https://surrogatemodels.com/ News, papers, datasets, and funding and deals in Physics AI for engineers: General PDEs. en https://surrogatemodels.com/papers/data-driven-discretizationhttps://surrogatemodels.com/papers/data-driven-discretizationThu, 12 Nov 2026 09:00:00 GMTA neural network reads the coarse solution around each cell boundary and outputs the finite-difference coefficients used to estimate the spatial derivatives there. The estimates feed a finite-volume flux, which keeps the scheme conservative, and training is end-to-end against coarse-grained high-resolution solutions of the known equation.General PDEs https://surrogatemodels.com/papers/ledhttps://surrogatemodels.com/papers/ledThu, 12 Nov 2026 09:00:00 GMTAn autoencoder compresses the full simulation state to a few latent variables and an LSTM advances them in time (Latent-LED). Following the equation-free approach, the scheme can also alternate between the latent model and short runs of the original solver started from the decoded state, which corrects the accumulated error (Multiscale-LED).Aero & CFDGeneral PDEs https://surrogatemodels.com/papers/adaledhttps://surrogatemodels.com/papers/adaledThu, 12 Nov 2026 09:00:00 GMTCouples a running solver with a surrogate trained online during the same run: a convolutional autoencoder and an ensemble of probabilistic LSTMs. The simulation switches to the surrogate when its error against the solver and the ensemble uncertainty are below set thresholds, and returns to the solver when the uncertainty rises.Aero & CFDGeneral PDEs https://surrogatemodels.com/papers/g-ledhttps://surrogatemodels.com/papers/g-ledThu, 12 Nov 2026 09:00:00 GMTThe latent state is the simulation snapshot subsampled to a coarse grid, with no trained encoder. An autoregressive attention model advances it in time, and a diffusion model conditioned on the latent state generates the full-resolution fields.Aero & CFDGeneral PDEs https://surrogatemodels.com/papers/function-space-certificationhttps://surrogatemodels.com/papers/function-space-certificationMon, 09 Nov 2026 09:00:00 GMTInterval bound propagation and CROWN are formulated in function space: upper and lower bound functions are propagated through each neural-operator layer with rigorous numerical integration, and extended to higher-order derivatives to bound PDE residuals and solution error.Weather & climateGeneral PDEs https://surrogatemodels.com/papers/universal-latent-operatorhttps://surrogatemodels.com/papers/universal-latent-operatorSun, 08 Nov 2026 09:00:00 GMTA functional autoencoder is pre-trained on analytic functions generated on the fly, without a solver or labels, and its latent space is frozen; neural operators are then trained on cached embeddings.General PDEs https://surrogatemodels.com/news/beyondmath-physics-factoryhttps://surrogatemodels.com/news/beyondmath-physics-factoryWed, 21 Oct 2026 09:00:00 GMTBeyondMath says its Physics Factory platform is now open to everyone. It has three tools: Cadgen2, where an AI engineer builds a described part, checks it against requirements and returns STEP and BREP; Forge, where a team of agents trains models on a user's simulation data, tests them on held-out cases and reports back; and Studio, which runs a released model on unseen geometry and returns fields and force coefficients in seconds. Users sign up and buy credits.General PDEs https://surrogatemodels.com/datasets/neuraldem-datahttps://surrogatemodels.com/datasets/neuraldem-dataWed, 12 Aug 2026 09:00:00 GMTSimulation data for training and evaluating surrogates of particulate flows: hopper discharge of about 250,000 particles across hopper angles and friction regimes, and a fluidized-bed reactor of about 500,000 particles with air injection at varying inlet velocities. Licence: CC-BY-NC 4.0.General PDEs https://surrogatemodels.com/news/mistral-industrial-engineeringhttps://surrogatemodels.com/news/mistral-industrial-engineeringThu, 28 May 2026 09:00:00 GMTAt its AI Now Summit in Paris on 28 May 2026, Mistral launched Mistral for Industrial Engineering, which combines advanced physics models, engineering expertise and robotics. Mistral names Airbus, BMW Group (for its Large Industry Model initiative) and ASML as partners. No performance figures are given in the announcement.General PDEs https://surrogatemodels.com/news/emmi-noether-frameworkhttps://surrogatemodels.com/news/emmi-noether-frameworkTue, 27 Jan 2026 09:00:00 GMTEmmi AI releases Noether, a deep learning framework for engineering simulation models built on transformer components. It covers the workflow from data through training and evaluation to application, from CFD meshes to volumetric data, with ready-to-use model architectures. The code is on GitHub under the Emmi Open Research License, which is free for research and piloting and requires a licence for commercial use. Emmi lists groups at JKU Linz, ETH Zürich, TUM and other universities as supporters, and Luminary, Rescale and SimScale as platform partners.General PDEs https://surrogatemodels.com/datasets/pdebenchhttps://surrogatemodels.com/datasets/pdebenchSat, 01 Oct 2022 09:00:00 GMT1D, 2D and 3D generic PDEs (advection, Burgers, Darcy, compressible and incompressible Navier–Stokes, shallow water) with varied initial and boundary conditions and parameters, plus code and baselines. Licence: CC BY 4.0.General PDEs