Pith. sign in

REVIEW 3 major objections 3 minor 62 references

Toroidal area-preserving parameterizations of genus-one closed surfaces

T0 review · 3 major / 3 minor · reviewed 2026-08-05 · deepseek-v4-flash

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

Pith's one-line read The paper claims that toroidal area-preserving parameterizations of genus-one closed surfaces can be computed by minimizing a stretch-energy functional over a power manifold of ring tori, and gives four Riemannian-optimization algorithms th

desk verdict The submission is un-reviewable as-is: the supplied full text is a different paper on whole-slide image classification, so only the abstract is usable—and the abstract leaves the central stretch-energy/area-preservation equivalence unsupported. read the letter →

arxiv 2508.05111 v1 pith:ZNTRE4YS submitted 2025-08-07 math.NA cs.NA

Marco Sutti , Mei-Heng Yueh This is my paper
classification math.NAcs.NA MSC 65D1853A0565K10
keywords area-preservingparameterizationgenus-onesurfacesstretchenergyringtoripowermanifoldRiemannianoptimizationsurfaceregistrationtexturemapping
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

The paper aims to compute area-preserving parameterizations of genus-one closed surfaces — surfaces that are topologically one-holed tori — by mapping them onto a standard ring torus in three-dimensional space without distorting area. Its central proposal is to minimize a stretch-energy functional, which measures how far a tentative map is from preserving area, while constraining the map to a power manifold of ring tori. The paper introduces four Riemannian-geometry algorithms — projected gradient descent, projected conjugate gradient, Riemannian gradient, and Riemannian conjugate gradient — to solve this constrained minimization, and reports numerical experiments on mesh models in support of the framework. If the claim holds, the resulting maps give distortion-controlled parameterizations that can serve as canonical domains for surface registration and texture mapping of genus-one shapes.

What carries the argument

The central object is the stretch-energy functional minimized over the power manifold of ring tori embedded in $\mathbb R^3$ — the constrained set of admissible ring-torus target geometries. A ring torus is the doughnut-shaped surface of revolution obtained by revolving a circle around a coplanar external axis. Minimizing the stretch energy over this manifold is the paper's proposed formulation of toroidal area-preserving parameterization. The four algorithms differ in how they carry out the constrained descent: two project updates from the ambient space back onto the manifold, and two use the manifold's intrinsic Riemannian structure.

What would settle it

Take a mesh of a Euclidean torus quotient (a square with opposite edges identified), run the proposed algorithms on it, and measure the ratio of each triangle's area on the ring torus to its area on the input surface. If, under mesh refinement, these per-triangle area ratios do not converge to a common constant value, then the minimizers of the stretch-energy functional on the power manifold are not area-preserving parameterizations.

Watch

Extended reading notes

Core claim

The central claim is that an area-preserving parameterization of a genus-one closed surface can be obtained by minimizing a stretch-energy functional over a power manifold of ring tori embedded in $\mathbb R^3$. In the paper's terms, the minimizers of this constrained optimization problem are effective toroidal area-preserving parameterizations. Four algorithms are proposed to reach those minimizers: projected gradient descent, projected conjugate gradient, Riemannian gradient, and Riemannian conjugate gradient. The paper presents numerical experiments on several mesh models as evidence that the algorithms compute the desired parameterizations, and it shows how the resulting maps can be used

Load-bearing premise

The whole method depends on the premise that minimizers of the stretch-energy functional on the power manifold of ring tori are exactly the area-preserving parameterizations; if that equivalence is false, the algorithms solve the wrong optimization problem.

Editorial extensions

If this is right

  • If the central claim is right, any genus-one closed surface can be reparametrized onto a common ring torus with area preserved, giving such surfaces a canonical coordinate system.
  • The four algorithms give implementable options that trade ambient-space projection against intrinsic Riemannian descent, so users can choose based on the geometry of their meshes.
  • The resulting parameterizations support texture mapping directly on the three-dimensional torus, avoiding the cuts and area distortion introduced by planar flattening.
  • Surface registration of genus-one models reduces to aligning toroidal coordinates, since the ring-torus image is shared from one model to another.
  • The paper's reported mesh experiments indicate that the constrained minimizers are numerically reachable, not merely theoretically characterized.

Reading between the lines

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

  • Editorial note on the supplied text: the manuscript body accompanying this record is an unrelated paper on whole-slide image classification, so the summary above is built from the title and abstract; the paper's own equations, references, and numerical tables were not available to inspect.
  • The area-preserving claim can be tested directly on a flat-torus quotient mesh: the Jacobian determinant of the computed map should be constant over every triangle, and the variance of per-triangle area ratios should vanish under refinement.
  • The same constrained-manifold formulation could extend to other canonical target surfaces, such as spheres or hyperbolic octagons, for surfaces of other genera, although the paper confines its experiments to genus-one meshes.
Share X Bluesky LinkedIn Reddit HN

Editorial analysis

A structured set of objections, weighed in public.

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

Referee Report

3 major / 3 minor

Summary. The submission, arXiv:2508.05111, is titled 'Toroidal area-preserving parameterizations of genus-one closed surfaces' and its abstract announces four Riemannian optimization algorithms—projected gradient descent, projected conjugate gradient, Riemannian gradient, and Riemannian conjugate gradient—that minimize a stretch energy functional on a power manifold of ring tori to obtain area-preserving parameterizations of genus-one closed surfaces. The abstract further claims numerical experiments on several mesh models demonstrate effectiveness and that the algorithms can be used for surface registration and texture mapping. However, the supplied full text is not this paper. It is arXiv:2508.05114, an unrelated manuscript on whole-slide image classification (AHDMIL). Consequently, the submission contains no mathematical definitions, no statement of the proposed algorithms, no theorems or proofs, and no experimental details. The central claim that stretch-energy minimization over the power manifold yields area-preserving parameterizations is therefore unverifiable from the submitted material.

Significance. If the announced framework were fully developed and its central equivalence theorem proved, the work could contribute to toroidal parameterization, surface registration, and texture mapping, where area preservation is a desirable property. The use of Riemannian optimization on a power manifold of ring tori is a potentially interesting idea. However, in the submitted form the paper provides no verifiable mathematical or computational content: there are no derivations, no proofs, no algorithm pseudocode, no convergence analysis, and no reproducible experiments. The claimed results are entirely unsupported, so no significance can be assigned beyond the plausibility of the abstract's premise.

major comments (3)
  1. [Full text / entire submission] The body of the submission is not the stated paper. The supplied full text is arXiv:2508.05114, an unrelated whole-slide-image classification manuscript, while the abstract describes a numerical analysis paper on toroidal parameterizations. None of the promised mathematical content appears: no definition of stretch energy, no description of the power manifold of ring tori, no algorithm descriptions, no theorems, and no experiments. This is a load-bearing omission because every claim in the abstract is unsupported by accessible evidence.
  2. [Abstract] The load-bearing premise is that minimizers of the stretch energy functional over the power manifold of ring tori are area-preserving parameterizations. This equivalence is not stated precisely, let alone proved. Constraining the target to ring tori does not by itself guarantee area preservation; stretch energy typically measures local distortion and may drive maps toward conformality or harmonicity rather than toward constant area density. A theorem establishing this equivalence is necessary before the four algorithms can be claimed to solve the advertised problem.
  3. [Abstract] The numerical claims are unsupported. The abstract states that 'numerical experiments on several mesh models demonstrate the effectiveness of the proposed framework,' but the submission provides no information about the meshes, the chosen error metrics, the comparison baselines, the convergence behavior, or the reproducibility artifacts. Without these details, the effectiveness claim cannot be assessed or reproduced.
minor comments (3)
  1. [Title/Abstract] The title and abstract describe a math.NA paper, but the full text is a different paper in computational pathology. The authors should either supply the correct manuscript or withdraw the submission; the mismatch must be resolved before any further reviewing.
  2. [Abstract] No references are provided for the stretch energy functional, the power manifold of ring tori, or prior work on toroidal parameterizations. Placement of the work in the existing literature is absent.
  3. [Abstract] If the correct manuscript is resubmitted, it should include precise notation, numbered sections, theorem statements with proofs or clear proof sketches, and full experimental settings including error measures and comparison methods.

Circularity Check

0 steps flagged · score 0.0 of 10

No circularity detected; supplied full text is a different paper, so no derivation chain is available to exhibit a circular reduction.

full rationale

The only substantive content from arXiv:2508.05111 is the abstract, which states that the objective is a stretch energy functional and the constraint is a power manifold of ring tori, with numerical experiments reported for several mesh models. The body text supplied is arXiv:2508.05114 (AHDMIL), a whole-slide image classification manuscript, not the toroidal parameterization paper. Consequently, there are no equations, theorems, or derivation steps from the target paper to compare against its inputs. Circularity can only be claimed when a specific reduction is exhibited, e.g., a fitted parameter being renamed a prediction or a self-citation carrying a uniqueness theorem. No such quote or equation-level equivalence can be provided from the available abstract. The skeptic's concern that stretch-energy minimizers may not be area-preserving is a substantive correctness or missing-support issue, but it is not a circularity argument. Accordingly, the score is 0: no circular reasoning is evident, and the lack of full text prevents identifying any hidden circular step without speculation, which the review rules prohibit.

Assumptions & free parameters 0 free parameters · 2 assumptions · 0 invented entities

Since the full text for the stated paper was not available (the supplied full text is a different manuscript), the ledger is inferred from the abstract. No free parameters are listed because none are disclosed. The axioms capture the geometric and optimization assumptions embedded in the problem formulation.

assumptions (2)
  • domain assumption Every closed genus-one surface admits a parameterization onto a ring torus.
    The problem statement presupposes the existence of such parameterizations for the surfaces considered. This is a standard topological fact for genus-one surfaces, but the computational tractability is assumed.
  • domain assumption Minimizing the stretch energy functional on the constrained manifold yields area-preserving maps.
    The method's validity rests on the equivalence between the optimization objective and the geometric property of area preservation. This equivalence is not proven in the abstract and is a load-bearing premise.

how reviews work

0 comments
Cite this review

Pith. "Pith review of Toroidal area-preserving parameterizations of genus-one closed surfaces." pith.science (2026). https://pith.science/paper/ZNTRE4YS

@misc{pith2026250805111,
  author       = {Pith},
  title        = {Pith review of: Toroidal area-preserving parameterizations of genus-one closed surfaces},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/ZNTRE4YS}},
  note         = {Machine review of arXiv:2508.05111}
}
read the original abstract

We consider the problem of computing toroidal area-preserving parameterizations of genus-one closed surfaces. We propose four algorithms based on Riemannian geometry: the projected gradient descent method, the projected conjugate gradient method, the Riemannian gradient method, and the Riemannian conjugate gradient method. Our objective function is based on the stretch energy functional, and the minimization is constrained on a power manifold of ring tori embedded in three-dimensional Euclidean space. Numerical experiments on several mesh models demonstrate the effectiveness of the proposed framework. Finally, we show how to use the proposed algorithms in the context of surface registration and texture mapping applications.

Discussion (0). Sign in to comment.

Reference graph

Works this paper leans on

62 extracted references · 44 canonical work pages

  1. [1]

    Nature biomedical engineering 5(6), 555–570 (2021)

    Lu, M.Y., Williamson, D.F., Chen, T.Y., Chen, R.J., Barbieri, M., Mahmood, F.: Data-efficient and weakly supervised com- putational pathology on whole-slide images. Nature biomedical engineering 5(6), 555–570 (2021)

    work page 2021

  2. [2]

    In: Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, pp

    Li, B., Li, Y., Eliceiri, K.W.: Dual-stream multiple instance learning network for whole slide image classification with self-supervised contrastive learning. In: Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, pp. 14318–14328 (2021)

    work page 2021

  3. [3]

    In: Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, pp

    Zhang, H., Meng, Y., Zhao, Y., Qiao, Y., Yang, X., Coupland, S.E., Zheng, Y.: Dtfd- mil: Double-tier feature distillation multiple instance learning for histopathology whole slide image classification. In: Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, pp. 18802– 18812 (2022)

    work page 2022

  4. [4]

    Advances in neural information processing systems 34, 2136–2147 (2021)

    Shao, Z., Bian, H., Chen, Y., Wang, Y., Zhang, J., Ji, X., et al.: Transmil: Trans- former based correlated multiple instance learning for whole slide image classification. Advances in neural information processing systems 34, 2136–2147 (2021)

    work page 2021

  5. [5]

    In: Interna- tional Conference on Medical Image Comput- ing and Computer-Assisted Intervention, pp

    Fillioux, L., Boyd, J., Vakalopoulou, M., Courn` ede, P.-H., Christodoulidis, S.: Struc- tured state space models for multiple instance learning in digital pathology. In: Interna- tional Conference on Medical Image Comput- ing and Computer-Assisted Intervention, pp. 594–604 (2023). Springer

    work page 2023

  6. [6]

    : Bi- directional weakly supervised knowledge dis- tillation for whole slide image classification

    Qu, L., Wang, M., Song, Z., et al. : Bi- directional weakly supervised knowledge dis- tillation for whole slide image classification. Advances in Neural Information Processing Systems 35, 15368–15381 (2022)

    work page 2022

  7. [7]

    Nature methods 14(9), 873– 876 (2017)

    Schapiro, D., Jackson, H.W., Raghuraman, S., Fischer, J.R., Zanotelli, V.R., Schulz, D., Giesen, C., Catena, R., Varga, Z., Boden- miller, B.: histocat: analysis of cell pheno- types and interactions in multiplex image cytometry data. Nature methods 14(9), 873– 876 (2017)

    work page 2017

  8. [8]

    Nature methods 16(12), 1233–1246 (2019)

    Moen, E., Bannon, D., Kudo, T., Graf, W., Covert, M., Van Valen, D.: Deep learning for cellular image analysis. Nature methods 16(12), 1233–1246 (2019)

    work page 2019

Show all 62 references
  1. [9]

    IEEE transactions on medical imaging 39(11), 3257–3267 (2019)

    Mahmood, F., Borders, D., Chen, R.J., McKay, G.N., Salimian, K.J., Baras, A., Durr, N.J.: Deep adversarial training for multi-organ nuclei segmentation in histopathology images. IEEE transactions on medical imaging 39(11), 3257–3267 (2019)

    2019

  2. [10]

    Medical image analysis 58, 101563 (2019)

    Graham, S., Vu, Q.D., Raza, S.E.A., Azam, A., Tsang, Y.W., Kwak, J.T., Rajpoot, N.: Hover-net: Simultaneous segmentation and classification of nuclei in multi-tissue his- tology images. Medical image analysis 58, 101563 (2019)

    2019

  3. [11]

    : Spa- tial organization and molecular correlation of tumor-infiltrating lymphocytes using deep learning on pathology images

    Saltz, J., Gupta, R., Hou, L., Kurc, T., Singh, P., Nguyen, V., Samaras, D., Shroyer, K.R., Zhao, T., Batiste, R., et al. : Spa- tial organization and molecular correlation of tumor-infiltrating lymphocytes using deep learning on pathology images. Cell reports 23(1), 181–193 (2018)

    2018

  4. [12]

    Medical image analysis 63, 101696 (2020)

    Javed, S., Mahmood, A., Fraz, M.M., Koohbanani, N.A., Benes, K., Tsang, Y.-W., Hewitt, K., Epstein, D., Snead, D., Rajpoot, N.: Cellular community detection for tissue phenotyping in colorectal cancer histology images. Medical image analysis 63, 101696 (2020)

    2020

  5. [13]

    arXiv preprint arXiv:2403.06800 (2024)

    Yang, S., Wang, Y., Chen, H.: Mambamil: Enhancing long sequence modeling with sequence reordering in computational pathol- ogy. arXiv preprint arXiv:2403.06800 (2024)

    2024 arXiv

  6. [14]

    IEEE Transactions on Medical Imaging 40(12), 3739–3747 (2021) 19

    Shao, W., Wang, T., Huang, Z., Han, Z., Zhang, J., Huang, K.: Weakly supervised deep ordinal cox model for survival prediction from whole-slide pathological images. IEEE Transactions on Medical Imaging 40(12), 3739–3747 (2021) 19

    2021

  7. [15]

    Chen, R.J., Lu, M.Y., Shaban, M., Chen, C., Chen, T.Y., Williamson, D.F., Mah- mood, F.: Whole slide images are 2d point clouds: Context-aware survival pre- diction using patch-based graph convolu- tional networks. In: Medical Image Com- puting and Computer Assisted Interventi...

    2021

  8. [16]

    Medical Image Analysis 65, 101789 (2020)

    Yao, J., Zhu, X., Jonnagaddala, J., Hawkins, N., Huang, J.: Whole slide images based can- cer survival prediction using attention guided deep multiple instance learning networks. Medical Image Analysis 65, 101789 (2020)

    2020

  9. [17]

    Jama 318(22), 2199–2210 (2017)

    Bejnordi, B.E., Veta, M., Van Diest, P.J., Van Ginneken, B., Karssemeijer, N., Litjens, G., Van Der Laak, J.A., Hermsen, M., Man- son, Q.F., Balkenhol, M., et al.: Diagnostic assessment of deep learning algorithms for detection of lymph node metastases in women with breast can...

    2017

  10. [18]

    In: Proceedings of the Computer Vision and Pattern Recog- nition Conference, pp

    Dong, J., Jiang, J., Jiang, K., Li, J., Zhang, Y.: Fast and accurate gigapixel pathological image classification with hierarchical distilla- tion multi-instance learning. In: Proceedings of the Computer Vision and Pattern Recog- nition Conference, pp. 30818–30828 (2025)

    2025

  11. [19]

    Zhu, W., Lou, Q., Vang, Y.S., Xie, X.: Deep multi-instance networks with sparse label assignment for whole mammogram classifica- tion. In: Medical Image Computing and Com- puter Assisted Intervention- MICCAI 2017: 20th International Conference, Quebec City, QC, Canada, Septemb...

    2017

  12. [20]

    In: Proceedings of the IEEE Con- ference on Computer Vision and Pattern Recognition, pp

    Pinheiro, P.O., Collobert, R.: From image- level to pixel-level labeling with convolutional networks. In: Proceedings of the IEEE Con- ference on Computer Vision and Pattern Recognition, pp. 1713–1721 (2015)

    2015

  13. [21]

    In: Proceedings of the AAAI Conference on Artificial Intelligence, vol

    Feng, J., Zhou, Z.-H.: Deep miml network. In: Proceedings of the AAAI Conference on Artificial Intelligence, vol. 31 (2017)

    2017

  14. [22]

    Bioinformat- ics 32(12), 52–59 (2016)

    Kraus, O.Z., Ba, J.L., Frey, B.J.: Classify- ing and segmenting microscopy images with deep multiple instance learning. Bioinformat- ics 32(12), 52–59 (2016)

    2016

  15. [23]

    Advances in neural information processing systems 10 (1997)

    Maron, O., Lozano-P´ erez, T.: A framework for multiple-instance learning. Advances in neural information processing systems 10 (1997)

    1997

  16. [24]

    Advances in neural infor- mation processing systems 3 (1990)

    Keeler, J., Rumelhart, D., Leow, W.: Inte- grated segmentation and recognition of hand- printed numerals. Advances in neural infor- mation processing systems 3 (1990)

    1990

  17. [25]

    In: ML-2000 Workshop Attribute-Value and Relational Learning (2000)

    Ramon, J.: Multi instance neural networks. In: ML-2000 Workshop Attribute-Value and Relational Learning (2000)

    2000

  18. [26]

    In: International Conference on Machine Learn- ing, pp

    Ilse, M., Tomczak, J., Welling, M.: Attention- based deep multiple instance learning. In: International Conference on Machine Learn- ing, pp. 2127–2136 (2018). PMLR

    2018

  19. [27]

    Advances in Neural Information Processing Systems 35, 35971– 35983 (2022)

    Gu, A., Goel, K., Gupta, A., R´ e, C.: On the parameterization and initialization of diago- nal state space models. Advances in Neural Information Processing Systems 35, 35971– 35983 (2022)

    2022

  20. [28]

    arXiv preprint arXiv:2312.00752 (2023)

    Gu, A., Dao, T.: Mamba: Linear-time sequence modeling with selective state spaces. arXiv preprint arXiv:2312.00752 (2023)

    2023 arXiv

  21. [29]

    In: International Con- ference on Machine Learning, pp

    Bolukbasi, T., Wang, J., Dekel, O., Saligrama, V.: Adaptive neural networks for efficient inference. In: International Con- ference on Machine Learning, pp. 527–536 (2017). PMLR

    2017

  22. [30]

    In: 2016 23rd International Conference on Pattern Recog- nition (ICPR), pp

    Teerapittayanon, S., McDanel, B., Kung, H.- T.: Branchynet: Fast inference via early exit- ing from deep neural networks. In: 2016 23rd International Conference on Pattern Recog- nition (ICPR), pp. 2464–2469 (2016). IEEE

    2016

  23. [31]

    Advances in Neural Information Processing Systems 37, 114765– 114796 (2024)

    Zhao, W., Tang, J., Han, Y., Song, Y., Wang, K., Huang, G., Wang, F., You, Y.: Dynamic 20 tuning towards parameter and inference effi- ciency for vit adaptation. Advances in Neural Information Processing Systems 37, 114765– 114796 (2024)

    2024

  24. [32]

    arXiv preprint arXiv:2202.07800 (2022)

    Liang, Y., Ge, C., Tong, Z., Song, Y., Wang, J., Xie, P.: Not all patches are what you need: Expediting vision transform- ers via token reorganizations. arXiv preprint arXiv:2202.07800 (2022)

    2022 arXiv

  25. [33]

    In: Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, pp

    Meng, L., Li, H., Chen, B.-C., Lan, S., Wu, Z., Jiang, Y.-G., Lim, S.-N.: Adavit: Adaptive vision transformers for efficient image recog- nition. In: Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, pp. 12309–12318 (2022)

    2022

  26. [34]

    Advances in neural information process- ing systems 34, 13937–13949 (2021)

    Rao, Y., Zhao, W., Liu, B., Lu, J., Zhou, J., Hsieh, C.-J.: Dynamicvit: Efficient vision transformers with dynamic token sparsifica- tion. Advances in neural information process- ing systems 34, 13937–13949 (2021)

    2021

  27. [35]

    Advances in Neural Information Processing Systems 34, 5770–5783 (2021)

    Song, L., Zhang, S., Liu, S., Li, Z., He, X., Sun, H., Sun, J., Zheng, N.: Dynamic grained encoder for vision transformers. Advances in Neural Information Processing Systems 34, 5770–5783 (2021)

    2021

  28. [36]

    Advances in neural infor- mation processing systems 34, 11960–11973 (2021)

    Wang, Y., Huang, R., Song, S., Huang, Z., Huang, G.: Not all images are worth 16x16 words: Dynamic transformers for efficient image recognition. Advances in neural infor- mation processing systems 34, 11960–11973 (2021)

    2021

  29. [37]

    Neural Networks 15(1), 57– 67 (2002)

    Sprecher, D.A., Draghici, S.: Space-filling curves and kolmogorov superposition-based neural networks. Neural Networks 15(1), 57– 67 (2002)

    2002

  30. [38]

    In: Artificial Neural Net- works—ICANN 2002: International Confer- ence Madrid, Spain, August 28–30, 2002 Proceedings 12, pp

    K¨ oppen, M.: On the training of a kol- mogorov network. In: Artificial Neural Net- works—ICANN 2002: International Confer- ence Madrid, Spain, August 28–30, 2002 Proceedings 12, pp. 474–479 (2002). Springer

    2002

  31. [39]

    Neural Computa- tion 5(1), 18–20 (1993)

    Lin, J.-N., Unbehauen, R.: On the realization of a kolmogorov network. Neural Computa- tion 5(1), 18–20 (1993)

    1993

  32. [40]

    arXiv preprint arXiv:2112.09963 (2021)

    Lai, M.-J., Shen, Z.: The kolmogorov super- position theorem can break the curse of dimensionality when approximating high dimensional functions. arXiv preprint arXiv:2112.09963 (2021)

    2021

  33. [41]

    Neural Net- works 152, 332–346 (2022)

    Fakhoury, D., Fakhoury, E., Speleers, H.: Exsplinet: An interpretable and expressive spline-based neural network. Neural Net- works 152, 332–346 (2022)

    2022

  34. [42]

    arXiv preprint arXiv:2404.19756 (2024)

    Liu, Z., Wang, Y., Vaidya, S., Ruehle, F., Halverson, J., Soljaˇ ci´ c, M., Hou, T.Y., Tegmark, M.: Kan: Kolmogorov-arnold networks. arXiv preprint arXiv:2404.19756 (2024)

    2024 arXiv

  35. [43]

    Neural net- works 137, 119–126 (2021)

    Schmidt-Hieber, J.: The kolmogorov–arnold representation theorem revisited. Neural net- works 137, 119–126 (2021)

    2021

  36. [44]

    arXiv preprint arXiv:2409.04290 (2024)

    Knottenbelt, W., Gao, Z., Wray, R., Zhang, W.Z., Liu, J., Crispin-Ortuzar, M.: Coxkan: Kolmogorov-arnold networks for inter- pretable, high-performance survival analysis. arXiv preprint arXiv:2409.04290 (2024)

    2024

  37. [45]

    arXiv preprint arXiv:2406.02486 (2024)

    Genet, R., Inzirillo, H.: A tempo- ral kolmogorov-arnold transformer for time series forecasting. arXiv preprint arXiv:2406.02486 (2024)

    2024 arXiv

  38. [46]

    arXiv preprint arXiv:2405.08790 (2024)

    Vaca-Rubio, C.J., Blanco, L., Pereira, R., Caus, M.: Kolmogorov-arnold networks (kans) for time series analysis. arXiv preprint arXiv:2405.08790 (2024)

    2024

  39. [47]

    arXiv preprint arXiv:2405.07344 (2024)

    Genet, R., Inzirillo, H.: Tkan: Temporal kolmogorov-arnold networks. arXiv preprint arXiv:2405.07344 (2024)

    2024 arXiv

  40. [48]

    arXiv preprint arXiv:2409.08605 (2024)

    Xu, A., Zhang, B., Kong, S., Huang, Y., Yang, Z., Srivastava, S., Sun, M.: Effective integration of kan for keyword spotting. arXiv preprint arXiv:2409.08605 (2024)

    2024 arXiv

  41. [49]

    arXiv preprint arXiv:2406.01034 (2024) 21

    Xu, J., Chen, Z., Li, J., Yang, S., Wang, W., Hu, X., Ngai, E.C.-H.: Fourierkan-gcf: Fourier kolmogorov-arnold network–an effec- tive and efficient feature transformation for graph collaborative filtering. arXiv preprint arXiv:2406.01034 (2024) 21

    2024 arXiv

  42. [50]

    arXiv preprint arXiv:2406.14495 (2024)

    Aghaei, A.A.: rkan: Rational kolmogorov- arnold networks. arXiv preprint arXiv:2406.14495 (2024)

    2024 arXiv

  43. [51]

    arXiv preprint arXiv:2406.13155 (2024)

    Bodner, A.D., Tepsich, A.S., Spol- ski, J.N., Pourteau, S.: Convolutional kolmogorov-arnold networks. arXiv preprint arXiv:2406.13155 (2024)

    2024 arXiv

  44. [52]

    arXiv preprint arXiv:2409.10594 (2024)

    Yang, X., Wang, X.: Kolmogorov-arnold transformer. arXiv preprint arXiv:2409.10594 (2024)

    2024 arXiv

  45. [53]

    arXiv preprint arXiv:2405.12832 (2024)

    Bozorgasl, Z., Chen, H.: Wav-kan: Wavelet kolmogorov-arnold networks. arXiv preprint arXiv:2405.12832 (2024)

    2024 arXiv

  46. [54]

    arXiv preprint arXiv:2406.11173 (2024)

    Ta, H.-T.: Bsrbf-kan: A combination of b-splines and radial basic functions in kolmogorov-arnold networks. arXiv preprint arXiv:2406.11173 (2024)

    2024 arXiv

  47. [55]

    In: Euro- pean Conference on Computer Vision, pp

    Herrmann, C., Bowen, R.S., Zabih, R.: Chan- nel selection using gumbel softmax. In: Euro- pean Conference on Computer Vision, pp. 241–257 (2020). Springer

    2020

  48. [56]

    arXiv preprint arXiv:2404.10518 (2024)

    Qin, D., Leichner, C., Delakis, M., Fornoni, M., Luo, S., Yang, F., Wang, W., Banbury, C., Ye, C., Akin, B., et al.: Mobilenetv4- universal models for the mobile ecosystem. arXiv preprint arXiv:2404.10518 (2024)

    2024 arXiv

  49. [57]

    IEEE Transactions on Medical Imaging (2025)

    Dong, J., Jiang, J., Jiang, K., Li, J., Cai, L., Zhang, Y.: Disentangled pseudo-bag aug- mentation for whole slide image multiple instance learning. IEEE Transactions on Medical Imaging (2025)

    2025

  50. [58]

    arXiv preprint arXiv:1412.6980 (2014)

    Kingma, D.P., Ba, J.: Adam: A method for stochastic optimization. arXiv preprint arXiv:1412.6980 (2014)

    2014 arXiv

  51. [59]

    Advances in neural information processing systems 25 (2012)

    Krizhevsky, A., Sutskever, I., Hinton, G.E.: Imagenet classification with deep convolu- tional neural networks. Advances in neural information processing systems 25 (2012)

    2012

  52. [60]

    In: Proceedings of the IEEE International Con- ference on Computer Vision, pp

    He, K., Zhang, X., Ren, S., Sun, J.: Delving deep into rectifiers: Surpassing human-level performance on imagenet classification. In: Proceedings of the IEEE International Con- ference on Computer Vision, pp. 1026–1034 (2015)

    2015

  53. [61]

    In: Proceedings of the Thirteenth International Conference on Artificial Intel- ligence and Statistics, pp

    Glorot, X., Bengio, Y.: Understanding the difficulty of training deep feedforward neural networks. In: Proceedings of the Thirteenth International Conference on Artificial Intel- ligence and Statistics, pp. 249–256 (2010). JMLR Workshop and Conference Proceed- ings

    2010

  54. [62]

    arXiv preprint arXiv:1802.03426 (2018) 22

    McInnes, L., Healy, J., Melville, J.: Umap: Uniform manifold approximation and projec- tion for dimension reduction. arXiv preprint arXiv:1802.03426 (2018) 22

    2018 arXiv

Pith tools

Reviewed August 5, 2026 · model on record in the stance chip above.