging the gap between physics-informed 4D generation and true world models.

Electromagnetics

Introduces a reformulation of the S-matrix approach for layered photonic structures, defining the S-matrix as an intrinsic layer property independent of neighboring interfaces to enhance numerical stability.

Electromagnetics

Proposes Neural Modal Decomposition (NMD) to decompose multi-port system response matrices into pole-residue form for neural network-based modeling of RF cavities and photonic devices.

Electromagnetics

Proposes a KAN-based single-parameterized surrogate model to accelerate drift kinetic equation solving for fast neoclassical toroidal viscosity torque calculation in tokamak plasma modeling.

ScientificML

Proposes a continuous delayed-memory stochastic gradient descent method combined with neural network-parameterized stochastic differential equations to model quasar brightness variations from astronomical survey time series.

CFD

Introduces mollified-sharp decomposition for probabilistic regularization of moving shocks in parametric reduced-order models, enhancing shock-capturing capabilities in CFD applications.

ScientificML

Proposes combining Artificial Representative Trees with leaf-wise Mondrian Conformal Predictive Systems to enable a single decision tree to provide continuous predictions, prediction intervals, and probabilistic outputs as an interpretable ML approach.

Acoustics

This paper investigates the acoustic dip phenomenon in the effective temperature of hot Brownian particles in compressible solvents, revealing how finite sound speed modifies the fluctuation-dissipation relation in the kinetic regime.

ScientificML

Proposes an adaptive memory gating mechanism enabling neural operators to dynamically adjust memory weights based on observation conditions, improving adaptability in solving time-dependent PDEs.

ScientificML

This paper proposes an improved periodic activation using complex exponential functions generating paired sine-cosine outputs for Physics-Informed Neural Networks, demonstrating superior performance over sine-activated MLPs in reconstructing temperature fields from sparse velocity data in convective flows.

Thermal Management

Proposes a validation-gated multi-agent governance framework enabling thermal-hydraulic surrogate models to adapt online during operating-regime shifts, preventing condition-locking issues in offline-pretrained models.

ScientificML

Proposes TASE-stabilized Time-Evolving Natural Gradient (TENG) method that projects time-discrete target states onto the neural-network manifold via local natural-gradient iterations to address stability constraints in stiff diffusion-dominated PDEs.

Multibody/Robotics

Introduces a computational framework for mechanical systems combining reservoir encoding and active inference, enabling adaptive learning in changing environments beyond traditional passive design and offline optimization constraints.

ScientificML

Proves the original BGN method for mean curvature flow converges with order h² using backward error analysis from ODE numerical analysis.

ScientificML

Proposes a nearly-instant latent space Bayesian optimization method that replaces expensive sequential decision-making with sampling overhead, making BO practical for generative model-driven de novo design pipelines with cheap virtual screens.

ScientificML

Proposes test-time trained flow matching for Bayesian filtering, enabling continuous distribution representation for state estimation in high-dimensional physical systems.

CFD

Proposes a Runge-Kutta aligned oscillation-eliminating DG framework that fixes the disruption of error structures caused by nonlinear stabilization, enabling superconvergence analysis for fully discrete shock-capturing DG schemes.

Materials

Presents a method using sparse differentiation to compute exact molecular Hessians from machine learning interatomic potentials, reducing computational complexity from O(N²) to O(N), overcoming limitations for large atomistic systems.

CFD/Fluid Dynamics

VIOT is a generative neural operator for incompressible density transport that predicts divergence-free velocity fields via feed-forward inference, replacing hour-scale optimization in adjoint fluid solvers and enabling rapid transport trajectory generation.

CFD

Establishes asymptotic-preserving property of exponential Euler method for stochastic reaction-diffusion-advection equations in R² under fast Hamiltonian advection with multiple critical points, reducing dynamics to a stochastic PDE on a noncompact metric graph.

Quantum/Chemical

Proposes iCRNNs to discover chemical reaction network structures and kinetics from time-series concentration data without prior mechanistic assumptions, balancing interpretability and computational efficiency.

ScientificML

Proposes NeSTR method that learns the S-Transform on a finite dimensional subspace of white noise test space to reconstruct stochastic solutions for Gaussian SPDEs.

ScientificML

PosteriorBench introduces a benchmark for evaluating how well generative inverse solvers capture true posterior distributions in scientific inverse problems, addressing limitations of pointwise accuracy metrics.

Multibody

LOInK learns a bi-Lipschitz invertible mapping from configuration space to a decoupled task/latent space, structuring the latent space so that cost-minimizing solutions are placed at the origin, enabling efficient sampling of approximately optimal inverse kinematics solutions.

ScientificML

Proposes a novel framework using set-based value function characterization and neural network approximation to estimate stabilization domains and stabilizing controllers for input-constrained discrete-time nonlinear systems.

ScientificML

Presents AKAPINN, a mesh-free KAN-PINN method for nonlinear strain-limiting PDEs with exact hard boundary ansatz and pseudo-Huber loss, evaluated across three material-parameter pairs.

Materials

Develops a general theory of rolling construction for arbitrary 2D crystals to create nanotubes, implementing an open-source library NTBuilder with desktop and web interfaces, providing a catalog of 20 million structures with exact unit cell calculations.

Multibody/Robotics

Proposes D3DWA, an adaptive DWA framework using Dueling Double Deep Q-Network to adjust weights and prediction horizon for improved robot navigation in varied environments.

ScientificML

Proposes neural network-based pulse deconvolution for FCC-ee's high-granularity dual-readout calorimeter, leveraging timing separation of Cherenkov and scintillation signals to improve energy resolution.

Electromagnetics

Proposes a coupled Maxwell-cable framework integrating FDTD solutions with Hodgkin-Huxley and Fitzhugh-Nagumo dynamics to simulate action potential propagation in branched neuronal geometries.

Multibody

This paper solves the four-dimensional HJB equation for two diffusively coupled Stuart-Landau-like oscillators to obtain full-state optimal feedback control, revealing a sharp transition in control strategy from phase synchronization to amplitude collapse at a critical coupling threshold.

CFD

This paper proposes a periodic activation function architecture using complex exponential to generate paired sine and cosine outputs, demonstrating improved performance over sine-activated MLPs for reconstructing convective flows in physics-informed neural networks.

Structural

This paper rigorously establishes inf-sup stability of the Scott-Vogelius finite element on three-dimensional Freudenthal meshes for polynomial degree k ≥ 4, completing the theoretical framework for this classical divergence-free mixed element used in Stokes flow computations.

ScientificML

GENERIC-FNO embeds the metriplectic degeneracy structure of nonequilibrium thermodynamics into Fourier neural operators, coupling reversible energy-conserving and irreversible entropy-producing dynamics via degeneracy conditions, extending structure-preserving neural operators beyond single conservation laws.

ScientificML

Investigates how uncertainty quantification approaches perform when adapted to geometry-conditioned neural surrogates for CAE, comparing closed-form and sampling-based methods across varying geometries and engineering quantities of interest.

CFD

Proposes p-FNM, a neural-operator framework using periodic Fourier mapping to predict unsteady pressure distributions and aeroelastic loads on turbine rotor blades, reducing high-fidelity CFD computational costs.

Quantum/Chemical

MARC-TS introduces a two-stage framework that learns endpoint-conditioned continuous reaction paths, queries at arbitrary resolution, and refines transition-state candidates via local path context, with T1x-IRC-8K dataset of 8,209 reactions and 1.09M path-resolved geometries.

Multibody

DeViGrasp introduces methods to systematically examine how occlusion, segmentation-mask dropout, depth noise, and target-localization jitter affect grasp reasoning and target tracking for quadruped manipulators, improving whole-body control robustness under degraded perception.

CFD

Proposes a subcell-refined entropy-residual-driven limiter for high-order DG methods on Legendre-Gauss-Lobatto nodes, enabling robust simulation of under-resolved nonlinear hyperbolic systems while preserving accuracy via nearest-neighbor pairwise dissipation.

Electromagnetics

Presents a 3D multiphysics FEM framework for high-frequency induction tube welding, coupling time-harmonic magnetic scalar potential with quasi-stationary heat transport for coupled electromagnetic-thermal simulation.

ScientificML

Proposes a physics-informed inverse modeling framework to infer governing parameters of the Moose-Wolf ecosystem in Isle Royale National Park from limited and noisy observational data, providing a data-driven approach for ecological dynamics modeling.

Materials

Neural networks trained on 4DSTEM diffraction patterns enable rapid identification of LiNiO2 and NiO material phases, providing a computationally efficient approach for analyzing large sets of materials image data.

ScientificML

This paper proposes PIRNN to enhance interpretability and accuracy of time series forecasting for physical processes by incorporating physical knowledge from historical data during training.

Multibody

CHOREO enables training-free composition of heterogeneous humanoid skills learned via reinforcement learning, motion imitation, and generative modeling, bridging incompatible representations, interfaces, and controllers through unified trajectories.

ScientificML

Proposes a tensor product multilevel method (TPML) combining Smolyak's sparse grids with kernel-based interpolation, coupling kernel scale to fill distance for uniform condition number bounds in high-dimensional function approximation.

ScientificML

Proposes Adaptive Operator Transformation (AOT) to address structural diversity challenges in multi-PDE pre-training by adapting target solution operators rather than increasing model capacity, improving neural operator generalization across equations.

Acoustics

A browser-based acoustic simulator and inverse design tool for wind instrument bores without tone holes, using Transfer Matrix Method for forward modeling and adjoint-based Tikhonov regularization for inverse design optimization.

CFD

An SO(3)-equivariant neural network predicts the full 6×6 grand resistance matrix for microparticles in Stokes flow, outperforming traditional scalar drag correlations by preserving orientation dependence and rotational response.

ScientificML

This paper introduces ICG-MTO, a framework that leverages numerical foundational models to improve inter-task coupling estimation in few-shot multi-task optimization scenarios, addressing negative transfer issues.

CFD

Numerical investigation of how stationary distributions concentrate onto large scales in the fast-advection limit of stochastically forced 2D incompressible Navier-Stokes equations at N=128 resolution.

Electromagnetics

A reduced-order physical model describes re-ignition dynamics between consecutive discharges at the same position in dielectric barrier discharge arrangements, enabling time prediction without machine learning.

Materials

Molecular dynamics simulations investigate how local chemical clustering affects structural and functional properties of Ba(Zr_xTi_{1-x})O_3 in ferroelectric and relaxor regimes, revealing mechanisms of chemical inhomogeneity on material performance.

Electromagnetics

A receiver-side physics-informed residual digital twin for inter-satellite optical links that integrates Gaussian optics and quadrant-photodetector physics with an autoregressive ridge model to predict and compensate residual line-of-sight motion under sensor noise and actuator latency.

Electromagnetics

A topology-aware GCN model for fault classification in distribution networks with inverter-induced harmonic distortion using IEEE 34-bus system topology.

CFD

Proposes a non-intrusive correction method using VAE latent space adaptation to reduce discrepancies between CFD predictions and experiments for open tip clearance flow in compressor cascades, trained on 166 parametrically sampled CFD total pressure loss fields.

ScientificML

Proposes block-triangular joint drifting to overcome the challenge of estimating conditional distributions in stochastic transition modeling, enabling one-step generative surrogate models for trajectory data prediction.

Multibody/Robotics

Proposes adaptive interaction strategies for companion robots using deep reinforcement learning, overcoming traditional fixed-position tracking limitations to enable flexible robot accompaniment in dynamic workspaces.

ScientificML

Proposes hybrid neural density estimation combining neural ratio estimation with flow-based models to address density normalization and efficient sampling in high-dimensional unbinned simulation-based inference.

Materials

Proposes a physics-informed ML framework combined with multi-objective optimization for inverse design of Co-based superalloy compositions to enhance γ' solvus temperature while considering alloy density.

Materials

Presents a physics-informed ML framework for inverse design of Co-based superalloys, using atomic size mismatch and other descriptors with multi-objective optimization to enhance gamma-prime solvus temperature while constraining density.

CFD/Fluid

A periodic physics-informed neural network (P-PINN) framework is proposed for solving parametric periodic flows and analyzing finite-amplitude aeroelastic responses in a purely physics-driven manner.

Power Systems

This paper develops a surrogate-assisted stochastic multiobjective framework for photovoltaic hosting-capacity planning, jointly optimizing penetration and total electrical loss under reliability constraints.

ScientificML

Proposes Interface Autoencoder (IAE) for learning neural operators on geometry-dependent PDEs, representing interface-function states in a fixed domain via finite projection codes.

CFD/Fluid

This study evaluates physics-informed neural networks for atmospheric potential-temperature forecasting, constrained by pressure-coordinate thermodynamic advection-source equations to address short-horizon predictions under sparse thermal observation conditions.

Thermal Management

This research applies Bayesian optimization and surrogate modeling to investigate urban canopy parameter ranges for heat mitigation, analyzing trade-offs among daytime air temperature, nighttime air temperature, and Universal Thermal Climate Index (UTCI) in urban environments.

ScientificML

A Bayesian tree-adjoining grammar framework for equation discovery and structure identification in general nonlinear systems.

Electromagnetics

Demonstrates a large-scale metasurface-based reflection cavity five-layer diffractive neural network achieving angle-multiplexed multi-task processing, addressing scalability limitations in all-optical computing.

Materials

This paper formulates optimal biasing as an ergodic optimal control problem, recast as a linear eigenvalue problem for the infinitesimal generator of unbiased dynamics to learn bias potentials that accelerate transitions between metastable states in molecular dynamics.

Materials / Soft Matter

Proposes active self-consistent field theory (ASCFT) for studying steady-state behavior of active Ornstein-Uhlenbeck polymers, extending classical polymer field theory to non-equilibrium systems.

Multibody

This paper proposes a separated-section constitutive law for Cosserat rod modeling of trimmed helicoid soft robot arms, evaluating each helix domain's mechanical response in its local frame and mapping it back to the global coordinate system for improved modeling accuracy.

ScientificML

GravINNs embeds post-Newtonian equations of motion directly into neural network training to learn binary orbital dynamics, including single-orbit surrogates and extensions across a four-dimensional parameter space.

ScientificML

This paper studies the convergence properties of the cyclic steepest descent (CSD) method for minimizing strictly convex quadratic functions in arbitrary dimensions, proving doubly exponential convergence rates and extending prior theoretical results limited to two-dimensional cases.

Electromagnetics

Presents a general method for producing stochastic synthetic axion waveforms by shaping Gaussian white noise in the frequency domain, targeting microwave cavity haloscope calibration and analysis pipeline validation for dark matter axion searches.

ScientificML

This paper reveals the multi-conformal symplectic structure preserved by Hamiltonian PDEs regularized via Brinkman penalization, proving an exact local conservation law under a compatibility condition linking the symplectic matrix and penalization projection.

ScientificML

Investigates how load information quality affects power grid control using a DC power flow model and Grid2Op-inspired simulator, exploring data assimilation combined with reinforcement learning strategies.

Multibody

Proposes a harmonic-regression combined ARMA model for fault detection in rotating machinery, addressing limitations of purely rational spectrum assumptions under varying operating conditions.

Acoustics

This paper applies meta-adaptive filtering to active noise control when acoustic paths vary over time, constructing physics-informed optimizer features from a secondary-path estimate to support online filter updates.

Electromagnetics

This paper develops an analytical and numerical framework for temporal scattering in electromagnetic metamaterials with shape-varying cylindrical inclusions, enabling dynamic control through abrupt geometric transformations of meta-atoms.

Fluid Mechanics

This study evaluates five deep learning architectures for univariate significant wave height forecasting using buoy observations. Across a 47-buoy, 37-year evaluation corpus, the model families converge to nearly identical performance, with an between-family standard deviation of 0.0014 square meters.

Quantum/Chemistry

Proposes using ordinary multilayer perceptrons with finite-group averaging to achieve exact equivariant surrogate models for correlated electron systems, enabling efficient force and transition energy evaluations.

Geophysics

This paper presents differentiable and hardware-accelerated implementations of the machine-learning surrogate models mlgw and mlgw-bns for gravitational-wave waveform modeling, enabling JIT compilation, vector parallelization, GPU acceleration, and automatic differentiation within JAX.

Materials

A region-held-out evaluation protocol for hydrogen embrittlement detection in 316L stainless steel, preventing data leakage from same specimen regions, comparing texture features vs deep learning for SEM micrograph classification.

ScientificML

This paper studies gradient conflicts and multi-task optimization in Physics-Informed Neural Networks, with implications for more stable convergence in stiff and high-frequency partial differential equations.

High-Energy Physics

A physics-informed self-supervised learning framework that jointly performs wire calibration and interaction position reconstruction in multi-wire parallel plate avalanche counters, enhancing adaptability to changing operating conditions and detector aging.

Multibody

Proposes a PINN-based approach for the weakly coupled two-compartment neuron model to infer unmeasurable somatic membrane potential dynamics from dendritic recordings.

CFD

Uses patch-based autoencoder and neural ODE to build a distributed reduced-order model for studying relative periodic orbits in 2D Kolmogorov turbulence.

Materials

A self-driving laboratory strategy that simultaneously and dynamically assesses composition-dependent synthetic accessibility and target properties via aggregated cost functions to accelerate optimization of multicomponent NASICON fast ionic conductors while maintaining phase purity.

ScientificML

This survey covers brain-to-language decoding techniques, including neural signal acquisition, representation learning, and text generation, applicable to brain-computer interfaces and speech restoration.

Electromagnetics

This paper establishes a foundation for physically interpretable RF sensing that disentangles reflector speed from geometry, jointly recovering speed, geometry factor, relative amplitude, and width of dominant Doppler ridges for improved applicability in unconstrained real-world settings.

Electromagnetics

A physics-informed neural operator learns the map between function spaces in magnetohydrodynamic simulations, enabling efficient state inference for magnetic reconnection at any parameter and time step, significantly reducing computational cost at high Lundquist numbers.

Materials

A physics-informed computational screening framework that decomposes effective second-harmonic response to optimize nonlinear optical properties of 2D semiconductor/organic molecule interfaces.

Multibody

A history-conditioned physics-informed flow-matching framework for probabilistic dynamics prediction of tendon-driven continuum robots, addressing uncertainties from material behavior, tendon transmission, friction, and contact.

Materials

Investigates ferroelectric switching modeling by comparing four cost-effective DFT methods against high-accuracy ab initio approaches, aiming to establish reliable benchmarks for active learning workflows.

Electromagnetics

Presents a fully implicit PIC method extended to electromagnetic cases with exact conservation of global energy and local charge, enabling plasma equilibrium simulations at high (up to extreme) β in open magnetic traps.

Framework Paper / Neural Operator

NVIDIA's PINN framework built on neural operators (FNO/DeepONet/AFNO) for end-to-end differentiable physics simulation, with multi-GPU training and deployment.

Annual Survey / Physics-Informed Learning

A systematic review of 2025-2026 progress in physics-informed ML - neural operators, diffusion for PDE solving, large benchmarks and industrial deployment.

Must-Read Benchmark Paper

Proposes a unified benchmark across 16 physical domains, arguing for cross-domain pretraining gains - a must-read for physics foundation models.

Frontier Paper

Proposes a physics-attention mechanism that projects meshes onto learnable physical-state tokens, improving efficiency and accuracy of PDE solving on large industrial geometries.