Pith. sign in

REVIEW 4 major objections 4 minor 57 references

HYDRA: Hyperbolic Dynamic Representation Architecture for Kolmogorov-Arnold Networks

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

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

Pith's one-line read This paper reports that a hyperbolic, low-rank variant of Kolmogorov-Arnold Networks matches or beats standard KANs and MLPs on all eight tabular benchmarks while using fewer trainable parameters.

desk verdict A clever low-rank hyperbolic KAN with a clean density proof for the full-rank case, but unreproducible parameter counts and internally inconsistent tables undermine the main empirical claim. read the letter →

arxiv 2608.12194 v1 pith:2WPY2D5U submitted 2026-08-12 cs.LG cs.AI

Zhao Su , Yuxin Xia , Haoran Li , Jun Shen , Qi Zhu , Qingguo Zhou , Binbin Yong This is my paper · ORCID
classification cs.LGcs.AI
keywords Kolmogorov-ArnoldNetworkshyperbolicrepresentationlearningPoincaréballlow-rankprototypeparameterefficiencytangent-spacesplineupdatesradiuscontroltabularbenchmarks
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 proposes HYDRA, a variant of Kolmogorov-Arnold Networks that keeps KAN-style spline function learning but moves it into the tangent space of a bounded hyperbolic (Poincaré-ball) representation. Its central claim is empirical: across eight tabular benchmarks, HYDRA achieves the strongest or tied-strongest primary metric—RMSE for regression, accuracy for classification—while using fewer trainable parameters than Euclidean KAN and MLP, with an average parameter reduction of 34.9% relative to KAN and 37.1% relative to MLP. If true, this matters because KANs' main practical bottleneck is the quadratic parameter growth of dense hidden-to-hidden spline blocks, and HYDRA cuts that cost from $O(d^2K)$ to $O(dr + r^2K)$ without sacrificing accuracy. The paper also argues that the final hyperbolic radius and latent path geometry give an inspectable, physically meaningful diagnostic, illustrated on the CCPP dataset where an ambient-temperature sweep moves the representation from small-radius to large-radius and reduces predicted power by 37.36 MW.

What carries the argument

The load-bearing mechanism is the low-rank prototype functional block executed in the tangent space of the Poincaré ball $\mathbb{B}^d_c = \{h \in \mathbb{R}^d : c\|h\|_2^2 < 1\}$, with exponential and logarithmic maps at the origin as the fixed coordinate chart. The block writes the tangent-space KAN update as $z_{l+1} = z_l + \alpha_l W_\uparrow \Phi_l(W_\downarrow z_l)$, so the dominant cost falls from $O(d^2K)$ to $O(dr + r^2K)$; rank $r$ is a compression knob the authors choose as the smallest value that preserves the primary metric. Two complementary controls stabilize the geometry: a hard projection $\Pi_{r_l}$ keeps every hidden state inside a per-layer radius budget, and a soft penalty $L_{\text{rad}}$ discourages outward drift before the projection activates, preventing the near-boundary distance and gradient amplification that otherwise lets a model win by pushing samples outward instead of learning smooth spline responses. The final hyperbolic radius and latent path length are then available as interpretability diagnostics.

What would settle it

Train HYDRA and a full-rank Euclidean KAN of matched width on a synthetic tabular target whose hidden interactions are deliberately high-dimensional, such as a sum of many independent nonlinear pair terms that no radial coordinate can absorb. If HYDRA needs $r$ close to $d$ to match the full-rank KAN's accuracy, then the low-effective-dimension premise fails and the reported savings cannot be attributed to the hyperbolic geometry.

Watch

Extended reading notes

Core claim

HYDRA's core claim is that a KAN's expensive hidden-to-hidden functional map can be replaced by a low-rank prototype update inside a hyperbolic representation without giving up predictive accuracy. Each block computes $z_l = \log_c^0(h_l)$ in tangent coordinates, projects down with $W_\downarrow \in \mathbb{R}^{r \times d}$, applies a spline block $\Phi_l$ to the $r$ prototype coordinates, lifts with $W_\uparrow$, adds a scaled residual, and re-embeds under a radius projection $\Pi_{r_l}$. The authors report this architecture achieves the strongest or tied-strongest primary metric on all eight datasets in Table 1—for instance, reducing Parkinsons Telemonitoring RMSE from 4.424 to 3.534 while cutting parameters from 2.4k to 1.4k—and they support the decomposition claim with ablations in which low-rank models use a median of 33.8% of full-rank HYDRA's parameters, and radius-constrained models reduce mean radius while improving the primary metric. The universal-approximation appendix shows that when $r \ge d$ the prototype block recovers the full tangent-space KAN update, so HYDRA is a parameter-efficient subfamily of a dense universal class.

Load-bearing premise

The claim rests on the premise that on these tabular tasks the hidden functional update is effectively low-dimensional, so a small prototype rank $r$ preserves what a full-rank KAN block would contribute—a premise supported only by post hoc rank-selection ablations, not by a prior characterization or theory of when low effective dimension holds.

Editorial extensions

If this is right

  • The dominant hidden-to-hidden cost of KAN-style networks drops from $O(d^2K)$ to $O(dr + r^2K)$, so wider hidden layers become affordable at a fixed spline budget—an architectural saving, not a pruning artifact.
  • Radius control is more than numerical safeguarding: ablations show constrained models reach smaller mean radii and better primary metrics than unconstrained models, indicating the representation itself changes.
  • The final hyperbolic radius and latent path length can serve as HYDRA-specific interpretability diagnostics; the CCPP case study links an ambient-temperature sweep to a 37.36 MW drop in predicted power with the expected radius and SHAP direction.
  • Universal approximation is retained within the low-rank family: the appendix shows $r \ge d$ recovers the full tangent-space KAN update and HYDRA networks are dense in $C(K,\mathbb{R}^m)$.
  • Across the eight datasets the average parameter count is 34.9% below Euclidean KAN and 37.1% below MLP while the primary metric is competitive or better, so the method, if correct, gives interpretable functional learning at a budget close to dense neural networks.

Reading between the lines

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

  • Beyond the paper: the low-rank prototype update is geometry-agnostic in its parameter counting, so the same bottleneck could be tested in Euclidean or spherical latent spaces; that comparison would isolate whether the gains come from curvature itself or from the radius constraint and rank compression.
  • Beyond the paper: the final hyperbolic radius could be tested as a lightweight difficulty or out-of-distribution signal—if radius consistently tracks error or low confidence, the interpretability diagnostic becomes a usable uncertainty estimate rather than a post hoc visualization.
  • Beyond the paper: the effective-rank assumption implies a measurable compressibility test on any new dataset—estimate the rank of the tangent-space update Jacobian on validation data and check whether the selected $r$ tracks it; datasets with high effective rank should demand larger $r$.
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 / 4 minor

Summary. HYDRA is a proposed KAN variant that stores hidden states on the Poincaré ball, performs KAN-style spline updates in the tangent space through a low-rank prototype bottleneck, and adds a radius-control penalty and projection. The paper claims that across eight tabular benchmarks HYDRA achieves the strongest or tied-strongest primary metric on every dataset while using fewer trainable parameters than Euclidean KAN and MLP, and that ablations support the low-rank and radius-control design choices. Appendices provide a universal approximation argument, hyperparameter tables, additional metrics, and a stated limitations section.

Significance. If the empirical claims were fully supported, HYDRA would be a meaningful parameter-efficiency improvement for KAN-style architectures, and the radial-coordinate diagnostic would be a useful interpretability addition. The low-rank parameter count in Eq. (9)-(10) is a clean and correct design contribution, and the radius-control mechanism is a sensible response to known Poincaré-boundary instability. The universal approximation attempt is non-circular in intent because it is checked against the Euclidean KAN realization rather than against HYDRA's fitted values. However, the central empirical claim is currently not auditable: the reported parameter counts do not follow from the stated configurations, identical primary metrics differ across tables, and all results rest on a single split and seed. These issues, together with a gap in the scaling argument of Appendix A, mean that the paper's headline conclusions are not yet established.

major comments (4)
  1. [Table 1 vs. Appendix D] The parameter-efficiency comparison cannot be reconstructed from the supplied configurations. For HYDRA on CCPP, Table 6 gives w=43, K=6, r=6, so a single low-rank block already contains 2dr + r^2(K+1) = 2*43*6 + 36*7 = 768 parameters; with a 4-dimensional linear embedding (172 parameters) and a linear readout (43 parameters), the reported 4.8k total is consistent only with an unreported number of blocks or additional structure. For KAN on CCPP, Table 7 gives w=51, K=12; a dense 51-to-51 KAN block has 51^2*(12+1) = 33,813 parameters, and even a single 4-to-51-to-1 KAN has 2,652 + 663 = 3,315 parameters, neither of which matches the reported 7.0k. Since no depth or layer count is given for any model, the claimed 34.9% parameter reduction relative to Euclidean KAN is not verifiable.
  2. [Tables 1, 2, 3, and 5] The same primary metric for the same model family takes different values in different tables. Energy Heating RMSE is 0.706 in Table 1 but 1.193 in Table 3 and 1.193/1.384 in Table 5; Phoneme accuracy is 0.885 in Table 1 but 0.856/0.845 in Tables 3 and 5. Moreover, the selected low-rank HYDRA results in Table 2 do not match Table 1 for several datasets, e.g., CCPP 3.919 vs. 3.604, Ionosphere 0.943 vs. 0.971, and QSAR 0.882 vs. 0.900. These are not auxiliary metrics; they are the same primary metrics used for the headline claim, so the reader cannot tell which configuration produced the reported performance.
  3. [Experimental Setup and Appendix E] All runs use a single fixed split and seed 42 with no standard deviations or repeated trials, yet the paper states that HYDRA 'achieves the strongest or tied-strongest primary metric on all eight datasets.' With a single split, and with rank and radius selected from multiple configurations as the paper itself acknowledges in Appendix E, the reported ranking is compatible with selection over many candidates. The claim requires either repeated runs with variance estimates or a clearly prespecified evaluation protocol, especially for datasets where the margins between models are small.
  4. [Appendix A, Eqs. (A6)-(A11)] The universal approximation proof contains a scaling step that is not justified for spline KANs. The proof scales the Euclidean realization to z_l^(α) = α z_l, then defines f_HYDRA = α^{-1} Q H^(α) and g_ϕ = α^{-1} Q G^(α). This assumes that rescaling the hidden coordinates rescales the Euclidean network's output by α, i.e., G^(α) = α g_ϕ, but KAN spline blocks are not homogeneous: φ(α z_j) is not generally α φ(z_j). The displayed O(R_α^3) bounds control only the exp/log discrepancy, not the discrepancy between the rescaled Euclidean network and the original g_ϕ. A repair would need to construct a Euclidean network with bounded hidden states directly and absorb all output scaling into the final readout.
minor comments (4)
  1. [Method, Eq. (1)] The notation for layer-dependent quantities is introduced loosely: the number of blocks L, the residual scale α_l, the radius budget r_l, and the soft threshold r_allow,l are used before their schedules are specified. Please state the layer-count and radius-budget definitions explicitly.
  2. [Table 1 caption] The caption says that the smallest and second-smallest parameter counts are bold and underlined, but no bold or underline appears inside the parentheses in the table as typeset. The formatting should be corrected or the caption amended.
  3. [Eq. (5)] The sentence following Eq. (5) contains a grammatical error: 'that leading to every input-output coordinate pair owns a separate spline' should be rephrased, e.g., 'so that every input-output coordinate pair owns a separate spline.'
  4. [Figures 2 and 3] The text refers to panel (b) in Figure 2 and to specific visual effects in Figure 3, but the figures appear only as placeholders in the manuscript. Please ensure the final figures are legible and that panels are labeled consistently with the text.

Circularity Check

0 steps flagged · score 0.0 of 10

No load-bearing circularity: HYDRA's formal, efficiency, and empirical claims are containment arguments, algebraic identities, or externally benchmarked results.

full rationale

The only formal derivation in the paper is the universal-approximation argument in Appendix A. That proof constructs HYDRA parameters that exactly reproduce a Euclidean KAN residual block (setting W_down=[I;0] and W_up=[A 0] with r>=d) and then invokes external spline/Stone-Weierstrass approximation; this is a containment argument, not a self-definitional reduction. The parameter-efficiency claim is the algebraic identity P_lr = 2dr + r^2(K+1) in Eq. (9) compared with P_full = d^2(K+1) in Eq. (8); no fitted parameter is relabeled as a prediction, and Appendix E explicitly states that prototype rank and radius remain validation-dependent hyperparameters, so the rank selection is disclosed tuning rather than a disguised fit. The low-rank and radius-control ablations compare HYDRA against its own full-rank or unconstrained variants; these are internal consistency checks, not predictions derived from the fitted values. The interpretability case study uses SHAP as an explicitly labeled post hoc reference and is further disclaimed in Appendix E as non-causal; it is a consistency check between two views of the same model, not a load-bearing derivation. Externally flagged issues such as the inconsistency of Phoneme accuracy between Table 1 (0.885) and Tables 3/5 (0.856/0.845) or the irreproducibility of Table 1 parameter counts from Appendix D widths are reproducibility and auditability defects, not circularity: they do not show that a claimed result reduces to its own inputs by construction. All cited foundations (KAN, hyperbolic embeddings, SHAP, GAMs) are external works, with no author self-citation chain carrying the argument. Accordingly, no circular step meets the quoting-and-reduction bar, and the correct finding is no significant circularity.

Assumptions & free parameters 3 free parameters · 4 assumptions · 0 invented entities

The central claim rests on three families of tuned hyperparameters (rank, radius budget, regularization) and on the untested low-effective-dimension assumption. The density proof itself is standard and self-contained.

free parameters (3)
  • prototype rank r = per dataset: 3, 3, 3, 6, 1, 7, 4, 1 (Table 2) and final values in Table 6
    Selected per dataset as the smallest rank that preserves the primary metric; the rank directly controls the parameter count and is not derived from theory.
  • radius budget and penalty = target ratios 0.88 to 0.90, penalty weights 0 to 0.03 (Table 6)
    Validation-dependent hyperparameters that set soft threshold r_allow and the hard radius r_l; they influence the learned geometry without a first-principles value.
  • spline regularization strength = not tabulated; combined with weight decay
    The objective includes sparse/smooth spline regularization (lambda_sp), whose strength is tuned but not explicitly reported.
assumptions (4)
  • standard math Stone-Weierstrass and spline approximation guarantee finite Euclidean spline/KAN networks approximate any continuous function on compact sets.
    Used in Appendix A to produce the Euclidean network g_phi that HYDRA then approximates.
  • domain assumption Euclidean KANs are universal approximators for continuous functions on compact domains.
    The proof inherits universal approximation from Liu et al. 2025 and Rudin 1976 without re-deriving it.
  • standard math Exponential and logarithmic maps on the Poincaré ball are continuous inverses on bounded balls and match the identity to third order near the origin.
    Appendix A uses this to bound the per-block difference between Euclidean and HYDRA updates by C_l R_alpha^3.
  • ad hoc to paper The tangent-space functional interactions in tabular data have low effective dimension, so a small prototype rank preserves accuracy.
    This is the load-bearing assumption for parameter efficiency; it is supported only by post hoc ablation, not by theory.

how reviews work

0 comments
Cite this review

Pith. "Pith review of HYDRA: Hyperbolic Dynamic Representation Architecture for Kolmogorov-Arnold Networks." pith.science (2026). https://pith.science/paper/2WPY2D5U

@misc{pith2026260812194,
  author       = {Pith},
  title        = {Pith review of: HYDRA: Hyperbolic Dynamic Representation Architecture for Kolmogorov-Arnold Networks},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/2WPY2D5U}},
  note         = {Machine review of arXiv:2608.12194}
}
read the original abstract

Kolmogorov-Arnold Networks (KANs) enhance nonlinear function approximation by replacing scalar weights with learnable univariate functions. However, assigning an independent function to every connection results in substantial parameter redundancy, limiting their scalability and efficiency. To reduce this redundancy, we introduce \textbf{HY}perbolic \textbf{D}ynamic \textbf{R}epresentation \textbf{A}rchitecture (HYDRA), a parameter-efficient hyperbolic extension of KAN that combines spline-based functional learning with representations in the Poincar\'e ball. HYDRA maps vector-valued inputs into a bounded hyperbolic latent space, performs KAN-style updates in tangent space, and employs a low-rank prototype block to share functional transformations across hidden dimensions. The resulting hyperbolic representations provide a structured radial coordinate for interpretation, while radius control improves training stability by preventing boundary saturation. Extensive experiments across eight benchmark datasets demonstrate that HYDRA consistently achieves competitive or superior predictive performance while improving parameter efficiency and representation interpretability.

Figures

Figures reproduced from arXiv: 2608.12194 by the authors.

Figure 1
Figure 1. Overview of HYDRA. The architecture maps vector-valued features into a bounded hyperbolic representation, applies [PITH_FULL_IMAGE:figures/full_fig_p003_1.png] view at source ↗
Figure 2
Figure 2. Low-rank ablation visualization. Selected ranks [PITH_FULL_IMAGE:figures/full_fig_p005_2.png] view at source ↗
Figure 3
Figure 3. Radius-control ablation visualization. Constrained [PITH_FULL_IMAGE:figures/full_fig_p006_3.png] view at source ↗
Figures from the paper (2 more)
Figure 4
Figure 4. Figure 4: CCPP interpretability case study linking SHAP [PITH_FULL_IMAGE:figures/full_fig_p007_4.png]
Figure 5
Figure 5. Figure 5: Additional interpretability cases beyond CCPP. Each panel traces how sweeping a representative feature changes both [PITH_FULL_IMAGE:figures/full_fig_p011_5.png]

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

57 extracted references · 36 canonical work pages

  1. [1]

    ACM SIGKDD Explorations Newsletter , volume=

    OpenML: Networked science in machine learning , author=. ACM SIGKDD Explorations Newsletter , volume=. 2013 , publisher=

    work page 2013

  2. [2]

    Communication, Simulation, and Intelligent Agents: Implications of Personal Intelligent Machines for Medical Education

    Clancey, William J. Communication, Simulation, and Intelligent Agents: Implications of Personal Intelligent Machines for Medical Education. Proceedings of the Eighth International Joint Conference on Artificial Intelligence (IJCAI-83)

  3. [3]

    Classification Problem Solving

    Clancey, William J. Classification Problem Solving. Proceedings of the Fourth National Conference on Artificial Intelligence

  4. [4]

    , title =

    Robinson, Arthur L. , title =. 1980 , doi =. https://science.sciencemag.org/content/208/4447/1019.full.pdf , journal =

    1980

  5. [5]

    New Ways to Make Microcircuits Smaller---Duplicate Entry

    Robinson, Arthur L. New Ways to Make Microcircuits Smaller---Duplicate Entry. Science

  6. [6]

    Clancey and Glenn Rennels , abstract =

    Diane Warner Hasling and William J. Clancey and Glenn Rennels , abstract =. Strategic explanations for a diagnostic consultation system , journal =. 1984 , issn =. doi:https://doi.org/10.1016/S0020-7373(84)80003-6 , url =

    doi:10.1016/s0020-7373(84)80003-6 1984

  7. [7]

    and Rennels, Glenn R

    Hasling, Diane Warner and Clancey, William J. and Rennels, Glenn R. and Test, Thomas. Strategic Explanations in Consultation---Duplicate. The International Journal of Man-Machine Studies

  8. [8]

    Poligon: A System for Parallel Problem Solving

    Rice, James. Poligon: A System for Parallel Problem Solving

Show all 57 references
  1. [9]

    Transfer of Rule-Based Expertise through a Tutorial Dialogue

    Clancey, William J. Transfer of Rule-Based Expertise through a Tutorial Dialogue

  2. [10]

    The Engineering of Qualitative Models

    Clancey, William J. The Engineering of Qualitative Models

  3. [11]

    2017 , eprint=

    Attention Is All You Need , author=. 2017 , eprint=

    2017

  4. [12]

    Pluto: The 'Other' Red Planet

    NASA. Pluto: The 'Other' Red Planet

  5. [13]

    International Conference on Learning Representations , year =

    Liu, Ziming and Wang, Yixuan and Vaidya, Sachin and Ruehle, Fabian and Halverson, James and Solja. International Conference on Learning Representations , year =. 2404.19756 , archivePrefix =

    arXiv

  6. [14]

    Nickel, Maximilian and Kiela, Douwe , booktitle =. Poincar. 2017 , eprint =

    2017

  7. [15]

    Advances in Neural Information Processing Systems , volume =

    Hyperbolic Neural Networks , author =. Advances in Neural Information Processing Systems , volume =. 2018 , eprint =

    2018

  8. [16]

    Advances in Neural Information Processing Systems , volume =

    Hyperbolic Graph Convolutional Neural Networks , author =. Advances in Neural Information Processing Systems , volume =. 2019 , eprint =

    2019

  9. [17]

    2021 , doi =

    Yang, Zebin and Zhang, Aijun and Sudjianto, Agus , journal =. 2021 , doi =. 2003.07132 , archivePrefix =

    2021 arXiv

  10. [18]

    Advances in Neural Information Processing Systems , volume =

    Neural Additive Models: Interpretable Machine Learning with Neural Nets , author =. Advances in Neural Information Processing Systems , volume =. 2021 , eprint =

    2021

  11. [19]

    2022 , eprint =

    Chang, Chun-Hao and Caruana, Rich and Goldenberg, Anna , booktitle =. 2022 , eprint =

    2022

  12. [20]

    Advances in Neural Information Processing Systems , volume =

    A Unified Approach to Interpreting Model Predictions , author =. Advances in Neural Information Processing Systems , volume =. 2017 , eprint =

    2017

  13. [21]

    2024 , eprint =

    Chebyshev Polynomial-Based Kolmogorov-Arnold Networks: An Efficient Architecture for Nonlinear Function Approximation , author =. 2024 , eprint =

    2024

  14. [22]

    2405.12832 , archivePrefix =

    Bozorgasl, Zavareh and Chen, Hao , year =. 2405.12832 , archivePrefix =

    arXiv

  15. [23]

    Proceedings of the AAAI Conference on Artificial Intelligence , volume =

    Arik, Sercan. Proceedings of the AAAI Conference on Artificial Intelligence , volume =. 2021 , doi =

    2021

  16. [24]

    International Conference on Learning Representations , year =

    Neural Oblivious Decision Ensembles for Deep Learning on Tabular Data , author =. International Conference on Learning Representations , year =

  17. [25]

    Advances in Neural Information Processing Systems , volume =

    Revisiting Deep Learning Models for Tabular Data , author =. Advances in Neural Information Processing Systems , volume =. 2021 , url =

    2021

  18. [26]

    Bayan and Goldstein, Tom , year =

    Somepalli, Gowthami and Goldblum, Micah and Schwarzschild, Avi and Bruss, C. Bayan and Goldstein, Tom , year =. 2106.01342 , archivePrefix =

    arXiv

  19. [27]

    International Conference on Learning Representations , year =

    Deep Compression: Compressing Deep Neural Networks with Pruning, Trained Quantization and Huffman Coding , author =. International Conference on Learning Representations , year =

  20. [28]

    Advances in Neural Information Processing Systems , volume =

    Tensorizing Neural Networks , author =. Advances in Neural Information Processing Systems , volume =. 2015 , url =

    2015

  21. [29]

    and Shen, Yelong and Wallis, Phillip and Allen-Zhu, Zeyuan and Li, Yuanzhi and Wang, Shean and Wang, Lu and Chen, Weizhu , booktitle =

    Hu, Edward J. and Shen, Yelong and Wallis, Phillip and Allen-Zhu, Zeyuan and Li, Yuanzhi and Wang, Shean and Wang, Lu and Chen, Weizhu , booktitle =. 2022 , url =

    2022

  22. [30]

    International Conference on Learning Representations , year =

    Riemannian Adaptive Optimization Methods , author =. International Conference on Learning Representations , year =

  23. [31]

    1976 , publisher =

    Principles of Mathematical Analysis , author =. 1976 , publisher =

    1976

  24. [32]

    International Conference on Machine Learning , pages =

    Learning Continuous Hierarchies in the Lorentz Model of Hyperbolic Geometry , author =. International Conference on Machine Learning , pages =. 2018 , url =

    2018

  25. [33]

    International Conference on Machine Learning , pages =

    Hyperbolic Entailment Cones for Learning Hierarchical Embeddings , author =. International Conference on Machine Learning , pages =. 2018 , url =

    2018

  26. [34]

    International Conference on Learning Representations , year =

    Hyperbolic Attention Networks , author =. International Conference on Learning Representations , year =

  27. [35]

    Advances in Neural Information Processing Systems , volume =

    Hyperbolic Graph Neural Networks , author =. Advances in Neural Information Processing Systems , volume =. 2019 , eprint =

    2019

  28. [36]

    International Conference on Learning Representations , year =

    Mixed-curvature Variational Autoencoders , author =. International Conference on Learning Representations , year =

  29. [37]

    Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition , pages =

    Hyperbolic Image Embeddings , author =. Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition , pages =. 2020 , url =

    2020

  30. [38]

    Proceedings of the 60th Annual Meeting of the Association for Computational Linguistics , pages =

    Fully Hyperbolic Neural Networks , author =. Proceedings of the 60th Annual Meeting of the Association for Computational Linguistics , pages =. 2022 , doi =

    2022

  31. [39]

    International Conference on Learning Representations , year =

    Hyperbolic Neural Networks++ , author =. International Conference on Learning Representations , year =

  32. [40]

    International Conference on Machine Learning , pages =

    Hyperbolic Representation Learning: Revisiting and Advancing , author =. International Conference on Machine Learning , pages =. 2023 , url =

    2023

  33. [41]

    International Journal of Computer Vision , volume =

    Hyperbolic Deep Learning in Computer Vision: A Survey , author =. International Journal of Computer Vision , volume =. 2024 , doi =

    2024

  34. [42]

    Proceedings of the ACM SIGKDD Conference on Knowledge Discovery and Data Mining , pages =

    Hypformer: Exploring Efficient Transformer Fully in Hyperbolic Space , author =. Proceedings of the ACM SIGKDD Conference on Knowledge Discovery and Data Mining , pages =. 2024 , doi =

    2024

  35. [43]

    Advances in Neural Information Processing Systems , volume =

    Learning Structured Representations with Hyperbolic Embeddings , author =. Advances in Neural Information Processing Systems , volume =. 2024 , url =

    2024

  36. [44]

    Advances in Neural Information Processing Systems , volume =

    Hyperbolic Fine-Tuning for Large Language Models , author =. Advances in Neural Information Processing Systems , volume =. 2025 , eprint =

    2025

  37. [45]

    IEEE Transactions on Neural Networks and Learning Systems , volume =

    Hyperbolic Binary Neural Network , author =. IEEE Transactions on Neural Networks and Learning Systems , volume =. 2025 , doi =

    2025

  38. [46]

    International Conference on Machine Learning , pages =

    Hyperbolic Image-text Representations , author =. International Conference on Machine Learning , pages =. 2023 , url =

    2023

  39. [47]

    International Conference on Learning Representations , year =

    Fully Hyperbolic Convolutional Neural Networks for Computer Vision , author =. International Conference on Learning Representations , year =

  40. [48]

    Proceedings of the ACM SIGKDD Conference on Knowledge Discovery and Data Mining , pages =

    Improving Robustness of Hyperbolic Neural Networks by Lipschitz Analysis , author =. Proceedings of the ACM SIGKDD Conference on Knowledge Discovery and Data Mining , pages =. 2024 , doi =

    2024

  41. [49]

    Advances in Neural Information Processing Systems , volume =

    Hyperbolic Embeddings of Supervised Models , author =. Advances in Neural Information Processing Systems , volume =. 2024 , url =

    2024

  42. [50]

    Proceedings of the ACM SIGKDD Conference on Knowledge Discovery and Data Mining , pages =

    Lorentzian Residual Neural Networks , author =. Proceedings of the ACM SIGKDD Conference on Knowledge Discovery and Data Mining , pages =. 2025 , doi =

    2025

  43. [51]

    Proceedings of the ACM SIGKDD Conference on Knowledge Discovery and Data Mining , pages =

    Hyperbolic Deep Learning for Foundation Models: A Survey , author =. Proceedings of the ACM SIGKDD Conference on Knowledge Discovery and Data Mining , pages =. 2025 , doi =

    2025

  44. [52]

    2024 , eprint =

    Kolmogorov-Arnold Networks are Radial Basis Function Networks , author =. 2024 , eprint =

    2024

  45. [53]

    2501.07032 , archivePrefix =

    Ta, Hoang-Thang and Thai, Duy-Quy and Tran, Anh and Sidorov, Grigori and Gelbukh, Alexander , year =. 2501.07032 , archivePrefix =

    arXiv

  46. [54]

    International Conference on Learning Representations , year =

    Measuring the Intrinsic Dimension of Objective Landscapes , author =. International Conference on Learning Representations , year =

  47. [55]

    2408.10205 , archivePrefix =

    Liu, Ziming and Ma, Pingchuan and Wang, Yixuan and Matusik, Wojciech and Tegmark, Max , year =. 2408.10205 , archivePrefix =

    arXiv

  48. [56]

    ACM Computing Surveys , volume =

    A Survey on Kolmogorov-Arnold Network , author =. ACM Computing Surveys , volume =. 2025 , doi =. 2411.06078 , archivePrefix =

    2025 arXiv

  49. [57]

    2512.09084 , archivePrefix =

    Eliasson, Oscar , year =. 2512.09084 , archivePrefix =

    arXiv

Pith tools

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