# Finite Element Method (FEM)

A numerical method that discretizes differential equations into systems of equations for computational engineering and scientific simulation.

This is one page of public article previews, not the complete archive. Follow Next page to continue. Summaries are not the original full articles.

## PSTO: A PDE-structured transport operator with shared implicit updates for multiscale CDR systems

DevFeed: [PSTO: A PDE-structured transport operator with shared implicit updates for multiscale CDR systems](<https://devfeed.tech/articles/psto-a-pde-structured-transport-operator-with-shared-implicit-updates-for-multiscale-cdr-systems-145077.md>)

Original publisher: [Read original article](<https://arxiv.org/abs/2611.25607>)

Author: Kaiwen Li, Dezhi Zhou, Kun Wu, Xingcai Lu

Published: 2026-11-20T05:00:00Z

Content type: article

Language: en

Sources: [arXiv: Machine Learning](<https://devfeed.tech/sources/arxiv-machine-learning.md>)

Topics: [Machine learning](<https://devfeed.tech/topics/machine-learning.md>), [Neural Network](<https://devfeed.tech/topics/neural-network.md>), [Finite Element Method (FEM)](<https://devfeed.tech/topics/finite-element-method.md>)

Tags: [2](<https://devfeed.tech/tags/2.md>), [2d](<https://devfeed.tech/tags/2d.md>), [ablation](<https://devfeed.tech/tags/ablation.md>), [abstract](<https://devfeed.tech/tags/abstract.md>), [after](<https://devfeed.tech/tags/after.md>), [all](<https://devfeed.tech/tags/all.md>), [analysis](<https://devfeed.tech/tags/analysis.md>), [architecture](<https://devfeed.tech/tags/architecture.md>), [arxiv](<https://devfeed.tech/tags/arxiv.md>), [benchmarks](<https://devfeed.tech/tags/benchmarks.md>), [block](<https://devfeed.tech/tags/block.md>), [canonical](<https://devfeed.tech/tags/canonical.md>), [case](<https://devfeed.tech/tags/case.md>), [cdr](<https://devfeed.tech/tags/cdr.md>), [cs-lg](<https://devfeed.tech/tags/cs-lg.md>), [operator](<https://devfeed.tech/tags/operator.md>), [structured](<https://devfeed.tech/tags/structured.md>), [systems](<https://devfeed.tech/tags/systems.md>), [transport](<https://devfeed.tech/tags/transport.md>), [updates](<https://devfeed.tech/tags/updates.md>), [with](<https://devfeed.tech/tags/with.md>)

### AI overview

PSTO learns convection-diffusion dynamics from reaction-free data, then applies them to full convection-diffusion-reaction systems without retraining. It shares one transport block across transported variables and uses relaxed fixed-point iterations for implicit-like latent corrections, aiming to improve stability without adding trainable parameters. Across three benchmarks--a 2D CDR system, a Robertson counterflow flame and a hydrogen planar jet flame--it reports at least 50% lower mean relative L2 error after 100 coupled steps than the strongest neural baseline. The abstract also reports maintained accuracy with a tenfold larger time step, while baseline accuracy decays rapidly. These results are limited to the stated benchmarks and comparisons.

### Source excerpt

arXiv:2611.25607v1 Announce Type: new Abstract: Multiscale convection-diffusion-reaction (CDR) systems couple spatial transport with local stiff reaction kinetics, making their numerical simulation computationally demanding. Conventional neural surrogates require costly coupled CDR training data, and their parameter counts scale with the number of transported variables, limiting scalability to high-dimensional CDR systems. Here we propose a PDE-Structured Transport Operator (PSTO) that learns convection--diffusion dynamics solely from reaction-free data. By embedding the PDE structure into the neural architecture, PSTO generalizes to full CDR systems at inference without retraining. A single transport operator block is shared across all transported variables and refined through relaxed fixed-point iterations, performing an implicit-like correction in latent space. This improves stability during recursive multi-step prediction without adding trainable parameters. We evaluate PSTO on three benchmarks: a canonical 2D CDR system, a multiscale Robertson counterflow flame, and a multiscale hydrogen planar jet flame. In every case, PSTO reduces the mean relative $L_2$ error after 100 coupled steps by at least 50% compared to the strongest neural baseline. Ablation studies on the fixed-point iterations and detailed error decomposition further confirm that PSTO achieves both high one-step accuracy and long-term stability in coupled rollouts. Time-step sensitivity analysis shows that PSTO maintains accuracy even with a tenfold larger time step, whereas baseline accuracies decay rapidly as the time step increases. This framework enables efficient, modular modeling of multiscale CDR dynamics by separating transport learning from the treatment of local kinetics.

## Uncertainty-aware damage identification in short-span bridges via physics-informed variational autoencoder

DevFeed: [Uncertainty-aware damage identification in short-span bridges via physics-informed variational autoencoder](<https://devfeed.tech/articles/uncertainty-aware-damage-identification-in-short-span-bridges-via-physics-informed-variational-autoencoder-143287.md>)

Original publisher: [Read original article](<https://arxiv.org/abs/2607.05025>)

Author: Ana Fernandez-Navamuel, A. Javier Omella, Diego Zamora-Sanchez, David Pardo

Published: 2026-11-19T05:00:00Z

Content type: article

Language: en

Sources: [arXiv: Machine Learning](<https://devfeed.tech/sources/arxiv-machine-learning.md>)

Topics: [VAE](<https://devfeed.tech/topics/vae.md>), [Physics-guided deep learning](<https://devfeed.tech/topics/physics-guided-deep-learning.md>), [Machine learning](<https://devfeed.tech/topics/machine-learning.md>), [Finite Element Method (FEM)](<https://devfeed.tech/topics/finite-element-method.md>)

Tags: [autoencoder](<https://devfeed.tech/tags/autoencoder.md>), [bridge](<https://devfeed.tech/tags/bridge.md>), [bridges](<https://devfeed.tech/tags/bridges.md>), [cs-lg](<https://devfeed.tech/tags/cs-lg.md>), [health-monitoring](<https://devfeed.tech/tags/health-monitoring.md>), [machine-learning](<https://devfeed.tech/tags/machine-learning.md>), [physics-informed](<https://devfeed.tech/tags/physics-informed.md>), [probabilistic](<https://devfeed.tech/tags/probabilistic.md>), [uncertainty](<https://devfeed.tech/tags/uncertainty.md>), [uncertainty-quantification](<https://devfeed.tech/tags/uncertainty-quantification.md>)

### AI overview

The proposed physics-informed Gaussian copula variational autoencoder estimates bridge damage and diagnostic uncertainty from vibration data. It embeds a numerical eigenvalue solver as its decoder, so posterior samples satisfy structural dynamics equations, and uses a Gaussian copula to capture correlations among stiffness reduction factors with fewer parameters than Gaussian mixtures. In a numerical short-span bridge study with elastic bearings, modeling errors, measurement noise, and a finer test discretization, its damage posteriors closely matched sequential Monte Carlo estimates. The method can indicate damage location while representing uncertainty, including cases where noise masks light damage or modal properties are weakly sensitive to damage.

### Source excerpt

arXiv:2607.05025v3 Announce Type: replace Abstract: Vibration-based damage identification in civil infrastructure is a challenging, ill-posed inverse problem due to measurement noise, sparse sensor arrays, and modeling errors. While deep learning is powerful for system identification, deterministic approaches lack reliable uncertainty quantification and can yield physically inconsistent results. This work proposes a probabilistic Scientific Machine Learning framework for structural health monitoring: a physics-informed Gaussian copula variational autoencoder. The methodology introduces two main contributions. First, a numerical eigenvalue solver is embedded directly into the variational autoencoder as its decoder, which replaces data-driven surrogates, ensures that every posterior sample satisfies the governing equations of structural dynamics, and reduces the number of trainable parameters. Second, a Gaussian copula describes the posterior distribution of the stiffness factors, capturing cross-correlations between components with fewer parameters than Gaussian mixtures in real time. We validate the method numerically on a short-span bridge supported on elastic bearings, whose response is dominated by vertical bending, accounting for modeling errors, measurement noise, and a finer finite element discretization for testing. The method estimates the posterior distribution of the damage condition, i.e., the stiffness reduction factors of the deck zones and bearings, and these posteriors closely match the exact ones computed with sequential Monte Carlo sampling. The results reveal that the proposed methodology offers a robust and interpretable decision-making tool, able not only to indicate the presence and location of damage, but also to inform on the uncertainty associated with the diagnosis, for instance, when light damage is masked by the effect of noise or when the measured modal properties present limited sensitivity to damage.

## Solver to System: LLM Reasoning Across Representations and Execution in Scientific Simulation

DevFeed: [Solver to System: LLM Reasoning Across Representations and Execution in Scientific Simulation](<https://devfeed.tech/articles/solver-to-system-llm-reasoning-across-representations-and-execution-in-scientific-simulation-141711.md>)

Original publisher: [Read original article](<https://arxiv.org/abs/2611.21964>)

Author: Eliza Berman, Rohan Subramanian, Shreya Jha, Lohit Jagarapu, Romain Lopez, Rajesh Ranganath

Published: 2026-11-18T05:00:00Z

Content type: article

Language: en

Sources: [arXiv: Machine Learning](<https://devfeed.tech/sources/arxiv-machine-learning.md>)

Topics: [llm-reasoning](<https://devfeed.tech/topics/llm-reasoning.md>), [Simulation](<https://devfeed.tech/topics/simulation.md>), [Large Language Model](<https://devfeed.tech/topics/llm.md>), [Finite Element Method (FEM)](<https://devfeed.tech/topics/finite-element-method.md>)

Tags: [cs-cl](<https://devfeed.tech/tags/cs-cl.md>), [cs-lg](<https://devfeed.tech/tags/cs-lg.md>), [execution](<https://devfeed.tech/tags/execution.md>), [llm](<https://devfeed.tech/tags/llm.md>), [partial-differential-equations](<https://devfeed.tech/tags/partial-differential-equations.md>), [reasoning](<https://devfeed.tech/tags/reasoning.md>), [scientific-simulation](<https://devfeed.tech/tags/scientific-simulation.md>)

### AI overview

Solver2System evaluates whether eight open-weight LLMs can reconcile PDE solvers with equations, natural-language descriptions, and numerical trajectories, and detect physically invalid code. When representations conflict, models favor some over others; lexical changes to code can also shift how they interpret unchanged representations. Executing code improves physical-validity judgments, but weaker agents may reach the right judgment without identifying the defect that caused it. The findings show that a correct verdict does not necessarily mean the model consistently reconciled evidence or diagnosed the underlying cause.

### Source excerpt

arXiv:2611.21964v1 Announce Type: new Abstract: Simulators based on partial differential equations (PDEs) are widely used to model physical systems, but generating and interpreting PDE solver code with large language models (LLMs) remains challenging. Scientific simulation code is a useful testbed for LLM scientific reasoning because (1) one system can be represented by code, equations, natural language, and numerical trajectories and (2) code can execute successfully while implementing invalid physics. We introduce Solver2System, a dataset of PDE solvers with multiple representations, controlled corruptions, and lexical perturbations, to evaluate whether LLMs can recover the underlying physical system through (1) cross-representation consistency and (2) static and agentic detection of physical invalidity. Across eight open-weight models, LLMs systematically favor some representations when evidence conflicts, and lexical perturbations to code alter interpretation of other unchanged representations. Execution improves physical-validity judgments, but weaker agents can judge correctly without identifying the causal defect. Thus, LLMs may reach correct judgments despite inconsistent representation reconciliation or causal misdiagnosis.

## Variationally correct physics-informed operator learning for nonlinear partial differential equations

DevFeed: [Variationally correct physics-informed operator learning for nonlinear partial differential equations](<https://devfeed.tech/articles/variationally-correct-physics-informed-operator-learning-for-nonlinear-partial-differential-equations-139175.md>)

Original publisher: [Read original article](<https://arxiv.org/abs/2611.21469>)

Author: Yuan Qiu, Wolfgang Dahmen, Peng Chen

Published: 2026-11-17T05:00:00Z

Content type: article

Language: en

Sources: [arXiv: Machine Learning](<https://devfeed.tech/sources/arxiv-machine-learning.md>)

Topics: [Finite Element Method (FEM)](<https://devfeed.tech/topics/finite-element-method.md>), [Machine learning](<https://devfeed.tech/topics/machine-learning.md>)

Tags: [cs-lg](<https://devfeed.tech/tags/cs-lg.md>), [cs-na](<https://devfeed.tech/tags/cs-na.md>), [diffusion](<https://devfeed.tech/tags/diffusion.md>), [element](<https://devfeed.tech/tags/element.md>), [math-na](<https://devfeed.tech/tags/math-na.md>), [neural](<https://devfeed.tech/tags/neural.md>), [numerical-analysis](<https://devfeed.tech/tags/numerical-analysis.md>), [operator-learning](<https://devfeed.tech/tags/operator-learning.md>), [partial-differential-equations](<https://devfeed.tech/tags/partial-differential-equations.md>), [physics-informed](<https://devfeed.tech/tags/physics-informed.md>), [reaction](<https://devfeed.tech/tags/reaction.md>)

### AI overview

For nonlinear PDEs written as first-order systems, the paper establishes that a least-squares residual can measure solution error in the PDE-compliant norm near a solution branch, provided the linearized residual is stable and its derivative is Lipschitz continuous. The authors verify these conditions for monotone reaction, solution- and gradient-dependent diffusion, and compressible Neo-Hookean hyperelasticity; the stability mechanisms and regions of equivalence differ by problem class. In a conforming finite element space, loss excess near the finite element solution is equivalent to squared distance from it. The authors use this loss to train a reduced basis neural operator, requiring solution snapshots only to build the basis; the finite element library assembles the loss and its symbolically derived gradient. At inference, loss estimates error. Numerically, the square root of loss excess tracks error within factors comparable to the linear case, and the operator is more accurate in the PDE-compliant norm than grid-based baselines evaluated in the same space.

### Source excerpt

arXiv:2611.21469v1 Announce Type: cross Abstract: Physics-informed operator learning trains neural operators on the residual of a parametric partial differential equation (PDE) instead of labeled solution data, but its reliability depends on a variationally correct residual: one that is equivalent to the error in a norm dictated by the equation, the PDE-compliant norm. For nonlinear equations written as first-order systems, we prove that the first-order system least-squares residual is equivalent to the error in this norm near a solution branch if the linearized residual is stable and the residual derivative is Lipschitz continuous, possibly in a norm stronger than the error norm. We verify these conditions for monotone reaction, solution- and gradient-dependent diffusion, and compressible Neo-Hookean hyperelasticity. The stability mechanisms vary across these classes, as do the regions where the equivalence holds, from arbitrary bounded sets to a neighborhood of the branch. In a conforming finite element space, we show that the loss excess of a state near the finite element solution is equivalent to its squared distance from that solution. We use this loss to train a reduced basis neural operator with values in this space, with solution snapshots needed only for the basis and the loss and its symbolically derived gradient assembled by the finite element library. At inference, this loss estimates the error. Numerically, the square root of the loss excess tracks the error within factors comparable to those of the linear case, and the operator achieves higher accuracy in the PDE-compliant norm than grid-based baselines evaluated in the same finite element space.

## PDEAgent-Bench: A Multi-Metric, Multi-Library Benchmark for PDE Solver Generation

DevFeed: [PDEAgent-Bench: A Multi-Metric, Multi-Library Benchmark for PDE Solver Generation](<https://devfeed.tech/articles/pdeagent-bench-a-multi-metric-multi-library-benchmark-for-pde-solver-generation-137253.md>)

Original publisher: [Read original article](<https://arxiv.org/abs/2605.09636>)

Author: Zhen Hang, Yushan Yashengjiang, Junhui Li, Huanshuo Dong, Yang Wei, Zhezheng Hao, Jiangtao Ma, Songlin Bai, Haozhong Kai, Xihang Yue, Gangzong Si, Dongming Jiang, Chao Yao, Zhanhua Hu, Jiangqing Zhang

Published: 2026-11-17T05:00:00Z

Content type: news

Language: en

Sources: [arXiv: Artificial Intelligence](<https://devfeed.tech/sources/arxiv-artificial-intelligence.md>)

Topics: [Finite Element Method (FEM)](<https://devfeed.tech/topics/finite-element-method.md>), [Benchmark](<https://devfeed.tech/topics/benchmark.md>), [Library](<https://devfeed.tech/topics/library.md>)

Tags: [agent](<https://devfeed.tech/tags/agent.md>), [benchmark](<https://devfeed.tech/tags/benchmark.md>), [code-generation](<https://devfeed.tech/tags/code-generation.md>), [cs-ai](<https://devfeed.tech/tags/cs-ai.md>), [efficiency](<https://devfeed.tech/tags/efficiency.md>), [evaluation](<https://devfeed.tech/tags/evaluation.md>), [libraries](<https://devfeed.tech/tags/libraries.md>), [testbed](<https://devfeed.tech/tags/testbed.md>)

### AI overview

PDEAgent-Bench evaluates generated PDE solvers on whether they run, meet numerical accuracy targets, and satisfy computational efficiency targets. It contains 645 instances spanning 6 mathematical categories and 11 PDE families, with cases for DOLFINx, Firedrake, and deal.II; each includes a problem specification, a reference solution on a prescribed grid, and case-specific accuracy and runtime targets. Experiments with representative LLMs and code agents found that runnable code was often achievable, but pass rates fell substantially when accuracy and efficiency were also required, indicating limits in producing reliable, efficient solvers.

### Source excerpt

arXiv:2605.09636v2 Announce Type: replace Abstract: PDE-to-solver code generation aims to automatically synthesize executable numerical solvers from partial differential equation (PDE) specifications. This task requires not only understanding the mathematical structure of PDEs, but also selecting appropriate discretization schemes and solver configurations, and correctly implementing the resulting formulations in finite-element method (FEM) libraries. Existing code generation benchmarks mainly evaluate syntactic correctness, or success on predefined test cases. To our knowledge, there is currently no publicly available benchmark specifically for PDE-to-solver code generation, and general-purpose code benchmarks do not fully capture the unique challenges of numerical PDE solution, such as ensuring solver accuracy, efficiency, and compatibility with professional FEM libraries. We introduce PDEAgent-Bench, to the best of our knowledge, the first multi-metric, multi-library benchmark for PDE-to-solver code generation. PDEAgent-Bench contains 645 instances across 6 mathematical categories and 11 PDE families, with common FEM libraries for DOLFINx, Firedrake, and deal.II. Each instance provides an agent-facing problem specification, a reference solution on a prescribed evaluation grid, and case-specific accuracy and runtime targets. PDEAgent-Bench adopts a staged evaluation framework in which generated solvers must sequentially pass executability, numerical accuracy, and computational efficiency checks. Experiments with representative LLMs and code agents show that models can often produce runnable code, but their pass rate drops substantially once accuracy and efficiency requirements are enforced. These results indicate that current agents remain limited in producing numerically reliable and efficient PDE solvers, and that PDEAgent-Bench provides a reproducible testbed grounded in the practical requirements of numerical PDE solving.

## Learning inadequacy of Bayesian neural operators for PDE-constrained inference and design

DevFeed: [Learning inadequacy of Bayesian neural operators for PDE-constrained inference and design](<https://devfeed.tech/articles/learning-inadequacy-of-bayesian-neural-operators-for-pde-constrained-inference-and-design-137397.md>)

Original publisher: [Read original article](<https://arxiv.org/abs/2611.19115>)

Author: Pratyush Kumar Singh, Lianghao Cao, E. Bruce Pitman, Danial Faghihi

Published: 2026-11-17T05:00:00Z

Content type: article

Language: en

Sources: [arXiv: Artificial Intelligence](<https://devfeed.tech/sources/arxiv-artificial-intelligence.md>)

Topics: [Artificial Intelligence](<https://devfeed.tech/topics/ai.md>), [Neural Network](<https://devfeed.tech/topics/neural-network.md>), [Finite Element Method (FEM)](<https://devfeed.tech/topics/finite-element-method.md>), [Inference](<https://devfeed.tech/topics/inference.md>), [Simulation and Design](<https://devfeed.tech/topics/simulation-and-design.md>)

Tags: [architectures](<https://devfeed.tech/tags/architectures.md>), [bayesian](<https://devfeed.tech/tags/bayesian.md>), [benchmark](<https://devfeed.tech/tags/benchmark.md>), [building](<https://devfeed.tech/tags/building.md>), [cs-ai](<https://devfeed.tech/tags/cs-ai.md>), [design](<https://devfeed.tech/tags/design.md>), [inference](<https://devfeed.tech/tags/inference.md>), [learning](<https://devfeed.tech/tags/learning.md>), [markov-chain-monte-carlo](<https://devfeed.tech/tags/markov-chain-monte-carlo.md>), [neural](<https://devfeed.tech/tags/neural.md>), [online](<https://devfeed.tech/tags/online.md>), [operators](<https://devfeed.tech/tags/operators.md>), [optimization](<https://devfeed.tech/tags/optimization.md>), [optimized](<https://devfeed.tech/tags/optimized.md>), [parameter](<https://devfeed.tech/tags/parameter.md>), [partial-differential-equations](<https://devfeed.tech/tags/partial-differential-equations.md>), [simulation](<https://devfeed.tech/tags/simulation.md>), [simulations](<https://devfeed.tech/tags/simulations.md>), [thermal](<https://devfeed.tech/tags/thermal.md>), [training](<https://devfeed.tech/tags/training.md>), [uncertainty](<https://devfeed.tech/tags/uncertainty.md>)

### AI overview

RI-BNO models Bayesian neural operator inadequacy by using a posteriori PDE error estimates to correct predictions and represent uncertainty; local error bounds are established under stability and smoothness assumptions. In the reported reaction-diffusion, glioma inference, and thermal-break design examples, it improved agreement with finite-element reference predictions, parameter posteriors, and optimized designs while using half as many full nonlinear or transient training simulations as vanilla BNO. For thermal-break design, online optimization took 18 minutes with RI-BNO versus 178 hours with the FE model. These results are limited to the evaluated examples.

### Source excerpt

arXiv:2611.19115v1 Announce Type: new Abstract: Partial differential equations (PDEs) underpin predictive modeling across science, engineering and medicine. Their use in digital twins, parameter inference and optimal design requires repeated solutions across parameters and scenarios. For nonlinear, transient and multiphysics systems, these calculations can make numerical simulation prohibitively expensive. Here we introduce the residual-informed Bayesian neural operator (RI-BNO), which learns operator inadequacy using a posteriori error estimates from the governing PDEs and incorporates its correction and uncertainty into Bayesian training and prediction. We establish local error bounds for the residual-informed estimates under stability and smoothness assumptions. We demonstrate RI-BNO on a non-linear reaction-diffusion benchmark with DeepONet and latent-subspace architectures, Markov chain Monte Carlo inference of glioma parameter fields from magnetic resonance imaging, and Monte Carlo risk-averse optimization of three-dimensional thermal breaks in building envelopes under manufacturing uncertainty. Across the evaluated examples, RI-BNO improves agreement with finite element (FE) reference predictions, parameter posteriors and optimized designs while using half as many full nonlinear or transient training simulations as the vanilla Bayesian neural operator (vBNO). For thermal break design, online optimization time decreases from 178 hours with the FE model to 18 minutes with RI-BNO. These results demonstrate the value of explicit inadequacy modeling in neural operators for reliable PDE-based inference and design.

## A mesh-free deep energy method for phase-field modeling of thermally driven brittle fracture

DevFeed: [A mesh-free deep energy method for phase-field modeling of thermally driven brittle fracture](<https://devfeed.tech/articles/a-mesh-free-deep-energy-method-for-phase-field-modeling-of-thermally-driven-brittle-fracture-135812.md>)

Original publisher: [Read original article](<https://arxiv.org/abs/2611.17630>)

Author: Han Zhang, Mehrisadat Makki Alamdari, Babak Shahbodagh, Mohammad Vahab, Cosmin Anitescu, Timon Rabczuk, Elena Atroshchenko

Published: 2026-11-16T05:00:00Z

Content type: article

Language: en

Sources: [arXiv: Machine Learning](<https://devfeed.tech/sources/arxiv-machine-learning.md>)

Topics: [Deep learning](<https://devfeed.tech/topics/deep-learning.md>), [Physics-guided deep learning](<https://devfeed.tech/topics/physics-guided-deep-learning.md>), [Simulation](<https://devfeed.tech/topics/simulation.md>), [Finite Element Method (FEM)](<https://devfeed.tech/topics/finite-element-method.md>)

Tags: [array](<https://devfeed.tech/tags/array.md>), [arrays](<https://devfeed.tech/tags/arrays.md>), [arrest](<https://devfeed.tech/tags/arrest.md>), [benchmark](<https://devfeed.tech/tags/benchmark.md>), [ceramic](<https://devfeed.tech/tags/ceramic.md>), [cond-mat-mtrl-sci](<https://devfeed.tech/tags/cond-mat-mtrl-sci.md>), [crack](<https://devfeed.tech/tags/crack.md>), [cs-lg](<https://devfeed.tech/tags/cs-lg.md>), [deep](<https://devfeed.tech/tags/deep.md>), [encoding](<https://devfeed.tech/tags/encoding.md>), [experiment](<https://devfeed.tech/tags/experiment.md>), [fields](<https://devfeed.tech/tags/fields.md>), [free](<https://devfeed.tech/tags/free.md>), [functional](<https://devfeed.tech/tags/functional.md>), [heat-equation](<https://devfeed.tech/tags/heat-equation.md>), [loading](<https://devfeed.tech/tags/loading.md>), [modeling](<https://devfeed.tech/tags/modeling.md>), [neural](<https://devfeed.tech/tags/neural.md>), [prediction](<https://devfeed.tech/tags/prediction.md>), [reference](<https://devfeed.tech/tags/reference.md>), [spacing](<https://devfeed.tech/tags/spacing.md>), [temperature](<https://devfeed.tech/tags/temperature.md>)

### AI overview

The proposed mesh-free deep energy method predicts spontaneous crack formation under transient thermal loading without prescribed crack paths, nucleation sites, or crack-adapted meshes. Separate multiresolution networks represent displacement and phase fields and temperature; energy terms are estimated at integration points redrawn each iteration. Restricting irreversibility to the crack set, carrying optimizer state across load steps, and decaying learning rates within steps help limit spurious damage. Reported results include initiation peak loads on two notched plates that closely match an independent finite element reference, and crack patterns in glass and ceramic cases consistent with stability predictions or published observations and benchmarks. However, crack arrays in a quenched slab and annulus are denser than finite element results on the same material field.

### Source excerpt

arXiv:2611.17630v1 Announce Type: new Abstract: Phase-field simulation of thermally driven fracture requires resolving localized crack bands whose paths are unknown in advance and thermal boundary layers whose thickness evolves during the transient. Neural energy-based computations of thermal fracture have so far started from defects placed in advance. This paper proposes a mesh-free deep energy method based on multiresolution neural representations that predicts spontaneous crack nucleation, crack spacing, and selective crack propagation and arrest under transient thermal loading. Two networks, each with its own multiresolution encoding, represent the displacement and phase fields and the transient temperature, and the mechanical energy and the incremental functional of the backward Euler heat equation are estimated on integration points redrawn at every iteration. Restricting irreversibility to the crack set, carrying the optimizer state across load steps and decaying the learning rates within each step help prevent spurious damage accumulation. On two notched plates the initiation peak loads agree closely with an independent finite element reference. On a quenched glass strip under an analytical temperature field the crack path changes from straight to oscillatory and then to branched, with an oscillation threshold close to the linear stability prediction. Under a prescribed boundary temperature a quenched ceramic slab forms a crack array of the type observed in experiment, with a spacing close to that of a published benchmark. On the slab and a quenched annulus the arrays are denser than those of a finite element computation on the same material field. No crack paths, nucleation sites or crack-adapted meshes are prescribed.

## Fuse-PINN: Combining Finite Element Methods and Physics-Informed Neural Networks to enable massively parallel simulations of non-linear PDEs

DevFeed: [Fuse-PINN: Combining Finite Element Methods and Physics-Informed Neural Networks to enable massively parallel simulations of non-linear PDEs](<https://devfeed.tech/articles/fuse-pinn-combining-finite-element-methods-and-physics-informed-neural-networks-to-enable-massively-parallel-simulations-of-non-linear-pdes-135898.md>)

Original publisher: [Read original article](<https://arxiv.org/abs/2611.18409>)

Author: Stanislas Pamela, Nicolas Pailliez, Emmanuel Franck, Victor Michel-Dansac, Laurent Navoret, Vignesh Gopakumar

Published: 2026-11-16T05:00:00Z

Content type: article

Language: en

Sources: [arXiv: Machine Learning](<https://devfeed.tech/sources/arxiv-machine-learning.md>)

Topics: [Artificial Intelligence](<https://devfeed.tech/topics/ai.md>), [Neural Network](<https://devfeed.tech/topics/neural-network.md>), [Finite Element Method (FEM)](<https://devfeed.tech/topics/finite-element-method.md>), [High-Performance Computing](<https://devfeed.tech/topics/high-performance-computing.md>)

Tags: [cs-lg](<https://devfeed.tech/tags/cs-lg.md>), [framework](<https://devfeed.tech/tags/framework.md>), [gpu](<https://devfeed.tech/tags/gpu.md>), [gradient](<https://devfeed.tech/tags/gradient.md>), [gradient-descent](<https://devfeed.tech/tags/gradient-descent.md>), [high-performance-computing](<https://devfeed.tech/tags/high-performance-computing.md>), [multi-gpu](<https://devfeed.tech/tags/multi-gpu.md>), [neural-networks](<https://devfeed.tech/tags/neural-networks.md>), [parallel](<https://devfeed.tech/tags/parallel.md>), [physics](<https://devfeed.tech/tags/physics.md>), [physics-informed](<https://devfeed.tech/tags/physics-informed.md>), [physics-plasm-ph](<https://devfeed.tech/tags/physics-plasm-ph.md>), [scaling](<https://devfeed.tech/tags/scaling.md>), [simulations](<https://devfeed.tech/tags/simulations.md>)

### AI overview

Fuse-PINN combines a finite-element formulation for neural networks (FEM-NN) with an Energy Natural Gradient Descent optimizer to scale physics-informed neural networks on GPU-based high-performance computing systems. The authors demonstrate it on magnetic-confinement-fusion models, including nonlinear convection with fine-scale dynamics, and report near-perfect scaling in multi-GPU runs as problem size increases. The supplied abstract does not give benchmark details or quantify convergence accuracy.

### Source excerpt

arXiv:2611.18409v1 Announce Type: new Abstract: Research in the domain of Physics-Informed Neural Networks applied to PDE modelling has advanced significantly in recent years, providing a wide ecosystem of novel methods. However, one frequent limitation is the inability of PINNs to scale out on High-Performance Computing architectures to address more ambitious challenges, particularly non-linear physics models with turbulent dynamics in large spatial domains. Typical studies restrain demonstrations to toy models with simple physics equations, few variables, and rarely any complex non-linear dynamics evolved over long timescales. Part of the AI community has even adopted the idea that "PINNs don't scale". This work proves otherwise. We present a mathematically robust formulation of Finite-Element-Method for Neural Networks (FEM-NN) implemented into a comprehensive, modular framework: Fuse-PINN. This formulation provides the basis for a sclable parallelisation of Physics-Informed Neural Networks on GPU-based HPC systems. Furthermore, an Energy Natural Gradient Descent optimiser is implemented to allow high-precision convergence. Fuse-PINN is demonstrated on physics models relevant to Magnetic Confinement Fusion, including non-linear convective models with fine-scale dynamics. Near-perfect scaling is achieved on multi-GPU runs with increasing problem size.

## MechFoundry: An agentic loop that poses and performs engineering analyses and optimization of real-world objects

DevFeed: [MechFoundry: An agentic loop that poses and performs engineering analyses and optimization of real-world objects](<https://devfeed.tech/articles/mechfoundry-an-agentic-loop-that-poses-and-performs-engineering-analyses-and-optimization-of-real-world-objects-132023.md>)

Original publisher: [Read original article](<https://arxiv.org/abs/2611.14571>)

Author: Xue-Ling Luo, Marco Conte, Peter Yichen Chen, Nikolaos Bouklas

Published: 2026-11-13T05:00:00Z

Content type: article

Language: en

Sources: [arXiv: Artificial Intelligence](<https://devfeed.tech/sources/arxiv-artificial-intelligence.md>)

Topics: [Loop Engineering](<https://devfeed.tech/topics/loop-engineering.md>), [Finite Element Method (FEM)](<https://devfeed.tech/topics/finite-element-method.md>), [Operations research and optimization](<https://devfeed.tech/topics/operations-research-and-optimization.md>)

Tags: [agentic](<https://devfeed.tech/tags/agentic.md>), [agentic-loop](<https://devfeed.tech/tags/agentic-loop.md>), [ai](<https://devfeed.tech/tags/ai.md>), [ai-agents](<https://devfeed.tech/tags/ai-agents.md>), [analysis](<https://devfeed.tech/tags/analysis.md>), [automate](<https://devfeed.tech/tags/automate.md>), [cs-ai](<https://devfeed.tech/tags/cs-ai.md>), [cs-ce](<https://devfeed.tech/tags/cs-ce.md>), [design](<https://devfeed.tech/tags/design.md>), [engineering](<https://devfeed.tech/tags/engineering.md>), [gpu](<https://devfeed.tech/tags/gpu.md>), [human-in-the-loop](<https://devfeed.tech/tags/human-in-the-loop.md>), [language-model](<https://devfeed.tech/tags/language-model.md>), [loading](<https://devfeed.tech/tags/loading.md>), [material](<https://devfeed.tech/tags/material.md>), [modeling](<https://devfeed.tech/tags/modeling.md>), [object](<https://devfeed.tech/tags/object.md>), [optimization](<https://devfeed.tech/tags/optimization.md>), [real-world](<https://devfeed.tech/tags/real-world.md>), [shape](<https://devfeed.tech/tags/shape.md>), [simulation](<https://devfeed.tech/tags/simulation.md>), [software](<https://devfeed.tech/tags/software.md>)

### AI overview

MechFoundry automates structural analysis and design improvement from an object's text description: it generates a shape, infers a use-related loading case, assigns materials to parts, solves nonlinear structural response, and optimizes the design for those loads. A differentiable GPU solver using an adaptive octree supports the transition from shape generation to simulation and optimization. Across 50 everyday and engineering objects, the system posed and solved a loading case in an average of 6.5 minutes and US$1.3 per object. Because there is no test suite to verify the engineering decisions, language-model reviewers graded sessions as an initial assessment of where the agent's judgment succeeds or fails.

### Source excerpt

arXiv:2611.14571v1 Announce Type: cross Abstract: AI agents can automate software engineering because a compiler or a test suite checks that the code does what the user or the agent intended. In engineering analysis, however, the steps that precede the computation have few easily verifiable checks: modeling a design or an object, deciding how it is used, where supports and loads are applied, and what material each part is made of. A human expert therefore makes these decisions, carries out the steps, and judges the result, one object at a time. We present MechFoundry, an agentic loop coupled with bespoke software that automates structural analyses of real-world objects without a human in the loop. From an object's lexical description, it generates a shape, poses a loading case relevant to how the object is used, assigns a material to each part, solves the nonlinear structural response, and improves the design against the loads it carries. A differentiable GPU solver built on an adaptive octree that need not conform to the surface enables the transition from shape generation to simulation and optimization. Across 50 everyday and engineering objects, the agent posed and solved a loading case in 6.5 minutes and for US$1.3 per object on average. Without a test suite to check against, we had language-model reviewers grade each session, as a first indication of where the agent's engineering judgment holds and where it fails.

## A hybrid FEM-PINN method for time-dependent partial differential equations

DevFeed: [A hybrid FEM-PINN method for time-dependent partial differential equations](<https://devfeed.tech/articles/a-hybrid-fem-pinn-method-for-time-dependent-partial-differential-equations-131839.md>)

Original publisher: [Read original article](<https://arxiv.org/abs/2409.02810>)

Author: Xiaodong Feng, Haojiong Shangguan, Tao Tang, Xiaoliang Wan

Published: 2026-11-13T05:00:00Z

Content type: article

Language: en

Sources: [arXiv: Artificial Intelligence](<https://devfeed.tech/sources/arxiv-artificial-intelligence.md>)

Topics: [Finite Element Method (FEM)](<https://devfeed.tech/topics/finite-element-method.md>)

Tags: [adaptive](<https://devfeed.tech/tags/adaptive.md>), [cs-ai](<https://devfeed.tech/tags/cs-ai.md>), [cs-na](<https://devfeed.tech/tags/cs-na.md>), [deep-learning](<https://devfeed.tech/tags/deep-learning.md>), [deep-neural-networks](<https://devfeed.tech/tags/deep-neural-networks.md>), [efficiency](<https://devfeed.tech/tags/efficiency.md>), [errors](<https://devfeed.tech/tags/errors.md>), [experiments](<https://devfeed.tech/tags/experiments.md>), [hybrid](<https://devfeed.tech/tags/hybrid.md>), [math-na](<https://devfeed.tech/tags/math-na.md>), [partial-differential-equations](<https://devfeed.tech/tags/partial-differential-equations.md>), [sampling](<https://devfeed.tech/tags/sampling.md>)

### AI overview

The method combines time finite elements with PINNs for evolution PDEs. It represents time dependence with finite element basis functions and uses a neural network to produce their space-dependent coefficients; Galerkin or collocation projection in time then yields a system of PDEs for those coefficients, approximated with PINNs. The authors say this avoids statistical error in time integrals and lets the network output serve as reduced spatial basis functions. To address high dimensionality and low regularity, adaptive sampling estimates the distribution induced by the PDE residual with an explicit density model and adds time-dependent random training samples drawn from it. The abstract reports effectiveness and efficiency in numerical experiments but gives no specific results.

### Source excerpt

arXiv:2409.02810v2 Announce Type: replace-cross Abstract: In this work, we present a hybrid numerical method for solving evolution partial differential equations (PDEs) by merging the time finite element method with deep neural networks. In contrast to the conventional deep learning-based formulation where the neural network is defined on a spatiotemporal domain, our methodology utilizes finite element basis functions in the time direction where the space-dependent coefficients are defined as the output of a neural network. We then apply the Galerkin or collocation projection in the time direction to obtain a system of PDEs for the space-dependent coefficients which is approximated in the framework of PINN. The advantages of such a hybrid formulation are twofold: statistical errors are avoided for the integral in the time direction, and the neural network's output can be regarded as a set of reduced spatial basis functions. To further alleviate the difficulties from high dimensionality and low regularity, we have developed an adaptive sampling strategy that refines the training set. More specifically, we use an explicit density model to approximate the distribution induced by the PDE residual and then augment the training set with new time-dependent random samples given by the learned density model. The effectiveness and efficiency of our proposed method have been demonstrated through a series of numerical experiments.

## Learned Lagrangian Models of PDEs via Euler-Lagrange Residual Minimization

DevFeed: [Learned Lagrangian Models of PDEs via Euler-Lagrange Residual Minimization](<https://devfeed.tech/articles/learned-lagrangian-models-of-pdes-via-euler-lagrange-residual-minimization-132779.md>)

Original publisher: [Read original article](<https://arxiv.org/abs/2605.07157>)

Author: Lyra Zhornyak, Eric Forgoston, M. Ani Hsieh

Published: 2026-11-13T05:00:00Z

Content type: article

Language: en

Sources: [arXiv: Machine Learning](<https://devfeed.tech/sources/arxiv-machine-learning.md>)

Topics: [Machine learning](<https://devfeed.tech/topics/machine-learning.md>), [Finite Element Method (FEM)](<https://devfeed.tech/topics/finite-element-method.md>)

Tags: [abstract](<https://devfeed.tech/tags/abstract.md>), [approach](<https://devfeed.tech/tags/approach.md>), [arxiv](<https://devfeed.tech/tags/arxiv.md>), [boundary](<https://devfeed.tech/tags/boundary.md>), [bypass](<https://devfeed.tech/tags/bypass.md>), [construction](<https://devfeed.tech/tags/construction.md>), [continuous](<https://devfeed.tech/tags/continuous.md>), [cs-lg](<https://devfeed.tech/tags/cs-lg.md>), [develop](<https://devfeed.tech/tags/develop.md>), [domain](<https://devfeed.tech/tags/domain.md>), [error](<https://devfeed.tech/tags/error.md>), [existing](<https://devfeed.tech/tags/existing.md>), [first](<https://devfeed.tech/tags/first.md>), [fixed](<https://devfeed.tech/tags/fixed.md>), [forecast](<https://devfeed.tech/tags/forecast.md>), [global](<https://devfeed.tech/tags/global.md>), [horizon](<https://devfeed.tech/tags/horizon.md>), [integration](<https://devfeed.tech/tags/integration.md>), [iteration](<https://devfeed.tech/tags/iteration.md>), [lagrange](<https://devfeed.tech/tags/lagrange.md>), [models](<https://devfeed.tech/tags/models.md>)

### AI overview

The method forecasts PDE dynamics from a learned continuous Lagrangian by minimizing squared Euler-Lagrange residuals on overlapping local space-time patches. This avoids the global implicit solve of fixed-discretization integrators and decouples model phase error from integration conservation error. Jacobi iteration gives linear scaling with domain size, and the approach imposes no structural requirements on the learned network. Tests on a double pendulum and one- and two-dimensional wave equations report long-horizon errors comparable to classical symplectic methods; the method also generalizes to spatially varying dynamics and arbitrary boundary conditions without retraining.

### Source excerpt

arXiv:2605.07157v2 Announce Type: replace Abstract: We present the first method to directly use a learned continuous Lagrangian to forecast the dynamics of systems governed by partial differential equations, exploiting the inherent conservative structure to achieve stable long-range predictions. We develop an integrator that minimizes the squared Euler-Lagrange residual via a discretization-agnostic, structure-preserving construction on local space-time patches. Different from integrators for analytical models, integrators for learned models should decouple model error (phase error) from integration error (conservation error). By minimizing the residual locally on overlapping patches, we bypass the global implicit solve inherent to fixed discretizations, which slows time- and space-stepping and complicates learning. Our method scales linearly with domain size via Jacobi iteration, and places no structural requirements on the learned network, allowing it to be coupled with existing physics-guided machine learning (ML) methods. We validate our approach on a learned representation of a double pendulum, a one-dimensional wave equation, and a two-dimensional wave equation. Our method achieves long-horizon predictions with error comparable to classical symplectic methods while generalizing to spatially varying dynamics and arbitrary boundary conditions without retraining.

## Accelerating Direct Sparse Solvers with Graph Computing Accelerators

DevFeed: [Accelerating Direct Sparse Solvers with Graph Computing Accelerators](<https://devfeed.tech/articles/accelerating-direct-sparse-solvers-with-graph-computing-accelerators-126735.md>)

Original publisher: [Read original article](<https://arxiv.org/abs/2611.11876>)

Author: Jiya Su, Fran\\c{c}ois-Henry Rouet, Jeff Beisheim, Robert Lucas, Yuqing Wang, Alexander Fell, Chris Mccarthy, David F. Gleich, Andrew A Chien

Published: 2026-11-10T05:00:00Z

Content type: news

Language: en

Sources: [arXiv: Distributed Computing](<https://devfeed.tech/sources/arxiv-distributed-computing.md>)

Topics: [Distributed Systems & Parallel Computing](<https://devfeed.tech/topics/distributed-systems-parallel-computing.md>), [High-Performance Computing](<https://devfeed.tech/topics/high-performance-computing.md>), [graph theory](<https://devfeed.tech/topics/graph-theory.md>), [Finite Element Method (FEM)](<https://devfeed.tech/topics/finite-element-method.md>), [GPU](<https://devfeed.tech/topics/gpu.md>)

Tags: [accelerators](<https://devfeed.tech/tags/accelerators.md>), [algorithm](<https://devfeed.tech/tags/algorithm.md>), [avx](<https://devfeed.tech/tags/avx.md>), [computer-aided-engineering](<https://devfeed.tech/tags/computer-aided-engineering.md>), [computing](<https://devfeed.tech/tags/computing.md>), [cs-dc](<https://devfeed.tech/tags/cs-dc.md>), [fpga](<https://devfeed.tech/tags/fpga.md>), [gaussian](<https://devfeed.tech/tags/gaussian.md>), [gpus](<https://devfeed.tech/tags/gpus.md>), [graph](<https://devfeed.tech/tags/graph.md>), [irregular](<https://devfeed.tech/tags/irregular.md>), [linear](<https://devfeed.tech/tags/linear.md>), [parallel](<https://devfeed.tech/tags/parallel.md>), [platforms](<https://devfeed.tech/tags/platforms.md>), [scaling](<https://devfeed.tech/tags/scaling.md>), [simulation](<https://devfeed.tech/tags/simulation.md>)

### AI overview

The abstract reports that the UpDown graph computing accelerator could reduce METIS nested dissection reordering time, a step that can bottleneck large sparse-system solvers because reordering scales less effectively than factorization and cannot exploit AVX or GPUs. Results from validated simulation and FPGA emulation indicate small UpDown systems speed METIS by about two orders of magnitude over a single CPU. The authors anticipate 4-7x speedups over ParMETIS running on hundreds of MPI ranks. These results concern small-scale systems, with commercial availability stated for 2027; the ParMETIS comparison is an anticipation, and the abstract frames faster finite-element simulations as a potential benefit.

### Source excerpt

arXiv:2611.11876v1 Announce Type: new Abstract: Solving large sparse linear systems by Gaussian elimination is a challenging computational task due to the introduction of non-zero elements, or fill-in. Practitioners use heuristics such as the nested dissection reordering algorithm to try to reduce fill-in. In practice, reordering codes do not achieve the same parallel scaling as factorization, nor can they exploit AVX instructions or GPUs. Thus, reordering can be the primary runtime bottleneck in large-scale computer-aided engineering applications. We consider the use of a novel accelerator, the UpDown system, to accelerate the METIS multilevel nested dissection reordering algorithm. The UpDown system is designed for a broad class of graph computing and irregular algorithms. We report results for small-scale UpDown systems that will be available commercially in 2027, using validated simulation and FPGA emulation platforms. Our findings indicate that small UpDown systems can speed up METIS by two-orders of magnitude versus a single CPU. Compared to ParMETIS on hundreds of MPI ranks, we anticipate the same small UpDown systems offer speedups of 4 - 7, providing the potential to significantly speed finite element simulations

## A Judge Agent Closes the Reliability Gap in AI-Generated Scientific Simulation

DevFeed: [A Judge Agent Closes the Reliability Gap in AI-Generated Scientific Simulation](<https://devfeed.tech/articles/a-judge-agent-closes-the-reliability-gap-in-ai-generated-scientific-simulation-127183.md>)

Original publisher: [Read original article](<https://arxiv.org/abs/2603.25780>)

Author: Chengshuai Yang

Published: 2026-11-10T05:00:00Z

Content type: news

Language: en

Sources: [arXiv: Software Engineering](<https://devfeed.tech/sources/arxiv-software-engineering.md>)

Topics: [Code quality](<https://devfeed.tech/topics/code-quality.md>), [Finite Element Method (FEM)](<https://devfeed.tech/topics/finite-element-method.md>), [Uncertainty quantification LLMs](<https://devfeed.tech/topics/uncertainty-quantification-llms.md>)

Tags: [agent](<https://devfeed.tech/tags/agent.md>), [ai](<https://devfeed.tech/tags/ai.md>), [automated](<https://devfeed.tech/tags/automated.md>), [benchmark](<https://devfeed.tech/tags/benchmark.md>), [certification](<https://devfeed.tech/tags/certification.md>), [code](<https://devfeed.tech/tags/code.md>), [convergence](<https://devfeed.tech/tags/convergence.md>), [cs-lg](<https://devfeed.tech/tags/cs-lg.md>), [cs-se](<https://devfeed.tech/tags/cs-se.md>), [data](<https://devfeed.tech/tags/data.md>), [domains](<https://devfeed.tech/tags/domains.md>), [error](<https://devfeed.tech/tags/error.md>), [failure](<https://devfeed.tech/tags/failure.md>), [language-models](<https://devfeed.tech/tags/language-models.md>), [pipeline](<https://devfeed.tech/tags/pipeline.md>), [reliability](<https://devfeed.tech/tags/reliability.md>), [simulation](<https://devfeed.tech/tags/simulation.md>), [specification](<https://devfeed.tech/tags/specification.md>), [test](<https://devfeed.tech/tags/test.md>)

### AI overview

The abstract reports that a Judge Agent automating mathematical checks--well-posedness, convergence, and error certification--reduced silent failures from 42% to 1.5% across 134 tests in 12 scientific domains. In a prospective benchmark of 72 blinded tasks from 12 scientists, success was 89% with automated error bounds versus 53% without the Judge (95% CI 80-95%). A clinical CT experiment, the only powered experiment (n=200), reached 99% of expert quality. Remaining failures concentrated at bifurcation points, where certifiability breaks down. The work also introduces the simulability class S and a structured, solver-independent specification format called spec.md.

### Source excerpt

arXiv:2603.25780v2 Announce Type: replace Abstract: Large language models can generate scientific simulation code, but the generated code silently fails on most non-textbook problems. We show that classical mathematical validation -- well-posedness, convergence, and error certification -- can be fully automated by a Judge Agent, reducing the silent-failure rate from 42% to 1.5% across 134 test cases spanning 12 scientific domains. The headline result comes from a prospective benchmark: 72 blinded tasks submitted by 12 independent scientists yield an 89% success rate (95% CI: [80%, 95%]) with automated error bounds, versus 53% without the Judge. On clinical CT (the only powered experiment, n = 200), the pipeline reaches 99% of expert quality. The residual 1.5% concentrates at bifurcation points where certifiability breaks down. We formalize this boundary through the simulability class S and introduce spec.md, a structured specification format that makes any scientific computation problem machine-readable and solver-independent. Code, data, and all 72 benchmark tasks are publicly archived.

## Deep backward multistep high-order regression methods for high dimensional nonlinear parabolic partial differential equations

DevFeed: [Deep backward multistep high-order regression methods for high dimensional nonlinear parabolic partial differential equations](<https://devfeed.tech/articles/deep-backward-multistep-high-order-regression-methods-for-high-dimensional-nonlinear-parabolic-partial-differential-equations-125221.md>)

Original publisher: [Read original article](<https://arxiv.org/abs/2611.10570>)

Author: Qiang Han, Shaolin Ji, Zihao Xv

Published: 2026-11-10T05:00:00Z

Content type: article

Language: en

Sources: [arXiv: Machine Learning](<https://devfeed.tech/sources/arxiv-machine-learning.md>)

Topics: [Deep learning](<https://devfeed.tech/topics/deep-learning.md>), [Artificial Intelligence](<https://devfeed.tech/topics/ai.md>), [Finite Element Method (FEM)](<https://devfeed.tech/topics/finite-element-method.md>)

Tags: [analysis](<https://devfeed.tech/tags/analysis.md>), [approximation](<https://devfeed.tech/tags/approximation.md>), [benchmarks](<https://devfeed.tech/tags/benchmarks.md>), [complexity](<https://devfeed.tech/tags/complexity.md>), [conditional](<https://devfeed.tech/tags/conditional.md>), [cs-lg](<https://devfeed.tech/tags/cs-lg.md>), [deep](<https://devfeed.tech/tags/deep.md>), [dimension](<https://devfeed.tech/tags/dimension.md>), [error](<https://devfeed.tech/tags/error.md>), [generators](<https://devfeed.tech/tags/generators.md>), [gradient](<https://devfeed.tech/tags/gradient.md>), [independent](<https://devfeed.tech/tags/independent.md>), [j](<https://devfeed.tech/tags/j.md>), [math-ap](<https://devfeed.tech/tags/math-ap.md>), [monte-carlo](<https://devfeed.tech/tags/monte-carlo.md>), [network](<https://devfeed.tech/tags/network.md>), [networks](<https://devfeed.tech/tags/networks.md>), [neural](<https://devfeed.tech/tags/neural.md>), [neural-network](<https://devfeed.tech/tags/neural-network.md>), [optimization](<https://devfeed.tech/tags/optimization.md>), [partial-differential-equations](<https://devfeed.tech/tags/partial-differential-equations.md>), [regression](<https://devfeed.tech/tags/regression.md>), [stat-ml](<https://devfeed.tech/tags/stat-ml.md>)

### AI overview

The proposed MHDBR method solves semilinear parabolic PDEs through their backward stochastic differential equation representations, combining multistep time discretization with regression targets based on conditional expectations. Under stated regularity, initialization, forward-approximation, and network-moment assumptions, its error bound separates time-discretization, conditional-sampling, and neural-network approximation errors. For generators independent of the gradient variable, constrained GroupSort networks and sufficiently large conditional sample sizes yield order J in the stated nodal L² error norm. The analysis assumes exact population minimization and excludes finite-sample generalization and optimization errors; network sizes depend on dimension, so dimension-independent complexity is not established. The abstract also reports one reproducible test satisfying the assumptions and three broader benchmarks whose connection to them is discussed.

### Source excerpt

arXiv:2611.10570v1 Announce Type: cross Abstract: We propose a deep backward multistep high-order regression (MHDBR) method for semilinear parabolic partial differential equations through their backward stochastic differential equation representations. The method combines multistep time discretization with regression targets formed by conditional expectations. Under explicit regularity, starting-value, forward-approximation and network-moment assumptions, we prove an error bound separating time discretization, conditional Monte Carlo sampling and neural-network approximation. The analysis assumes exact population minimization; finite-training-sample generalization and optimization errors are not included. For generators independent of the gradient variable, suitably constrained GroupSort networks and sufficiently large conditional sample sizes give order $J$ in the stated nodal $L^2$ error norm. The required network sizes depend on dimension, so this result does not establish dimension-independent complexity. We present a reproducible time-discretization test satisfying the analytical assumptions and three broader neural-network benchmarks, whose relation to the assumptions is discussed explicitly.

## ALETHEIA: A Multi-Frequency Eddy Current Pulsed Thermography Dataset for Neural Operator Learning in Nondestructive Testing

DevFeed: [ALETHEIA: A Multi-Frequency Eddy Current Pulsed Thermography Dataset for Neural Operator Learning in Nondestructive Testing](<https://devfeed.tech/articles/aletheia-a-multi-frequency-eddy-current-pulsed-thermography-dataset-for-neural-operator-learning-in-nondestructive-testing-125346.md>)

Original publisher: [Read original article](<https://arxiv.org/abs/2611.11252>)

Author: Changbin Sun, Jie Xiao, Xiaotian Chen, Yuankai Wu

Published: 2026-11-10T05:00:00Z

Content type: article

Language: en

Sources: [arXiv: Machine Learning](<https://devfeed.tech/sources/arxiv-machine-learning.md>)

Topics: [Machine learning](<https://devfeed.tech/topics/machine-learning.md>), [Artificial Intelligence](<https://devfeed.tech/topics/ai.md>), [Neural Network](<https://devfeed.tech/topics/neural-network.md>), [Open Source Models & Datasets](<https://devfeed.tech/topics/open-source-models-datasets.md>), [Finite Element Method (FEM)](<https://devfeed.tech/topics/finite-element-method.md>)

Tags: [3d](<https://devfeed.tech/tags/3d.md>), [benchmark](<https://devfeed.tech/tags/benchmark.md>), [code](<https://devfeed.tech/tags/code.md>), [cs-lg](<https://devfeed.tech/tags/cs-lg.md>), [dataset](<https://devfeed.tech/tags/dataset.md>), [defects](<https://devfeed.tech/tags/defects.md>), [evolution](<https://devfeed.tech/tags/evolution.md>), [heat-equation](<https://devfeed.tech/tags/heat-equation.md>), [infrared](<https://devfeed.tech/tags/infrared.md>), [irregular](<https://devfeed.tech/tags/irregular.md>), [machine-learning](<https://devfeed.tech/tags/machine-learning.md>), [operator-learning](<https://devfeed.tech/tags/operator-learning.md>), [partial-differential-equations](<https://devfeed.tech/tags/partial-differential-equations.md>), [regular](<https://devfeed.tech/tags/regular.md>), [tasks](<https://devfeed.tech/tags/tasks.md>)

### AI overview

Aletheia is a 3D benchmark for learning neural PDE solvers in nondestructive testing. It provides over 4,700 high-resolution simulated samples across 10 excitation frequencies (1-100 kHz), with volumetric heat-source and transient temperature fields. The dataset supports forward temperature prediction and inverse reconstruction of internal heat sources or defects from surface infrared measurements. It also includes thermography data from cracked rail specimens for calibration and comparison, and defines four benchmark tasks on regular and irregular grids. The abstract describes benchmarking with various neural operators but gives no comparative results. Dataset and code are publicly available.

### Source excerpt

arXiv:2611.11252v1 Announce Type: new Abstract: Learning neural solvers for spatiotemporal partial differential equations (PDEs) under real-world constraints remains a key challenge in scientific machine learning, especially for inverse tasks with sparse and noisy boundary observations. We present the Aletheia dataset, the first 3D benchmark for learning data-driven solvers in the context of nondestructive testing (NDT). The dataset simulates eddy-current-induced heating in conductive solids and models the resulting transient heat propagation governed by the heat equation. Aletheia contains over 4,700 high-resolution samples across 10 excitation frequencies (1-100,kHz), each providing volumetric heat source and temperature fields over time. It supports both forward prediction of temperature evolution and inverse reconstruction of internal heat sources or defects from surface infrared measurements. Real infrared thermography data from cracked rail specimens are included for calibration and comparison. We define four canonical tasks on both regular and irregular grids and benchmark them using various neural operators. Aletheia establishes a unified platform for evaluating neural PDE solvers under realistic NDT conditions, enabling progress in reliable, data-driven inverse modeling. The dataset and code are publicly available at https://github.com/sherlock-0327/ALETHEIA.

## Learnable Hodge-Laplacian Driven Mixture of Experts for Steady-PDE Surrogates on Coarse Unstructured Meshes

DevFeed: [Learnable Hodge-Laplacian Driven Mixture of Experts for Steady-PDE Surrogates on Coarse Unstructured Meshes](<https://devfeed.tech/articles/learnable-hodge-laplacian-driven-mixture-of-experts-for-steady-pde-surrogates-on-coarse-unstructured-meshes-123228.md>)

Original publisher: [Read original article](<https://arxiv.org/abs/2610.20585>)

Author: Stavros Nousias, Akis Nousias, Konstantinos Gkrispanis

Published: 2026-11-09T05:00:00Z

Content type: article

Language: en

Sources: [arXiv: Machine Learning](<https://devfeed.tech/sources/arxiv-machine-learning.md>)

Topics: [Mixture of Experts (MoE)](<https://devfeed.tech/topics/mixture-of-experts-moe.md>), [Finite Element Method (FEM)](<https://devfeed.tech/topics/finite-element-method.md>)

Tags: [cs-cv](<https://devfeed.tech/tags/cs-cv.md>), [cs-lg](<https://devfeed.tech/tags/cs-lg.md>), [dataset](<https://devfeed.tech/tags/dataset.md>), [error](<https://devfeed.tech/tags/error.md>), [fields](<https://devfeed.tech/tags/fields.md>), [mesh](<https://devfeed.tech/tags/mesh.md>), [mixture-of-experts](<https://devfeed.tech/tags/mixture-of-experts.md>), [partial-differential-equations](<https://devfeed.tech/tags/partial-differential-equations.md>), [routing](<https://devfeed.tech/tags/routing.md>), [spatial](<https://devfeed.tech/tags/spatial.md>), [supervision](<https://devfeed.tech/tags/supervision.md>), [transformer](<https://devfeed.tech/tags/transformer.md>)

### AI overview

The proposed surrogate combines transformer experts for steady PDE solutions on coarse unstructured meshes. A controller uses learnable Laplacian matrices to weight expert outputs at each mesh point, and a decoder predicts the solution fields. During training, repeated smoothing with a mesh Laplacian decomposes reference fields into spatial scales for multiscale supervision. On remeshed AirFRANS data with triangular meshes of about 4.9k vertices, the authors report lower mean relative L2 error and interior mean squared error than the models compared in their study.

### Source excerpt

arXiv:2610.20585v4 Announce Type: replace Abstract: Surrogate models approximate the solutions of partial differential equations given a certain geometry and boundary conditions. Traditional transformer-based approaches employ a single mixing mechanism to the whole solution. although the solution fields exhibit variations across multiple spatial scales. We propose a mixture of transformer experts that combines a routing mechanism with multiscale supervision. Specifically, the individual experts are routed through a controller that utilizes learnable Laplacian matrices to compute weights that are subsequently used to define a weighted sum of the experts for the output feature vectors at each point of the domain. Subsequently, a decoder receives the weighted sum of the experts output and predicts the solution fields. During training, the reference fields are decomposed into components at different spatial scales by repeated smoothing with a mesh Laplacian. We evaluate our approach on the official AirFRANS dataset using remeshed solution domains leading to unstructured triangular meshes of $4.9$k vertices. The experimental outcomes reveal lower mean relative $L_2$ error and interior mean squared errors among the compared models of this study.

## Asymptotic-Preserving Tensor Neural Networks for 3D High-Field Kinetic Equations with Applications to Uncertainty Quantification

DevFeed: [Asymptotic-Preserving Tensor Neural Networks for 3D High-Field Kinetic Equations with Applications to Uncertainty Quantification](<https://devfeed.tech/articles/asymptotic-preserving-tensor-neural-networks-for-3d-high-field-kinetic-equations-with-applications-to-uncertainty-quantification-120605.md>)

Original publisher: [Read original article](<https://arxiv.org/abs/2611.07364>)

Author: Shi Jin, Tian-ai Zhang

Published: 2026-11-06T05:00:00Z

Content type: article

Language: en

Sources: [arXiv: Machine Learning](<https://devfeed.tech/sources/arxiv-machine-learning.md>)

Topics: [Neural Network](<https://devfeed.tech/topics/neural-network.md>), [Finite Element Method (FEM)](<https://devfeed.tech/topics/finite-element-method.md>), [Artificial Intelligence](<https://devfeed.tech/topics/ai.md>)

Tags: [3d](<https://devfeed.tech/tags/3d.md>), [abstract](<https://devfeed.tech/tags/abstract.md>), [acceleration](<https://devfeed.tech/tags/acceleration.md>), [added](<https://devfeed.tech/tags/added.md>), [applications](<https://devfeed.tech/tags/applications.md>), [approximation](<https://devfeed.tech/tags/approximation.md>), [arxiv](<https://devfeed.tech/tags/arxiv.md>), [can](<https://devfeed.tech/tags/can.md>), [canonical](<https://devfeed.tech/tags/canonical.md>), [challenges](<https://devfeed.tech/tags/challenges.md>), [closure](<https://devfeed.tech/tags/closure.md>), [costs](<https://devfeed.tech/tags/costs.md>), [cs-lg](<https://devfeed.tech/tags/cs-lg.md>), [cs-na](<https://devfeed.tech/tags/cs-na.md>), [devices](<https://devfeed.tech/tags/devices.md>), [field](<https://devfeed.tech/tags/field.md>), [math-na](<https://devfeed.tech/tags/math-na.md>), [networks](<https://devfeed.tech/tags/networks.md>), [neural](<https://devfeed.tech/tags/neural.md>), [neural-networks](<https://devfeed.tech/tags/neural-networks.md>), [uncertainty](<https://devfeed.tech/tags/uncertainty.md>), [uncertainty-quantification](<https://devfeed.tech/tags/uncertainty-quantification.md>), [with](<https://devfeed.tech/tags/with.md>)

### AI overview

APTNN addresses the high dimensionality and multiscale behavior of high-field kinetic equations using a canonical polyadic tensor network to approximate solutions. Because equation residuals retain this decomposition, high-dimensional loss integrals can be evaluated as combinations of one-dimensional integrals, reducing computational and storage growth from exponential to polynomial with dimension. For improved accuracy in the high-field limit, the method adds a first-moment equation term to the asymptotic-preserving loss, constraining the limiting flux closure while retaining dimension-wise evaluation. Reported numerical results show improved high-field prediction accuracy with lower computational cost and memory use; training a 3D3V VPFP problem with multiple random inputs took under 30 minutes on a single GPU.

### Source excerpt

arXiv:2611.07364v1 Announce Type: cross Abstract: Kinetic equations in the high-field regime describe particle transport dominated by electric-field acceleration and collisional relaxation. These equations are widely used in simulations of plasmas and semiconductor devices. Typical models include the Vlasov--Poisson--Fokker--Planck (VPFP) and Boltzmann--Poisson systems. Their high dimensionality and multiscale nature pose two of the major challenges to accurate and efficient numerical solution. To address these challenges, we propose an asymptotic-preserving tensor neural network (APTNN) method. Specifically, to reduce high-dimensional computational costs and memory requirements, we adopt the canonical polyadic (CP) decomposition structure of TNNs to approximate their solutions. We show that the equation residuals retain this decomposition form, allowing high-dimensional loss integrals to be evaluated through combinations of one-dimensional integrals. The resulting computational and storage costs are reduced from exponential to polynomial growth with dimension. To improve the prediction accuracy in the high-field limit, we enhance the baseline AP loss by incorporating the first moment equation. The added loss term directly constrains the limiting flux closure to reduce its approximation error in the high-field regime. The enhanced AP loss can still be evaluated dimension-wise. Numerical results show that APTNN improves prediction accuracy in the high-field regime with significantly reduced computational costs and memory usage. For a 3D3V VPFP problem with multiple random inputs, training takes less than 30 minutes on a single GPU.

## Physics-Informed Functional Tucker Method with RKHS Factors for Sparse Spatiotemporal Reconstruction

DevFeed: [Physics-Informed Functional Tucker Method with RKHS Factors for Sparse Spatiotemporal Reconstruction](<https://devfeed.tech/articles/physics-informed-functional-tucker-method-with-rkhs-factors-for-sparse-spatiotemporal-reconstruction-117361.md>)

Original publisher: [Read original article](<https://arxiv.org/abs/2611.03717>)

Author: Zhengyang Liu, Jianguo Huang, Junxi Hu, Yucheng Mi, Haojie Cai, Wenxuan Shen, Zhikun Zhang, Renfeng Peng, Guangtao Zhang, Zhiqiang Liu

Published: 2026-11-04T05:00:00Z

Content type: article

Language: en

Sources: [arXiv: Artificial Intelligence](<https://devfeed.tech/sources/arxiv-artificial-intelligence.md>)

Topics: [Finite Element Method (FEM)](<https://devfeed.tech/topics/finite-element-method.md>), [Image processing](<https://devfeed.tech/topics/image-processing.md>), [VAE](<https://devfeed.tech/topics/vae.md>)

Tags: [3d-reconstruction](<https://devfeed.tech/tags/3d-reconstruction.md>), [continuous](<https://devfeed.tech/tags/continuous.md>), [cs-ai](<https://devfeed.tech/tags/cs-ai.md>), [cs-cv](<https://devfeed.tech/tags/cs-cv.md>), [experiments](<https://devfeed.tech/tags/experiments.md>), [kernel](<https://devfeed.tech/tags/kernel.md>), [physics-informed](<https://devfeed.tech/tags/physics-informed.md>), [sensors](<https://devfeed.tech/tags/sensors.md>), [spatial](<https://devfeed.tech/tags/spatial.md>)

### AI overview

PhysFTM addresses reconstruction from sparse sensors and missing frames by combining a Tucker representation with reproducing kernel Hilbert space (RKHS) mode factors and PDE constraints. It adds finite-difference residuals at off-sensor collocation points to provide physical constraints where observations are absent. For observed support frames, it alternately optimizes the Tucker core and RKHS factors; it then interpolates a continuous trajectory in core space and decodes it with the learned spatial representation to reconstruct missing frames. The abstract reports accurate 2D and 3D reconstruction across Allen-Cahn, Navier-Stokes, 3D Heat equations, and gridded Copernicus ocean reanalysis under two missing-frame protocols.

### Source excerpt

arXiv:2611.03717v1 Announce Type: cross Abstract: Reconstructing physical dynamics from fixed sparse sensors and missing frames is a fundamental challenge in scientific research. Sparsity leaves large regions of the field unconstrained, making the inverse problem ill-posed: a model may fit the observed sensors while producing inaccurate off-sensor structures or unreliable predictions between support frames. To this end, we propose PhysFTM---a physics-informed functional Tucker model that parameterizes the Tucker mode factors in reproducing kernel Hilbert spaces (RKHSs) and incorporates PDE constraints through finite-difference residuals evaluated at off-sensor collocation points. These residuals provide additional physical information in regions without direct observational supervision. PhysFTM first reconstructs continuous spatial fields at support frames via an alternating minimization scheme, optimizing the Tucker core tensor and the RKHS-based factors separately. To enable missing-frame reconstruction, PhysFTM constructs a continuous trajectory in Tucker-core space, which is decoded by the learned spatial RKHS-FTM representation to recover the full spatiotemporal field. Experiments on Allen--Cahn, Navier--Stokes, three-dimensional Heat equations, and a three-dimensional gridded Copernicus ocean reanalysis demonstrate accurate 2D/3D reconstruction under two missing-frame protocols. Code is available at https://github.com/liusirzzz/PhysFTM.

## CloakFEM: Agentic FEM Code Generation for Red-Teaming Computational Resource Exhaustion

DevFeed: [CloakFEM: Agentic FEM Code Generation for Red-Teaming Computational Resource Exhaustion](<https://devfeed.tech/articles/cloakfem-agentic-fem-code-generation-for-red-teaming-computational-resource-exhaustion-116148.md>)

Original publisher: [Read original article](<https://arxiv.org/abs/2611.04356>)

Author: Li Zheng, Simin Chen

Published: 2026-11-04T05:00:00Z

Content type: article

Language: en

Sources: [arXiv: Artificial Intelligence](<https://devfeed.tech/sources/arxiv-artificial-intelligence.md>), [arXiv: Cryptography and Security](<https://devfeed.tech/sources/arxiv-cryptography-and-security.md>), [arXiv: Software Engineering](<https://devfeed.tech/sources/arxiv-software-engineering.md>)

Topics: [Finite Element Method (FEM)](<https://devfeed.tech/topics/finite-element-method.md>), [computational complexity](<https://devfeed.tech/topics/computational-complexity.md>), [Adversarial attacks](<https://devfeed.tech/topics/adversarial-attacks.md>), [Large Language Model](<https://devfeed.tech/topics/llm.md>)

Tags: [17](<https://devfeed.tech/tags/17.md>), [adversarial](<https://devfeed.tech/tags/adversarial.md>), [agentic](<https://devfeed.tech/tags/agentic.md>), [amplify](<https://devfeed.tech/tags/amplify.md>), [attack-surface](<https://devfeed.tech/tags/attack-surface.md>), [code-generation](<https://devfeed.tech/tags/code-generation.md>), [cs-ai](<https://devfeed.tech/tags/cs-ai.md>), [cs-cr](<https://devfeed.tech/tags/cs-cr.md>), [cs-se](<https://devfeed.tech/tags/cs-se.md>), [denial-of-service](<https://devfeed.tech/tags/denial-of-service.md>), [efficiency](<https://devfeed.tech/tags/efficiency.md>), [framework](<https://devfeed.tech/tags/framework.md>), [llm](<https://devfeed.tech/tags/llm.md>), [red-teaming](<https://devfeed.tech/tags/red-teaming.md>), [resource](<https://devfeed.tech/tags/resource.md>), [search](<https://devfeed.tech/tags/search.md>), [simulation](<https://devfeed.tech/tags/simulation.md>)

### AI overview

CloakFEM uses LLM-guided search, an FEM-specific intermediate representation, and a deterministic validity gate to generate runnable FEM programs that increase solver workload. Across three FEM stacks, 17 problem classes, and 1,068 seeds, it found valid, solvable inputs that amplified computation; the largest reported increase was from 43 to 247,367 solver iterations (5,752.72x). The abstract reports that CloakFEM outperformed five generic black-box baselines and retained an advantage over evolutionary search under a controlled bounded action space. Cross-solver analysis identified both problem-intrinsic and solver-specific vulnerability sources. These results indicate a potential computational denial-of-service attack surface; the supplied evidence is an abstract, not a full paper.

### Source excerpt

arXiv:2611.04356v1 Announce Type: new Abstract: Finite element method (FEM) solvers are foundational to scientific and engineering simulation, yet their susceptibility to \emph{efficiency attacks} remains largely unexplored. An adversary can craft syntactically valid and runnable FEM programs that complete successfully while forcing excessive computational work, enabling computational denial-of-service (DoS) attacks. Finding such inputs is challenging because FEM programs occupy a vast, structured space constrained by numerical, geometric, and constitutive requirements. We introduce CloakFEM, a domain-guided LLM-agent framework for automatically generating runnable adversarial FEM programs that amplify solver computation. CloakFEM combines LLM-guided search with an FEM-specific intermediate representation and a deterministic validity gate that enforces runnable constraints. Across three FEM stacks, 17 problem classes, and 1,068 seeds, CloakFEM discovers adversarial inputs that substantially amplify solver work while remaining valid and solvable. In the most extreme case, solver iterations increase from 43 to 247,367, a \(5{,}752.72\times\) amplification. CloakFEM outperforms five generic black-box baselines and retains an advantage over evolutionary search under a controlled bounded action space. Cross-solver analysis further reveals both problem-intrinsic and solver-specific sources of vulnerability. Our results uncover an underexplored efficiency-attack surface in FEM solvers and show that runnable FEM programs can serve as adversarial inputs.

## Embedding a de Rham Complex Structure into Eig-PIELM for Maxwell Eigenvalue Problems

DevFeed: [Embedding a de Rham Complex Structure into Eig-PIELM for Maxwell Eigenvalue Problems](<https://devfeed.tech/articles/embedding-a-de-rham-complex-structure-into-eig-pielm-for-maxwell-eigenvalue-problems-114686.md>)

Original publisher: [Read original article](<https://arxiv.org/abs/2611.02337>)

Author: Hao Kang, Liang Li, Xing Li

Published: 2026-11-03T05:00:00Z

Content type: article

Language: en

Sources: [arXiv: Machine Learning](<https://devfeed.tech/sources/arxiv-machine-learning.md>)

Topics: [Finite Element Method (FEM)](<https://devfeed.tech/topics/finite-element-method.md>)

Tags: [arxiv](<https://devfeed.tech/tags/arxiv.md>), [assembly](<https://devfeed.tech/tags/assembly.md>), [boundary](<https://devfeed.tech/tags/boundary.md>), [construction](<https://devfeed.tech/tags/construction.md>), [convergence](<https://devfeed.tech/tags/convergence.md>), [cs-lg](<https://devfeed.tech/tags/cs-lg.md>), [cs-na](<https://devfeed.tech/tags/cs-na.md>), [curl](<https://devfeed.tech/tags/curl.md>), [design](<https://devfeed.tech/tags/design.md>), [develop](<https://devfeed.tech/tags/develop.md>), [dimension](<https://devfeed.tech/tags/dimension.md>), [dr](<https://devfeed.tech/tags/dr.md>), [electric](<https://devfeed.tech/tags/electric.md>), [example](<https://devfeed.tech/tags/example.md>), [experiments](<https://devfeed.tech/tags/experiments.md>), [extension](<https://devfeed.tech/tags/extension.md>), [field](<https://devfeed.tech/tags/field.md>), [fixed](<https://devfeed.tech/tags/fixed.md>), [form](<https://devfeed.tech/tags/form.md>), [gradient](<https://devfeed.tech/tags/gradient.md>), [math-na](<https://devfeed.tech/tags/math-na.md>), [maxwell](<https://devfeed.tech/tags/maxwell.md>), [structure](<https://devfeed.tech/tags/structure.md>)

### AI overview

dR-Eig-PIELM addresses gradient-contaminated modes that can remain in Eig-PIELM because PEC boundary conditions alone do not remove the Maxwell curl-curl operator's gradient nullspace. It combines compatible tensor-product grad-curl trial spaces with a constrained null-space construction that restricts the sampled PEC space to the mass-orthogonal complement of the represented gradient image before solving the eigenproblem. Symmetric weak-form Gram assembly produces a symmetric-definite reduced problem without an eigenvalue cutoff, under the stated sampling and material-positivity assumptions. For rectangles and cuboids, the authors establish fixed-degree exactness at the PEC vector-field space and derive rank and dimension identities. In reported numerical experiments, the method excludes spurious modes with significant gradient components seen in the boundary-only baseline, converges rapidly with polynomial degree, and has competitive work-precision performance in tested configurations. A mapped-waveguide example demonstrates a constrained local geometry update using computed cutoff frequencies and sensitivities.

### Source excerpt

arXiv:2611.02337v1 Announce Type: cross Abstract: The perfect electric conductor (PEC) boundary condition alone does not eliminate the gradient nullspace of the Maxwell curl-curl operator. Consequently, a boundary-only Maxwell implementation of Eig-PIELM, an eigenvalue extension of the physics-informed extreme learning machine (PIELM), may retain represented gradient modes in its reduced trial space, producing nonphysical, gradient-contaminated modes. We develop dR-Eig-PIELM, a finite-dimensional structure-preserving extension that retains the linear-coefficient, boundary-null-space, and one-shot multimode architecture of Eig-PIELM. It combines compatible tensor-product grad--curl trial spaces with a constrained null-space construction to restrict the sampled PEC space to the mass-orthogonal complement of the represented gradient image before the eigensolve. Symmetric weak-form Gram assembly yields a symmetric-definite reduced eigenproblem without an eigenvalue cutoff under the stated sampling and material-positivity assumptions. For rectangles and cuboids, we establish fixed-degree exactness at the PEC vector-field space and derive rank and dimension identities. Numerical experiments show that dR-Eig-PIELM excludes the spurious eigenmodes with significant gradient components observed in the boundary-only baseline. The method exhibits rapid convergence with increasing polynomial degree and competitive work--precision performance in the tested configurations. A covariantly mapped waveguide example illustrates the potential of dR-Eig-PIELM for integration into design optimization workflows through a constrained local geometry update driven by computed cutoff frequencies and their sensitivities. This work provides a mathematical and computational framework for separating represented gradient modes from physical Maxwell eigenmodes within Eig-PIELM and supports its use in waveguide spectral analysis and geometry design.

## From 1D Benchmarks to Complex 3D Geometries: Can Agents Reliably Specify Problems for Automated PDE Simulation?

DevFeed: [From 1D Benchmarks to Complex 3D Geometries: Can Agents Reliably Specify Problems for Automated PDE Simulation?](<https://devfeed.tech/articles/from-1d-benchmarks-to-complex-3d-geometries-can-agents-reliably-specify-problems-for-automated-pde-simulation-112049.md>)

Original publisher: [Read original article](<https://arxiv.org/abs/2610.41267>)

Author: Akshat Jani, Sakhinana Sagar Srinivas, Venkataramana Runkana

Published: 2026-11-02T05:00:00Z

Content type: news

Language: en

Sources: [arXiv: Artificial Intelligence](<https://devfeed.tech/sources/arxiv-artificial-intelligence.md>)

Topics: [Artificial Intelligence](<https://devfeed.tech/topics/ai.md>), [Simulation](<https://devfeed.tech/topics/simulation.md>), [agentic workflows](<https://devfeed.tech/topics/agentic-workflows.md>), [Finite Element Method (FEM)](<https://devfeed.tech/topics/finite-element-method.md>), [3D](<https://devfeed.tech/topics/3d.md>)

Tags: [3d](<https://devfeed.tech/tags/3d.md>), [abstract](<https://devfeed.tech/tags/abstract.md>), [agent](<https://devfeed.tech/tags/agent.md>), [agents](<https://devfeed.tech/tags/agents.md>), [arxiv](<https://devfeed.tech/tags/arxiv.md>), [automate](<https://devfeed.tech/tags/automate.md>), [automated](<https://devfeed.tech/tags/automated.md>), [benchmarks](<https://devfeed.tech/tags/benchmarks.md>), [boundary](<https://devfeed.tech/tags/boundary.md>), [but](<https://devfeed.tech/tags/but.md>), [can](<https://devfeed.tech/tags/can.md>), [combines](<https://devfeed.tech/tags/combines.md>), [complete](<https://devfeed.tech/tags/complete.md>), [completion](<https://devfeed.tech/tags/completion.md>), [complex](<https://devfeed.tech/tags/complex.md>), [component](<https://devfeed.tech/tags/component.md>), [conditional](<https://devfeed.tech/tags/conditional.md>), [consistency](<https://devfeed.tech/tags/consistency.md>), [cs-ai](<https://devfeed.tech/tags/cs-ai.md>), [cs-ce](<https://devfeed.tech/tags/cs-ce.md>), [problems](<https://devfeed.tech/tags/problems.md>), [simulation](<https://devfeed.tech/tags/simulation.md>)

### AI overview

The workflow combines model extraction, deterministic scientific and capability checks, explicit clarification, and validated local patches to build executable JSON specifications for automated PDE simulations. On 120 template-related requests, it achieved 98.3% correct first-turn specification and decision outcomes, versus 92.5% for schema-assisted single shot and 82.5% for model-only, with no false-ready decisions among 42 requests needing clarification or rejection. The study separately evaluates specification correctness, execution, and numerical verification; downstream cases include 1D-3D problems and complex wing geometries. Nine of 12 selected live user-to-results tests succeeded jointly across specification, execution, export, and numerical reference. The findings support engineer-reviewed assistance within the tested scope, while fresh-response regressions point to consistency as an area for improvement.

### Source excerpt

arXiv:2610.41267v1 Announce Type: new Abstract: Turning an engineering request into a simulation requires consistent geometry, physics, boundary conditions, and numerical settings. Recent agent-assisted systems automate simulation setup, but successful execution alone does not establish whether the specification preserves the intended problem or correctly handles missing information. We study whether agents can reliably construct complete executable JSON specifications for automated geometry creation, meshing, PDE solving, and result export. Our workflow combines model extraction with deterministic scientific and capability checks, explicit clarification, and validated local patches that preserve unrelated accepted fields. We evaluate specification correctness, execution completion, and numerical verification separately. On 120 template-related requests, the full workflow achieves 98.3% correct first-turn specification and decision outcomes, compared with 92.5% for schema-assisted single shot and 82.5% for model-only: gains of 5.8 and 15.8 percentage points, respectively. It makes no false-ready decisions among the 42 requests requiring clarification or rejection. Component ablations identify contributions from scientific checks and conditional repair. Separate downstream studies cover 1D, 2D, and 3D steady and unsteady, linear and nonlinear problems, including flow and elasticity with complex three-dimensional wing geometries. A 30-case gallery demonstrates automated execution from retained JSON specifications, while 9 of 12 selected live user-to-results tests achieve joint specification, execution, export, and numerical-reference success (75.0%). These results support traceable, engineer-reviewed simulation assistance within the tested scope, while fresh-response regressions identify consistency as a priority for further improvement.

## Meshless Latent Dynamics Models for Model Order Reduction of time-dependent Partial Differential Equations

DevFeed: [Meshless Latent Dynamics Models for Model Order Reduction of time-dependent Partial Differential Equations](<https://devfeed.tech/articles/meshless-latent-dynamics-models-for-model-order-reduction-of-time-dependent-partial-differential-equations-111537.md>)

Original publisher: [Read original article](<https://arxiv.org/abs/2610.41043>)

Author: Lorenzo Tomada, Federico Pichi, Gianluigi Rozza

Published: 2026-11-02T05:00:00Z

Content type: article

Language: en

Sources: [arXiv: Machine Learning](<https://devfeed.tech/sources/arxiv-machine-learning.md>)

Topics: [Finite Element Method (FEM)](<https://devfeed.tech/topics/finite-element-method.md>), [World models](<https://devfeed.tech/topics/world-models.md>)

Tags: [analysis](<https://devfeed.tech/tags/analysis.md>), [architectures](<https://devfeed.tech/tags/architectures.md>), [cs-lg](<https://devfeed.tech/tags/cs-lg.md>), [cs-na](<https://devfeed.tech/tags/cs-na.md>), [error](<https://devfeed.tech/tags/error.md>), [experiments](<https://devfeed.tech/tags/experiments.md>), [grid](<https://devfeed.tech/tags/grid.md>), [learning](<https://devfeed.tech/tags/learning.md>), [math-na](<https://devfeed.tech/tags/math-na.md>), [mechanism](<https://devfeed.tech/tags/mechanism.md>), [mesh](<https://devfeed.tech/tags/mesh.md>), [models](<https://devfeed.tech/tags/models.md>), [neural](<https://devfeed.tech/tags/neural.md>), [numerical-analysis](<https://devfeed.tech/tags/numerical-analysis.md>), [partial-differential-equations](<https://devfeed.tech/tags/partial-differential-equations.md>), [reduction](<https://devfeed.tech/tags/reduction.md>), [uat](<https://devfeed.tech/tags/uat.md>)

### AI overview

The paper presents meshless latent dynamics models for reducing time-dependent PDEs, using neural fields conditioned on a latent state to represent spatially continuous evolution without tying training and inference to a fixed grid. It classifies conditioning as weak, strong, or pure, proves a universal approximation result, and derives Lipschitz stability bounds. The analysis identifies an expressivity-stability trade-off: more expressive conditioning may impair temporal extrapolation. In experiments on Burgers' equation, flow past a cylinder, and linear advection, meshless models reportedly match or exceed mesh-based predictive accuracy with 2% of their trainable parameters and nearly constant error across spatial discretizations. The abstract also reports grid-independent Lipschitz bounds for meshless models, unlike mesh-based counterparts.

### Source excerpt

arXiv:2610.41043v1 Announce Type: cross Abstract: Latent dynamics learning has emerged as a prominent approach for Model Order Reduction (MOR) of time-dependent, parameterized Partial Differential Equations (PDEs). Traditional autoencoder-based approaches are intrinsically tied to fixed spatial discretizations during both training and inference, motivating the development of meshless methodologies. In this context, conditioning provides a natural mechanism for representing temporal evolution through spatially continuous functions, enabling mesh-agnostic Latent Dynamics Models (LDMs). In this work, we introduce a unified mathematical framework for mesh-agnostic LDMs based on conditioning in Neural Fields (NeFs). Leveraging a taxonomy of conditioning mechanisms (weak, strong, and pure), we formulate meshless LDMs as NeFs conditioned on a latent state. We establish a Universal Approximation Theorem (UAT) for this class of models and analyze their stability by deriving Lipschitz bounds for different conditioning mechanisms. Our analysis reveals an expressivity--stability trade-off, providing insight into why more expressive conditioning, despite its higher representational capacity, can potentially degrade performance during temporal extrapolation. Finally, we provide a comprehensive comparison of meshless and mesh-based architectures in the continuous limit of the spatial discretization, showing that meshless models guarantee grid-independent Lipschitz bounds, unlike mesh-based counterparts. Our experiments across Burgers' equation, Navier--Stokes flow past a cylinder, and linear advection show that meshless LDMs can achieve predictive accuracy comparable to or better than their mesh-based counterparts while using only 2% of the trainable parameters and exhibiting nearly constant error across spatial discretizations.

## Threshold Compatibility and Support-Selection Diagnostics in Neural Discovery of Partial Differential Equations

DevFeed: [Threshold Compatibility and Support-Selection Diagnostics in Neural Discovery of Partial Differential Equations](<https://devfeed.tech/articles/threshold-compatibility-and-support-selection-diagnostics-in-neural-discovery-of-partial-differential-equations-108580.md>)

Original publisher: [Read original article](<https://arxiv.org/abs/2610.37726>)

Author: Zeen Song

Published: 2026-10-30T04:00:00Z

Content type: article

Language: en

Sources: [arXiv: Machine Learning](<https://devfeed.tech/sources/arxiv-machine-learning.md>)

Topics: [Neural Network](<https://devfeed.tech/topics/neural-network.md>), [Finite Element Method (FEM)](<https://devfeed.tech/topics/finite-element-method.md>)

Tags: [compatibility](<https://devfeed.tech/tags/compatibility.md>), [cs-lg](<https://devfeed.tech/tags/cs-lg.md>), [diagnostics](<https://devfeed.tech/tags/diagnostics.md>), [discovery](<https://devfeed.tech/tags/discovery.md>), [neural](<https://devfeed.tech/tags/neural.md>), [partial-differential-equations](<https://devfeed.tech/tags/partial-differential-equations.md>), [selection](<https://devfeed.tech/tags/selection.md>), [support](<https://devfeed.tech/tags/support.md>)

### AI overview

The study finds that accurate coefficient regression can still produce incorrect PDE support after thresholding. For a median/sample-standard-deviation threshold, it derives a two-signal retention boundary and a complete-vector perturbation bound; exact advection-diffusion libraries nevertheless lose diffusion. Refitting the true support reduces diffusion error in 40 saved states but does not restore all derivative accuracy. A sequential threshold-ridge comparator recovers exact support in 60 ideal controls but adds terms in every learned state. Validation rejects all-fold correct-support candidates, and post hoc energy pooling retains their higher minimum loss in all 60 states. Supplied-support estimates beat unchanged selected equations in 240 physical-residual comparisons across four unused analytic trajectories. These finite diagnostics separate threshold incompatibility, refit damage, and score rejection, but do not establish automatic correction, general recovery, or predictive stability.

### Source excerpt

arXiv:2610.37726v1 Announce Type: new Abstract: Accurate regression need not yield the correct equation after support selection. For a median/sample-standard-deviation threshold in neural PDE discovery, we derive an exact two-signal retention boundary and a complete-vector perturbation bound. Exact advection-diffusion libraries lose diffusion despite coefficient recovery to numerical precision. True-support refits of forty saved sparse states reduce diffusion error without restoring all derivative accuracy. A fixed sequential threshold ridge comparator recovers exact support in sixty ideal controls but selects extra terms in every learned state. All-fold correct-support candidates are rejected by conditional validation; post hoc energy pooling preserves their higher minimum loss in all sixty states. Supplied-support estimates outperform unchanged selected equations in all 240 physical-residual comparisons on four unused analytic trajectories. With coefficients, observation points and weights fixed, exact replacement of the learned library and target reverses their learned-row ordering in 59/60 states. Weighting both losses also fails the specified Burgers diffusion-benefit criterion at every tested budget. These finite diagnostics distinguish threshold incompatibility, refit damage and score rejection; they establish neither automatic correction nor general recovery or predictive stability.

## NAGO: Noise-Aware Generative Operator via Flow Matching

DevFeed: [NAGO: Noise-Aware Generative Operator via Flow Matching](<https://devfeed.tech/articles/nago-noise-aware-generative-operator-via-flow-matching-108676.md>)

Original publisher: [Read original article](<https://arxiv.org/abs/2610.38273>)

Author: Guanyu Chen, Shengze Xu, Pengwei Liu, Xingyu Ren, Pengkai Wang, Antonios Armaou, Tieyong Zeng, Dong Ni

Published: 2026-10-30T04:00:00Z

Content type: article

Language: en

Sources: [arXiv: Machine Learning](<https://devfeed.tech/sources/arxiv-machine-learning.md>)

Topics: [Machine learning](<https://devfeed.tech/topics/machine-learning.md>), [Finite Element Method (FEM)](<https://devfeed.tech/topics/finite-element-method.md>)

Tags: [analysis](<https://devfeed.tech/tags/analysis.md>), [covariance](<https://devfeed.tech/tags/covariance.md>), [cs-lg](<https://devfeed.tech/tags/cs-lg.md>), [design](<https://devfeed.tech/tags/design.md>), [evaluations](<https://devfeed.tech/tags/evaluations.md>), [field](<https://devfeed.tech/tags/field.md>), [fields](<https://devfeed.tech/tags/fields.md>), [flow](<https://devfeed.tech/tags/flow.md>), [flow-matching](<https://devfeed.tech/tags/flow-matching.md>), [functional](<https://devfeed.tech/tags/functional.md>), [gaussian](<https://devfeed.tech/tags/gaussian.md>), [generative](<https://devfeed.tech/tags/generative.md>), [guidance](<https://devfeed.tech/tags/guidance.md>), [infinite](<https://devfeed.tech/tags/infinite.md>), [invariant](<https://devfeed.tech/tags/invariant.md>), [noise](<https://devfeed.tech/tags/noise.md>), [operator](<https://devfeed.tech/tags/operator.md>), [partial-differential-equations](<https://devfeed.tech/tags/partial-differential-equations.md>), [sampling](<https://devfeed.tech/tags/sampling.md>), [source](<https://devfeed.tech/tags/source.md>), [space](<https://devfeed.tech/tags/space.md>)

### AI overview

NAGO treats the sampling source measure as a design choice for flow-based generation of SPDE invariant measures. Its analysis links finite-band source-to-target covariance mismatch to spectral stiffness, leading to a rule: choose a source rough enough for well-posed infinite-dimensional flows, then minimize covariance mismatch on resolved Fourier modes. Across three stochastic-field benchmarks, this design improved few-step spectral accuracy and other distributional and physical statistics. In a controlled Matérn-field ablation, it reduced the accurate RK4 budget from about 30 steps for Gaussian functional flow matching to 3-5 steps, or 6-10x fewer neural ODE evaluations. Reported gains were strongest for stationary multiscale fields and depend on metric, resolution, and guidance protocol.

### Source excerpt

arXiv:2610.38273v1 Announce Type: new Abstract: We introduce NAGO, a noise-aware generative operator for efficient sampling from invariant measures of stochastic partial differential equations (SPDEs) in function space. Our central idea is to treat the source measure as a design variable rather than a fixed Gaussian default. We establish admissibility conditions for well-defined infinite-dimensional flows and show that the finite-band source--target covariance ratio controls the spectral stiffness of the marginal velocity field. This analysis yields an admissibility-constrained spectral-matching rule: choose a source that is sufficiently rough for well-posedness, then minimize covariance mismatch over the resolved Fourier modes. On three representative stochastic-field benchmarks, this source design improves few-step spectral accuracy and several complementary distributional and physical statistics. In a controlled Matern-field ablation, it reduces the accurate RK4 budget from about 30 steps for Gaussian functional flow matching to 3--5 steps, a $6$--$10\times$ reduction in neural ODE evaluations. These gains are strongest for stationary multiscale fields and depend on the metric, resolution, and guidance protocol.

[Next page](<https://devfeed.tech/topics/finite-element-method.md?cursor=WyIyMDI2LTEwLTMwVDA0OjAwOjAwKzAwOjAwIiwgImM0MTU2ZmJmLTA2YWMtNDU5YS1iZTA2LTdjMzVkZTgxNGVhMCJd>)