A team led by Hyojin Park, Nam-Hyun Yoo and Jinhong Yang has now released CD-CureNO, an open-source software package published in the journal SoftwareX that wraps neural-operator surrogates for composite curing in an unusually rigorous verification and provenance framework. The software, versioned at v0.0.3 and archived in Software Heritage under an Apache 2.0 license, is written in Python 3.10 through 3.12 using PyTorch, NumPy, SciPy and pandas, and it runs its audits and test suites on ordinary CPUs, with CUDA reserved for optional model training. What distinguishes the project is not a claim of new mathematics—the authors explicitly disclaim novelty for Fourier operators, joint temperature-and-cure prediction, or physics-based losses—but a domain-specific integration and verification layer built around inspectable scientific controls.

The motivating problem came from a legacy residual Fourier neural operator that stacked up to 51 independently trained positional models for curing analysis. Rather than silently replacing that code, CD-CureNO keeps two reproduction paths separately named: one that reproduces the legacy behavior exactly, and one that corrects known defects. The team regression-tested nine legacy defects, including normalization leakage and incomplete field outputs, so that the historical behavior is documented rather than erased. From that audited foundation, the software extends to joint temperature-and-degree-of-cure fields, causal response, conservative label generation, and transfer of learned knowledge across spatial dimensions.

The architecture is organized around explicit development and validation phases, labeled P0 through P6. P0 audits legacy behavior, P1 reproduces and corrects it, P2 develops joint one-plus-one-dimensional operators over the thickness-time field, P3 supplies a one-dimensional solver stage, P4 generates a true two-dimensional benchmark with two spatial coordinates evolving in time, P5 tests transfer, restriction and causality, and P6 is reserved for a future confirmatory held-out campaign. Each stage emits inspectable artifacts—configuration snapshots, Git and environment records, checkpoints, predictions and machine-readable receipts—that the next stage consumes, with cross-cutting guardrails verified again during postflight aggregation and reporting.

The authors define a result as auditable when its input, code, configuration, split, seed and checkpoint are all identifiable; when its reported metric can be recalculated from saved case outputs; when integrity violations or missing prerequisites fail explicitly; and when every claim can be traced back to those artifacts. Hashes, manifests, tensor mappings and receipts supply the observable evidence. Failed and superseded runs remain on record instead of being replaced by result-selected retries, and complete cure cycles or simulation cases—not individual grid points—define the samples. Crucially, the team distinguishes its independent postflight calculations, which are separate from the training and execution path, from validation by an unaffiliated research group or an independent physical model. A new NumPy-only example verifier imports neither training nor model code and reconstructs metrics from saved fields; a deliberately altered metric is rejected even after its file checksum is updated, and a failed recheck replaces any earlier success receipt.

Two technical contracts give the software its distinctive character. The first is restriction preservation: for laterally uniform coefficients, initial conditions and boundary data with zero lateral flux, an x-independent solution should reduce to the same through-thickness problem at every lateral position. The two-dimensional target therefore implements the lifted one-dimensional field plus a residual correction, with copied source tensors and zero-residual lateral branches that recover the source mapping exactly at initialization up to floating-point ordering. A verifier reconstructs the target state from the checkpoint, since output equality alone would not prove provenance. The second contract is causality: the causal model replaces temporal Fourier transforms with dilated convolutions, and perturbing all future input channels must leave every preceding output field unchanged. A nonnegative latent cure rate guarantees bounded, monotone cure evolution, and finite-volume-compatible fluxes handle material interfaces.

The verification evidence is unusually detailed. A GitHub Actions run on Ubuntu 24.04 passed 224 tests with 30 explicitly documented skips in about 42 seconds, and a fresh Windows run on a separate commit passed the same counts in about 48 seconds; continuous integration even exposed an input-stride-dependent CPU rounding difference, fixed by making shared target channels contiguous, with a new regression test added. The historical two-dimensional benchmark holds 512 temperature-and-cure cases of roughly 918 megabytes, anchored by 27 prespecified solver checks, zero failed cases, a maximum relative global energy residual of 1.345 times ten to the minus twelve, and exact float32 slice-hash replay for selected cases. Fresh regeneration of all 512 cases took 214.5 seconds on the documented Windows machine, and independent hashing confirmed that all eight output arrays match the frozen originals byte for byte. The team is candid that strict Linux plan validation rejected 33 heat-transfer values differing from the frozen Windows plan by at most 2.842 times ten to the minus fourteen, so canonical replay requires the documented compatible environment.

New in the revision is an independent mathematical verification of the production solver against elementary continuous reference solutions. Four synthetic case families—an anisotropic insulated transient, steady manufactured fields exercising all four Robin boundary faces with uniform and then discontinuous conductivity, and a solvable cure specialization coupled to reaction heating—were specified in an archived protocol with acceptance criteria fixed before execution. All 17 scheduled configurations and 72 recorded acceptance items passed. Measured spatial convergence orders clustered near two, temporal orders near one, consistent with backward Euler, and the synthetic cure equation gave orders near four, consistent with the production fourth-order Runge-Kutta update. Finest temperature relative errors ranged from 5.82 times ten to the minus five to 5.10 times ten to the minus four, and the finest cure maximum error was 1.80 times ten to the minus eight. The authors stress the boundaries: these studies verify anisotropic diffusion, boundary treatment and a special reaction-heat coupling, but not general kinetic validity, proprietary solver parity or experimental validation.

A fully public end-to-end example demonstrates the entire workflow without any private data or checkpoints. It generates eight true two-dimensional cases with laterally varying Robin coefficients, trains production causal operators with a fixed seed for 32 source and 48 target epochs, transfers knowledge through tensor copying and zero-residual initialization, and then recomputes held-out metrics in a separate NumPy-only postflight process. The two held-out cases yielded temperature relative errors of 0.006408 and 0.009741, temperature root-mean-square errors of 2.199 and 3.385 kelvin, and cure errors near 0.028, with future-prefix change exactly zero and independently recomputed metrics agreeing to within 4.44 times ten to the minus sixteen. The complete run took under ten seconds on a desktop CPU. The authors deliberately label this a functionality example, not evidence of superiority, and the figures show visible prediction errors rather than flattering comparisons.

Looking forward, the team has published a registered plan for a confirmatory P6 campaign: 40 held-out out-of-distribution cases across five target seeds, paired with a generic causal transfer baseline and analyzed with 10,000 crossed paired bootstrap replicates. A confirmatory claim would require a registered positive 95 percent improvement-interval bound, at least a 10 percent case-median improvement, physical-error guardrails and complete verified execution without test-based selection—and the authors note plainly that 40 cases across five seeds are not 200 independent cases and do not guarantee adequate statistical power. That discipline is the real story of CD-CureNO. In a field where surrogate models often arrive with cherry-picked demos and unverifiable numbers, the package demonstrates that neural operators for manufacturing can ship with the same provenance discipline as the experiments they hope to replace: every number traceable to a run, every failed attempt preserved, and every claim bounded by the evidence that produced it.

Subject of Research: Auditable restriction-preserving causal neural operators for thermochemical composite curing simulation

Article Title: CD-CureNO: Auditable software for restriction-preserving causal neural operators in composite curing

Article References: Park, H., Yoo, N.-H., & Yang, J. (2026). CD-CureNO: Auditable software for restriction-preserving causal neural operators in composite curing. SoftwareX, 36, Article 103086. https://doi.org/10.1016/j.softx.2026.103086

Image Credits: AI Generated

DOI: Not provided

Keywords: neural operators, composite curing, Fourier neural operator, software verification, reproducibility, provenance, causality, thermochemical modeling, surrogate models, manufacturing simulation, open-source software, convergence testing

The review, led by Mohammad Nooraiepour of the University of Oslo with colleagues from the University of Bergen, Siemens Energy AI Lab, and Friedrich-Schiller-University Jena, does not declare a winner in the contest between traditional solvers and neural networks. Instead, it delivers a more unsettling and more useful conclusion: the two paradigms are epistemologically different kinds of machinery, and confusing them leads to both overhyped claims about artificial intelligence and underuse of what it genuinely offers. Classical methods are deductive, meaning their errors can be bounded by quantities derived directly from the structure of the equation and the parameters of the discretization. Machine learning methods are inductive, meaning their accuracy depends on statistical proximity to the data they were trained on. That single distinction, the authors argue, determines not which approach is superior but what each can certify, and therefore where each should be deployed.

To make the comparison rigorous, the review organizes the entire field around six fundamental computational challenges: high dimensionality, nonlinearity, geometric complexity, discontinuities, multiscale phenomena, and multiphysics coupling. These six axes function as a common testing ground. Classical methods, including finite difference, finite element, finite volume, and spectral discretizations, are assessed on each axis for their structure-preserving properties, their rigorous convergence theory, and their scalable solver design. The verdict is nuanced. On problems of moderate dimension with well-behaved geometry, classical methods remain unmatched, delivering certified accuracy with error estimates that engineers can take to a regulator or a certification authority. But the review is candid about their persistent weaknesses: the curse of dimensionality crushes grid-based methods as the number of variables grows, and geometrically complex domains can make mesh generation a bottleneck that consumes more human effort than the solve itself.

Machine learning approaches are then subjected to exactly the same critical assessment, which is itself a departure from much of the existing literature, where neural methods are typically benchmarked only against other neural methods. The review builds a taxonomy organized by the degree to which physical knowledge is incorporated into the learning process. At one end sit purely data-driven surrogates; in the middle sit physics-informed neural networks, which embed the residual of the governing PDE directly into the loss function so that the network is penalized for violating the physics; and further along sit neural operators, which learn mappings between function spaces rather than pointwise values, promising generalization across families of related problems. Graph-based architectures, transformers, and generative models round out the landscape, alongside hybrid frameworks that interleave classical solvers with learned components.

The critical findings here are sobering for anyone hoping that neural networks will simply replace finite element solvers. Physics-informed neural networks, despite enormous popularity, struggle with stiff problems, sharp gradients, and the kind of discontinuous behavior that dominates real multiphysics applications such as fracture propagation or salt precipitation during carbon dioxide storage. Their training is often fragile, their convergence theory is far less mature, and their accuracy guarantees are fundamentally statistical rather than deterministic. Neural operators fare better on generalization across parameter ranges, but the review emphasizes that their reliability is always conditional on the training distribution. A neural surrogate that performs beautifully within the range of parameters it has seen can fail catastrophically, and silently, the moment it is asked to extrapolate. Classical methods never fail silently in quite the same way, because their error bounds degrade predictably with refinement.

What elevates the review above a simple scorecard is its treatment of complementarity. The authors identify three genuine synergies between the paradigms. First, machine learning can accelerate the inner workings of classical solvers, for example by learning preconditioners, coarse-grid corrections, or adaptive mesh indicators, while the outer classical framework retains its convergence guarantees. Second, classical methods can generate the high-fidelity training data and validation benchmarks that neural surrogates require, creating a pipeline in which the deductive method certifies the inductive one. Third, learned models can capture constitutive behavior or subgrid physics that is too expensive or too poorly understood to resolve explicitly, feeding these closures into structure-preserving classical schemes. In each case, the division of labor follows the epistemological split: the learned component supplies speed and flexibility where statistics suffice, and the classical component supplies certification where proof is required.

The review goes further and develops design principles for hybrid methods, including a framework for what the authors call the structure inheritance problem, that is, how to ensure that a learned component does not destroy the conservation laws, stability properties, or symmetry structures that a classical discretization was carefully constructed to preserve. They also propose an error budget decomposition that separates the total error of a hybrid simulation into three contributions: the discretization error of the classical scheme, the approximation error of the neural component, and the coupling error introduced at their interface. This decomposition matters because it turns hybrid design from an art into an accounting exercise. If a coupled simulation goes wrong, the framework tells practitioners where to look, and it prevents the common failure mode in which a neural component’s unquantified error is silently absorbed into an apparently converged classical result.

Looking forward, the authors assess emerging frontiers with the same skeptical lens. Foundation models for scientific computing, inspired by the large language models that transformed natural language processing, are evaluated against the structural constraints of PDE solving, and the review asks whether the scaling laws that powered language models can plausibly transfer to problems governed by conservation laws and boundary conditions. Differentiable programming for inverse design, in which entire simulation pipelines are made end-to-end differentiable so that gradients can flow from design objectives back through the solver, is judged promising but dependent on the same error-accounting discipline. Quantum algorithms for PDEs are treated with particular caution: the review distinguishes between barriers that are fundamental and those that are merely contingent on engineering progress, and finds that for several quantum approaches the current obstacles look structural rather than incidental. Exascale co-design, the joint development of algorithms and hardware at extreme scale, is identified as a domain where hybrid classical-learned solvers may find their most immediate practical impact.

The timing of this synthesis is significant. Scientific machine learning has grown from a niche curiosity into a field with its own conferences, journals, and funding programs, and the hype cycle has produced both genuine breakthroughs and inflated claims. Reviews that merely catalog methods multiply yearly, but rigorous evaluations that apply identical criteria to both paradigms remain rare. By insisting that epistemological character, rather than computational speed, is the primary criterion for responsible method selection, the review provides something the community has lacked: a principled basis for deciding, problem by problem, whether a learned surrogate, a certified classical solver, or a carefully engineered hybrid is the right tool. The stakes are concrete. Applications such as geological carbon storage, where the authors’ own research is anchored through projects on salt precipitation and convective mixing during CO2 injection, demand simulations that are simultaneously fast enough for uncertainty quantification over thousands of realizations and trustworthy enough to inform regulatory decisions about subsurface storage integrity.

The conclusion the authors reach is not that one paradigm supersedes the other, but that their principled integration offers the most credible path toward computationally tractable solutions to the most consequential open problems that neither approach can address alone. For decades, computational science advanced by refining discretizations, solvers, and hardware within a single deductive tradition. The last several years added a second tradition with fundamentally different guarantees. The review’s lasting contribution may be its refusal to choose between them, and its demonstration that the future of simulation lies in knowing precisely what each tradition can promise, what it cannot, and how to combine them without letting the strengths of one conceal the weaknesses of the other. In an era when artificial intelligence is routinely described as a replacement for existing tools, this is a refreshingly precise message: the equations that govern the physical world will be solved not by machines alone, nor by classical mathematics alone, but by an alliance whose terms are now, finally, being written down.

Subject of Research: A critical synthesis of classical numerical, machine learning, and hybrid methods for solving partial differential equations

Article Title: Partial differential equations in the age of machine learning: a critical synthesis of classical, machine learning, and hybrid methods

Article References: Nooraiepour, M., Both, J. W., Kadeethum, T., & Sadeghnejad, S. (2026). Partial differential equations in the age of machine learning: a critical synthesis of classical, machine learning, and hybrid methods. Artificial Intelligence Review. https://doi.org/10.1007/s10462-026-11677-5

Image Credits: AI Generated

DOI: 10.1007/s10462-026-11677-5

Keywords: partial differential equations, machine learning, physics-informed neural networks, neural operators, finite element methods, hybrid computational methods, scientific machine learning, structure-preserving discretization, error quantification, multiphysics simulation, foundation models, carbon dioxide storage

Neural operators have quietly become one of the most consequential tools in computational science. Instead of learning a mapping between finite vectors, they learn mappings between functions, which makes them natural solvers for partial differential equations. Given a family of PDEs describing, say, fluid flow through porous rock or heat diffusing through a material, a trained neural operator can predict the full solution field for a new set of parameters almost instantly, bypassing the expensive numerical solvers that would otherwise be required. Classical architectures such as DeepONet and the Fourier neural operator have transformed this landscape, but their quantum counterparts have lagged behind, hampered by processing overheads and theoretical gaps about what quantum circuits can actually represent.

The central obstacle has been scaling. Many quantum machine-learning paradigms demand qubit counts or circuit depths that grow so quickly with problem size that they collapse into impracticality on near-term hardware. The Shanghai team, led by Ruocheng Wang, Xiaoqiu Zhong, Zhuo Xia and Junchi Yan, attacked the problem from two directions at once: they built a leaner architecture and, crucially, they proved something rigorous about its power.

The theoretical centerpiece is an extension of the universal approximation theorem to the quantum domain for continuous nonlinear operators. Universal approximation, first established for classical neural networks in the 1990s, guarantees that sufficiently wide networks can represent a broad class of functions; the operator version underpins DeepONet. Proving an analogous guarantee for quantum circuits closes a foundational gap, but the team went further. Departing from the conventional view that quantum advantage must flow from the exponential size of Hilbert space, they showed that the architecture’s density matrix implicitly constructs a quadratic feature frame.

That quadratic frame yields a striking expressivity bound. For operators, the implicit feature space scales as O(p²) in the number of parameters p, circumventing the O(p) linear capacity limits of matched classical models. In plain terms, each additional parameter in the quantum model buys roughly a square’s worth of representational capacity compared with a classical model of the same size. This is a subtle but meaningful advantage: it does not rely on exotic claims about exponentially large state spaces, but on a concrete, provable property of how the quantum circuit encodes features.

The second innovation addresses a practical bottleneck known as spectrum alignment. Quantum models encode input data through frequency modulation, and capturing high-frequency components of a solution, the sharp gradients and fine oscillations that matter in real PDEs, typically requires either deep circuits or many parameters. The researchers introduced a trainable-frequency strategy, dubbed TF-QuanONet, in which the base frequencies of the encoding adaptively space themselves during training. The network effectively learns which frequencies to emphasize, capturing relatively high-frequency structure without inflating the parameter count or the circuit depth.

The benchmarks are where the claims meet reality. Across extensive experiments, TF-QuanONet notably outperformed competing quantum baselines and achieved accuracy competitive with classical frameworks under strictly matched-parameter conditions, a fairness constraint that quantum machine-learning comparisons often fail to honor. More intriguing still is what happened as the problems grew. In high-dimensional scaling regimes, with latent dimensions p approaching 256, the quantum architecture exhibited superior optimization robustness, consistently converging to its intrinsic error floor while classical baselines suffered from high variance. The quantum model was not just accurate; it was reliably trainable where its classical competitors became erratic.

The team also tested the architecture on real IBM quantum processors, including the ibm_fez device, as a qualitative proof of concept. Comparisons between noise-free simulations and physical hardware demonstrated that QuanONet retains functional resilience on near-term machines, a nontrivial achievement given that real qubits decohere, gates misfire, and measurement noise corrupts outputs. The experiments spanned a range of canonical problems, including dynamical systems, advection equations and Darcy flow, with visualizations confirming that predictions track ground-truth solutions across varying input frequencies.

The work arrives amid a broader reckoning in quantum machine learning, where researchers have grown wary of claims that evaporate under fair comparison or realistic hardware assumptions. By grounding its architecture in a proven expressivity theorem, keeping resource requirements modest, and validating on commercial hardware, the study offers a template for what credible quantum advantage in scientific machine learning might look like. The code has been released publicly on GitHub and archived on Zenodo, and all datasets were generated directly from it, inviting the community to scrutinize and extend the results.

Whether the quadratic expressivity edge translates into decisive practical wins as quantum hardware improves remains an open question, but the study reframes the debate. Rather than waiting for fault-tolerant machines or betting everything on exponential Hilbert spaces, it demonstrates that carefully designed quantum architectures, with theory and engineering aligned, can already hold their own against strong classical baselines on the noisy processors available today.

Subject of Research: A quantum neural operator architecture with a proven quadratic expressivity bound for solving partial differential equations on near-term quantum hardware.

Article Title: Quantum neural operators with implicit quadratic frame and expressivity advantages

Article References: Wang, R., Zhong, X., Xia, Z., & Yan, J. (2026). Quantum neural operators with implicit quadratic frame and expressivity advantages. Nature Machine Intelligence. https://doi.org/10.1038/s42256-026-01289-7

Image Credits: AI Generated

DOI: 10.1038/s42256-026-01289-7

Keywords: quantum machine learning, neural operators, partial differential equations, QuanONet, NISQ era, universal approximation, expressivity, trainable frequencies, IBM quantum processors, DeepONet, quantum computing, scientific machine learning

The survey’s central argument is that weather and climate represent an unusually demanding, and unusually revealing, testbed for artificial intelligence. Unlike image recognition or language modelling, atmospheric science confronts machine-learning systems with petabyte-scale spatiotemporal data spread across a rotating sphere, governed by partial differential equations and constrained by global observational and reanalysis archives. On top of that sits a challenge that most mainstream AI benchmarks simply do not have: climate non-stationarity. The statistical properties of the atmosphere are themselves shifting as the planet warms, which means models trained on the past may be silently invalidated by the very future they are asked to predict. The authors argue that this combination of scale, physics and drift makes meteorology a proving ground whose lessons generalize far beyond forecasting.

To organize an enormous and sometimes chaotic literature, the survey classifies the major AI architectures by their physical inductive biases — the built-in assumptions each design makes about the structure of the world. Convolutional neural networks, the workhorses of early deep-learning weather prediction, assume local spatial structure and translation invariance. Graph neural networks treat the atmosphere as an irregular mesh, naturally handling the geometry of the sphere and unstructured computational grids. Transformers bring global attention mechanisms that can capture long-range teleconnections such as the El Niño–Southern Oscillation and the Madden–Julian oscillation, which link weather patterns across entire hemispheres. Generative models, including generative adversarial networks and diffusion-based approaches, address a different problem entirely: producing realistic ensembles and downscaling coarse global fields to fine local detail.

Two unifying mathematical themes run through this taxonomy. The first is geometric deep learning, the program of designing networks whose internal operations respect the symmetries of the underlying space — in this case, rotation and translation on a sphere rather than on a flat plane. The second is operator learning, exemplified by neural operators such as the Fourier neural operator and its spherical variant, which learn mappings between entire function spaces rather than between individual data points. This distinction matters because weather and climate models are fundamentally functions of functions: they map one continuous field of temperature, pressure and wind onto another. Architectures that respect the spherical geometry of the planet and learn operators rather than fixed-resolution maps are, the authors argue, better positioned to generalize across resolutions and physical regimes.

Beyond architecture, the survey gives systematic treatment to three areas it considers underappreciated. Representation learning and foundation models — large networks pre-trained on vast atmospheric archives and then adapted to many downstream tasks — are examined as the emerging backbone of the field. Uncertainty quantification receives extensive attention, covering techniques from Bayesian approaches to ensemble generation, all aimed at answering the question operational forecasters care most about: not just what will happen, but how confident we should be. And causal discovery is framed as a complement to pure prediction, a way of using machine learning not merely to reproduce correlations in reanalysis data but to probe the physical mechanisms connecting them — a distinction that becomes critical when the climate itself is changing.

The paper’s most provocative contribution, however, is its naming and dissection of what the authors call evaluation pathologies in scientific machine learning. The first is the RMSE smoothness bias. Because root mean square error and mean-squared-error training objectives penalize sharp, spatially displaced features more harshly than blurry, averaged ones, models optimized on these metrics are systematically pushed toward smooth, blurred forecasts. A prediction that gets the shape of a storm exactly right but places it a few dozen kilometers off can score worse than a smeared, featureless field that is wrong everywhere but mildly. The practical consequence is that the metrics used to declare AI models superior to numerical weather prediction may be quietly selecting for aesthetically smooth mediocrity while penalizing the crisp, high-impact detail that matters most to forecasters and the public.

The second pathology the authors identify is ERA5 training–evaluation circularity. ERA5, the European Centre for Medium-Range Weather Forecasts’ flagship reanalysis, is the de facto training ground for most AI weather models — but it is also the reference against which those models are scored. A model trained to reproduce ERA5 and then evaluated against ERA5 is, in a meaningful sense, being graded on its own homework. The circularity inflates apparent skill, obscures the reanalysis’s own biases, and makes it difficult to know how models would perform against genuinely independent observations. The third pathology, benchmark overfitting, compounds the problem: as the community iterates on a small set of standard test cases, models increasingly specialize to those cases, and leaderboard gains stop translating into real-world forecasting skill.

The survey does not stop at diagnosis. It frames the field’s open problems as scientific machine learning challenges that extend well beyond meteorology. Distribution shift under a non-stationary climate is the paradigm case: any AI system deployed over years must cope with input statistics that drift, potentially violating the stationarity assumptions baked into training. Physical consistency of learned operators — whether a neural network’s predictions obey conservation laws and dynamical constraints even far from its training distribution — remains unsolved. Sample efficiency in data-sparse regimes, such as the ocean interior, polar regions and the developing world’s observation networks, tests whether foundation-model approaches can transfer knowledge to places with few measurements. And the authors argue for intrinsic interpretability: not post-hoc explanations bolted onto a black box, but models whose internal reasoning is transparent enough for scientists to trust and interrogate, a theme connected to the explainable AI research program the work was partly funded to advance.

Why does this matter now? Because the operational stakes are rising fast. Deep-learning weather prediction systems are already being trialed by major forecasting centers, and skill on benchmarks is being cited as evidence they can replace or supplement physics-based simulation. If the benchmarks reward blur and circularity, the field risks institutionalizing models that look excellent on paper while underperforming on the rare, extreme events — hurricanes, heat waves, flash floods — where forecasts save lives. The survey’s argument is that verification against rare extremes, using metrics such as the fractions skill score and the continuous ranked probability score alongside traditional correlation measures, must become central rather than peripheral to how AI forecasters are judged.

The broader lesson, the authors contend, is that weather and climate offer scientific machine learning a uniquely honest mirror. The domain combines massive data, hard physics, distribution drift and unforgiving operational verification — a combination that strips away the comfortable assumptions of mainstream AI benchmarking. The survey, published open access with funding support from the Austrian Science Fund and the European Union’s Horizon Europe RI-SCALE project, is intended as both a map and a challenge: a structured account of where AI methods for the atmosphere stand today, and a warning that the path forward runs through better evaluation, not just bigger models. If the field heeds it, the same rigor that makes forecasting trustworthy in a changing climate could reshape how machine learning is validated across the sciences.

Subject of Research: Artificial intelligence methods, benchmarking, and scientific machine learning challenges for weather and climate prediction

Article Title: Artificial intelligence for weather and climate: a survey of methods, benchmarks, and scientific machine learning challenges

Article References: Holzinger, A., Fiore, S., Degiacomi, T., Antonio, F., & Müller, H. (2026). Artificial intelligence for weather and climate: a survey of methods, benchmarks, and scientific machine learning challenges. Artificial Intelligence Review. https://doi.org/10.1007/s10462-026-11690-8

Image Credits: AI Generated

DOI: 10.1007/s10462-026-11690-8

Keywords: artificial intelligence, weather forecasting, climate modeling, machine learning, neural operators, geometric deep learning, uncertainty quantification, benchmark design, distribution shift, spatiotemporal learning, foundation models, scientific machine learning