Pith. sign in

REVIEW 4 major objections 5 minor 1 cited by

Modelling Mosquito Population Dynamics using PINN-derived Empirical Parameters

T0 review · 4 major / 5 minor · reviewed 2026-08-11 · deepseek-v4-flash

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

Pith's one-line read The paper claims that a neural-network replacement for the pupa-development parameter in a mechanistic mosquito model, learned from weather and two years of trap counts, beats the fixed-form model on an eight-year validation.

desk verdict A genuine temporal holdout and a sensible choice of what to learn, but the same validation years also picked the architecture, so the headline peak metrics are conditional until a nested selection is run. read the letter →

arxiv 2412.07514 v3 pith:DM722C5A submitted 2024-12-10 physics.bio-ph cs.LGcs.NAmath.NA

Branislava Lalic , Dinh Viet Cuong , Mina Petric , Vladimir Pavlovic , Ana Firanj Sremac , Mark Roantree This is my paper · ORCID
classification physics.bio-phcs.LGcs.NAmath.NA
keywords mosquitopopulationdynamicsphysics-informedneuralnetworksinverseparameterestimationmechanisticODEmodelspupadevelopmentrateCulexpipienspeakpredictionhybriddynamicmodel
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

This paper argues that a physics-informed neural network (PINN) can replace one fixed empirical parameter in a mechanistic mosquito-population model with a data-learned function of weather, and that the resulting hybrid model predicts adult Culex pipiens abundance and peaks better than the temperature-only mechanistic model. The authors train a neural network to infer the pupa development rate $f_P$—the parameter their sensitivity analysis identifies as the most influential—from two years of weekly adult trap counts together with daily temperature, humidity, and precipitation. On an eight-year out-of-sample simulation, the hybrid Hy PopMosq model achieves lower population RMSE (0.18 vs 0.26) and better peak detection (precision 0.63 vs 0.09, recall 0.56 vs 0.13) than the baseline Dy PopMosq model. An ablation study reports that a branched Fourier-feature MLP with a smoothed absolute-value output activation is the best-performing configuration. If correct, the result would make weather-driven, machine-learned parameterisation a practical route to more accurate vector-abundance forecasts.

What carries the argument

The load-bearing object is the parameter network $\Theta(m; W_\theta)$, a neural network that maps a vector of daily and seven-day historical meteorological conditions plus the day of year to the pupa development rate $f_P$. It is trained jointly with a state network $U(t; W_U)$ by minimising a loss with a data term (weekly adult counts, with unobserved stages masked) and a physics term (residuals of the ten ODEs at collocation points). After training, $\Theta$ is detached from $U$ and supplies $f_P$ to the original ODE system for forward simulation. The architecture is a two-branch FourierMLP with GELU hidden activations and a SoftAbs output activation, chosen by ablations that show this configuration best avoids trivial-zero convergence and best captures annual periodicity.

What would settle it

Train the PINN with several random initialisations and two or more different architectures, then compare the inferred $f_P$ functions: if functions that fit the two training years equally well diverge strongly outside the training period, the inverse problem is non-identifiable and the validation gains are not robust. Alternatively, hold out one of the eight validation years or a second trap site entirely, retrain only on the first site's two years, and check whether the RMSE and peak-F1 advantages persist in the held-out data.

Watch

Extended reading notes

Core claim

The central claim is that the pupa development rate $f_P$ can be learned as a function of meteorological inputs by a PINN, and that substituting this learned parameter for the empirical temperature-only formula in the ten-stage ODE system gives better simulations of the adult blood-seeking population ($A_{b1}+A_{b2}$) over 2000–2007. The trained parameter network is the only change to the underlying Dy PopMosq model; once trained, it is frozen and used as a parameter-supplying function in forward ODE simulations. Across all years Hy PopMosq has lower RMSE for population and for growth rate, a standard deviation closer to the observed value, and higher peak recall and precision. The two worst years (2002 and 2007) remain difficult for both models, and the paper attributes those failures to environmental factors outside weather, such as temporary water retention. The architecture study attributes the hybrid's performance to Fourier features combined with a multi-branch structure, and to the SoftAbs activation preventing collapse to the trivial zero solution.

Load-bearing premise

The load-bearing assumption is that a unique, transferable pupa-development function can be recovered from two years of weekly adult trap counts at one site, with off-site weather data, and that this function stays valid for the eight-year validation period.

Editorial extensions

If this is right

  • Weather-driven learning of $f_P$ lowers population RMSE from 0.26 to 0.18 and raises peak F1 from 0.11 to 0.57 over the eight-year validation.
  • Because only the parameter function changes, the hybrid retains the interpretable stage-structured ODE description of mosquito biology.
  • The peak-detection gains imply the hybrid model is better positioned to time vector-control interventions than the temperature-only baseline.
  • The ablation results show that the two-branch Fourier feature structure, not simply network capacity, drives the advantage in capturing annual periodicity.
  • Both models miss sudden ecological events such as the 2002 and 2007 peaks, so the learned parameterisation does not remove the need for exogenous disturbance information.

Reading between the lines

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

  • A direct test of transferability would be to train the parameter network at Petrovaradin and apply it, without retraining, to a second Culex site with its own trap counts; the current validation is at the same location across years.
  • The same PINN-inverse procedure could be applied to the larva development rate $f_L = 1.65 f_P$ or to mortality rates; the paper's sensitivity analysis suggests $f_P$ is the most influential, but a multi-parameter version would test whether learned interactions improve or destabilise the ODE system.
  • The 2002 and 2007 failures point to an implicit assumption that weather alone drives the learned parameter; incorporating flood-retention or land-cover proxies is a natural, testable extension suggested by the paper's own discussion.
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 presents Hy PopMosq, a hybrid model that replaces the fixed empirical pupa development rate fP in the Dy PopMosq ODE model with a neural-network parameter function Θ(m; Wθ) learned via a physics-informed neural network from meteorological data and adult trap counts. The parameter network is trained on 2016-2017 Petrovaradin data and then used with the Dy PopMosq ODE to simulate adult abundance for 2000-2007. The paper reports lower RMSE (0.18 vs 0.26) and higher peak recall/precision/F1 (0.56/0.63/0.57 vs 0.13/0.09/0.11) for Hy PopMosq. Section 4 presents an ablation of architectures and output activations, motivating the choice of Branched FourierMLP and SoftAbs(ε=10^-4) that define the Hy PopMosq configuration.

Significance. If the comparison were clean, the paper would provide useful evidence that a weather-driven neural parameterisation of a single development rate can improve a mechanistic mosquito model's out-of-sample abundance and peak forecasts. The temporal holdout (train 2016-2017, test 2000-2007) is a genuine out-of-sample design, and the ablation is systematic. The authors should be credited for using a real multi-year dataset and for testing a hypothesis that is falsifiable. However, the reported test period was used to select the architecture and activation function, and the baseline comparison is not fully controlled because Dy PopMosq and Hy PopMosq use different initial conditions. These issues are fixable, but they currently make the headline quantitative claims conditional.

major comments (4)
  1. [Section 4 (Tables 2 and 3) vs Table 1] The Hy PopMosq configuration reported in Table 1 is the result of a model-selection process that used the 2000-2007 test-period metrics. In Sec. 4.1 the four architectures are compared "using the same experimental procedure as in Sec. 2", and Table 2 shows peak recall ranging from 0.146 (MLP) to 0.563 (Branched FourierMLP) and F1 from 0.101 to 0.567; Table 3 shows F1 ranging from 0.063 (ReLU) to 0.567 (SoftAbs with ε=10^-4). The chosen Branched FourierMLP and SoftAbs(ε=10^-4) are exactly the best performers on those test-period metrics, so Table 1's RMSE 0.18, recall 0.56, precision 0.63, and F1 0.57 are not an out-of-sample evaluation of a pre-specified model. The "generally outperforms" claim therefore needs a nested protocol: hold out a portion of 2000-2007 for architecture/activation selection, or select using training-period metrics, and only then evaluate the selected model on the remaining holdout. Reporting the test-period best as the headline result overstates the evidence.
  2. [Appendix C and Sec. 3.3] The comparison between Hy PopMosq and Dy PopMosq is not controlled with respect to initial conditions. The PINN simulations use initial conditions derived from the trained state network U, while Dy PopMosq uses a fixed initial condition of 300 for every state component, as stated in Appendix C and Sec. 2.2. Since the reported RMSE and peak metrics are computed on the simulated adult abundance, the hybrid model's advantage could partly reflect more favorable initialisation rather than the learned fP parameterisation. The authors should either run Dy PopMosq with the same data-derived initial conditions (or a range of initial conditions) and report the resulting metrics, or explicitly justify that the spin-up period (about 200 days) makes the initial condition irrelevant for the 7-8 year evaluation.
  3. [Sec. 4.2, final paragraph] The checkpoint-selection procedure is not applied on equal footing across activation functions. Models using ReLU and Softplus were selected at 11,000 and 12,000 training steps out of 300,000, when training loss had not been fully minimised, while Identity, Abs, and SoftAbs were selected at about 250,000 steps. This means the activation comparison in Table 3 confounds the activation choice with the number of training steps and the stopping criterion. Moreover, selecting the checkpoint by "best RMSE when simulating with PINN-learned parameters" on the training period is a training-based selection, which does not justify selecting the architecture or epsilon on the test period. The authors should report results for a fixed checkpoint rule (e.g., final step or best training loss) across all activations, and should clearly separate training-based selection from test-based evaluation.
  4. [Sec. 3.2 and Fig. 3 caption / Appendix C] The peak-detection metrics are computed with a 7-day, 0.2-prominence detector, but no sensitivity analysis or alternative peak definition is provided. Because the strongest improvement over Dy PopMosq is in peak recall/precision/F1 (0.56/0.63/0.57 vs 0.13/0.09/0.11), and because these metrics also drive the architecture and activation selection in Tables 2-3, the authors should report how peak metrics vary with detector parameters (window, prominence threshold) and with the normalization used. This would establish that the peak advantage is not an artifact of a particular peak-finding configuration.
minor comments (5)
  1. [Sec. 3.1 vs Table D.5] The text says validation uses daily data from 2000-2007, but Table D.5 lists annual metrics only for 2001-2007; please reconcile or clarify whether 2000 is included in the average.
  2. [Table 1] The row 'No. Peaks' is not defined in the validation methodology; please state whether it is the average number of detected peaks per year and how it is computed.
  3. [Eq. (4)] Equation (4) uses the symbol D (θ D Θ(m; Wθ)) which appears to be a typo for equality; also define all symbols in Eq. (5), where the letter 'e' appears in the Fourier feature definition.
  4. [Appendix C] Appendix C states the validation dataset has 7 years, while Sec. 3.1 and Table D.5 say the period is 2000-2007; please ensure the period counts are consistent throughout the manuscript.
  5. [Sec. 2.2] The phrase 'No bias reduction is applied in the original model' is unclear; please specify whether this refers to bias correction of meteorological inputs or statistical bias in the ODE simulation.

Circularity Check

1 steps flagged · score 4.0 of 10

The 2000-2007 test period is used to select the reported architecture and activation function, so the headline peak metrics are partly selection criteria rather than clean out-of-sample predictions.

  1. fitted input called prediction [Sec. 4.1-4.2 (Tables 2-3) vs Sec. 3.3 (Table 1)]
    "Using the same experimental procedure as in Sec. 2, we now investigate how the proposed addition of neural network architectures and the activation function affect the performance of the framework. Specifically, we first train PINNs using data from the training period ... Then, the trained parameter network is used to predict parameters for the test period ... The Branched FourierMLP used in Sec. 2 achieves the best overall performance, exhibiting the lowest RMSE and the highest scores in peak detection metrics."

    The validation period 2000-2007 is the same period on which the ablation study computes its metrics. Sec. 4.1 compares four architectures and Sec. 4.2 compares seven activation settings on the test period, selecting Branched FourierMLP and SoftAbs(eps=1e-4) as superior because they give the best RMSE and peak scores there. Table 1 then reports exactly this selected configuration as Hy PopMosq, with the same test-period peak recall 0.56, precision 0.63 and F1 0.57. Thus the headline peak-detection numbers are the criteria used to choose the model, not an independent out-of-sample evaluation; the test data have influenced which model is reported.

full rationale

The core inverse-modelling chain is not circular: the pupa development rate fP(m) is learned from the 2016-2017 training data via the PINN loss (Sec. 2.3.2) and then inserted into the Dy PopMosq ODE to simulate the separate 2000-2007 period (Sec. 3.1-3.3), so the temporal holdout is genuine and no equation reduces the reported simulation to its inputs. The self-citation to [28] provides ODE normalisation, gradient balancing and the general PINN construction; it is methodology, not the load-bearing evidence for the performance claim, so it does not constitute circularity. The main defect is validation circularity at the model-selection level: the ablation study in Sec. 4 evaluates architectures and activation functions on the test period, and the configuration reported in Sec. 2 and Table 1 was chosen because it scored best on that period. Consequently the headline peak-detection metrics are partly fitted to the test set rather than predicted from it. The RMSE advantage is more robust across configurations, so the central claim retains independent content, but the strongest peak-detection advantage is not a clean out-of-sample result. Score 4 reflects this partial, non-definitional circularity.

Assumptions & free parameters 5 free parameters · 5 assumptions · 1 invented entities

The central comparison rests on the Dy PopMosq model from [29] and the inverse PINN training from [28], plus a meteorological transfer regression and a tuned epsilon. The learned fP network absorbs most of the model's freedom; its identifiability and transferability are assumed rather than demonstrated.

free parameters (5)
  • PINN parameter-network weights W_theta = not reported (neural network with thousands of weights)
    Map weather inputs to fP; fitted on 2016-2017 trap counts, with no uncertainty or parameter values reported (Sec. 2.3, App. C).
  • SoftAbs epsilon = 1e-4 selected after comparing 1e-6 and 1e-2 in Sec. 4.2
    The 'rigorous validation' that fixes epsilon uses the same test-period simulation metrics later reported as performance; a tuned hyperparameter.
  • Linear regression transfer from Rimski Sancevi to Petrovaradin = coefficients not reported
    Reconstructs validation-period meteorology; fitted on 2016-2017 paired data (Sec. 3.1) with no error bars.
  • Training checkpoint step = 250,000 for Identity/Abs/SoftAbs; 11,000-12,000 for ReLU/Softplus (of 300,000)
    Final models are early-stopped checkpoints selected by training-period simulation RMSE, not converged networks; confounds the activation comparison (Sec. 4.2).
  • Initial conditions for validation simulations = taken from trained network U, values not reported
    First-year initial conditions come from the trained state network rather than the baseline 300 (App. C); different initial conditions would change the long-horizon comparisons.
assumptions (5)
  • domain assumption The 10-stage Dy PopMosq ODE system, with literature parameters and temperature-only empirical formulas, is a valid mechanistic description of Culex pipiens dynamics.
    Adopted from [29] as the baseline; the hybrid model only replaces one parameter (fP), so any structural error in the ODE is attributed to fP.
  • domain assumption The weather-to-fP relationship learned from 2016-2017 is stationary and transfers to 2000-2007.
    Validation applies the trained parameter network to historical weather without retraining or drift correction (Sec. 3.3).
  • domain assumption Weekly counts of adult blood-seeking stages (Ab1+Ab2) suffice to identify the pupa development rate in a 10-state inverse problem.
    All other stages are masked in the data loss (Eq. 9); the identifiability of fP from this partial observation is assumed, not analyzed.
  • standard math PINN optimization reaches a useful local minimum of the composite data plus physics loss.
    No convergence guarantee is given; the authors rely on checkpoint selection and gradient balancing (App. C).
  • domain assumption The linear regression between Rimski Sancevi and Petrovaradin meteorological data is unbiased for the validation period.
    Used to create the 2000-2007 weather forcing (Sec. 3.1); regression error is not propagated into the simulations.
invented entities (1)
  • Parameter network Theta(m; W_theta)
    purpose: Replaces the empirical pupa development-rate formula with a learned function of temperature, humidity, precipitation, and day of year.
    Introduced in Eq. 4 as the core novelty; its output is fitted to one site's data, so the network itself has no falsifiable handle outside this dataset.

how reviews work

0 comments
Cite this review

Pith. "Pith review of Modelling Mosquito Population Dynamics using PINN-derived Empirical Parameters." pith.science (2026). https://pith.science/paper/DM722C5A

@misc{pith2026241207514,
  author       = {Pith},
  title        = {Pith review of: Modelling Mosquito Population Dynamics using PINN-derived Empirical Parameters},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/DM722C5A}},
  note         = {Machine review of arXiv:2412.07514}
}
read the original abstract

Vector-borne diseases continue to pose a significant health threat globally with more than 3 billion people at risk each year. Despite some limitations, mechanistic dynamic models are a popular approach to representing biological processes using ordinary differential equations where the parameters describe the different development and survival rates. Recent advances in population modelling have seen the combination of these mechanistic models with machine learning. One approach is physics-informed neural networks (PINNs) whereby the machine learning framework embeds physical, biological, or chemical laws into neural networks trained on observed or measured data. This enables forward simulations, predicting system behaviour from given parameters and inputs, and inverse modelling, improving parameterisation of existing parameters and estimating unknown or latent variables. In this paper, we focus on improving the parameterisation of biological processes in mechanistic models using PINNs to determine inverse parameters. In comparing mechanistic and PINN models, our experiments offer important insights into the strengths and weaknesses of both approaches but demonstrated that the PINN approach generally outperforms the dynamic model. For a deeper understanding of the performance of PINN models, a final validation was used to investigate how modifications to PINN architectures affect the performance of the framework. By varying only a single component at a time and keeping all other factors constant, we are able to observe the effect of each change.

Figures

Figures reproduced from arXiv: 2412.07514 by the authors.

Figure 1
Figure 1. Hybrid Mosquito Population Model. Hy PopMosq enhances Dy PopMosq model by replacing its parameters with a neural network Θ learned through a PINN process. Two neural networks are trained jointly: state network U(t; WU ) and the parameter network Θ(m; WΘ), where the driver m = M(t) are provided. The training minimizes a data loss on observations Du = {(ui , ti)} and a physics loss on collocation times tj . Once train… view at source ↗
Figure 2
Figure 2. Multi-branch Fourier-MLP PINN Framework. Each branch receives a distinct group of inputs and passes them through a FourierMLP, which is a Fourier-feature layer followed by fully-connected layers. The branch outputs are summed and passed through the S o f tAbs activation function. function as the output activation function ϕo which we have called S o f tAbs. This novel function is shown in Eq. (6) where ϵ is determin… view at source ↗
Figure 3
Figure 3. Observed vs. simulated adult-mosquito abundance. Daily, normalised counts of blood-seeking adults (Ab1 + Ab2) are plotted for each calendar year. Circles mark peaks identified with a 7-day, 0.2-prominence detector. 9 [PITH_FULL_IMAGE:figures/full_fig_p009_3.png] view at source ↗
Figures from the paper (3 more)
Figure 4
Figure 4. Figure 4: Rate of adult mosquito population growth. Each panel shows the first difference (week-to-week increment) of the normalised adult population timeseries depicted in [PITH_FULL_IMAGE:figures/full_fig_p011_4.png]
Figure 5
Figure 5. Figure 5: Non-negativity Activation Functions. The Soft Abs functions with ϵ = 10−6 and ϵ = 10−4 appear very close in the plot, closely resembling the positive part of ReLU and the identity function. To address this issue, we propose using the absolute value function to impose n…
Figure 6
Figure 6. Figure 6: Parameter fP predictions with different output activation functions, training period the optimised models for these two activation functions may reside at local minima in the loss landscape. In contrast, models employing the Identity, Abs, and SoftAbs activation functi…

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. Redefining Influenza Transmission Seasonality Using the Novel Seasonality Index

    q-bio.QM 2025-01 reject novelty 4.0 of 10

    A temperature-range seasonality index shows only moderate, seasonally confounded correlations with influenza infection rates, and the claimed strong link between winter length and flu oscillation period is contradicte...

Reference graph

Works this paper leans on

49 extracted references · 42 canonical work pages · cited by 1 Pith paper

  1. [1]

    O. H. et. al., Mosquito-borne diseases and their control strategies: An overview focused on green synthesized plant-based metallic nanopar- ticles., Insects 14(3) (2023) PMID: 36975906. URL doi:10.3390/insects14030221

    work page doi:10.3390/insects14030221 2023

  2. [2]

    Becker, D

    N. Becker, D. Petri ´c, M. Zgomba, C. Boase, M. B. Madon, C. Dahl, A. Kaiser, Mosquitoes: identification, ecology and control, Springer Nature, 2020

    work page 2020

  3. [3]

    Erickson, S

    R. Erickson, S. Presley, L. Allen, K. Long, S. Cox, A stage-structured, aedes albopictus population model, Ecological Modelling 221 1273–1282

    work page

  4. [4]

    Cailly, A

    P. Cailly, A. Tran, T. Balenghien, G. L’Ambert, C. Toty, P. Ezanno, A climate-driven abundance model to assess mosquito control strategies, Ecological Modelling 227 7–17

    work page

  5. [5]

    Virgillito, M

    C. Virgillito, M. Manica, G. Marini, B. Caputo, A. Della Torre, R. Ros `a, Modelling arthropod active dispersal using partial di fferential equations: the case of the mosquito aedes albopictus, Ecological Modelling 456 109658

    work page

  6. [6]

    Frantz, H

    R. Frantz, H. Godinez, K. Martinez, W. Cuello, C. Manore, Age structured partial differential equations model for culex mosquito abundance, Ecological Modelling 494 110764

    work page

  7. [7]

    Otero, H

    M. Otero, H. Solari, N. Schweigmann, A stochastic population dynamics model for aedes aegypti: Formulation and application to a city with temperate climate, Bulletin of Mathematical Biology 68 1945–1974

    work page 1945

  8. [8]

    Edwards, E

    C. Edwards, E. Crone, Estimating abundance and phenology from transect count data with glms, Oikos 130 1335–1345

    work page

Show all 49 references
  1. [9]

    Tsantalidou, E

    A. Tsantalidou, E. Parselia, G. Arvanitakis, K. Kyratzi, S. Gewehr, A. Vakali, C. Kontoes, Mamoth: An earth observational data-driven model for mosquitoes abundance prediction, Remote Sensing 13 2557

  2. [10]

    Kinney, S

    A. Kinney, S. Current, J. Lega, Aedes-ai: Neural network models of mosquito abundance, PLoS Computational Biology 17 1009467

  3. [11]

    Joshi, C

    A. Joshi, C. Miller, Review of machine learning techniques for mosquito control in urban environments, Ecological Informatics 61 101241

  4. [12]

    Zhang, X

    M. Zhang, X. Wang, S. Tang, Integrating dynamic models and neural networks to discover the mechanism of meteorological factors on aedes population, PLoS Computational Biology 20 1012499

  5. [13]

    A. Tran, G. L’Ambert, G. Lacour, R. Beno ˆıt, M. Demarchi, M. Cros, P. Cailly, M. Aubry-Kientz, T. Balenghien, P. Ezanno, A rainfall- and temperature-driven abundance model for aedes albopictus populations, International Journal of Environmental Research and Public Health 10 1698–1719

  6. [14]

    Erraguntla, D

    M. Erraguntla, D. Dave, J. Zapletal, K. Myles, Z. Adelman, T. Pohlenz, M. Lawley, Predictive model for microclimatic temperature and its use in mosquito population modeling, Scientific Reports 11

  7. [15]

    Yamana, E

    T. Yamana, E. B. Eltahir, Incorporating the e ffects of humidity in a mechanistic model of anopheles gambiae mosquito population dynamics in the sahel region of africa, Parasites & Vectors 6

  8. [16]

    Da Re, W

    D. Da Re, W. Bortel, F. Reuss, R. M ¨uller, S. Boyer, F. Montarsi, S. Ciocchetta, D. Arnoldi, G. Marini, A. Rizzoli, G. L’Ambert, G. Lacour, C. Koenraadt, S. Vanwambeke, M. Marcantonio, Dynamaedes: a unified modelling framework for invasive aedes mosquitoes, Parasites & Vectors 15

  9. [17]

    Yamashita, S

    W. Yamashita, S. Das, G. Chapiro, Numerical modeling of mosquito population dynamics of aedes aegypti, Parasites & Vectors 11

  10. [18]

    Da Re, G

    D. Da Re, G. Marini, C. Bonannella, F. Laurini, M. Manica, N. Anicic, A. Albieri, P. Angelini, D. Arnoldi, F. Bertola, B. Caputo, C. Liberato, A. Della Torre, E. Flacio, A. Franceschini, F. Gradoni, P. Kadriaj, V . Lencioni, I. Del Lesto, F. Russa, R. Lia, F. Montarsi, D. Otra...

  11. [19]

    Steindorf, H

    V . Steindorf, H. B, N. Stollenwerk, A. Cevidanes, J. Barandika, P. Vazquez, A. Garc ´ıa-P´erez, M. Aguiar, Forecasting invasive mosquito abundance in the basque country, spain using machine learning techniques, Parasites & Vectors 18

  12. [20]

    Athni, M

    T. Athni, M. Childs, C. Glidden, E. Mordecai, Temperature dependence of mosquitoes: Comparing mechanistic and machine learning approaches, PLoS Neglected Tropical Diseases 18 0012488

  13. [21]

    Ferraguti, S

    M. Ferraguti, S. Argany, C. Mora-Rubio, D. Bravo-Barriga, F. Lope, A. Marzal, Landscape and climatic factors shaping mosquito abundance and species composition in southern spain: A machine learning approach to the study of vector ecology, Ecological Informatics 102860

  14. [22]

    Karniadakis, I

    G. Karniadakis, I. Kevrekidis, L. Lu, P. Perdikaris, S. Wang, L. Yang, Physics-informed machine learning, Nature Reviews Physics 3 422–440

  15. [23]

    Wesselkamp, N

    M. Wesselkamp, N. Moser, M. Kalweit, J. Boedecker, C. Dormann, Process-informed neural networks: A hybrid modelling approach to improve predictive performance and inference of neural networks in ecology and beyond, Ecology Letters 27

  16. [24]

    Oneto, D

    L. Oneto, D. Chicco, Eight quick tips for biologically and medically informed machine learning, PLoS Computational Biology 21 1012711

  17. [25]

    Aatif, J

    M. Aatif, J. Tiwari, A facile approach for enhancing device performance of excitonic solar cells with an innovative sno2/tcne electron transport layer, AIP Advances 10. 16

  18. [26]

    Lotfollahi, S

    M. Lotfollahi, S. Rybakov, K. Hrovatin, S. Hediyeh-Zadeh, C. Talavera-L´opez, A. Misharin, F. Theis, Biologically informed deep learning to query gene programs in single-cell atlases, Nature Cell Biology

  19. [27]

    X. Meng, G. Karniadakis, A composite neural network that learns from multi-fidelity data: Application to function approximation and inverse pde problems, Journal of Computational Physics 401 109020

  20. [28]

    Viet Cuong, B

    D. Viet Cuong, B. Lali ´c, M. Petri´c, N. Thanh Binh, M. Roantree, Adapting physics-informed neural networks to improve ode optimization in mosquito population dynamics, PLOS ONE 19 (12) (2024) 1–30. doi:10.1371/journal.pone.0315762. URL https://doi.org/10.1371/journal.pone.0315762

    2024 doi

  21. [29]

    Petric, Modelling the influence of meteorological conditions on mosquito vector population dynamics (diptera, culicidae

    M. Petric, Modelling the influence of meteorological conditions on mosquito vector population dynamics (diptera, culicidae

  22. [30]

    Focks, D

    D. Focks, D. Haile, E. Daniels, G. Mount, Dynamic life table model for aedes aegypti (diptera: Culicidae): Analysis of the literature and model development, Journal of Medical Entomology 30 1003–1017

  23. [31]

    Ezanno, M

    P. Ezanno, M. Aubry-Kientz, S. Arnoux, P. Cailly, G. L’Ambert, C. Toty, T. Balenghien, A. Tran, A generic weather-driven model to predict mosquito population dynamics applied to species of anopheles, culex and aedes genera of southern france, Preventive Veterinary Medicine 120 39–50

  24. [32]

    E. B. Vinogradova, Experimental investigation of the ecological factors causing diapause of the adults of blood-sucking mosquitoes (diptera, culicldae), Entomol. Obozr. 39 (1960) 327–340

    1960

  25. [33]

    Raissi, P

    M. Raissi, P. Perdikaris, G. Karniadakis, Physics-informed neural networks: A deep learning framework for solving forward and inverse problems involving nonlinear partial differential equations, Journal of Computational Physics 378 686–707

  26. [34]

    Tancik, P

    M. Tancik, P. Srinivasan, B. Mildenhall, S. Fridovich-Keil, N. Raghavan, U. Singhal, R. Ramamoorthi, J. Barron, R. Ng, Fourier features let networks learn high frequency functions in low dimensional domains, Neural Information Processing Systems 33 7537–7547

  27. [35]

    S. Wang, H. Wang, P. Perdikaris, On the eigenvector bias of fourier feature networks: From regression to solving multi-scale pdes with physics-informed neural networks, Computer Methods in Applied Mechanics and Engineering 384 113938

  28. [36]

    Hendrycks, K

    D. Hendrycks, K. Gimpel, Gaussian error linear units (gelus, preprint at. URL http://arxiv.org/abs/1606.08415

    arXiv

  29. [37]

    Kingma, J

    D. Kingma, J. Ba, A method for stochastic optimisation, preprint at. URL http://arxiv.org/abs/1412.6980

    arXiv

  30. [38]

    Paszke, S

    A. Paszke, S. Gross, F. Massa, A. Lerer, J. Bradbury, G. Chanan, T. Killeen, Z. Lin, N. Gimelshein, L. Antiga, A. Desmaison, A. Kopf, E. Yang, Z. DeVito, M. Raison, A. Tejani, S. Chilamkurthy, B. Steiner, L. Fang, J. Bai, C. S, Pytorch: An imperative style, preprint at. URL ht...

    1912 arXiv

  31. [39]

    Beck-Johnson, W

    L. Beck-Johnson, W. Nelson, K. Paaijmans, A. Read, M. Thomas, O. Bjørnstad, The e ffect of temperature on anopheles mosquito population dynamics and the potential for malaria transmission, PLoS ONE 8 79276

  32. [40]

    Loetti, N

    V . Loetti, N. Schweigmann, N. Burroni, Development rates, larval survivorship and wing length ofculex pipiens(diptera: Culicidae) at constant temperatures, Journal of Natural History 45 2203–2213

  33. [41]

    Pielke, Mesoscale meteorological modeling, Academic Press, New York, N

    P. Pielke, Mesoscale meteorological modeling, Academic Press, New York, N. Y

  34. [42]

    A. S. Krishnapriyan, A. Gholami, S. Zhe, R. M. Kirby, M. W. Mahoney, Characterizing possible failure modes in physics-informed neural networks (2021). arXiv:2109.01050

    2021 arXiv

  35. [43]

    Glorot, A

    X. Glorot, A. Bordes, Y . Bengio, Deep sparse rectifier neural networks, in: Proceedings of the fourteenth international conference on artificial intelligence and statistics, JMLR Workshop and Conference Proceedings, 2011, pp. 315–323

    2011

  36. [44]

    L. Lu, Y . Shin, Y . Su, G. E. Karniadakis, Dying relu and initialization: Theory and numerical examples, arXiv preprint arXiv:1903.06733 (2019)

    2019 arXiv

  37. [45]

    X. Jia, J. Willard, A. Karpatne, J. Read, J. Zwart, M. Steinbach, V . Kumar, Physics-guided machine learning for scientific discovery: An application in simulating lake temperature profiles, ACM/IMS Transactions on Data Science 2 1–26

  38. [46]

    Nisar, M

    K. Nisar, M. Anjum, M. Raja, M. Shoaib, Design of a novel intelligent computing framework for predictive solutions of malaria propagation model, PLoS ONE 19 0298451

  39. [47]

    Petri ´c, R

    D. Petri ´c, R. Bellini, E.-J. Scholte, L. Rakotoarivony, F. Scha ffner, Monitoring population and insecticide resistance of aedes albopictus in europe, Parasite. Vector 13 1–20

  40. [48]

    SFI /12/RC/2289 P2) and by the European Union (Grant Agreement No

    Acknowledgement This research is partially supported by a grant from Research Ireland (Grant No. SFI /12/RC/2289 P2) and by the European Union (Grant Agreement No. 101136578). The views and opinions expressed in this publication are solely those of the author(s) and do not nec...

    2024

  41. [49]

    List of Figures 1 Hybrid Mosquito Population Model

    Author contributions BL- design of dynamical systems hybridisation concept and main paper writing; DVC-contributed to solution design and development, performed all PINN numerical experiments, contributed to main paper writing; BL and DVC equally contributed to the realisation...

    2020

Pith tools

Reviewed August 11, 2026 · model on record in the stance chip above.