Pith. sign in

REVIEW 3 major objections 4 minor 3 cited by

On the Universal Statistical Consistency of Expansive Hyperbolic Deep Convolutional Neural Networks

T0 review · 3 major / 4 minor · reviewed 2026-08-12 · deepseek-v4-flash

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

Pith's one-line read This paper claims that a deep convolutional network built on the Poincaré disc is strongly universally consistent: its truncated empirical risk minimizer converges almost surely to the Bayes-optimal regression function as sample size grows.

desk verdict Theorem 11 is false: the network's bounded Poincaré-disc output cannot approximate unbounded real regression targets, so the main consistency claim collapses. read the letter →

arxiv 2411.10128 v1 pith:KR6D6OG3 submitted 2024-11-15 stat.ML cs.LG

Sagar Ghosh , Kushal Bose , Swagatam Das This is my paper · ORCID
classification stat.MLcs.LG MSC 62G0868T0762G20
keywords hyperbolicdeepconvolutionalneuralnetworksPoincarédiscexpansiveconvolutionuniversalconsistencyempiricalriskminimizationregressioncoveringnumberpseudo-dimension
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 an expansive hyperbolic deep convolutional network (eHDCNN) on the Poincaré disc and argues that its empirical risk minimizer is strongly universally consistent: as the sample size grows, the expected squared error converges almost surely to the Bayes-optimal regression error for every distribution whose target is square-integrable after logarithmic projection. The proof extends the Euclidean consistency framework for deep convolutional networks by replacing linear convolution with a hyperbolic operation that maps inputs to the tangent space, convolves there, and maps back through the exponential map. If the proof holds, hyperbolic convolutional architectures inherit the statistical guarantees of Euclidean ones while using curved geometry to represent hierarchical data. The authors also report experiments on synthetic and real datasets showing faster error reduction than Euclidean counterparts, with the best performance at low positive curvature.

What carries the argument

The argument is carried by three pieces. First, hyperbolic convolution is defined by transporting inputs to the tangent space at $0$ via $\log_c^0$, applying ordinary expansive convolution there, and returning via $\exp_c^0$; layer outputs are combined with Möbius addition and ReLU, so the network lives in nested Poincaré discs of increasing dimension. Second, a capacity bound (Lemma 16) controls the metric entropy of the truncated function class $\pi_M \mathcal{H}_{L,s}$ in terms of depth $L$, filter length $s$, dimension $d$, and truncation level $M$, reused from the Euclidean setting. Third, a concentration inequality (Lemma 17) together with a hyperbolic universal approximation lemma (Lemma 19), which approximates continuous functions by composing Euclidean approximators with $\log_c^0$, yields almost-sure convergence of the truncated empirical error to the truncated generalization error. The truncation operator $\pi_M$ is needed because the projected target $\log_c^0(y)$ grows like $\tanh^{-1}(M\sqrt{c})$ as samples approach the boundary of the disc.

What would settle it

Compute the metric-entropy covering number of the truncated hyperbolic class $\pi_M \mathcal{H}_{L,s}$ at positive curvature $c$ and compare it with the bound claimed in Lemma 16: if the bound is exceeded for any allowed $L$, $s$, $d$, and $M$, the proof collapses. Alternatively, simulate the empirical risk minimizer on a regression problem with a known hyperbolic regression function and check whether $E(\pi_{M_m} f_{D,L_m,s}) - E(f_\rho)$ fails to converge to zero when the growth conditions of Theorem 11 are satisfied, which would disprove the theorem.

Watch

Extended reading notes

Core claim

The central claim is Theorem 11: for filter length $2 \le s \le d$ and depth $L_m \to \infty$, the truncated empirical risk minimizer $\pi_{M_m} f_{D,L_m,s}$ of the eHDCNN class $\mathcal{H}_{L,s}$ satisfies $\lim_{m\to\infty} E(f_m) - E(f_\rho) = 0$ almost surely, provided the truncation level $M_m \to 1/\sqrt{c}$, the sample-size constraint $m^{-\theta} M_m^2 [1 + 1/(M_m \sqrt{c} \tanh^{-1}(M_m \sqrt{c}))]^2 \to 0$, and the growth condition in equation (VI.1) hold for $\theta \in (0,1/2)$. Here $f_\rho$ is the hyperbolic regression function, the conditional expectation of $\log_c^0(Y)$ given $X$, which minimizes the hyperbolic generalization error. In words, the architecture asymptotically recovers the optimal regression function under any Borel distribution with $\log_c^0(Y)$ square-integrable, and the result reduces to the Euclidean theorem when $c \to 0$. The paper also introduces a fully hyperbolic 1-D expansive convolution operation on the Poincaré disc as a generalization of Euclidean convolution.

Load-bearing premise

The proof assumes that the complexity and approximation estimates proved for Euclidean convolutional networks transfer unchanged to the hyperbolic function class, even though the hyperbolic layers use nonlinear Möbius addition and are not simple conjugates of Euclidean layers; if that transfer fails, Theorem 11 is not established.

Editorial extensions

If this is right

  • When the curvature $c$ tends to $0$, Theorem 11 reduces to the Euclidean universal consistency result, so the hyperbolic theorem is a strict generalization of the flat-space statement.
  • The eHDCNN architecture is claimed to asymptotically match the Bayes-optimal regression function without any prior on the distribution beyond square-integrability of the projected target.
  • Lemma 18 states that the truncated empirical error converges to the truncated generalization error faster for hyperbolic convolution than for Euclidean convolution, which the paper ties to requiring fewer training iterations.
  • The result is currently limited to 1-D expansive convolution with one input channel and ReLU activation; the paper states that higher-dimensional and contractive cases remain open.
  • Experiments across synthetic and four real-world datasets indicate that low positive curvature improves error convergence, while very high curvature shrinks the Poincaré disc and degrades performance.

Reading between the lines

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

  • Extending beyond the paper: the unproved transfer of the covering-number bound from Euclidean to hyperbolic layers could be tested numerically, and if it fails, a corrected bound with curvature-dependent factors might still preserve consistency under modified growth conditions on depth and truncation.
  • Extending beyond the paper: because all model hyperbolic spaces are isometric via the Killing-Hopf theorem, the same consistency result should carry over to hyperboloid or Klein-Beltrami models if the log/exp maps are replaced accordingly.
  • Extending beyond the paper: the faster-convergence claim suggests that hyperbolic convolution could be most advantageous for regression targets with hierarchical or tree-structured dependencies, where curved geometry captures the structure in lower dimensions.
  • Extending beyond the paper: the open higher-dimensional case may be approachable by factorizing 2-D convolution into sequences of 1-D expansive convolutions, but the paper does not establish this.
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

3 major / 4 minor

Summary. The paper proposes a hyperbolic deep convolutional network (eHDCNN) on the Poincaré disc, defines hyperbolic expansive convolution via log/exp maps and Möbius operations, and claims strong universal consistency of the truncated empirical risk minimizer (Theorem 11). The proof follows the three-part strategy of Lin et al. [5]: a covering-number bound for the hyperbolic class (Lemma 16), a concentration inequality (Lemmas 17-18), and a universal approximation lemma (Lemma 19). Experiments on synthetic and real regression/classification data are reported. The central theorem is not established: the architecture outputs points in the Poincaré disc, while the regression targets are real-valued and unbounded, so the risk minimizer cannot converge to arbitrary regression functions.

Significance. If Theorem 11 were correct, the paper would provide the first statistical consistency result for a fully hyperbolic convolutional architecture, a meaningful contribution to the theoretical understanding of hyperbolic neural networks. The paper also ships experiments and an anonymous implementation, which is a strength. However, the central claim is false as stated because of the range mismatch between the network output and the regression target, and the two transfer lemmas (capacity and approximation) are unproved, with Lemma 19 actually false. The claimed contribution therefore is not established; the paper would require a substantially different architecture or a restricted target class to make the theorem viable.

major comments (3)
  1. [Theorem 11, Definition 5 (V.5)] Theorem 11 is false as stated because every f in the class H_{L,s} takes values in the Poincaré disc D^1_c. By Eq. (V.5), the final output is h_L(x)=exp_0[a_L·log_0(h_L(x))], so |f(x)|<1/sqrt(c) for all x. The risk is defined against targets log_0(y) that are arbitrary real numbers, and the regression function f_ρ(x)=E[log_0(Y)|x] can exceed 1/sqrt(c). For a concrete counterexample, take c=1, d=2, s=2, X={0} subset of D^2_1, and the deterministic distribution ρ with Y=exp_0(2) almost surely. Then log_0(Y)=2 is in L^2 and f_ρ is identically 2. For every f in H_{L,s}, |f(0)|<1, so E(f)-E(f_ρ)=(f(0)-2)^2≥1. The truncation π_{M_m} does not help because M_m→1/sqrt(c)=1, so π_{M_m} f_m remains bounded by 1. Consequently, the almost-sure limit asserted in Theorem 11 cannot hold for this distribution. This is an internal inconsistency of the stated result, not merely a missing proof detail.
  2. [Lemma 16] The covering-number bound for the hyperbolic class is asserted without proof. The text states that Lemmas 2-4 of [5] are results on general metric spaces and then simply states Lemma 4 in the hyperbolic context. This does not justify the transfer: the class H_{L,s} is defined through Möbius addition, layer-wise exp/log maps, and ReLU in the tangent/log domain, none of which appears in the Euclidean derivation of [5]. The proof of Lemma 18 uses this exact entropy bound to obtain the exponential decay after the covering-number term, so if the hyperbolic entropy differs from the Euclidean one, the proof of Theorem 11 collapses. The bounded range of H_{L,s} makes the standard Euclidean comparison particularly delicate, since truncation interacts with the range in a different way.
  3. [Lemma 19] The universal approximation lemma is false in the stated form. Its proof defines g(y)=f(exp_0(y)), takes a Euclidean DCNN approximator g^{w,b}_L from Theorem 1 of [8], and sets f^{w,b}_L(x)=g^{w,b}_L(log_0(x)). The authors do not prove that this composed function can be realized by the recursive hyperbolic architecture (V.4)-(V.5); in particular, the Möbius operations and layer-wise exp/log mappings are not eliminated by a single conjugacy with log_0. Moreover, the claim for arbitrary f in C(Ω) is contradicted by the range constraint: with c=1 and Ω={0}, the constant function f≡2 cannot be approximated in sup norm by any element of H_{L,s}, since all such elements satisfy |f(0)|<1. Lemma 19 is used in the proof of Theorem 11 through the term S_7, so the proof fails at this point as well.
minor comments (4)
  1. [Lemma 9] The statement of Lemma 9 omits a square: the proof gives E(f)-E(f_ρ)=∫(f-f_ρ)^2 dρ_X, which is ∥f-f_ρ∥^2_{L^2_{ρ_X}}, not the norm itself as written.
  2. [Definition 10] Definition 10 quantifies over 'every Borel probability distribution λ' but the rest of the paper uses ρ for the joint distribution; the condition log_0(Y)∈L^2 should be stated for the joint distribution on X×Y to be unambiguous.
  3. [Section VII] The experiments do not specify how real-valued inputs are mapped into the Poincaré disc before applying the hyperbolic architecture, which makes it difficult to connect the empirical results to the theoretical setup where X is a compact subset of D^d_c.
  4. [Remark 20] Remark 20 claims the proof shows faster convergence than the Euclidean equivalent, but the displayed inequalities only bound errors by different expressions involving M_m and tanh^{-1}; no rate comparison between the hyperbolic and Euclidean estimators is actually derived.

Circularity Check

0 steps flagged · score 0.0 of 10

No circularity: the consistency theorem rests on external benchmarks [5], [8], [44]; no fitted parameter is renamed as a prediction, and no self-citation is load-bearing.

full rationale

The paper's central claim, Theorem 11, is an asymptotic consistency statement for an empirical risk minimizer over the class H_{L,s}. The proof chain is: Lemma 16 borrows a covering-number bound from the external Euclidean paper [5] (Lin et al.); Lemma 17 is a concentration inequality from [44]/[5]; Lemma 18 combines these; Lemma 19 invokes the external universality theorem of [8] (Zhou) and transfers it to the hyperbolic class by composing with log/exp maps. None of these dependencies are self-citations of the present authors, and the target quantity E(f) - E(f_rho) is not defined in terms of a fitted parameter, so no step reduces by construction to its own input. The most serious issues are mathematical correctness gaps, not circularity: Lemma 16 asserts without proof that the Euclidean covering-number bound holds unchanged for the hyperbolic class, and Lemma 19 asserts that g^{w,b}_L(log_c^0(x)) lies in H_{L,s}, which is not established and is in tension with Definition 5 (V.5), where every h_L(x) = exp_0[a_L · log_0(h_L(x))] lies in D^1_c, so |h_L| < 1/sqrt(c) while f_rho can be any real number. A false or unsupported theorem is a validity defect, not a circularity; per the review rules, correctness concerns should not inflate the circularity score. Since the paper's claimed results are advertised as following from external, independent theorems rather than from its own fitted values or self-citations, the circularity score is 0.

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

The central theorem rests on imported results from Euclidean theory whose transfer to the hyperbolic function class is asserted, not proved. No parameters are fitted to data in the theorem itself; the free-parameter list is empty because the growth conditions are assumptions, not fitted constants, and the empirical hyperparameters do not enter the derivation.

assumptions (4)
  • ad hoc to paper Lemma 4 of [5] (covering number bound for the Euclidean expansive DCNN class) applies unchanged to the hyperbolic class H_{L,s}
    Lemma 16 states this without proof. The hyperbolic class uses Möbius addition and exp/log maps, so the Euclidean pseudo-dimension bound is not obviously inherited.
  • ad hoc to paper The Euclidean DCNN universal approximation theorem (Zhou [8], Theorem 1) transfers to the hyperbolic architecture via conjugation with log/exp maps
    Lemma 19 constructs f^{w,b}_L = g^{w,b}_L ∘ log_c^0 and calls it an element of H_{L,s}, but the hyperbolic layer operations are not the conjugation of the Euclidean layer operations because Möbius addition is nonlinear.
  • ad hoc to paper The network output is treated as a real-valued function in the risk, despite the final layer formally outputting a point in D_c^1
    Definition 5 and the hyperbolic generalization error in Section VI are inconsistent unless the estimator is implicitly log_c^0 of the final output; this is never stated.
  • standard math Standard hyperbolic geometry facts: Möbius gyrogroup axioms, Killing-Hopf isometry, global diffeomorphism of log/exp maps
    These are used throughout for the Poincaré disc operations and the coordinate changes; they are well-established results in the literature.

how reviews work

0 comments
Cite this review

Pith. "Pith review of On the Universal Statistical Consistency of Expansive Hyperbolic Deep Convolutional Neural Networks." pith.science (2026). https://pith.science/paper/KR6D6OG3

@misc{pith2026241110128,
  author       = {Pith},
  title        = {Pith review of: On the Universal Statistical Consistency of Expansive Hyperbolic Deep Convolutional Neural Networks},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/KR6D6OG3}},
  note         = {Machine review of arXiv:2411.10128}
}
read the original abstract

The emergence of Deep Convolutional Neural Networks (DCNNs) has been a pervasive tool for accomplishing widespread applications in computer vision. Despite its potential capability to capture intricate patterns inside the data, the underlying embedding space remains Euclidean and primarily pursues contractive convolution. Several instances can serve as a precedent for the exacerbating performance of DCNNs. The recent advancement of neural networks in the hyperbolic spaces gained traction, incentivizing the development of convolutional deep neural networks in the hyperbolic space. In this work, we propose Hyperbolic DCNN based on the Poincar\'{e} Disc. The work predominantly revolves around analyzing the nature of expansive convolution in the context of the non-Euclidean domain. We further offer extensive theoretical insights pertaining to the universal consistency of the expansive convolution in the hyperbolic space. Several simulations were performed not only on the synthetic datasets but also on some real-world datasets. The experimental results reveal that the hyperbolic convolutional architecture outperforms the Euclidean ones by a commendable margin.

Figures

Figures reproduced from arXiv: 2411.10128 by the authors.

Figure 1
Figure 1. Test Root Mean Squared error for f(x) and g(x) plotted using (a) eDCNN architecture curvature 0 (i.e., Euclidean space) and (b) eHDCNN architecture with curvature 1. the number of layers is 4. Both models are trained for 100 epochs over the training set. The test Root Mean Squared Loss (RMSE) is recorded after the completion of training and presented in [PITH_FULL_IMAGE:figures/full_fig_p003_1.png] view at source ↗
Figure 2
Figure 2. The complete workflow of expansive hyperbolic [PITH_FULL_IMAGE:figures/full_fig_p006_2.png] view at source ↗
Figure 3
Figure 3. The performance analysis of eHDCNN with varying space curvatures (a) for [PITH_FULL_IMAGE:figures/full_fig_p010_3.png] view at source ↗
Figures from the paper (2 more)
Figure 4
Figure 4. Figure 4: The performance analysis of eHDCNN with varying space curvatures (a) for Superconductivity (b) for Wave [PITH_FULL_IMAGE:figures/full_fig_p011_4.png]
Figure 5
Figure 5. Figure 5: Various experiments were performed on the Superconductivity dataset by varying filter length and number [PITH_FULL_IMAGE:figures/full_fig_p012_5.png]

Discussion (0). Continue with ORCID to comment.

Forward citations

Cited by 3 Pith papers

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

  1. Cartan Networks: Group theoretical Hyperbolic Deep Learning

    cs.LG 2025-05 reject novelty 6.0 of 10

    Cartan networks compose solvable-group homomorphisms with isometries to define hyperbolic layers, and the paper reports competitive benchmark performance.

  2. Navigation through Non-Compact Symmetric Spaces: a mathematical perspective on Cartan Neural Networks

    cs.LG 2025-07 conditional novelty 5.0 of 10

    It derives a covariant, activation-free layer map for neural networks on noncompact symmetric spaces and works out explicit hyperbolic and class-switching examples.

  3. Transformers Are Universally Consistent

    cs.LG 2025-05 reject novelty 4.0 of 10

    HyT, a hyperbolic Transformer, is claimed to be universally consistent for L2 regression, but the proof is invalidated by an algebraic error and circular reasoning.

Reference graph

Works this paper leans on

46 extracted references · 35 canonical work pages · cited by 3 Pith papers

  1. [5]

    Universal consistency of deep convolutional neural networks,

    S.-B. Lin, K. Wang, Y . Wang, and D.-X. Zhou, “Universal consistency of deep convolutional neural networks,” IEEE Transactions on Information Theory, vol. 68, no. 7, pp. 4610–4617, 2022

    work page 2022

  2. [8]

    Universality of deep convolutional neural networks,

    D.-X. Zhou, “Universality of deep convolutional neural networks,” Applied and computational harmonic analysis , vol. 48, no. 2, pp. 787–794, 2020

    work page 2020

  3. [1]

    Gradient-based learning applied to document recognition,

    Y . LeCun, L. Bottou, Y . Bengio, and P. Haffner, “Gradient-based learning applied to document recognition,” Proceedings of the IEEE , vol. 86, no. 11, pp. 2278–2324, 1998

    1998

  4. [2]

    Application of deep convolution neural network,

    J. Yang and J. Li, “Application of deep convolution neural network,” in 2017 14th International Computer Conference on Wavelet Active Media Technology and Information Processing (ICCWAMTIP) . IEEE, 2017, pp. 229–232

    work page 2017

  5. [3]

    Enhanced feature representation in detection for optical remote sensing images,

    K. Fu, Z. Chen, Y . Zhang, and X. Sun, “Enhanced feature representation in detection for optical remote sensing images,” Remote Sensing, vol. 11, no. 18, p. 2095, 2019

    work page 2019

  6. [4]

    Autonomous car using cnn deep learning algorithm,

    I. Sonata, Y . Heryadi, L. Lukas, and A. Wibowo, “Autonomous car using cnn deep learning algorithm,” in Journal of Physics: Conference Series, vol. 1869, no. 1. IOP Publishing, 2021, p. 012071

    work page 2021

  7. [6]

    Deep residual learning for image recognition,

    K. He, X. Zhang, S. Ren, and J. Sun, “Deep residual learning for image recognition,” in Proceedings of the IEEE conference on computer vision and pattern recognition , 2016, pp. 770–778

    2016

  8. [7]

    Very deep convolutional networks for large-scale image recognition,

    K. Simonyan, “Very deep convolutional networks for large-scale image recognition,” arXiv preprint arXiv:1409.1556 , 2014

    arXiv 2014

Show all 46 references
  1. [9]

    Contrastive multi-view hyperbolic hierarchical clustering,

    F. Lin, B. Bai, K. Bai, Y . Ren, P. Zhao, and Z. Xu, “Contrastive multi-view hyperbolic hierarchical clustering,” arXiv preprint arXiv:2205.02618, 2022

    2022 arXiv

  2. [10]

    Hyperbolic k-means for traffic-aware clustering in cloud and virtualized rans,

    H. Djeddal, L. Touzari, A. Giovanidis, C.-D. Phung, and S. Secci, “Hyperbolic k-means for traffic-aware clustering in cloud and virtualized rans,” Computer Communications, vol. 176, pp. 258–271, 2021

    2021

  3. [11]

    Cross-modal scalable hyperbolic hierarchical clustering,

    T. Long and N. van Noord, “Cross-modal scalable hyperbolic hierarchical clustering,” in Proceedings of the IEEE/CVF international conference on computer vision , 2023, pp. 16 655–16 664. 17

    2023

  4. [12]

    Hyperbolic neural networks,

    O. Ganea, G. Becigneul, and T. Hofmann, “Hyperbolic neural networks,” in Advances in Neural Information Processing Systems , S. Bengio, H. Wallach, H. Larochelle, K. Grauman, N. Cesa-Bianchi, and R. Garnett, Eds., vol. 31. Curran Associates, Inc., 2018. [Online]. Available: ht...

    2018

  5. [14]

    Fully hyperbolic convolutional neural networks for computer vision,

    A. Bdeir, K. Schwethelm, and N. Landwehr, “Fully hyperbolic convolutional neural networks for computer vision,” arXiv preprint arXiv:2303.15919, 2023

    2023 arXiv

  6. [15]

    Gromov hyperbolic spaces,

    J. V ¨ais¨al¨a, “Gromov hyperbolic spaces,” Expositiones Mathematicae, vol. 23, no. 3, pp. 187–231, 2005

    2005

  7. [16]

    Hyperbolic image embeddings,

    V . Khrulkov, L. Mirvakhabova, E. Ustinova, I. Oseledets, and V . Lempitsky, “Hyperbolic image embeddings,” in Proceedings of the IEEE/CVF conference on computer vision and pattern recognition , 2020, pp. 6418–6428

    2020

  8. [18]

    A wrapped normal distribution on hyperbolic space for gradient-based learning,

    Y . Nagano, S. Yamaguchi, Y . Fujita, and M. Koyama, “A wrapped normal distribution on hyperbolic space for gradient-based learning,” in International Conference on Machine Learning . PMLR, 2019, pp. 4693–4702

    2019

  9. [19]

    Hyperbolic image segmentation,

    M. G. Atigh, J. Schoep, E. Acar, N. Van Noord, and P. Mettes, “Hyperbolic image segmentation,” in Proceedings of the IEEE/CVF conference on computer vision and pattern recognition , 2022, pp. 4453–4462

    2022

  10. [20]

    Poincar ´e embeddings for learning hierarchical representations,

    M. Nickel and D. Kiela, “Poincar ´e embeddings for learning hierarchical representations,” Advances in neural information processing systems, vol. 30, 2017

    2017

  11. [21]

    Learning continuous hierarchies in the lorentz model of hyperbolic geometry,

    ——, “Learning continuous hierarchies in the lorentz model of hyperbolic geometry,” in International conference on machine learning . PMLR, 2018, pp. 3779–3788

    2018

  12. [22]

    Representation tradeoffs for hyperbolic embeddings,

    F. Sala, C. De Sa, A. Gu, and C. R ´e, “Representation tradeoffs for hyperbolic embeddings,” in International conference on machine learning. PMLR, 2018, pp. 4460–4469

    2018

  13. [23]

    Low distortion delaunay embedding of trees in hyperbolic plane,

    R. Sarkar, “Low distortion delaunay embedding of trees in hyperbolic plane,” in International symposium on graph drawing . Springer, 2011, pp. 355–366

    2011

  14. [24]

    Hyperbolic neural networks++,

    R. Shimizu, Y . Mukuta, and T. Harada, “Hyperbolic neural networks++,” arXiv preprint arXiv:2006.08210 , 2020

    2006 arXiv

  15. [25]

    Nested hyperbolic spaces for dimensionality reduction and hyperbolic nn design,

    X. Fan, C.-H. Yang, and B. C. Vemuri, “Nested hyperbolic spaces for dimensionality reduction and hyperbolic nn design,” in Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition , 2022, pp. 356–365

    2022

  16. [26]

    Lorentzian fully hyperbolic generative adversarial network,

    E. Qu and D. Zou, “Lorentzian fully hyperbolic generative adversarial network,” arXiv preprint arXiv:2201.12825 , 2022

    2022 arXiv

  17. [27]

    Poincare resnet,

    M. van Spengler, E. Berkhout, and P. Mettes, “Poincare resnet,” in Proceedings of the IEEE/CVF International Conference on Computer Vision, 2023, pp. 5419–5428

    2023

  18. [28]

    Hgcf: Hyperbolic graph convolution networks for collaborative filtering,

    J. Sun, Z. Cheng, S. Zuberi, F. P ´erez, and M. V olkovs, “Hgcf: Hyperbolic graph convolution networks for collaborative filtering,” in Proceedings of the Web Conference 2021 , 2021, pp. 593–601

    2021

  19. [29]

    Hrcf: Enhancing collaborative filtering via hyperbolic geometric regularization,

    M. Yang, M. Zhou, J. Liu, D. Lian, and I. King, “Hrcf: Enhancing collaborative filtering via hyperbolic geometric regularization,” in Proceedings of the ACM Web Conference 2022 , 2022, pp. 2462–2471

    2022

  20. [30]

    Hypersorec: Exploiting hyperbolic user and item representations with multiple aspects for social-aware recommendation,

    H. Wang, D. Lian, H. Tong, Q. Liu, Z. Huang, and E. Chen, “Hypersorec: Exploiting hyperbolic user and item representations with multiple aspects for social-aware recommendation,” ACM Transactions on Information Systems (TOIS) , vol. 40, no. 2, pp. 1–28, 2021

    2021

  21. [31]

    Modeling scale-free graphs with hyperbolic geometry for knowledge- aware recommendation,

    Y . Chen, M. Yang, Y . Zhang, M. Zhao, Z. Meng, J. Hao, and I. King, “Modeling scale-free graphs with hyperbolic geometry for knowledge- aware recommendation,” in Proceedings of the fifteenth ACM international conference on web search and data mining , 2022, pp. 94–102

    2022

  22. [32]

    Hyperbolic hypergraphs for sequential recommendation,

    Y . Li, H. Chen, X. Sun, Z. Sun, L. Li, L. Cui, P. S. Yu, and G. Xu, “Hyperbolic hypergraphs for sequential recommendation,” in Proceedings of the 30th ACM International Conference on Information & Knowledge Management , ser. CIKM ’21. New York, NY , USA: Association for Compu...

    2021

  23. [33]

    Batch normalization: Accelerating deep network training by reducing internal covariate shift,

    S. Ioffe, “Batch normalization: Accelerating deep network training by reducing internal covariate shift,” arXiv preprint arXiv:1502.03167, 2015

    2015 arXiv

  24. [34]

    Differentiating through the fr ´echet mean,

    A. Lou, I. Katsman, Q. Jiang, S. Belongie, S.-N. Lim, and C. De Sa, “Differentiating through the fr ´echet mean,” in International conference on machine learning . PMLR, 2020, pp. 6393–6403

    2020

  25. [35]

    The numerical stability of hyperbolic representation learning,

    G. Mishne, Z. Wan, Y . Wang, and S. Yang, “The numerical stability of hyperbolic representation learning,” in International Conference on Machine Learning . PMLR, 2023, pp. 24 925–24 949

    2023

  26. [36]

    Clipped hyperbolic classifiers are super-hyperbolic classifiers,

    Y . Guo, X. Wang, Y . Chen, and S. X. Yu, “Clipped hyperbolic classifiers are super-hyperbolic classifiers,” in Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition , 2022, pp. 11–20

    2022

  27. [37]

    Continuous hierarchical representations with poincar ´e variational auto-encoders,

    E. Mathieu, C. Le Lan, C. J. Maddison, R. Tomioka, and Y . W. Teh, “Continuous hierarchical representations with poincar ´e variational auto-encoders,” Advances in neural information processing systems , vol. 32, 2019

    2019

  28. [38]

    L. W. Tu, Differential geometry: connections, curvature, and characteristic classes . Springer, 2017, vol. 275

    2017

  29. [39]

    do Carmo, Riemannian Geometry , ser

    M. do Carmo, Riemannian Geometry , ser. Mathematics (Boston, Mass.). Birkh ¨auser, 1992. [Online]. Available: https: //books.google.co.in/books?id=uXJQQgAACAAJ

    1992

  30. [40]

    Lang, Differential and Riemannian manifolds

    S. Lang, Differential and Riemannian manifolds . Springer Science & Business Media, 1995

    1995

  31. [41]

    J. M. Lee, Riemannian manifolds: an introduction to curvature . Springer Science & Business Media, 2006, vol. 176

    2006

  32. [42]

    Ungar, A gyrovector space approach to hyperbolic geometry

    A. Ungar, A gyrovector space approach to hyperbolic geometry . Springer Nature, 2022

    2022

  33. [43]

    A geometric interpretation of ungar’s addition and of gyration in the hyperbolic plane,

    J. Vermeer, “A geometric interpretation of ungar’s addition and of gyration in the hyperbolic plane,” Topology and its Applications , vol. 152, no. 3, pp. 226–242, 2005

    2005

  34. [44]

    Gy ¨orfi, M

    L. Gy ¨orfi, M. Kohler, A. Krzyzak, and H. Walk, A Distribution-Free Theory of Nonparametric Regression , ser. Springer Series in Statistics. Springer New York, 2002. [Online]. Available: https://books.google.co.in/books?id=Ovmb9oGBlo0C

    2002

  35. [45]

    House price prediction based on machine learning: A case of king county,

    Y . Wang and Q. Zhao, “House price prediction based on machine learning: A case of king county,” in 2022 7th International Conference on Financial Innovation and Economic Development (ICFIED 2022) . Atlantis Press, 2022, pp. 1547–1555

    2022

  36. [46]

    A data-driven statistical model for predicting the critical temperature of a superconductor,

    K. Hamidieh, “A data-driven statistical model for predicting the critical temperature of a superconductor,” Computational Materials Science, 2018. [Online]. Available: https://api.semanticscholar.org/CorpusID:55069173

    2018

  37. [47]

    Optimization of power take-off system settings and regional site selection procedure for a wave energy converter,

    H. Mehdipour, E. Amini, S. T. O. Naeeni, M. Neshat, and A. H. Gandomi, “Optimization of power take-off system settings and regional site selection procedure for a wave energy converter,” Energy Conversion and Management: X , vol. 22, p. 100559, 2024

    2024

  38. [48]

    Activity recognition using cell phone accelerometers,

    J. R. Kwapisz, G. M. Weiss, and S. A. Moore, “Activity recognition using cell phone accelerometers,” ACM SigKDD Explorations Newsletter, vol. 12, no. 2, pp. 74–82, 2011. 18

    2011

Pith tools

Reviewed August 12, 2026 · model on record in the stance chip above.