pith. machine review for the scientific record. sign in

arxiv: 2604.18083 · v1 · submitted 2026-04-20 · 💻 cs.LG · cs.AI

Recognition: unknown

Implicit neural representations as a coordinate-based framework for continuous environmental field reconstruction from sparse ecological observations

Agnieszka Pregowska , Hazem M. Kalaji
Authors on Pith no claims yet

Pith reviewed 2026-05-10 05:48 UTC · model grok-4.3

classification 💻 cs.LG cs.AI
keywords implicit neural representationsenvironmental field reconstructionspecies distribution modelingcontinuous fieldssparse ecological datacoordinate-based learningphenological dynamicsbiodiversity informatics
0
0 comments X

The pith

Implicit neural representations reconstruct continuous environmental fields directly from sparse coordinate observations.

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

The paper tests implicit neural representations as a coordinate-based framework that learns continuous spatial and spatio-temporal fields straight from input coordinates rather than from gridded data. It applies this approach to three ecological tasks drawn from open biodiversity records: species distribution reconstruction, phenological dynamics, and morphological segmentation. The authors report that the resulting representations remain stable, support resolution-independent queries, and carry predictable computational cost. These properties let the method sit alongside classical smoothers and tree-based models inside existing environmental pipelines. A sympathetic reader cares because many ecological datasets arrive sparse and irregular, where grid-based methods scale poorly and generalise unevenly across domains.

Core claim

Implicit neural representations can act as a flexible coordinate-based modelling layer that reconstructs continuous environmental fields from sparse and heterogeneous ecological observations, yielding stable interpolations and predictable run-time behaviour across species distribution, phenological, and segmentation scenarios without domain-specific architectural changes.

What carries the argument

Implicit neural representations (INRs) that map raw coordinate inputs through a neural network to output field values, enabling direct learning of continuous functions from irregular point observations.

If this is right

  • INRs can be dropped into environmental modelling pipelines as a representation layer for large, irregularly sampled datasets.
  • Resolution-independent querying becomes available for downstream tasks such as habitat suitability mapping or phenology forecasting.
  • Computational cost scales predictably with network size rather than with grid resolution, simplifying workflow planning.
  • The method complements rather than displaces classical smoothers and tree-based approaches in biodiversity informatics.
  • Exploratory analysis of spatio-temporal fields becomes feasible even when observations arrive at uneven densities.

Where Pith is reading between the lines

These are editorial extensions of the paper, not claims the author makes directly.

  • If the approach holds, it could extend to fusing multi-source ecological data streams where grid alignment is impractical.
  • Adding uncertainty estimates or physics-informed constraints to the basic INR architecture would be a natural next test for ecological reliability.
  • The same coordinate framework might transfer to other sparse environmental domains such as soil property mapping or pollutant dispersion without major redesign.
  • Scalability claims would be strengthened by direct timing comparisons against established methods on datasets exceeding current biodiversity sizes.

Load-bearing premise

The inductive bias of standard implicit neural representations is sufficient to capture ecological field structure from sparse, heterogeneous observations without domain-specific architectural changes or post-hoc regularization.

What would settle it

A clear failure would be if INR interpolations in held-out regions far from observations produce visibly incoherent spatial patterns or higher error than baseline smoothers on the same biodiversity datasets.

Figures

Figures reproduced from arXiv: 2604.18083 by Agnieszka Pregowska, Hazem M. Kalaji.

Figure 1
Figure 1. Figure 1: Boundary-F1 as a function of spatial tolerance for Leafsnap (global threshold). [PITH_FULL_IMAGE:figures/full_fig_p023_1.png] view at source ↗
Figure 2
Figure 2. Figure 2: SDM for Quercus robur based on GBIF records. Panels compare implicit neural [PITH_FULL_IMAGE:figures/full_fig_p025_2.png] view at source ↗
Figure 3
Figure 3. Figure 3: Two-dimensional phenology maps based on iNaturalist flowering observations. Im [PITH_FULL_IMAGE:figures/full_fig_p027_3.png] view at source ↗
Figure 4
Figure 4. Figure 4: Three-dimensional phenology model in lon-lat-DOY space learned from iNaturalist [PITH_FULL_IMAGE:figures/full_fig_p028_4.png] view at source ↗
read the original abstract

Reconstructing continuous environmental fields from sparse and irregular observations remains a central challenge in environmental modelling and biodiversity informatics. Many ecological datasets are heterogeneous in space and time, making grid-based approaches difficult to scale or generalise across domains. Here, we evaluate implicit neural representations (INRs) as a coordinate-based modelling framework for learning continuous spatial and spatio-temporal fields directly from coordinate inputs. We analyse their behaviour across three representative modelling scenarios: species distribution reconstruction, phenological dynamics, and morphological segmentation derived from open biodiversity data. Beyond predictive performance, we examine interpolation behaviour, spatial coherence, and computational characteristics relevant for environmental modelling workflows, including scalability, resolution-independent querying, and architectural inductive bias. Results show that neural fields provide stable continuous representations with predictable computational cost, complementing classical smoothers and tree-based approaches. These findings position coordinate-based neural fields as a flexible representation layer that can be integrated into environmental modelling pipelines and exploratory analysis frameworks for large, irregularly sampled datasets.

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, simulated authors' rebuttal, and a circularity audit. Tearing a paper down is the easy half of reading it; the pith above is the substance, this is the friction.

Referee Report

2 major / 0 minor

Summary. The manuscript evaluates implicit neural representations (INRs) as a coordinate-based framework for learning continuous spatial and spatio-temporal environmental fields directly from sparse, irregular coordinate inputs. It examines behavior across three scenarios—species distribution reconstruction, phenological dynamics, and morphological segmentation from open biodiversity data—focusing on interpolation, spatial coherence, and computational traits such as scalability and resolution-independent querying, and concludes that neural fields yield stable representations with predictable costs that complement classical smoothers and tree-based methods.

Significance. If the empirical claims are substantiated, the work could position coordinate-based neural fields as a flexible, scalable representation layer for large irregularly sampled ecological datasets, offering advantages in generalization and querying over grid-based approaches in biodiversity informatics and environmental modeling.

major comments (2)
  1. Abstract: The evaluation across three scenarios is described without any quantitative metrics, error bars, baseline comparisons, network architecture details, training procedures, or data splits, rendering the claims of 'stable continuous representations' and 'complementing classical smoothers' unverifiable from the provided text.
  2. Abstract (central claim): The assumption that the inductive bias of standard INRs is sufficient to capture non-stationary spatial autocorrelation and multi-scale structure in sparse ecological fields is not tested via ablations, domain-specific adaptations, or comparisons to adapted models; this is load-bearing for the complementarity conclusion but unsupported by evidence in the manuscript.

Simulated Author's Rebuttal

2 responses · 0 unresolved

We thank the referee for their constructive comments on our manuscript. We address each major comment point by point below, indicating where revisions will be made to the next version.

read point-by-point responses

  1. Referee: Abstract: The evaluation across three scenarios is described without any quantitative metrics, error bars, baseline comparisons, network architecture details, training procedures, or data splits, rendering the claims of 'stable continuous representations' and 'complementing classical smoothers' unverifiable from the provided text.

    Authors: We acknowledge that the abstract, being a concise summary, omits specific quantitative details. The full manuscript reports these elements in the Methods (Sections 3.1-3.3) and Results (Sections 4.1-4.3), including RMSE and correlation metrics with standard errors, comparisons against kriging, thin-plate splines, and random forests, SIREN/MLP architectures, Adam training schedules, and 70/30 spatial data splits. To improve verifiability, we will revise the abstract to include one or two key quantitative highlights and a brief mention of the evaluation protocol. revision: yes

  2. Referee: Abstract (central claim): The assumption that the inductive bias of standard INRs is sufficient to capture non-stationary spatial autocorrelation and multi-scale structure in sparse ecological fields is not tested via ablations, domain-specific adaptations, or comparisons to adapted models; this is load-bearing for the complementarity conclusion but unsupported by evidence in the manuscript.

    Authors: The manuscript evaluates unmodified standard INR architectures (as described in Section 3.1) to assess their baseline applicability, with empirical complementarity demonstrated through direct performance and scalability comparisons to classical methods rather than through claims of theoretical sufficiency. We did not conduct ablations on adapted variants because the study scope was limited to off-the-shelf INRs. We agree this leaves the handling of non-stationary multi-scale structure as an open question and will revise the Discussion to explicitly acknowledge the limitation, clarify that complementarity is observed rather than proven via inductive bias alone, and outline directions for domain-specific positional encodings. revision: partial

Circularity Check

0 steps flagged

No circularity in empirical evaluation of existing INR framework

full rationale

The manuscript is an empirical evaluation applying standard implicit neural representations to three ecological reconstruction scenarios using open biodiversity data. It reports interpolation behavior, spatial coherence, and computational characteristics without any derivation chain, first-principles prediction, or fitted quantity that reduces to its own inputs. No equations, self-definitional steps, or load-bearing self-citations appear in the abstract or described structure; the central claim rests on observed performance of an off-the-shelf coordinate-based model rather than on any constructed equivalence.

Axiom & Free-Parameter Ledger

0 free parameters · 1 axioms · 0 invented entities

The central claim rests on the domain assumption that ecological variables form continuous fields approximable by coordinate-based neural networks from sparse samples; no free parameters or invented entities are introduced in the abstract.

axioms (1)
  • domain assumption Ecological phenomena can be represented as continuous functions of spatial and temporal coordinates.
    This premise is required for the coordinate-based INR framework to be applicable to species distribution, phenology, and segmentation tasks.

pith-pipeline@v0.9.0 · 5466 in / 1171 out tokens · 36556 ms · 2026-05-10T05:48:06.906277+00:00 · methodology

discussion (0)

Sign in with ORCID, Apple, or X to comment. Anyone can read and Pith papers without signing in.

Reference graph

Works this paper leans on

42 extracted references · 35 canonical work pages

  1. [1]

    Bellard, C., Bertelsmeier, C., Leadley, P., Thuiller, W., & Courchamp, F. (2012). Impacts of climate change on the future of biodiversity. Ecology Letters, 15(4), 365-377. https://doi.org/10.1111/j.1461-0248.2011.01736.x

    work page doi:10.1111/j.1461-0248.2011.01736.x 2012

  2. [2]

    E., Mathieu B

    Boone, M. E., Mathieu B. (2019) Using INaturalist to Contribute Your Nature Observations to Science: WEC413 UW458, 6 2019, EDIS 2019 (4). Gainesville, FL:5. https://doi.org/10.32473/edis-uw458-2019

    work page doi:10.32473/edis-uw458-2019 2019

  3. [3]

    J., Morris, S

    Briscoe, N. J., Morris, S. D., Mathewson, P. D., Buckley, L. B., Jusup, M., Levy, O., Maclean, I. M. D., Pincebourde, S., Riddell, E. A., Roberts, J. A., Schouten, R., Sears, M. W., Kearney, M. R. (2023). Mechanistic forecasts of species responses to climate change: The promise of biophysical ecology. Global Change Biology 29, 1451--1470. https://doi.org/...

    work page doi:10.1111/gcb.16557 2023

  4. [4]

    Broussin, J., Mouchet, M., Goberville, E. (2024). Generating pseudo-absences in the ecological space improves the biological relevance of response curves in species distribution models, Ecological Modelling 498, 110865. https://doi.org/10.1016/j.ecolmodel.2024.110865

    work page doi:10.1016/j.ecolmodel.2024.110865 2024

  5. [5]

    E., Hari, C., Pellissier, L., and Karger, D

    Brun, P., Zimmermann, N. E., Hari, C., Pellissier, L., and Karger, D. N.(2022) Global climate-related predictors at kilometer resolution for the past and future, Earth Syst. Sci. Data, 14, 5573--5603, https://doi.org/10.5194/essd-14-5573-2022

    work page doi:10.5194/essd-14-5573-2022 2022

  6. [6]

    J., Barve, V., Belitz, M

    Di Cecco, G. J., Barve, V., Belitz, M. W. (2021) Brian J Stucky, Robert P Guralnick, Allen H Hurlbert, Observing the Observers: How Participants Contribute Data to iNaturalist and Implications for Biodiversity Science, BioScience 71, 11, 79--1188. https://doi.org/10.1093/biosci/biab093

    work page doi:10.1093/biosci/biab093 2021

  7. [7]

    Dyderski, M. K. (2022). Species Distribution. ACADEMIA. The magazine of the Polish Academy of Sciences, 2(74), 38-41. https://doi.org/10.24425/academiaPAS.2022.143444

    work page doi:10.24425/academiapas.2022.143444 2022

  8. [8]

    https://www.inaturalist.org/

  9. [9]

    and Roy, D.B

    Isaac, N.J.B., van Strien, A.J., August, T.A., de Zeeuw, M.P. and Roy, D.B. (2014), Statistics for citizen science: extracting signals of change from noisy ecological data. Methods Ecol Evol, 5, 1052--1060. https://doi.org/10.1111/2041-210X.12254

    work page doi:10.1111/2041-210x.12254 2014

  10. [10]

    J., Anderson, R

    Phillips, S. J., Anderson, R. P., Schapire, R. E. (2006) Maximum entropy modeling of species geographic distributions, Ecological Modelling, 190, 3--4, 231--259. https://doi.org/10.1016/j.ecolmodel.2005.03.026

    work page doi:10.1016/j.ecolmodel.2005.03.026 2006

  11. [11]

    Bootstrap methods for standard errors, confidence intervals, and other measures of statistical accuracy,

    Hastie, T., Tibshirani R. (1986) Generalized Additive Models, Statist. Sci. 1(3), 297--10, https://doi.org/10.1214/ss/1177013604

    work page doi:10.1214/ss/1177013604 1986

  12. [12]

    Wood, S.N. (2017). Generalized Additive Models: An Introduction with R, Second Edition (2nd ed.). Chapman and Hall/CRC. https://doi.org/10.1201/9781315370279

    work page doi:10.1201/9781315370279 2017

  13. [13]

    Machine Learning , author =

    Breiman, L. Random Forests. Machine Learning 45, 5–32 (2001). https://doi.org/10.1023/A:1010933404324

    work page doi:10.1023/a:1010933404324 2001

  14. [14]

    (2006) Gaussian Processes for Machine Learning

    Rasmussen, C.E., Williams, C.K.I. (2006) Gaussian Processes for Machine Learning

    2006

  15. [15]

    Rue, H., Martino, S., Chopin, N. (2009). Approximate Bayesian Inference for Latent Gaussian models by using Integrated Nested Laplace Approximations, Journal of the Royal Statistical Society Series B: Statistical Methodology, 71, 2, 319--392. https://doi.org/10.1111/j.1467-9868.2008.00700.x

    work page doi:10.1111/j.1467-9868.2008.00700.x 2009

  16. [16]

    Lindgren, F., Rue, H., Lindström, J. (2011) An Explicit Link between Gaussian Fields and Gaussian Markov Random Fields: The Stochastic Partial Differential Equation Approach, Journal of the Royal Statistical Society Series B: Statistical Methodology, 73, 4, 423--498. https://doi.org/10.1111/j.1467-9868.2011.00777.x

    work page doi:10.1111/j.1467-9868.2011.00777.x 2011

  17. [17]

    (2019) Applications for deep learning in ecology

    Christin, S., Hervet, E., Lecomte, N. (2019) Applications for deep learning in ecology. Methods Ecol Evol. 10, 1632--1644. https://doi.org/10.1111/2041-210X.13256

    work page doi:10.1111/2041-210x.13256 2019

  18. [18]

    Elith, J., & Leathwick, J. R. (2009). Species distribution models: Ecological explanation and prediction across space and time. Annual Review of Ecology, Evolution, and Systematics, 40, 677–697. https://doi.org/10.1146/annurev.ecolsys.110308.120159

    work page doi:10.1146/annurev.ecolsys.110308.120159 2009

  19. [19]

    Gaston, K. J. (2003). The Structure and Dynamics of Geographic Ranges. Oxford University Press

    2003

  20. [20]

    https://www.gbif.org/what-is-gbif

  21. [22]

    Gelfand, A. E. (2022). Spatial modeling for the distribution of species in plant communities. Spatial Statistics, 50, 100582. https://doi.org/10.1016/j.spasta.2021.100582

    work page doi:10.1016/j.spasta.2021.100582 2022

  22. [23]

    S., Cortese, D., Cotgrove, L., Jolles, J

    Killen, S. S., Cortese, D., Cotgrove, L., Jolles, J. W., Munson, A., & Ioannou, C. C. (2021). The Potential for Physiological Performance Curves to Shape Environmental Effects on Social Behavior. Frontiers in Physiology, 12, Article 754719. https://doi.org/10.3389/fphys.2021.754719

    work page doi:10.3389/fphys.2021.754719 2021

  23. [24]

    Kumar, N. et al. (2012). Leafsnap: A Computer Vision System for Automatic Plant Species Identification. In: Fitzgibbon, A., Lazebnik, S., Perona, P., Sato, Y., Schmid, C. (eds) Computer Vision - ECCV 2012. ECCV 2012. Lecture Notes in Computer Science, vol 7573. Springer, Berlin, Heidelberg. https://doi.org/10.1007/978-3-642-33709-3\_36

    work page doi:10.1007/978-3-642-33709-3 2012

  24. [25]

    Mildenhall B., Srinivasan P.P., Tancik M., Barron J.T., Ramamoorthi R., Ng R. 2021. NeRF: representing scenes as neural radiance fields for view synthesis, Commun. ACM 65, 1, 99--106. https://doi.org/10.1145/3503250

    work page doi:10.1145/3503250 2021

  25. [26]

    (2008) Global warming and flowering times in Thoreau's Concord: a community perspective

    Miller-Rushing, A.J., Primack, R.B. (2008) Global warming and flowering times in Thoreau's Concord: a community perspective. Ecology 89(2), 332--41. https://doi.org/ 10.1890/07-0068.1

    work page doi:10.1890/07-0068.1 2008

  26. [27]

    J., Alves de Andrade, A

    Mendes, P., Elias Velazco, S. J., Alves de Andrade, A. F., De Marco, P. (2020). Dealing with overprediction in species distribution models: How adding distance constraints can improve model accuracy, Ecological Modelling 431, 109180. https://doi.org/10.1016/j.ecolmodel.2020.109180

    work page doi:10.1016/j.ecolmodel.2020.109180 2020

  27. [28]

    Mescheder L., Oechsle M., Niemeyer M., Nowozin S., Geiger A. 2019. Occupancy Networks: Learning 3D Reconstruction in Function Space,IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR), Long Beach, CA, USA, pp. 4455--4465. https://doi.org/10.1109/CVPR.2019.00459

    work page doi:10.1109/cvpr.2019.00459 2019

  28. [29]

    Müller T., Evans A., Schied C., Keller A. 2022. Instant neural graphics primitives with a multiresolution hash encoding. ACM Trans. Graph. 41, 4, 102. https://doi.org/10.1145/3528223.3530127

    work page doi:10.1145/3528223.3530127 2022

  29. [30]

    Park J.J., Florence P., Straub J., Newcombe R., Lovegrove S. 2019. DeepSDF: Learning Continuous Signed Distance Functions for Shape Representation, IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR), Long Beach, CA, USA, pp. 165--174. https://doi.org/10.1109/CVPR.2019.00025

    work page doi:10.1109/cvpr.2019.00025 2019

  30. [31]

    Parmesan, C., & Yohe, G. (2003). A globally coherent fingerprint of climate change impacts across natural systems. Nature, 421, 37–42. https://doi.org/10.1038/nature01286

    work page doi:10.1038/nature01286 2003

  31. [32]

    T., Araújo, M

    Pecl, G. T., Araújo, M. B., Bell, J. D., Blanchard, J., Bonebrake, T. C., Chen, I.-C., Clark, T. D., Colwell, R. K., Danielsen, F., Sorte, C. J. B., et al. (2017). Biodiversity redistribution under climate change: Impacts on ecosystems and human well-being. Science, 355(6332), eaai9214. https://doi.org/10.1126/science.aai9214

    work page doi:10.1126/science.aai9214 2017

  32. [33]

    Perez-Harguindeguy, N., Diaz, S., Garnier, E., Lavorel, S., Poorter, H., et al. (2013). New handbook for standardised measurement of plant functional traits worldwide. Australian Journal of Botany, 61(3), 167–234. https://doi.org/10.1071/BT12225

    work page doi:10.1071/bt12225 2013

  33. [34]

    A., Fu, Y., Dai, J., Liu, L., Lian, X., Shen, M., & Zhu, X

    Piao, S., Liu, Q., Chen, A., Janssens, I. A., Fu, Y., Dai, J., Liu, L., Lian, X., Shen, M., & Zhu, X. (2019). Plant phenology and global climate change: Current progress and future challenges. Glob Change Biol. 25, 1922--1940. https://doi.org/10.1111/gcb.14619

    work page doi:10.1111/gcb.14619 2019

  34. [35]

    Rahaman, N., Baratin, A., Arpit, D., Dräxler, F., Lin, M., Hamprecht, F.A., Bengio, Y., Courville, A.C. (2018). On the Spectral Bias of Neural Networks. International Conference on Machine Learning

    2018

  35. [36]

    (2021) Want to model a species niche? A step-by-step guideline on correlative ecological niche modelling, Ecological Modelling 456, 109671

    Sillero, N., Arenas-Castro, S., Enriquez-Urzelai, U., Gomes Vale, C., Sousa-Guedes, D., Martinez-Freiría, F., Real, R., Barbosa, A.M. (2021) Want to model a species niche? A step-by-step guideline on correlative ecological niche modelling, Ecological Modelling 456, 109671. https://doi.org/10.1016/j.ecolmodel.2021.109671

    work page doi:10.1016/j.ecolmodel.2021.109671 2021

  36. [37]

    Implicit Neural Representations with Periodic Activation Functions, in: Advances in Neural Information Processing Systems, Larochelle H., Ranzato M., Hadsell R., Balcan M.F., Lin H

    Sitzmann V., Martel J., Bergman A., Lindell D., Wetzstein G., 2020. Implicit Neural Representations with Periodic Activation Functions, in: Advances in Neural Information Processing Systems, Larochelle H., Ranzato M., Hadsell R., Balcan M.F., Lin H. (Eds.), Curran Associates, Inc., Article 33, 7462--7473

    2020

  37. [38]

    Strümpler Y., Postels J., Yang R., Van Gool L., Tombari F. 2022. Implicit Neural Representations for Image Compression, in: Computer Vision – ECCV 2022: 17th European Conference, Tel Aviv, Israel, October 23--27, 2022, Proceedings, Part XXVI. Springer-Verlag, Berlin, Heidelberg, pp. 74--91. https://doi.org/10.1007/978-3-031-19809-0\_5

    work page doi:10.1007/978-3-031-19809-0 2022

  38. [39]

    P., Mildenhall B., Fridovich-Keil S., Raghavan N., Singhal U., Ramamoorthi R, Barron J

    Tancik M., Srinivasan P. P., Mildenhall B., Fridovich-Keil S., Raghavan N., Singhal U., Ramamoorthi R, Barron J. T., Ng R., 2020. Fourier features let networks learn high frequency functions in low dimensional domains, in: Proceedings of the 34th International Conference on Neural Information Processing Systems (NIPS '20). Curran Associates Inc., Red Hook...

    2020

  39. [40]

    B., Sykes, M

    Thuiller, W., Lavorel, S., Araújo, M. B., Sykes, M. T., & Prentice, I. C. (2005). Climate change threats to plant diversity in Europe. PNAS, 102(23), 8245-8250. https://doi.org/10.1073/pnas.0409902102

    work page doi:10.1073/pnas.0409902102 2005

  40. [41]

    Urban, M. C. (2015). Accelerating extinction risk from climate change. Science, 348(6234), 571-573. https://doi.org/10.1126/science.aaa4984

    work page doi:10.1126/science.aaa4984 2015

  41. [42]

    Violle, C., Navas, M.L., Vile, D., Kazakou, E., Fortunel, C., Hummel, I., & Garnier, E. (2007). Let the concept of trait be functional! Oikos, 116(5), 882-892. https://doi.org/10.1111/j.0030-1299.2007.15559.x

    work page doi:10.1111/j.0030-1299.2007.15559.x 2007

  42. [43]

    T., Srinivasan, P

    Verbin, D., Hedman, P., Mildenhall, B., Zickler, T., Barron J. T., Srinivasan, P. P. (2025) Ref-NeRF: Structured View-Dependent Appearance for Neural Radiance Fields, IEEE Transactions on Pattern Analysis and Machine Intelligence 47(11), 9426--9437. https://doi.org/10.1109/TPAMI.2024.3360018

    work page doi:10.1109/tpami.2024.3360018 2025