Pith. sign in

REVIEW 4 major objections 5 minor 1 cited by

ChemKANs for Combustion Chemistry Modeling and Acceleration

T0 review · 4 major / 5 minor · reviewed 2026-08-16 · deepseek-v4-flash

Challenge this review Re-run · record.json Download PDF Read on arXiv ↗ Review JSON

Pith's one-line read This paper claims that a 344-parameter physics-structured neural network can learn all hydrogen-air combustion chemistry and run stiff ODE integration twice as fast as a detailed mechanism.

desk verdict A solid chemistry-structured KAN-ODE paper with real but modest results; the homogeneous-reactor demonstrations hold up, but the abstract's turbulent-flow generalization claim is untested and the paper's own temperature grid shows why that matters. read the letter →

arxiv 2504.12580 v2 pith:COGSDACT submitted 2025-04-17 cs.LG physics.chem-ph

Benjamin C. Koenig , Suyong Kim , Sili Deng This is my paper · ORCID
classification cs.LGphysics.chem-ph
keywords ChemKANKolmogorov-Arnoldnetworksneuralordinarydifferentialequationscombustionkineticshydrogen-airstiffchemistrysurrogatenoiserobustnessoperatorsplitting
verification ladder T0 review T1 audit T2 compute T3 formal

The pith

A machine-rendered reading of the paper's core claim, the machinery that carries it, and where it could break.

The reading

ChemKAN is a neural ordinary differential equation whose gradient getter is a Kolmogorov-Arnold network reshaped to mirror combustion kinetics: a kinetic core maps the current thermochemical state to species production rates, and a thermodynamic superstructure forms the temperature rate as a linear combination of those rates plus a small correction term. The paper's central claim is that this chemistry-specific inductive bias makes the network unusually parameter-lean and robust: a 344-parameter ChemKAN reproduces all nine species and temperature profiles for hydrogen-air homogeneous reactors across a range of initial temperatures and equivalence ratios, and it does so roughly twice as fast as the detailed GRI-Mech 3.0 mechanism in an ODE solver. In a separate model-inference task, the same ODE-coupled architecture recovers the underlying biodiesel transesterification kinetics from data with up to 15% added noise, showing no overfitting even when the network is overparameterized, while a DeepONet baseline overfits. The payoff, if the claims hold, is that kinetic surrogates can be trained on cheap homogeneous-reactor data and then replace the stiff chemistry term in larger reacting-flow simulations, where chemistry evaluation is typically the dominant computational cost.

What carries the argument

The load-bearing object is the ChemKAN architecture itself: a KAN-ODE whose learnable gradient function is split into a kinetic core, $\mathrm{KAN}_{kin}$, mapping the full state $u=[Y_1,\dots,Y_m,T]$ to the $m$ species production rates, and a thermodynamic superstructure that computes $dT/dt$ as a linear combination of those rates, mirroring the energy equation, plus a single-layer KAN correction for thermophysical parameter variation. The kinetic core stacks an additive KAN layer and a LeanKAN layer whose multiplicative sublayer captures the products of concentrations that appear in Arrhenius rate laws. Training proceeds in two stages, kinetics first and thermodynamics second, uses forward sensitivity analysis instead of adjoint differentiation to avoid stiffness-driven instability, and optionally adds an element-conservation penalty to the loss. This structure carries the argument by forcing temperature evolution to be built out of species production, sharing information across all species and outputs, and letting the outer ODE integrator smooth over noisy data.

What would settle it

Replace detailed GRI-Mech 3.0 chemistry with the 344-parameter ChemKAN in a freely propagating hydrogen-air laminar flame calculation under operator splitting and compare the predicted laminar flame speed and extinction limits with the detailed mechanism; disagreement beyond the homogeneous-reactor error levels would refute the transferability claim, as would failure to ignite at the 987.5 K conditions where the paper already reports order-of-magnitude larger errors.

Watch

Extended reading notes

Core claim

The central claim is that encoding the known kinetic-thermodynamic coupling into a KAN-ODE produces a source-term surrogate that is simultaneously accurate, sparse, and resistant to overfitting. For hydrogen-air combustion trained on 35 homogeneous-reactor conditions (one condition withheld), a single 344-parameter ChemKAN predicts temperature and mass fractions of every species, including the low-concentration radicals H, HO2, and H2O2 that the MLP-based ChemNODE baseline omitted, and it reproduces ignition delays across the studied range. The same network generalizes over most of a finer 441-condition grid, with the paper explicitly reporting degraded performance at colder, slower-igniting conditions around 987.5 K. For the biodiesel model-inference case, the kinetic core alone extracts smooth underlying profiles from sparse noisy data, and its test error stays nearly flat as parameters are added, where the DeepONet comparator's test error diverges. The paper describes the measured 2x ODE-integration speedup as a conservative lower bound and argues that the single-network full-state output makes the surrogate suitable for coupling to flow solvers.

Load-bearing premise

The promise of accelerating real combustion simulations depends on the untested premise that a network trained on zero-dimensional homogeneous-reactor data remains accurate in multi-dimensional flows under operator splitting, where the source term is evaluated on the local thermochemical state alone, and that the 35-condition training grid adequately covers the thermochemical manifold.

Editorial extensions

If this is right

  • A single ChemKAN forward pass returns the source term for the entire thermochemical state, so surrogate models can keep all species, including minor radicals, rather than dropping them as the ChemNODE baseline did.
  • The 2x per-step ODE speedup should compound in multi-dimensional reacting flows, where chemical source-term evaluation is usually the dominant cost, if the homogeneous-reactor surrogate transfers under operator splitting.
  • The two-stage training recipe and the linear thermodynamic coupling generalize as a template for other stiff kinetic systems; the kinetic core alone already suffices for isothermal problems such as biodiesel transesterification.
  • The optional element-conservation penalty provides a physics-consistency check that should reduce mass-fraction drift in long-horizon integrations and in repeated operator-split calls.

Reading between the lines

Editorial extensions of the paper, not claims the author makes directly.

  • A test the paper leaves open is to embed the trained 344-parameter ChemKAN in a laminar or turbulent reacting flow under operator splitting and compare flame speed, extinction limits, and ignition against detailed chemistry; the homogeneous-reactor 2x speedup could shrink or grow depending on per-cell evaluation cost.
  • The degraded 987.5 K results suggest the 35-condition grid is thinnest exactly where ignition chemistry is most temperature-sensitive, so a non-uniform or adaptive training grid, rather than a denser global grid, would likely recover accuracy in that regime.
  • The thermodynamic superstructure is not hydrogen-specific: any reacting system whose temperature equation is a linear combination of species production rates plus a thermophysical correction could reuse the same split, so the architecture should transfer to other fuels, pyrolysis, or thermal-runaway problems.
  • Because the outer ODE integrator is what smooths the noisy training data, noise robustness may scale with trajectory length or stiffness; varying the integration window in the biodiesel task would test that mechanism directly.
Share X Bluesky LinkedIn Reddit HN

Signed reviews

No signed human review yet.

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, and a circularity audit.

Referee Report

4 major / 5 minor

Summary. The paper introduces ChemKANs, a variant of Kolmogorov-Arnold Network ordinary differential equations (KAN-ODEs) in which the gradient getter is split into a kinetic core and an optional thermodynamic superstructure, mirroring the structure of Eqs. (1) and (2). Two applications are studied: (i) inference of a biodiesel transesterification kinetic model from synthetic noisy data, compared against DeepONets, and (ii) acceleration of zero-dimensional hydrogen-air homogeneous reactor simulations, where a single 344-parameter ChemKAN is reported to reproduce all nine species and temperature profiles and to run about 2x faster than the detailed GRI-Mech-based chemistry in Arrhenius.jl. The paper claims robustness to added noise, resistance to overfitting, and generalizability of the learned source-term surrogate to larger-scale flow simulations.

Significance. If the results hold, the architecture is a meaningful contribution to scientific machine learning for combustion: it encodes known kinetic/thermodynamic coupling in the network structure, uses a single information-sharing network instead of ChemNODE's m+1 separate networks, and demonstrates strong parameter efficiency. The two-stage training procedure and forward sensitivity analysis are practical and clearly described. The paper also provides a useful demonstration that KAN-ODEs can be applied to stiff, multi-species systems rather than only small dynamical systems. However, the claimed generality to multidimensional turbulent combustion and the quantitative speedup are not demonstrated by the experiments, which are limited to homogeneous reactors; the paper's own Fig. 8(A) reveals fragile generalization at unseen low-temperature conditions.

major comments (4)
  1. [Sec. I and Sec. III B] The central acceleration claim is extrapolated from zero-dimensional homogeneous reactors to multidimensional operator-split flow solvers without any demonstration. The Introduction states that 'such a surrogate can be directly coupled to existing CFD or machine learning-based flow solvers', and Sec. III B asserts generalizability 'to other simulation conditions when coupled to flow solvers, including simple laminar flames and complex 2-D and 3-D turbulent combustion conditions', but no laminar flame, 2-D, or 3-D simulation is performed. The measured speedup is for 36 homogeneous reactor integrations only. This is a load-bearing gap because the abstract's claim that the solver is 'generalizable to larger-scale turbulent flow simulations' rests entirely on this untested transfer.
  2. [Sec. III B, Fig. 8(A)] Fig. 8(A) shows that at 987.5 K, an unseen condition lying inside the stated 950-1200 K initial-temperature range, reconstruction errors are roughly an order of magnitude worse than at neighboring training conditions (about 10^-3 versus 10^-4). This directly weakens the premise that the 35-condition training grid covers the thermochemical manifold needed for generalization to other conditions. The limitation paragraph in Sec. III C acknowledges that a non-uniform training grid with denser sampling in the cooler regions would be needed for CFD applications, but the abstract and Sec. III B still assert generalizability; the claims should be qualified to match this evidence.
  3. [Sec. III B, Table I] The reported 2.0x speedup is based on a single comparison of average solve times for 36 homogeneous reactor conditions in Arrhenius.jl. No error bars, no repeated trials, no wall-clock breakdown between gradient evaluation and integrator overhead, and no comparison at different tolerances or state distributions are reported. Since the speedup is a central quantitative claim and the paper explicitly connects it to 'potential for substantial acceleration unlocked by ChemKANs' in larger simulations, the measurement needs to be characterized more carefully, or the claim should be limited to the homogeneous-reactor configuration actually measured.
  4. [Sec. III A and Fig. 5(A)] The abstract's statement that ChemKANs exhibit 'no overfitting or model degradation in any of these training cases' is stronger than the evidence in Fig. 5(A). The noise-free testing MSE increases by roughly a factor of two from 0% to 15% added noise, which is a mild but real degradation, and the claim of 'no degradation' is therefore not supported by the reported metric. The narrative should be revised to say the degradation is small and much less than that of the DeepONet baseline.
minor comments (5)
  1. [Sec. II C 4, Eq. (22)] Eq. (22) contains an extra closing parenthesis in 'u_pred_k(t_j, θ))' that should be removed.
  2. [Sec. III B, Fig. 8 caption] The caption says '405 additional testing locations between the initial 36', while the text says 441 total initial conditions with 406 unseen; please reconcile these counts explicitly (35 training plus 1 withheld plus 405 additional gives 441, but this accounting should be stated).
  3. [Data availability] The data availability statement says the data are 'available on request'; for a machine-learning paper whose main claims are empirical, releasing code, trained models, and data generation scripts would greatly strengthen reproducibility.
  4. [Sec. II D 2] The text refers to 'the H2/O2 mechanism from GRI-Mech 3.0 (9 species, 29 reactions)', but GRI-Mech 3.0 is a 53-species mechanism; please clarify that a subset is used and specify which reactions are retained.
  5. [Sec. II C 5] The bullet 'T raining stage 1' has an erroneous space in 'Training'; this is a typo that should be corrected.

Circularity Check

0 steps flagged · score 0.0 of 10

No circularity found: ChemKAN is a trained surrogate whose accuracy and speedup are measured against independent Cantera/GRI-Mech reference solutions and external benchmarks.

full rationale

The paper does not claim a first-principles derivation of combustion chemistry; it trains a learned source-term surrogate. The governing equations (Eqs. 1–2) are used only as architectural inductive bias: Eqs. 13–15 mirror the kinetic/thermodynamic split and dT/dt is expressed as a linear combination of species rates plus a correction term. The network parameters are fit to externally generated Cantera/GRI-Mech 3.0 homogeneous-reactor trajectories, and the reported accuracy, ignition-delay comparisons, robustness-to-noise results, and 2.0× speedup are evaluated against independent reference data and solvers, including held-out conditions and the MLP-based ChemNODE comparison. No fitted parameter is renamed as a prediction, no target quantity is defined in terms of the network parameters, and no uniqueness theorem is imported to force the architecture choice. The self-citations to KAN-ODEs [29] and LeanKAN [35] supply the baseline framework and layer implementation, but the novel ChemKAN structure is the contribution and its performance is externally benchmarked. The admitted poor generalization at 987.5 K in Fig. 8(A) is a limitation of the training-grid coverage and a correctness concern for downstream CFD use, not a circular step.

Assumptions & free parameters 7 free parameters · 5 assumptions · 0 invented entities

The paper introduces no new physical entities; all inductive biases come from known kinetic and thermodynamic laws. The main contributed cost is the network architecture and training procedure, while the physical models used to generate data are taken from the cited literature.

free parameters (7)
  • Kinetic core network weights θkin = not disclosed, 344 parameters total in H2 case
    All weights in KANkin (Eq. 13) are learned from trajectory data; individual values are not reported.
  • Thermodynamic linear mapping θthermo = not disclosed, m scalar coefficients
    Learned coefficients in Eq. 14 approximate the enthalpy and specific-heat ratios in the energy equation.
  • Thermodynamic correction weights θcor = not disclosed
    Learned weights in the single-output KAN correction (Eq. 15) that compensate for temperature-mixture dependence of cp.
  • Multiplication hyperparameter nmu = 2 (biodiesel), 3 (H2)
    Hand-chosen per case; sets how many inputs are multiplied in the LeanKAN layer (Eq. 9).
  • Element-conservation weight αPINN = 1e-4
    Hand-chosen weighting of the PINN loss term in Eq. 18; used only in the H2 case.
  • RBF grid size N = 3
    Number of basis functions per activation (Eq. 11), chosen by hand.
  • Hidden layer width = 4 nodes (biodiesel), 3 nodes (H2)
    Bottleneck size of the kinetic core; chosen to target parameter sparsity.
assumptions (5)
  • domain assumption Zero-dimensional homogeneous reactor model describes the chemical kinetics (Eq. 1).
    The surrogate learns dYi/dt = f(Y,T) with no transport; this standard operator-splitting premise is needed for coupling to CFD.
  • domain assumption Energy conservation through Eq. 2, with temperature derivative a linear combination of species production rates weighted by enthalpies and specific heats.
    This structure is encoded into the thermodynamic superstructure (Eqs. 14-15).
  • standard math Kolmogorov-Arnold representation theorem ensures that univariate-function compositions can approximate the source term.
    Basis for using AddKAN and LeanKAN layers as gradient getters.
  • domain assumption The detailed mechanisms (GRI-Mech 3.0 for H2, the biodiesel mechanism of Ref. [48]) are ground truth for generating training data.
    Accuracy is measured against synthetic Cantera data, not against experimental combustion data.
  • standard math Forward sensitivity analysis with the Tsit5 integrator yields accurate gradients through stiff ODE solves.
    The training relies on this to backpropagate through the integrated state profiles.

how reviews work

0 comments
Cite this review

Pith. "Pith review of ChemKANs for Combustion Chemistry Modeling and Acceleration." pith.science (2026). https://pith.science/paper/COGSDACT

@misc{pith2026250412580,
  author       = {Pith},
  title        = {Pith review of: ChemKANs for Combustion Chemistry Modeling and Acceleration},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/COGSDACT}},
  note         = {Machine review of arXiv:2504.12580}
}
read the original abstract

Efficient chemical kinetic model inference and application in combustion are challenging due to large ODE systems and widely separated time scales. Machine learning techniques have been proposed to streamline these models, though strong nonlinearity and numerical stiffness combined with noisy data sources make their application challenging. Here, we introduce ChemKANs, a novel neural network framework with applications both in model inference and simulation acceleration for combustion chemistry. ChemKAN's novel structure augments the generic Kolmogorov Arnold Network Ordinary Differential Equations (KAN-ODEs) with knowledge of the information flow through the relevant kinetic and thermodynamic laws. This chemistry-specific structure combined with the expressivity and rapid neural scaling of the underlying KAN-ODE algorithm instills in ChemKANs a strong inductive bias, streamlined training, and higher accuracy predictions compared to standard benchmarks, while facilitating parameter sparsity through shared information across all inputs and outputs. In a model inference investigation, we benchmark the robustness of ChemKANs to sparse data containing up to 15% added noise, and superfluously large network parameterizations. We find that ChemKANs exhibit no overfitting or model degradation in any of these training cases, demonstrating significant resilience to common deep learning failure modes. Next, we find that a remarkably parameter-lean ChemKAN (344 parameters) can accurately represent hydrogen combustion chemistry, providing a 2x acceleration over the detailed chemistry in a solver that is generalizable to larger-scale turbulent flow simulations. These demonstrations indicate the potential for ChemKANs as robust, expressive, and efficient tools for model inference and simulation acceleration for combustion physics and chemical kinetics.

Figures

Figures reproduced from arXiv: 2504.12580 by the authors.

Figure 1
Figure 1. FIG. 1. Comparison of three kinetic modeling approaches discussed in this work. (A) DeepONet, where the initial ( [PITH_FULL_IMAGE:figures/full_fig_p005_1.png] view at source ↗
Figure 2
Figure 2. FIG. 2. Proposed ChemKAN structure to embed physical knowledge in KAN-ODEs for chemical kinetic modeling. The grey [PITH_FULL_IMAGE:figures/full_fig_p007_2.png] view at source ↗
Figure 3
Figure 3. FIG. 3. Ground truth and prediction by the learned ChemKAN model for an unseen test case. Each column shows the evolution [PITH_FULL_IMAGE:figures/full_fig_p010_3.png] view at source ↗
Figures from the paper (5 more)
Figure 4
Figure 4. Figure 4: FIG. 4. Neural convergence comparison between ChemKAN and DeepONet (no noise). (A) Training MSE results with varying [PITH_FULL_IMAGE:figures/full_fig_p011_4.png]
Figure 5
Figure 5. Figure 5: FIG. 5. ChemKAN and DeepONet training results with increasing amounts of synthetic noise added to the training data. (A) [PITH_FULL_IMAGE:figures/full_fig_p013_5.png]
Figure 6
Figure 6. Figure 6: FIG. 6. ChemKAN and DeepONet model results trained with 15% noise, compared against both the noisy training data (blue [PITH_FULL_IMAGE:figures/full_fig_p014_6.png]
Figure 7
Figure 7. Figure 7: FIG. 7. KAN-ODE reconstruction of homogeneous reactor results. (A, B, C) temperature, species reconstructed here that [PITH_FULL_IMAGE:figures/full_fig_p015_7.png]
Figure 8
Figure 8. Figure 8: FIG. 8. Evaluation of the proposed ChemKAN framework for various conditions. (A) ChemKAN reconstruction error at 35 [PITH_FULL_IMAGE:figures/full_fig_p016_8.png]

Discussion (0). Continue with ORCID to comment.

Forward citations

Cited by 1 Pith paper

Reviewed papers in the Pith corpus that reference this work. Sorted by Pith novelty score. Full citation record

  1. SPIN-ODE: Stiff Physics-Informed Neural ODE for Chemical Reaction Rate Estimation

    cs.LG 2025-05 conditional novelty 6.0 of 10

    SPIN-ODE recovers chemical reaction rate coefficients from stiff ODE concentration trajectories using a three-stage neural ODE plus chemical reaction neural network pipeline.

Reference graph

Works this paper leans on

69 extracted references · 35 canonical work pages · cited by 1 Pith paper

  1. [1]

    II A) through a physics-inspired separation between the system’s parameter- 4 ization and the solution coordinate (Fig

    Deep Operator Network (DeepONet) As one of the more popular architectures for combustion applications [14, 18, 21, 22], DeepONets [20] learn the chemical kinetic system (outlined in Sec. II A) through a physics-inspired separation between the system’s parameter- 4 ization and the solution coordinate (Fig. 1(A)). Specifically, the solution is learned using...

    work page

  2. [2]

    3, rather than the direct thermochemical states over time u(t)

    ChemNODEs In contrast to the DeepONet approach, ChemNODE [17] aims to learn the source terms f in Eq. 3, rather than the direct thermochemical states over time u(t). However, training challenges were observed [17] when testing a single MLP network withm+1 inputs and m+1 outputs (m for Y and 1 forT ). This led to the development of a segregated model for e...

    work page

  3. [3]

    Vanilla KAN-ODEs The KAN-ODEs framework was proposed by Koenig et al. [29] to model a dynamical system in the form of differential equations, where the gradient function is replaced by a KAN network of L layers, du dt = KAN (u (t), θ) = (ΨL−1◦ ΨL−2◦···◦ Ψ1◦ Ψ0) (u (t)), (6) where KAN is the KAN representation of the system equation, parameterized by θ. Th...

    work page

  4. [4]

    on” or “off

    Novel chemistry KAN (ChemKAN) architecture Here, we design a novel ChemKAN architecture through a unique composition of AddKAN and LeanKAN layers that shows invariance to the number of species by combining all model behavior into a single network architecture (see Fig. 1). The standard KAN-ODE architecture has the same dimensions of inputs and outputs ( m...

    work page

  5. [5]

    II C 1 to leverage both the physics-guided ChemKAN architecture and the proven dynamical system benefits offered by the KAN-ODE integration

    Network structure details The ChemKAN structure proposed here is combined with the KAN-ODE framework detailed in Sec. II C 1 to leverage both the physics-guided ChemKAN architecture and the proven dynamical system benefits offered by the KAN-ODE integration. For nomenclature consistency, we henceforth refer to both the newly designed network 7 + 𝒅𝒀𝒎/𝒅𝒕𝒅𝒀𝟐...

    work page

  6. [6]

    m + 1, in Stage 2

    Loss function We define the loss function used for ChemKAN training as L (θ) =LMSE (θ) +LPINN (θ) = 1 n∗ NtX j=1 n∗ X k=1 ˆupred k (tj, θ)− ˆuobs k (tj) 2 | {z } MSE, variable n∗ +αPINN NeX i=1 NtX j=1 mX k=1 Nk iWi(Y pred k,j −Y pred k,1 ) Wk | {z } optional element conservation , where n∗ = m, in Stage 1. m + 1, in Stage 2. (18) 8 Here ˆuk denotes the k...

    work page

  7. [7]

    Training Process We highlight again that this architecture incorporates significant inductive bias by separating the inference of the kinetic behavior and thermodynamic behavior into distinct layers and activations within the complete structure, giving the network an explicit functional coupling between the temperature rate and the species rates to ease i...

    work page

  8. [8]

    [48] and previously modeled in Refs

    Model inference from noisy data – biodiesel production Biodiesel production involves the transesterification of branched triglyceride molecules (TG) with methanol into straight-chain methyl ester molecules, described in Ref. [48] and previously modeled in Refs. [5, 49]. A motivating goal here is to evaluate the ChemKAN framework’s capability to learn the ...

    work page

Show all 69 references
  1. [9]

    Model acceleration – hydrogen-air combustion Despite its relatively simple chemistry, hydrogen-air combustion is subject to numerical stiffness, leading to high computational costs and numerical instability in the ODE solver. The main goal of the ChemKAN here is to achieve hig...

  2. [10]

    13) of the ChemKAN comprises two layers: one AddKAN and one LeanKAN, as discussed in Sec

    Model performance of the proposed ChemKAN architecture The kinetic core (Eq. 13) of the ChemKAN comprises two layers: one AddKAN and one LeanKAN, as discussed in Sec. II C 2. Note that the thermodynamic superstructure (Eq. 15) is not applied in this problem as the process is i...

  3. [11]

    Here, different pa- rameter sizes were investigated by changing the number of nodes in the hidden layers

    Neural scaling with noise-free data Neural scaling is an effective method of measuring parameter efficiencies in neural networks. Here, different pa- rameter sizes were investigated by changing the number of nodes in the hidden layers. Training and test losses were evaluated f...

  4. [12]

    Phy-ChemNODE

    Benefits of latent-dynamics modeling: noisy inference comparison To gain a deeper understanding of how ChemKAN outperforms DeepONet, we further test the training behaviors of ChemKAN and DeepONet with noisy data. Early stopping is carried out for both network structures at 10,...

  5. [13]

    C. W. Gao, J. W. Allen, W. H. Green, R. H. West, Reaction Mechanism Generator: Automatic construction of chemical kinetic mechanisms, Computer Physics Communications 203 (2016) 212–225. doi:10.1016/j.cpc.2016.02.013

    2016 doi

  6. [14]

    T. Lu, C. K. Law, Toward accommodating realistic fuel chemistry in large-scale computations, Progress in Energy and Combustion Science 35 (2) (2009) 192–215. doi:10.1016/j.pecs.2008.10.002

    2009 doi

  7. [15]

    S. Kim, W. Ji, S. Deng, Y. Ma, C. Rackauckas, Stiff Neural Ordinary Differential Equations, Chaos: An Interdisciplinary Journal of Nonlinear Science 31 (9) (2021) 093122. doi:10.1063/5.0060697

    2021 doi

  8. [16]

    M. Ihme, W. T. Chung, A. A. Mishra, Combustion machine learning: Principles, progress and prospects, Progress in Energy and Combustion Science 91 (2022) 101010. doi:10.1016/j.pecs.2022.101010

    2022

  9. [17]

    W. Ji, S. Deng, Autonomous Discovery of Unknown Reaction Pathways from Data by Chemical Reaction Neural Network, The Journal of Physical Chemistry A 125 (4) (2021) 1082–1092. doi:10.1021/acs.jpca.0c09316

    2021 doi

  10. [18]

    B. C. Koenig, P. Zhao, S. Deng, Accommodating physical reaction schemes in DSC cathode thermal stability analysis using chemical reaction neural networks, Journal of Power Sources 581 (2023) 233443. doi:10.1016/j.jpowsour.2023.233443

    2023

  11. [19]

    B. C. Koenig, H. Chen, Q. Li, P. Zhao, S. Deng, Uncertain lithium-ion cathode kinetic decomposition modeling via Bayesian chemical reaction neural networks, Proceedings of the Combustion Institute 40 (1) (2024) 105243. doi:10.1016/j.proci. 2024.105243

    2024

  12. [20]

    B. C. Koenig, P. Zhao, S. Deng, Comprehensive thermal-kinetic uncertainty quantification of lithium-ion battery thermal runaway via bayesian chemical reaction neural networks, Chemical Engineering Journal 507 (2025) 160402. doi:10.1016/ j.cej.2025.160402

    2025

  13. [21]

    Hoffmann, C

    M. Hoffmann, C. Fr¨ ohner, F. No´ e, Reactive SINDy: Discovering governing reactions from concentration data, The Journal of Chemical Physics 150 (2) (2019) 025101. doi:10.1063/1.5066099

    2019 doi

  14. [22]

    Langary, Z

    D. Langary, Z. Nikoloski, Inference of chemical reaction networks based on concentration profiles using an optimization framework, Chaos: An Interdisciplinary Journal of Nonlinear Science 29 (11) (2019) 113121. doi:10.1063/1.5120598

    2019 doi

  15. [23]

    S. Kim, S. Deng, Learning reaction-transport coupling from thermal waves, Nature Communications 15 (1) (2024) 9930. doi:10.1038/s41467-024-54177-2

    2024 doi

  16. [24]

    S. Kim, S. Deng, Inference of chemical kinetics and thermodynamic properties from constant-volume combustion of ener- getic materials, Chemical Engineering Journal 469 (2023) 143779. doi:10.1016/j.cej.2023.143779

    2023

  17. [25]

    Q. Li, H. Chen, B. C. Koenig, S. Deng, Bayesian chemical reaction neural network for autonomous kinetic uncertainty quantification, Physical Chemistry Chemical Physics 25 (5) (2023) 3707–3717. doi:10.1039/D2CP05083H

    2023 doi

  18. [26]

    B. C. Koenig, S. Deng, Multi-target active subspaces generated using a neural network for computationally efficient turbulent combustion kinetic uncertainty quantification in the flamelet regime, Combustion and Flame 258 (2023) 113015. doi:10.1016/j.combustflame.2023.113015

    2023

  19. [27]

    Alqahtani, T

    S. Alqahtani, T. Echekki, A data-based hybrid model for complex fuel chemistry acceleration at high temperatures, Combustion and Flame 223 (2021) 142–152. doi:10.1016/j.combustflame.2020.09.022

    2021 doi

  20. [28]

    K. S. Jung, B. S. Soriano, J. H. Chen, M. Khalil, A Hessian-based transfer learning approach for artificial neural networks based chemical kinetics with a sparse dataset, Proceedings of the Combustion Institute 40 (1) (2024) 105390. doi: 10.1016/j.proci.2024.105390

    2024

  21. [29]

    Owoyele, P

    O. Owoyele, P. Pal, ChemNODE: A neural ordinary differential equations framework for efficient chemical kinetic solvers, Energy and AI 7 (2022) 100118. doi:10.1016/j.egyai.2021.100118

    2022

  22. [30]

    Kumar, T

    A. Kumar, T. Echekki, Combustion chemistry acceleration with DeepONets, Fuel 365 (2024) 131212. doi:10.1016/j. fuel.2024.131212

    2024

  23. [31]

    R. T. Q. Chen, Y. Rubanova, J. Bettencourt, D. Duvenaud, Neural Ordinary Differential Equations, arxiv preprint arXiv:1806.07366 (2019). doi:10.48550/arXiv.1806.07366

    2019 doi arXiv

  24. [32]

    L. Lu, P. Jin, G. Pang, Z. Zhang, G. E. Karniadakis, Learning nonlinear operators via DeepONet based on the universal ap- proximation theorem of operators, Nature Machine Intelligence 3 (3) (2021) 218–229. doi:10.1038/s42256-021-00302-5

    2021 doi

  25. [33]

    B. C. Koenig, W. Ji, S. Deng, Kinetic subspace investigation using neural network for uncertainty quantification in nonpremixed flamelets, Proc. Combust. Inst. 39 (4) (2023) 5229–5238

    2023

  26. [34]

    Goswami, A

    S. Goswami, A. D. Jagtap, H. Babaee, B. T. Susi, G. E. Karniadakis, Learning stiff chemical kinetics using extended deep neural operators, Computer Methods in Applied Mechanics and Engineering 419 (2024) 116674. doi:10.1016/j.cma. 2023.116674

    2024

  27. [35]

    Z. Liu, Y. Wang, S. Vaidya, F. Ruehle, J. Halverson, M. Soljaˇ ci´ c, T. Y. Hou, M. Tegmark, KAN: Kolmogorov-Arnold Networks, arXiv preprint arXiv:2404.19756 (2024). doi:10.48550/arXiv.2404.19756

    2024 doi arXiv

  28. [36]

    Patra, S

    S. Patra, S. Panda, B. K. Parida, M. Arya, K. Jacobs, D. I. Bondar, A. Sen, Physics Informed Kolmogorov-Arnold Neural Networks for Dynamical Analysis via Efficent-KAN and WAV-KAN, arxiv preprint arXiv:2407.18373 (2024). doi: 10.48550/arXiv.2407.18373. 20

    2024 doi arXiv

  29. [37]

    C. Guo, L. Sun, S. Li, Z. Yuan, C. Wang, Physics-informed Kolmogorov-Arnold Network with Chebyshev Polynomials for Fluid Mechanics, arxiv preprint arXiv:2411.04516 (2024). doi:10.48550/arXiv.2411.04516

    2024 doi arXiv

  30. [38]

    A. A. Howard, B. Jacob, S. H. Murphy, A. Heinlein, P. Stinis, Finite basis Kolmogorov-Arnold networks: domain decom- position for data-driven and physics-informed problems, arxiv preprint arXiv:2406.19662 (2024). doi:10.48550/arXiv. 2406.19662

    2024 doi

  31. [39]

    Kundu, A

    A. Kundu, A. Sarkar, A. Sadhu, Kanqas: Kolmogorov-arnold network for quantum architecture search, EPJ Quantum Technology 11 (1) (2024) 76

    2024

  32. [40]

    Kashefi, Kolmogorov–arnold pointnet: Deep learning for prediction of fluid fields on irregular geometries, Computer Methods in Applied Mechanics and Engineering 439 (2025) 117888

    A. Kashefi, Kolmogorov–arnold pointnet: Deep learning for prediction of fluid fields on irregular geometries, Computer Methods in Applied Mechanics and Engineering 439 (2025) 117888

    2025

  33. [41]

    B. C. Koenig, S. Kim, S. Deng, KAN-ODEs: Kolmogorov–Arnold network ordinary differential equations for learning dynamical systems and hidden physics, Computer Methods in Applied Mechanics and Engineering 432 (2024) 117397. doi:10.1016/j.cma.2024.117397

    2024

  34. [42]

    Azimi, L

    S. Azimi, L. Spekking, K. Staˇ nkov´ a, Kolmogorov-Arnold Networks and Evolutionary Game Theory for More Personalized Cancer Treatment, arXiv preprint arXiv:2501.07611 (2025). doi:10.48550/arXiv.2501.07611

    2025 doi arXiv

  35. [43]

    H. R. Sezavar, S. Hasanzadeh, Integrating Weibull analysis and KAN-ODEs for enhanced flashover prediction in contam- inated composite insulators, Electric Power Systems Research 244 (2025) 111584. doi:10.1016/j.epsr.2025.111584

    2025

  36. [44]

    H. Shen, C. Zeng, J. Wang, Q. Wang, Reduced Effectiveness of Kolmogorov-Arnold Networks on Functions with Noise, arXiv preprint arXiv:2407.14882 (2024). doi:10.48550/arXiv.2407.14882

    2024 doi arXiv

  37. [45]

    R. J. Kee, M. E. Coltrin, P. Glarborg, Chemically reacting flow: theory and practice, John Wiley & Sons, 2005

    2005

  38. [46]

    Li, Kolmogorov-Arnold Networks are Radial Basis Function Networks, arxiv preprint arXiv:2405.06721 (2024)

    Z. Li, Kolmogorov-Arnold Networks are Radial Basis Function Networks, arxiv preprint arXiv:2405.06721 (2024). doi: 10.48550/arXiv.2405.06721

    2024 doi arXiv

  39. [47]

    B. C. Koenig, S. Kim, S. Deng, LeanKAN: A Parameter-Lean Kolmogorov-Arnold Network Layer with Improved Memory Efficiency and Convergence Behavior, Neural Networks 192 (2025) 107883. doi:https://doi.org/10.1016/j.neunet. 2025.107883

    2025

  40. [48]

    Q. Qiu, T. Zhu, H. Gong, L. Chen, H. Ning, ReLU-KAN: New Kolmogorov-Arnold Networks that Only Need Matrix Addition, Dot Multiplication, and ReLU, arXiv preprint arXiv:2406.02075 (2024). doi:10.48550/arXiv.2406.02075

    2024 doi arXiv

  41. [49]

    H.-T. Ta, A. Tran, AF-KAN: Activation Function-Based Kolmogorov-Arnold Networks for Efficient Representation Learn- ing, arXiv preprint arXiv:2503.06112 (2025). doi:10.48550/arXiv.2503.06112

    2025 doi arXiv

  42. [50]

    Ta, D.-Q

    H.-T. Ta, D.-Q. Thai, A. B. S. Rahman, G. Sidorov, A. Gelbukh, FC-KAN: Function Combinations in Kolmogorov-Arnold Networks, arXiv preprint arXiv:2409.01763 (2025). doi:10.48550/arXiv.2409.01763

    2025 doi

  43. [51]

    Z. Liu, P. Ma, Y. Wang, W. Matusik, M. Tegmark, KAN 2.0: Kolmogorov-Arnold Networks Meet Science, arxiv preprint arXiv:2408.10205 (2024). doi:10.48550/arXiv.2408.10205

    2024 doi arXiv

  44. [52]

    Retrieved from https://github.com/Blealtan/efficient-kan (2024)

    Blealtan, efficient-kan, GitHub repository. Retrieved from https://github.com/Blealtan/efficient-kan (2024)

    2024

  45. [53]

    Puri, KolmogorovArnold.jl, GitHub repository

    V. Puri, KolmogorovArnold.jl, GitHub repository. Retrieved from https://github.com/vpuri3/KolmogorovArnold.jl. (2024)

    2024

  46. [54]

    Ramachandran, B

    P. Ramachandran, B. Zoph, Q. V. Le, Searching for Activation Functions, arXiv preprint arXiv:1710.05941 (2017). doi: 10.48550/arXiv.1710.05941

    2017 doi arXiv

  47. [55]

    Raissi, P

    M. Raissi, P. Perdikaris, G. E. Karniadakis, Physics-informed neural networks: A deep learning framework for solving forward and inverse problems involving nonlinear partial differential equations, Journal of Computational Physics 378 (2019) 686–707. doi:10.1016/j.jcp.2018.10.045

    2019 doi

  48. [56]

    Kumar, A

    T. Kumar, A. Kumar, P. Pal, A physics-informed autoencoder-neuralode framework (phy-chemnode) for learning complex fuel combustion kinetics, in: Proceedings of the NeurIPS Workshop on Machine Learning and the Physical Sciences, New Orleans, LA, USA, 2024

    2024

  49. [57]

    W. Ji, W. Qiu, Z. Shi, S. Pan, S. Deng, Stiff-PINN: Physics-Informed Neural Network for Stiff Chemical Kinetics, The Journal of Physical Chemistry A 125 (36) (2021) 8098–8106. doi:10.1021/acs.jpca.1c05102

    2021 doi

  50. [58]

    Tsitouras, Runge–Kutta pairs of order 5(4) satisfying only the first column simplifying assumption, Computers & Mathematics with Applications 62 (2) (2011) 770–775

    C. Tsitouras, Runge–Kutta pairs of order 5(4) satisfying only the first column simplifying assumption, Computers & Mathematics with Applications 62 (2) (2011) 770–775. doi:10.1016/j.camwa.2011.06.002

    2011 doi

  51. [59]

    D. P. Kingma, J. Ba, Adam: A Method for Stochastic Optimization, arXiv preprint arXiv:1412.6980 (2017). doi:10. 48550/arXiv.1412.6980

    2017 doi arXiv

  52. [60]

    Darnoko, M

    D. Darnoko, M. Cheryan, Kinetics of palm oil transesterification in a batch reactor, Journal of the American Oil Chemists’ Society 77 (12) (2000) 1263–1267. doi:10.1007/s11746-000-0198-y

    2000 doi

  53. [61]

    S. C. Burnham, D. P. Searson, M. J. Willis, A. R. Wright, Inference of chemical reaction networks, Chemical Engineering Science 63 (4) (2008) 862–873. doi:10.1016/j.ces.2007.10.010

    2008 doi

  54. [62]

    D. G. Goodwin, H. K. Moffat, I. Schoegl, R. L. Speth, B. W. Weber, Cantera: An object-oriented software toolkit for chemical kinetics, thermodynamics, and transport processes, https://www.cantera.org, version 3.0.0 (2023). doi: 10.5281/zenodo.8137090

    2023 doi

  55. [63]

    G. P. Smith, D. M. Golden, M. Frenklach, N. W. Moriarty, B. Eiteneer, M. Goldenberg, C. T. Bowman, R. K. Hanson, S. Song, W. C. Gardiner, Jr., V. V. Lissianski, Z. Qin, Gri-mech 3.0, http://combustion.berkeley.edu/gri-mech/

  56. [64]

    W. Ji, X. Su, B. Pang, S. J. Cassady, A. M. Ferris, Y. Li, Z. Ren, R. Hanson, S. Deng, Arrhenius.jl: A Differentiable Combustion SimulationPackage, arXiv preprint arxiv:2107.06172 (2021). doi:10.48550/arXiv.2107.06172

    2021 doi arXiv

  57. [65]

    Lee, J.-K

    S. Lee, J.-K. Kim, J. Kim, T. Kim, J. Lee, HiPPO-KAN: Efficient KAN Model for Time Series Analysis, arXiv preprint arXiv:2410.14939 (2024). doi:10.48550/arXiv.2410.14939

    2024 doi arXiv

  58. [66]

    Z. Chen, X. Zhang, LSS-SKAN: Efficient Kolmogorov-Arnold Networks based on Single-Parameterized Function, arXiv preprint arXiv:2410.14951 (2024). doi:10.48550/arXiv.2410.14951. 21

    2024 doi

  59. [67]

    Huang, J

    W.-H. Huang, J. Jia, Y. Kong, F. Waqar, T.-H. Wen, M.-F. Chang, S. Yu, Hardware Acceleration of Kolmogorov- Arnold Network (KAN) for Lightweight Edge Inference, in: Proceedings of the 30th Asia and South Pacific Design Automation Conference, ASPDAC ’25, Association for Computi...

    2025

  60. [68]

    Moradzadeh, L

    A. Moradzadeh, L. Wawrzyniak, M. Macklin, S. G. Paliwal, UKAN: Unbound Kolmogorov-Arnold Network Accompanied with Accelerated Library, arXiv preprint arXiv:2408.11200 (2024). doi:10.48550/arXiv.2408.11200

    2024 doi

  61. [69]

    R. Qiu, Y. Miao, S. Wang, L. Yu, Y. Zhu, X.-S. Gao, PowerMLP: An Efficient Version of KAN, arXiv preprint arXiv:2412.13571 (2024). doi:10.48550/arXiv.2412.13571

    2024 doi arXiv

Pith tools

Reviewed August 16, 2026 · model on record in the stance chip above.