Papers for

environmental modelers

Papers whose findings have a practical use for this group, as judged from the abstract. Open a paper to read what it means in practice.

New particle method predicts dynamic reaction and diffusion processes accurately

Score-based neural ordinary differential equations for reaction-diffusions

Abstract: A diffusion equation can be rewritten as a transport equation driven by the score function of the evolving density. Conventional probability-flow formulations, which move particles deterministically with this score-driven velocity, typically conserve total mass. In this work, we develop a Lagrangian method for reaction-diffusion equations in which each particle carries a position and a physical weight. The characteristic ODEs incorporate source terms into the evolution of the log-density and its gradient, while particle weights track changes in mass. We combine this representation with a learned velocity field and weighted particle training to approximate these dynamics. For non-increasing reaction rates, a relative-entropy inequality supports the mass-normalized self-consistency residual as the training objective. This objective can control the error of the induced density when the growth rate is bounded. The algorithm divides the time interval into short windows. In each window, a frozen velocity generates the particle data for a regression of the network velocity. We draw the particles from period-adapted initial sampling distributions and weight them in the loss partly by their predicted future mass, because the reaction can make regions with little initial mass important later. We numerically investigate the method on the Ornstein--Uhlenbeck (OU) equation with a linear source term, Fisher--KPP equations, and Wasserstein--Fisher--Rao gradient flows.

Fri 13 NovMachine Learning
The gist
Simulating how chemicals or substances spread and react over time is important but challenging. This paper introduces a method where each particle carries information about its position and how much 'stuff' it carries, allowing simulations to track both movement and growth or decay. The authors use neural networks to learn and predict how these particles move and change, improving accuracy in complex reactions. They tested their method on several mathematical models to show it works well.
Open → 2611.18467v1

Random forest methods improve analysis of compositional data types

Compositional random forest

Abstract: Compositional data have been analysed in the literature from three main points of view: directly on the simplex, in log-ratio coordinates, and as directional data on the unit sphere. In this paper, we construct a random forest framework for compositional responses that accommodates all three. Random forests are ensembles of decision trees, and the split criterion determines how each tree grows. We use the log-likelihood gain as the split criterion and consider distributions defined on the different spaces in which compositions can be represented. The Dirichlet likelihood has been studied in the literature, but its covariance is fixed by its mean up to a single scalar, and its maximum likelihood estimator has no closed form. To obtain more flexible criteria, we use two well-known transformations. First, the square-root transformation maps compositions onto the unit sphere, where existing directional distributions require numerical estimation. We introduce a tangent-normal (TN) distribution, which shares its parameters with the Kent distribution, and construct a split criterion whose gain is computed from additive sufficient statistics: the shape parameters are maximised out in closed form, up to an exponentially small truncation constant, so that no parameter is estimated iteratively. Second, the log-ratio transformation maps compositions to Euclidean space, where the multivariate logistic-normal distribution has a flexible covariance and a closed-form maximum log-likelihood, also based on additive sufficient statistics. Both criteria respond to a shift of the mean, measured relative to the covariance, and to a change in the covariance itself. In simulations, the most accurate forest depended on the covariance structure of the composition. The TN forest was the most accurate in two of the three designs.

Tue 10 NovMachine Learning
The gist
Compositional data represent parts of a whole, like percentages that sum to 100%, and can be tricky to analyze properly. The authors propose new random forest techniques that work on three common ways to represent such data, using special mathematical transformations and distributions. These methods help decision trees split data more accurately by considering how parts relate to each other as well as their variability. In tests, the best method depended on the data’s complexity, with one new approach often performing best.
Open → 2611.14275v1

Hydrograd improves water flow modeling with fast gradient calibration

Hydrograd v2.0: a differentiable two-dimensional shallow-water model on unstructured meshes with steady and unsteady adjoints for calibration and closure learning

Abstract: We present \textit{Hydrograd} v2.0, a significantly improved differentiable 2D shallow-water model written in Python with JAX. The model solves the shallow-water equations on the unstructured meshes of production models such as SRH-2D and HEC-RAS 2D with a well-balanced, positivity-preserving finite-volume scheme that handles wetting and drying. Gradients are tailored to the type of problem. Steady states are found by Newton's method with an exact sparse Jacobian obtained by graph colouring, and their gradients follow from the implicit function theorem at the cost of one linear solve. Unsteady flows are differentiated in reverse mode with checkpointing, or with an implicit time integrator whose gradient costs about two forward runs. All gradients agree with finite differences, to about $10^{-8}$ for steady flows. On the Savannah River case, the steady water surface differs from that of SRH-2D by 4.7\,mm (root mean square). Steady closure learning was 96 times faster than differentiating through time marching and reached a lower loss. An inflow hydrograph was inverted from stage records with a peak error of 2\,\%. In an unsteady roughness calibration, a singular value decomposition analysis of the sensitivity matrix showed that the stage records at the gauges are 70 to 160 times more sensitive to the total resistance of the reach than to its distribution among the roughness zones, which explains why only a gauge layout with gauges in every zone recovered all five roughness values. On one GPU, a mesh of 512,000 cells takes 7.6\,s for a forward run and 51\,s for a gradient.

Sun 8 NovComputational Engineering, Finance, and Science
The gist
Predicting how water moves in rivers and floodplains helps manage floods and rivers better, but making such models faster and more accurate is hard. The authors improved a computer model called Hydrograd that can simulate water flows on complex maps and calculate how changing certain factors affects the results efficiently. This helps quickly adjust model settings to match real water measurements and better understand what influences water behavior. Their method works faster and more precisely than before, making it easier to calibrate models and learn unknown details.
Open → 2611.11540v1

Hybrid models improve separate estimates in evapotranspiration studies

Non-Interfering Asymmetric AdamW Updates for Constrained Exploration of Equifinality in Hybrid Models

Abstract: Hybrid models let neural networks supply the unresolved parameters of a process-based simulator, and are fitted to observations that constrain aggregate fluxes and storages but not the internal partitioning that produces them. Such a fit is equifinal: configurations indistinguishable in the observed aggregates can disagree substantially about the unobserved internals, and therefore about the mechanism they attribute to the same data. We study evapotranspiration, whose total flux is observed while its split into transpiration and evaporation is not. A model ensemble built from random seeds alone spans only a narrow band of the partitionings the data admit. We produce the spread deliberately, training members under auxiliary repulsion on transpiration, subject to the fit to observations not degrading. Non-interference is usually arranged by making the auxiliary gradient orthogonal to the main one, but under AdamW the applied step is a preconditioned momentum direction, so orthogonality between raw gradients no longer implies it. We introduce Non-interfering Asymmetric Network Optimization (NANO), an update whose first-order effect on the main loss is exactly that of the main objective alone, whose auxiliary step is bounded, and which concentrates repulsion in the coordinates the main gradient leaves idle.

Tue 3 NovMachine Learning
The gist
Scientists want to understand parts of a process they can’t directly observe, like how water leaves plants versus the ground. The authors look at models that combine real-world knowledge with neural networks to estimate these hidden parts from overall measurements. They found that simply training these models can miss many different possible hidden explanations that still fit the observed data. To fix this, the authors developed a special way to train the models so they explore more of the possible hidden patterns without messing up the fit to what we see. Their method carefully balances learning from main data and exploring alternatives, improving understanding of hidden processes.
Open → 2611.04756v1

Efficient deep gaussian process method speeds up predictions for large datasets

Efficient Deep Gaussian Process Surrogates via Mini-Batch Latent MAP Estimation

Abstract: Deep Gaussian processes (DGPs) are flexible surrogate models for computer experiments with nonstationary response surfaces, but inference over their latent representations remains computationally challenging. Existing methods typically rely on variational inference or Markov chain Monte Carlo (MCMC), which impose variational-family assumptions or require costly posterior sampling. We propose the latent maximum-a-posteriori deep Gaussian process (LaMP-DGP), which uses gradient-based optimization to obtain MAP estimates of the latent representations and enables efficient probabilistic prediction through linked-GP calculations. To scale LaMP-DGP to large datasets, we develop a mini-batch optimization algorithm and, under a polynomial mini-batch bias-decay condition, establish a stationarity bound for latent-state optimization that motivates a growing batch-size schedule. Across numerical experiments and real-data analyses, LaMP-DGP provides consistently competitive predictive performance relative to doubly stochastic variational inference, stochastic gradient Hamiltonian Monte Carlo, and fully Bayesian Vecchia DGP inference. Computationally, LaMP-DGP is substantially faster than the Vecchia MCMC approach while remaining broadly competitive with the stochastic-gradient alternatives. These results position LaMP-DGP as a competitive and computationally efficient alternative for DGP inference. An R package, "LaMPDGP", is available for implementing the proposed method.

Tue 27 OctMachine Learning
The gist
Predicting outcomes in complex experiments can be slow and tricky when using deep Gaussian processes, a type of computer model. The authors propose a new way to quickly estimate the hidden parts of these models using a technique called maximum-a-posteriori (MAP) estimation. Their approach called LaMP-DGP uses small data batches to speed up computation without losing accuracy. The method performs well compared to current best techniques and runs faster, making it useful for big data tasks. The authors also provide an easy-to-use software package in R.
Open → 2610.38038v1

FastC-GC speeds up time series graph estimation by hundreds of times

fastC-GC and fastC-GC*: Efficient RCCP-Guided Graph Estimation for Time Series

Abstract: We introduce fastC-GC and fastC-GC$^*$, computationally efficient methods for time-series graph estimation using the two-test rule motivated by Reichenbach's common cause principle (RCCP) and the conditioning sets of causalised Granger causality (c-GC and c-GC$^*$). The methods use standardized data, ridge-residual correlation, and complete circular shift enumeration evaluated by Fourier transforms. Shared precision calculations and Schur downdates avoid repeated regression fits. On a single-threaded benchmark, fastC-GC was 68-277 times faster than the compared c-GC implementation, and fastC-GC$^*$ was 192-886 times faster than c-GC$^*$. Across four suites with reference graphs, aggregate mean area under the precision-recall curve (AUPRC) differed from the corresponding baseline by at most 0.019, but F1 was lower at the selected thresholds in every suite. The computational gains therefore accompany a change in estimator and selected graphs. Among eight methods, Bagged PCMCI+ had the highest mean AUPRC on sparse VAR, an autocorrelated structural process, and DREAM3; CUTS+ led on Lorenz-96. Wind-tunnel and river-discharge applications illustrate differences between inferred graphs without establishing causal accuracy. We prove the computational identities and a restricted population adjacency result for fastC-GC$^*$. Causal interpretation requires the stated graphical assumptions; validity of the implemented tests additionally requires a suitable shift-invariant null law. Empirical null rejection rates informed the conditional threshold $β= 0.005$, without establishing general test validity or graph-wide error control.

Tue 27 OctSocial and Information Networks
The gist
Finding connections in time-series data helps understand how variables influence each other over time, but current methods can be slow. The authors introduce fastC-GC and fastC-GC*, which are much faster techniques that use mathematical tricks like circular shifts and Fourier transforms to speed things up. These new methods run hundreds of times faster than existing approaches while producing similar accuracy on average. However, the faster speed comes with changes in the detailed results, so the graphs they produce differ somewhat from previous methods. The authors also provide mathematical proofs behind their approach and caution that certain assumptions are needed to interpret the results causally.
Open → 2610.37883v1

Evolving benchmark reveals where environmental models fail over time

EvoBench: An Evolving, Diagnostic Benchmark Framework for Environmental Modeling

Abstract: Machine learning models are now widely used to predict environmental variable dynamics in domains like hydrology, climate science, agriculture, and ecology. These models are usually compared on fixed benchmark datasets using a single average accuracy score. As models improve, their scores become very close when averaged over large regions and long time periods, so the benchmark can no longer tell them apart, and a single number hides \emph{where} and \emph{when} a model fails. Three failures are common across environmental domains: models get worse as the climate shifts (non-stationarity), models fail at unmonitored sites (spatial extrapolation), and models miss rare but important regimes such as floods and droughts. We present \textbf{EvoBench}, a benchmark framework that shows where a model fails, tracks failures over time, and evolves to focus on model capability gaps. We apply EvoBench to daily streamflow prediction as a case study (\textbf{CAMELS-Evolve}, about 9{,}000 U.S. basins, 1985--2024) and compare an operational physical model (NWM), a conceptual reference (HBV), and several data-driven models. No model wins on every testing scenarios (e.g., in terms of average accuracy, performance at low flows and unmonitored basins). Also we observe that all data-driven models under-predict floods. Additionally, we show that targeted fix based on the reported failures could mitigate weaknesses the benchmark identified in certain scenarios. EvoBench turns model evaluation from a saturated ranking into a moving map of model limitations.

Mon 26 OctArtificial IntelligenceMachine Learning
The gist
Environmental models predict things like streamflows and climate, but they often fail under changing conditions, new locations, or rare events like floods. The authors created EvoBench, a tool that shows exactly when and where different models struggle instead of just giving one average score. They tested this on daily streamflow predictions from thousands of U.S. basins and found no single model is best in all situations. EvoBench also helps identify and fix model weaknesses by tracking errors as they change through time.
Open → 2610.35464v1

Python tool finds equations behind changing systems with confidence

py-ARGOS: A Python package for automatic, statistically principled discovery of ordinary differential equations from data

Abstract: Recovering the governing equations of a dynamical system from data is a central problem in data-driven science, yet the available Python tooling forces a compromise among automation, statistical rigor, and reliability. We present py-ARGOS, a Python package implementing two automatic algorithms, bootstrap ARGOS and Bayesian-ARGOS, behind a uniform interface. Both are statistically principled throughout: automated model selection reduces the candidate library, each coefficient is reported with a bootstrap confidence or Bayesian credible interval that quantifies its uncertainty, and a term is retained when this interval excludes zero. Built-in diagnostics additionally flag unreliable fits. On chaotic benchmark systems, the two algorithms improve on the SINDy family's identification performance.

Sun 25 OctMathematical Software
The gist
Understanding how things change over time often relies on finding equations that describe their behavior. The authors present py-ARGOS, a Python package that can automatically discover these equations from data, doing so in a way that carefully checks how certain the results are. It uses two smart algorithms that not only pick which parts of an equation matter but also show how sure we can be about each part. This helps avoid mistakes and works well even for chaotic or complex systems.
Open → 2610.34458v1

Stable methods improve estimation of unknown forces in dynamic systems

Stable and Calibrated Latent Force ODE Filters

Abstract: Estimating unobserved forces in mechanistic systems is a central task of scientific machine learning. Approximate Gaussian filtering methods that synthesize mechanistic and empirical information have shown remarkable capacity in jointly identifying latent processes together with the system state in a single forward-backward pass. The resulting latent force ODE filter framework channels information from observational data specifically into the unknown parts of the system. This work addresses open questions about calibration and stability of the method. We show that the joint posterior over state and force is the reweighted conventional latent force posterior. This uncovers a crucial dependence of the stability of the method on the prior variance of the system state, which effectively controls the flow of information from observations to the latent force through the mechanistic system. From this insight, we construct an improved algorithm that provides stable and calibrated plug-and-play inference. We demonstrate the improved framework on several simulated experimental settings.

Sat 24 OctMachine Learning
The gist
When machines or natural systems change over time, we often want to guess hidden forces affecting them. The authors study a way to combine known system rules with observed data to estimate these hidden forces more reliably. They found that how the method handles uncertainty strongly affects its stability and accuracy. Using this insight, they made a better version of the method that gives more stable and trustworthy estimates, tested on simulated examples.
Open → 2610.33939v1

Multifidelity methods improve estimates for complex high-dimensional data

Multifidelity Monte Carlo Estimator for High-Dimensional Quantities of Interest

Abstract: Multifidelity Monte Carlo (MFMC) reduces the cost of estimating the statistics of quantities of interest (QoIs) computed via expensive high-fidelity model by using cheaper low-fidelity models as control variates. We extend MFMC to vector-valued QoIs by replacing the scalar control-variate coefficients with matrices and minimizing a weighted mean-squared error. The optimal full-matrix coefficients are available in closed form, and their variance reduction is summarized by an effective correlation that takes the place of the squared Pearson correlation in the closed-form MFMC design. Restricting the coefficients to block-diagonal, diagonal, and scalar classes gives a nested hierarchy of estimators with increasing variance reduction and increasing pilot-data demand, with closed-form coefficients for compatible weightings. For a scalar QoI obtained by linearly projecting the state, correcting the full state before projecting is never worse at the population level. Experiments on a Gaussian problem, a reduced-order model of an advection--diffusion problem, and a coarse-grid two-phase flow problem show that the best class depends on the pilot size relative to the number of covariance quantities it must estimate, and that reusing the pilot samples lowers the mean-squared error.

Sun 18 OctMachine Learning
The gist
Making accurate predictions from complex models can be very expensive to run. The authors extended a method called multifidelity Monte Carlo that uses cheaper, simpler models to help reduce the work needed for high-dimensional results. They show how to smartly combine information from multiple model types to get better estimates. Their new approach works better for vector data (multiple numbers at once) and can choose simpler or more complex combinations depending on data availability. Tests on different problems confirm the approach reduces errors efficiently.
Open → 2610.25016v1

Structured corrections improve gaussian process predictions in science

Learn Only Where You Can't Compute: Structured Corrections to Gaussian Process Posteriors

Abstract: Physics-informed Gaussian processes (GPs) provide a principled framework for scientific inference by expressing uncertainty over empirical measurements, boundary conditions, and governing equations. However, they struggle to scale to larger datasets and nonlinear information. Conversely, diffusion models excel at learning complex distributions over functions from massive datasets, but struggle in scientific contexts where training data are scarce and physical or observational information must be incorporated. We propose a structured posterior-correction framework that combines a tractable Gaussian reference with data-driven corrections via generative learning. After a mean correction and covariance calibration, a diffusion model learns a targeted correction in a low-dimensional subspace selected to capture discrepancies from the reference, while the complement retains its Gaussian law. Across three scientific inference tasks, these corrections recover hidden modes, improve sample structure, and reduce prediction error relative to the Gaussian reference. Comparisons with full-field generative models reveal task-dependent benefits and tradeoffs between predictive accuracy and computational cost.

Thu 15 OctMachine Learning
The gist
Gaussian processes are a way to predict based on imperfect scientific data but can struggle with lots of data or complex patterns. Diffusion models learn complicated patterns better but need much data and struggle with scientific rules. The authors created a method that combines a simple starting prediction with a focused correction learned by a diffusion model, improving predictions where the simple method falls short. This combined approach helps recover hidden details and reduces errors in scientific tasks compared to using just the simple method alone.
Open → 2610.21977v1

Forecast-error entropy reveals hidden predictive links in complex time series

Beyond RMSE and Granger Gain: Multi-Horizon Forecast-Error Entropy for Nonlinear Predictive Coupling

Abstract: We investigate whether the past of an auxiliary time series provides predictive information about a target beyond that contained in the target's own history. We compare matched out-of-sample target-only and target-plus-source forecasts and introduce the forecast-error distribution entropy (FEDE) predictive-causality framework, where predictive causality is understood in the Granger sense: auxiliary history is informative when it improves prediction beyond the target's own past. FEDE quantifies how the full signed forecast-error distribution changes across prediction horizons. Controlled nonlinear benchmarks hide source-target dependence behind nonlinear transformations not supplied to the predictors. LightGBM recovers these hidden channels, whereas Ridge does not. Across four equal-power nonlinearities and four noise levels, the mean response-sensitivity statistic for FEDE remains 29.00, 27.47, 23.22, and 28.67, compared with 14.09, 10.68, 12.67, and 12.71 for mean log-RMSE gain. Directional and source controls separate the true source-to-target channel from the reverse direction and from independent periodic, quasiperiodic, chaotic, and autoregressive auxiliaries. Chronological Granger-style comparisons show strong model-form dependence: a degree-2 polynomial model is exceptionally sensitive to the matched quadratic channel but becomes weak or oppositely aligned for other nonlinearities, while basic convergent cross mapping is weak or directionally inconsistent in this stochastic, partly non-invertible setting. FEDE remains positively aligned across all 108 tested combinations of window length, history length, and maximum forecast horizon. Together, these results establish FEDE as a sensitive and robust horizon-resolved tool that extracts predictive information beyond conventional error-magnitude measures when the nonlinear interaction form is unknown. Code available in Zenodo.

Tue 13 OctMachine Learning
The gist
It can be hard to tell if one changing thing helps predict another when relationships are complicated or nonlinear. This paper introduces a new way to measure how much extra information one time series provides about another by looking at the whole pattern of prediction errors over time. The authors show their approach works better than traditional error-based methods even when known models or simple assumptions fail. Their method can spot hidden interactions that others miss, making it a useful tool for understanding complex systems.
Open → 2610.18390v1

Earth science model calibration delegated to knowledge-guided language models

Delegating Earth science model calibration to knowledge-guided LLM agents

Abstract: Applying process-based Earth science models to new sites and questions often requires expert calibration decisions. Reliable delegation to language-model agents must preserve scientific competence and reproducibility. Here we develop a general calibration harness (GCH) for end-to-end delegation from a calibration request to assessed results. Constrained by a knowledge infrastructure (KI) and a validation convention, agents define parameter domains, observation mappings and objectives, and generate model-specific code. GCH checks the formulations against the same declarations and freezes them before numerical search and assessment, which require no further agent judgement. Agent-authored workflows executed across six model families, with assessment identifying unmet criteria. In snow-model tests, five of five knowledge-guided specifications executed on the first attempt, compared with none of five given only the model name, achieving validation Nash-Sutcliffe efficiency (NSE) of 0.81-0.94 against 0.57 for defaults. Under a common objective, independent GR4J formulations yielded scores within 0.001 NSE; re-evaluating eleven saved parameter vectors reproduced their metrics. In ten paired VIC tests with matched budgets, fixed search was non-inferior to continuing agent proposals (margin, -0.02 NSE; median difference, -0.003). This separation enables end-to-end calibration delegation with inspectable scientific choices and reproducible numerical evaluation.

Tue 13 OctArtificial IntelligenceComputational Engineering, Finance, and Science
The gist
Earth science models need careful tuning of parameters to work well at new locations or for new questions, which usually requires expert input. The authors created a system where intelligent language models, guided by scientific knowledge and rules, automatically handle this calibration from start to finish. Their system checks all steps to make sure decisions are sound and results can be reliably repeated. Tests showed the approach works well for different models, improving accuracy and consistency. This helps scientists delegate complex tuning without losing scientific rigor.
Open → 2610.18378v1

Cross-view method improves rare event generation in time series

Cross-View Time-Frequency Diffusion for Extreme-Aware Time-Series Generation

Abstract: Generating realistic time-series requires faithfully reproducing rare but high-impact extreme events. Existing approaches often rely on specialized representations or extreme-focused training objectives, which may restrict the diversity of captured extremes or compromise overall distributional fidelity. We propose Cross-View Time-Frequency Diffusion (CrossTF), a framework that exploits discrepancies between time- and frequency-view denoisers for extreme-aware time-series generation. Our key empirical finding is extreme-sensitive cross-view asymmetry: moving along the direction connecting the two views' clean-sample estimates substantially changes extreme fidelity while having limited effects on overall generation quality. We further identify a late-stage cross-view window in which this discrepancy stabilizes and cross-view adjustments have a particularly strong effect on extreme fidelity. These findings motivate a staged optimization strategy in CrossTF: after training and freezing both denoisers, it only needs to learn step-specific fusion parameters within this window using an Extreme Value Theory-based objective under an energy-distance constraint. This formulation supports interpolation between and extrapolation beyond the view-specific estimates while limiting degradation in overall generation quality. Extensive experiments on seven real-world datasets against 11 baselines show that CrossTF improves extreme fidelity by 26.1% and overall performance by 7.4% relative to the strongest baselines. These results highlight cross-view discrepancy as an effective basis for low-dimensional optimization in extreme-aware time-series generation.

Mon 12 OctMachine LearningArtificial Intelligence
The gist
Generating realistic time-series data is hard when it involves rare but important extreme events, like sudden spikes. The authors found that looking at data from both time and frequency perspectives reveals differences that help better capture these extremes. They created a method called CrossTF that combines these views at key moments to improve the accuracy of extreme event generation without hurting overall data quality. Testing on seven real datasets showed their method makes rare event predictions more reliable and improves overall realism compared to other techniques.
Open → 2610.17407v1

Graph partitioning improves interpretable spatial prediction

Interpretable Graph Partitioning for Spatial Prediction

Abstract: Spatial prediction estimates unobserved values from spatially distributed observations. Although flexible predictive models are increasingly available, their practical use remains limited when interpretability is essential. In this study, we formulate a graph partitioning problem to construct a global surrogate model that approximates black-box predictions with simpler representations. Specifically, we select connected subgraphs to minimize within-cluster residuals in vertex-level predictions. Although this formulation can be viewed as a variant of spatial regionalization, it operates on model-based predictions at individual locations rather than predefined areal units. With many locations, this point-level formulation may produce artificial connections between regions, reducing interpretability. We therefore introduce a constraint that excludes such connections while retaining flexible partition boundaries. We theoretically analyze how this constraint affects the resulting partition. We also develop a computational procedure for adjusting partition boundaries. Experimental results show that the proposed method yields spatially coherent partitions with flexible boundaries.

Mon 12 OctMachine Learning
The gist
Predicting values in places where you have little data is tricky, especially when you want the results to be easy to understand. The authors present a method that groups nearby points on a map into connected regions so the predictions can be simplified but still follow complex borders. They add rules to avoid strange connections that can make the groups confusing. Their approach makes it easier to explain predictions while keeping regions meaningful and spatially clear.
Open → 2610.16063v1

Hydrology model simulates rainfall runoff with interpretable water flow states

A Gated Catchment Hydrology Model (GCHM) for Physically-Interpretable Rainfall-Evapotranspiration-Runoff Simulation Across Diverse Hydroclimatic Regimes

Abstract: Recent advances in machine learning have improved rainfall-runoff prediction, yet most data-driven models remain difficult to interpret and do not explicitly enforce physical conservation laws. We present the Gated Catchment Hydrology Model (GCHM), a mass-conserving architecture based on the Mass-Conserving Perceptron (MCP). GCHM uses five storage states representing canopy interception, snow, soil moisture, groundwater, and streamflow generation. Constrained gates govern precipitation partitioning, evapotranspiration, infiltration, seepage, runoff, and baseflow while learning context-dependent process relationships. We evaluate GCHM across 50 CAMELS catchments spanning snowmelt-dominated, recent-rainfall, historical-rainfall, and arid regimes, using NOAA AORC meteorological forcing and remotely sensed vegetation and land-surface data. Performance is assessed using the Jawad-Kling-Gupta Efficiency (JKGE) metric, which accounts for stationary and non-stationary hydrologic behavior. We compare the full architecture with ablated variants and recurrent neural network benchmarks. GCHM reproduces streamflow dynamics across diverse hydroclimatic regimes, including arid catchments, while maintaining interpretable states and fluxes. Ablation experiments show that simplified architectures may perform adequately under wet conditions but struggle to represent flow persistence and recession during dry periods, when groundwater storage sustains low flows. Learned gates recover hydrologically plausible relationships governing snowmelt, runoff, seepage, and evapotranspiration. GCHM combines competitive predictive performance with mass conservation and interpretability, providing a foundation for hydrologic modeling and counterfactual analysis.

Sat 10 OctMachine Learning
The gist
Predicting how rainfall turns into river flow is important but usually hard to explain and make sure water is conserved. The authors built a model called GCHM that tracks water through different parts of a landscape like snow and soil, making sure water amounts add up correctly. It works well across many different climates and pours, and the model helps people understand how rainfall, evaporation, and runoff interact. This makes it easier to trust and use in studying water systems.
Open → 2610.14493v1

Adaptive physics-informed solvers improve accuracy and efficiency for complex equations

AB-PIELMs: Adaptive-Basis Physics-Informed Extreme Learning Machines for Residual-Driven Domain Decomposition

Abstract: Physics-informed neural solvers often struggle with multiscale behavior and sharp gradients due to the use of global approximation spaces and nonconvex training. We introduce AB-PIELM, an adaptive-basis physics-informed extreme learning machine that combines domain decomposition with joint optimisation of spatial and spectral hyperparameters while retaining deterministic normal-equation training. The method adapts both the number and placement of subdomains and the locality of radial basis functions, enabling problem-dependent allocation of representational capacity without modifying the underlying solver. Experiments on oscillatory function approximation and singularly perturbed advection-diffusion equations demonstrate accurate solutions across a wide range of stiffness regimes while using substantially fewer neurons than existing neural and PIELM-based approaches. The learned hyperparameters are interpretable, concentrating resolution near sharp gradients. The framework also supports inverse problems, successfully recovering diffusion coefficients from sparse noisy observations using Bayesian optimisation. These results indicate that adaptive-basis PIELM formulations provide an efficient and stable alternative to gradient-trained neural PDE solvers and naturally extend to broader classes of differential equations.

Thu 8 OctComputational Engineering, Finance, and Science
The gist
Solving complex physics equations with traditional neural networks can be tricky when the problems have sharp changes or happen at multiple scales. The authors developed a method called AB-PIELM that breaks the problem into smaller pieces and adjusts how it approximates each piece. This approach uses fewer computational resources while still giving accurate results, even for problems that are usually hard to solve. It can also help find unknown physical parameters from noisy data, making it useful in practical settings.
Open → 2610.11673v1

Physically consistent reconstruction improves ocean surface ph estimates

REACT: Physically and Chemically Consistent Reconstruction of Marine Active Tracers

Abstract: Reconstructing global sea surface pH from sparse observations is critical for monitoring ocean acidification and understanding marine carbon cycling. Traditional assimilation and inverse models are physically grounded but costly for large-scale reconstruction. Recent black-box and physics-guided AI models improve efficiency, but are mainly designed for passive tracers, where the reconstructed variable is also the transported inventory. In contrast, pH is an active carbonate tracer: it is the prediction target, while dissolved inorganic carbon (DIC) is the conserved carbon inventory. This mismatch can produce low pH error while violating carbonate closure and source-free carbon conservation. To address this, we introduce \textbf{REACT}, a carbon-first reconstruction framework that decouples transport, active correction, and chemical decoding. REACT transports a latent carbonate state with a conservative advection--diffusion solver, captures non-conservative carbon-cycle variations with a source module, decodes the corrected state into pH, and constrains the output through carbonate equilibrium. This design keeps pH as the target while enforcing consistency on the underlying carbon state. On simulation data, REACT reduces pH NRMSE by (14.7%) and chemical consistency error by (24.0%) over the best baseline. Cross-temporal-scale evaluations show robustness against error accumulation from coarse to fine temporal scales, and ablation studies validate the effectiveness of each component.

Fri 2 OctArtificial IntelligenceMachine Learning
The gist
Measuring the acidity (pH) of the ocean's surface is important but hard because data is limited. The authors created a method called REACT that uses carbon measurements to better estimate surface pH. By focusing on the carbon that drives changes in ocean chemistry, their approach keeps the math consistent with real ocean processes and reduces errors. This makes monitoring ocean acidification more reliable over large areas.
Open → 2610.03888v1

StyleGAN2 improves reservoir property modeling with better data matching

Parameterization method of reservoir properties for ensemble-based data assimilation using intermediate latent space of StyleGAN

Abstract: Ensemble smoothers are the most successful and efficient techniques currently available for history matching. However, because these methods rely on Gaussian assumptions, their performance is severely degraded when the prior geology is described in terms of complex facies distributions (non-Gaussian). In this way, for these methods, we need to apply efficient parameterization techniques. Currently, the most efficient methods for performing parameterization are deep learning models. However, given the variety of existing deep learning models, studies have not identified which is most suitable for use with ensemble-based methods, although some important models had already been evaluated. Based on a recent literature review, the most promising models selected were VAE-GAN, Latent Diffusion, and StyleGAN models. As a novel aspect of this work, data assimilation with the second generation of StyleGAN (StyleGAN2) model was performed using the latent z-space and intermediate w-space, separately. They were applied in two 2D case studies: one categorical (three facies) and the other continuous. The results demonstrated that all three models are highly efficient, with the StyleGAN2 model standing out for generating samples with geological realism and achieving excellent data matching in the cases studied. Our findings show that performing data assimilation with StyleGAN2 using the intermediate space (w-space) yielded better results than the traditional application in the latent space (z-space). This is due to the fact that ESMDA uses linear updates and the w-space is much more linear and disentangled than the highly entangled z-space, thereby ensuring that the updated vectors remain close to realistic geological patterns. These results were validated using main geostatistical and history matching metrics.

Wed 30 SeptMachine LearningArtificial Intelligence
The gist
When trying to improve predictions about underground reservoirs, current methods struggle if the underground patterns are complex. The authors studied new ways to simplify these patterns using advanced AI models called StyleGAN2 and others. They found that using a special internal part of StyleGAN2, called the intermediate space, led to better results than traditional approaches. This makes it easier to update their models to match real data while keeping the underground patterns realistic.
Open → 2609.39626v1

Geometry influences PDE learning accuracy and efficiency across domains

Geometry-physics confounding impairs PDE learning across varying domains

Abstract: Learning partial differential equation (PDE) dynamics across varying domains is central to predictive modelling and data-driven discovery of governing equations. However, geometric variation alters both field representation and the governing differential operators, confounding geometric effects with intrinsic physical properties in the observed dynamics. This work identifies geometry-physics confounding as a unified failure mechanism for PDE learning across varying domains. In forward operator learning, this confounding increases the burden of inferring geometry-dependent operator changes from finite data, reducing data efficiency and generalisation. In equation discovery, omitting geometry-induced operators misspecifies the candidate library, leading to biased parameters, missed governing terms and spurious terms. We propose a de-confounding framework that makes the known geometry-to-operator transformation explicit. Geometry-induced coefficient fields improve prediction and data efficiency across five operator-learning benchmarks, while geometry-complete candidate libraries recover the generating equations and reduce held-out PDE residuals by more than two orders of magnitude in both evolving-domain systems. By separating known geometric action from intrinsic physics, the proposed framework supports more reliable and data-efficient PDE learning across scientific and engineering problems with varying geometries.

Tue 29 SeptMachine Learning
The gist
When trying to teach computers how physical systems change over different shapes or spaces, variations in shapes can confuse the learning process. The authors found that these shape changes affect both how the data looks and the mathematical rules behind it, mixing up geometry with true physical behavior. They created a method that clearly separates the effects of geometry from the underlying physics, which helps computers learn more accurately and use less data. This approach improved predictions and helped identify the correct physical equations much better than before.
Open → 2609.38623v1

Physics-guided adapter improves flood prediction model transfer to new coasts

A Model-Agnostic Physics-Guided Adapter for Few-Shot Transfer of Coastal Flood Prediction Models to Unseen Regions

Abstract: Deep learning surrogates can produce high-resolution coastal flood maps orders of magnitude faster than physics-based hydrodynamic simulators, yet transferring them to new coastal regions remains costly, since generating target-region data for fine-tuning typically requires numerous time-consuming simulations. To tackle this bottleneck, we introduce the Physics Adapter (PA), a compact, architecture-agnostic adaptation interface that enables efficient few-shot transfer of flood prediction models across diverse coastal regions. PA predicts peak water level through a differentiable wet/dry response that compares terrain elevation against a learned water level, and blends this physics-structured prediction with a data-driven branch through a learned gate. Unlike physics-informed formulations, PA imposes no PDE-residual or conservation losses and instead exploits elevation as an architectural inductive bias, adding a negligible number of trainable parameters. We integrate PA into 12 heterogeneous models, and evaluate them on two coastal regions with markedly distinct geometries, topographies, and shoreline protection configurations. The performance of PA is benchmarked against a no-physics baseline, full fine-tuning, and standard parameter-efficient fine-tuning (PEFT) methods, considering both within-region generalization to unseen sea level rise values and between-region transfer. In low-shot regime (K=3), and averaged over all backbones and transfer settings, adding PA reduces root mean square error by 11.5% when only the output head is adapted on a frozen backbone, by 15.4% when combined with PEFT methods, and by 22.9% under full fine-tuning, compared to matched configurations without PA. Taken together, the findings of this work offer practitioners a concrete recipe for extending DL-based coastal flood predictors to new, data-scarce regions.

Tue 29 SeptMachine Learning
The gist
Flood prediction models based on deep learning are fast but often need lots of new data to work well in new coastal areas. The authors created a small add-on called the Physics Adapter that uses the shape of the land to help these models quickly adapt with very little new data. This adapter works with many kinds of models and improves accuracy when predicting floods in unfamiliar coastal regions. Their tests show that adding this adapter reduces prediction errors significantly, even when only a few examples are available for learning.
Open → 2609.37565v1

Machine learning speeds up complex dynamic system simulations accurately

Accelerated surrogate dynamics for dynamical, stochastic system evolution

Abstract: Dynamic simulations are an entrenched way of gaining insight into the evolution of system dynamics. Their computational cost however is often prohibitively high, especially in cases of stochastic frameworks. Machine learning algorithms are especially suited as simulation surrogates. Nevertheless, they face some very distinct limitations. Firstly, the sheer dimensionality of these systems, however, precludes the use of traditional time series models who struggle with high dimensional feature spaces. Additionally, traditional time series focus exclusively on either long or short range effects, causing local or global drift given enough time. In this paper, we propose a framework that addresses those limitations. Our framework combines a Variational Autoencoder, with a convolutional or graph basis that reduces the dimensionality of the system. This latent vector is propagated in time using a Temporal Fusion Transformer model, which includes both long range and short range effect encoding, as well as static covariate support. We test our framework on three distinct cases, to prove its robustness and in all three we have achieved practically identical to the simulation results at a fraction of the time. Further, our framework is flexible enough to be adapted to any new system and provides an inbuilt uncertainty quantification for targeted experiment design.

Tue 29 SeptArtificial Intelligence
The gist
Simulating how complex systems change over time can take a very long time, especially when randomness is involved. The authors found a way to make these simulations much faster by combining machine learning tools that simplify the system’s state and predict its future changes. Their approach keeps track of both short-term and long-term effects to avoid errors building up over time. They tested this approach on three different types of systems and got results that match detailed simulations but much quicker.
Open → 2609.37184v1

Benchmark platform measures physical field model accuracy across patterns

PDE-OBS: Controlled Evaluation Across Observation Patterns

Abstract: Physical-field reconstruction and forecasting depend on both measurement density and spatial layout, yet evaluation under a single observation pattern does not characterize performance when that pattern changes. We introduce PDE-OBS, an integrated benchmarking platform spanning numerical data generation, model training, and inference and evaluation under varying observation conditions. It combines 560,000 fields and trajectories from seven partial differential equation families with configurable observation operators and seven adapted baseline methods for stationary reconstruction and short-horizon forecasting. Separating observation construction from physical records allows users to specify parameterized patterns and deterministic mixtures for training and testing while preserving prediction targets and data splits. The evaluation protocol uses references trained for each test pattern to compare models on identical test observations and targets, alongside equal-count groups for spatial-layout comparisons. On a 14,000-record subset, we evaluate 441 trained models under nine test patterns, yielding 3,969 evaluations. Mean cross-pattern error exceeds mean matched-pattern error in all 49 PDE-method pairs, and this finding persists in a configuration-matched subset of 117 models. Denser test observations do not consistently reduce error for a fixed model. Mixed-pattern training on five completed pairs reduces large single-pattern transfer errors, although destination-trained references usually remain more accurate. Together, the benchmark and findings support systematic evaluation of observation-pattern sensitivity and provide a reusable workflow for developing methods under changing measurement conditions. Code: https://github.com/ru1ch3n/PDE-OBS.

Tue 29 SeptMachine Learning
The gist
Predicting physical fields like weather or fluids depends on where and how often you take measurements. The authors created PDE-OBS, a platform that tests how well different models predict these fields when the measurement locations and patterns change. They tested many models on thousands of scenarios and found that models trained for one pattern often do worse when tested on another. Training with mixed measurement patterns helped reduce errors, but models trained directly for a certain pattern usually worked best. This platform helps builders test and improve models under varying measurement conditions.
Open → 2609.36521v1

Deep learning predicts complex system changes from few snapshots

DisKO: Deep Koopman Learning in Distribution Space from Unpaired Snapshots

Abstract: Many complex systems are observed only through temporally unpaired distribution snapshots, making trajectory-based dynamical learning difficult without additional assumptions. We therefore formulate the problem directly in distribution space, treating the distribution itself as the dynamical state. The challenge is that distribution space is infinite-dimensional, making compact and approximately closed representations difficult to learn from finite snapshots. We introduce DisKO, which extends deep Koopman learning to distribution dynamics by jointly learning predictive distributional observables, a finite-dimensional Koopman representation, and a generative map back to the full distribution. Across seven diverse benchmarks, DisKO achieves state-of-the-art extrapolation performance, with substantially slower error accumulation on long-horizon prediction tasks. DisKO further recovers leading Koopman eigenvalues and eigenfunctions on systems with analytic spectra, revealing meaningful dynamical structure in the learned representation.

Mon 28 SeptMachine Learning
The gist
Many systems change over time but we only get to see scattered snapshots of how they look. This makes it hard to understand how they evolve because we don’t see smooth timelines. The authors created DisKO, a method that looks at these snapshots as whole distributions instead of individual steps, and learns patterns to predict future changes more accurately. DisKO does this by turning complex data into simpler mathematical forms and then turning those back into full pictures, helping it predict far ahead with less error.
Open → 2609.34629v1

New way to solve physics problems with less computation

Efficient Message Passing for Partial Differential Equation Priors

Abstract: Prior information for real-world physical quantities is most elegantly expressed via partial differential equations (PDEs). In this paper, we propose a novel way to solve PDEs using probabilistic inference on a factor graph. In general, factor graphs provide a natural way to encode prior knowledge into a model as explicit factors; here, this knowledge is provided by a governing PDE, which narrows the solution space, while observed data further shape the posterior over the parameters. The approximate parameter posterior is inferred using message passing based on moment matching, without posterior sampling or global gradient-based optimization. We demonstrate our approach on the first-order advection and the second-order semi-linear Fisher-KPP equations, where it achieves predictive accuracy comparable to a standard baseline while providing structured predictive uncertainty. Moreover, the inferred posterior marginal means and uncertainty structure match more closely those obtained using Hamiltonian Monte Carlo than the evaluated mean-field variational inference baseline, while requiring up to 10x less training time in our experiments, with inference speed comparable to variational inference.

Sat 26 SeptMachine LearningArtificial Intelligence
The gist
Solving physical problems often involves complicated equations called partial differential equations (PDEs) that describe how things like heat or fluids behave. The authors present a new method to solve these equations using a graphical model called a factor graph, which helps combine the rules from the equations with observed data. Their approach finds approximate solutions faster than some common methods and also provides a way to understand the uncertainty in the solutions. They tested their method on two example PDEs and found it gave accurate and reliable predictions more efficiently than traditional techniques.
Open → 2609.32956v1

Stable neural method improves solving stiff physics equations

NEXT: Physics-Informed Neuro-Spectral Exponential Time Differencing Architectures

Abstract: Physics-Informed Neural Networks (PINNs) build neural representations of time-dependent PDE solutions, naturally incorporating physics knowledge and observational data, which makes them well suited to both forward and inverse PDE problems. PINNs, however, are known to suffer from spectral bias and lack of causality. Neuro-Spectral Architectures (NeuSA), a recently proposed alternative to PINNs, mitigate both issues, but their numerical integration becomes unstable for stiff differential equations arising in many relevant physical problems. This study proposes Neuro-Spectral Exponential Time Differencing Architectures (NEXT), which combines the spectral representation of the PDE solution in NeuSA with high-order exponential integrators. Within this approach, the linear stiff part of the vector field induced by the PDE is integrated exactly through matrix exponentials, while the possibly nonlinear remainder is modeled by a neural network. The effectiveness of NEXT is verified through benchmark experiments on a set of stiff PDEs, in which NEXT is stable and accurate while NeuSA diverges numerically. It is also shown that NEXT can be applied to inverse problems, where the model has to learn unknown parameters or boundary conditions from sparse data. All code used in this work is publicly available at: https://github.com/marcioh2m/next.git .

Fri 25 SeptMachine Learning
The gist
Some computer methods struggle to solve physics problems involving certain tricky equations known as stiff PDEs because these problems are hard to compute accurately and stably. The authors propose a new approach called NEXT that combines a special neural network way to represent solutions with advanced math techniques to handle the tough parts exactly. This makes their method more stable and accurate on difficult physics problems and also lets it learn unknown parameters from limited data. Their tests showed that NEXT outperforms previous related methods that became unstable.
Open → 2609.31539v1

Predicting joint outcomes with multivariate quantile regression networks

Multivariate quantile regression via Kolmogorov-Arnold Networks

Abstract: This paper introduces a novel algorithm for predicting conditional joint distributions of vector-valued targets in stochastic systems whose randomness is intrinsic rather than arising from observation errors or additive noise. Multivariate quantile regression also involves modeling conditional joint distributions but represents a less challenging task. It predicts the probability that vector-valued targets fall within predefined regions, identifies regions corresponding to predefined probability levels, or performs both tasks simultaneously. The proposed identification technique employs ensembles of Kolmogorov--Arnold networks (KANs) as flexible function approximators. Although the suggested technique is not theoretically restricted to KANs, KANs are particularly well suited to the proposed construction and are therefore used throughout this study. In addition to the training procedure, this work introduces a new discrepancy measure for joint distributions and a goodness-of-fit (GoF) test based on it. This GoF test was initially developed to validate and calibrate the proposed identification technique and is used here in an ad hoc manner. Although the test could be tabulated for broader use, such a tabulation is not pursued in this work. The test is also applicable more generally.

Sun 20 SeptMachine Learning
The gist
Predicting the probability of where several related outcomes might happen together is challenging, especially when randomness is part of how the system works rather than measurement flaws. This paper presents a new way to model such joint outcomes using special neural networks called Kolmogorov-Arnold networks (KANs). These networks help predict regions where outcomes fall with certain probabilities, improving understanding of complex randomness. The authors also introduce a new way to check how well their predictions match real data.
Open → 2609.23906v1

Physical knowledge improves historic data forecasting of groundwater levels

Physical knowledge on historical data matters more than enforcing physical constraints on the forecast

Abstract: Time series forecasting has seen signicant advancements with the emergence of new deep learning models. However, forecasting time series in applications involving physical processes remains a major challenge. Despite the apparition of Physics Informed Neural Networks (PINN), recent models do not estimate unobservable intermediate physical variables, which are important for domain experts to understand the target behavior. To this end, we propose a Physics Informed Recurrent Neural Network (PIRNN) which predicts, along the target, unobservable variables on both historic data and forecast target. This approach enhances the model robustness and results interpretation using domain knowledge. Our method is easily adaptable to any physical model using several equations, each having its own set of unobservable variables, to describe it-self. As a case study, we incorporate physical equations used for groundwater levels predictions by the physical model called Gardenia. This model uses transfers equations between reservoirs, optimized with data assimilation, to simulate the evolution of groundwater levels. Evaluation includes several well known neural network models and the Gardenia model compared on twelve real world datasets. In addition, we study the impact of each component through an ablation study. Our model outperforms other models on ve out of the twelve datasets and our ablation study underlines the importance of having a physical background in our time series forecasting task. Finally, the coherence of the physical variables predicted by our neural network is assessed by a domain expert.

Thu 17 SeptArtificial Intelligence
The gist
Forecasting things that depend on physical processes, like groundwater levels, is hard because some important parts can't be measured directly. The authors created a new method that predicts these hidden physical variables alongside the usual forecast, using knowledge about how the physical process works. This makes their predictions more reliable and easier to understand. They tested their approach on real groundwater data and found it often works better than other models, and experts confirmed the predictions make physical sense.
Open → 2609.19871v1

Physics-structured model restores corrupted soil loss factors

PhyRestore: Physics-Structured Latent-Factor Restoration

Abstract: Estimating temporal soil-loss change is challenging when physically meaningful input factors are noisy or corrupted, particularly because substantial changes are rare relative to the large number of locations exhibiting little change. We study this problem through the Revised Universal Soil Loss Equation (RUSLE) and introduce PhyRestore, a physics-structured latent-factor restoration framework. Rather than directly predicting soil-loss change or correcting a degraded physical estimate, PhyRestore restores corrupted physical factors and reconstructs temporal change through the known physical relationship. We evaluate PhyRestore in a watershed-scale bitemporal raster setting under isolated and simultaneous corruption of rainfall erosivity and cover management, comparing it with the degraded RUSLE estimate and Direct RF, XGBoost, MLP, and CNN models. Factor restoration improves high-magnitude recovery when the corrupted factors remain identifiable, but its advantage weakens under joint corruption, sparse positive extremes, and factor values outside the training support.

Thu 17 SeptMachine Learning
The gist
Predicting how soil loss changes over time is hard when the factors used in calculations are noisy or broken. The researchers introduce PhyRestore, a method that fixes these corrupted factors by using knowledge of the physics behind soil loss. Instead of guessing soil loss directly, PhyRestore repairs the basic factors and then calculates the change from those. Their tests show that this method works best when the corrupted factors are still recognizable, but less well when multiple factors are corrupted, or when unusual extreme values occur.
Open → 2609.19776v1

Accurate nitrous oxide emission predictions improved with hybrid neural network

Enhanced Agriculture-informed Neural Network by Domain Knowledge

Abstract: Accurate prediction of nitrous oxide (N2O) emissions from agriculture is important for assessing environmental impacts and supporting sustainable farming. However, prediction remains difficult because N2O emissions result from complex interactions among soil properties, climate, biochemical processes, and management practices, while high-quality observations are limited. Deep learning models can capture nonlinear relationships but often lack physical interpretability and may generalize poorly across environmental conditions. We propose the Knowledge-enhanced Agriculture-informed Neural Network (KAINN), a hybrid neural-mechanistic framework that extends the Agriculture-informed Neural Network by incorporating domain knowledge about fertilizer diffusion, soil respiration, and water-filled porosity. We evaluate KAINN using CNN, LSTM, and Transformer architectures across multiple growing seasons and input-feature configurations. The results show that KAINN generally provides lower root mean square error and mean absolute error and higher R-squared values than purely data-driven models and the original AINN. Analysis of the learned interfaces also shows smoother and more physically consistent parameter trajectories with reduced uncertainty. These findings demonstrate that incorporating environmental knowledge into neural networks can improve the reliability, interpretability, and generalization of agricultural N2O-emission predictions.

Wed 16 SeptMachine Learning
The gist
Predicting nitrous oxide emissions from farms is hard because many factors like soil, climate, and farming methods interact in complex ways, and data is scarce. The authors created a new kind of neural network, called KAINN, which uses both data and knowledge about soil and fertilizer behavior to make better predictions. Their tests showed it predicts emissions more accurately and consistently than models that only use data. This approach helps make predictions easier to understand and more reliable across different environments.
Open → 2609.19466v1