Pith. sign in

REVIEW 3 major objections 3 minor 55 references

Supervised Quadratic Feature Analysis: Information Geometry Approach for Dimensionality Reduction

T0 review · 3 major / 3 minor · reviewed 2026-08-09 · deepseek-v4-flash

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

Pith's one-line read The paper argues that linear filters learned by maximizing Fisher-Rao distances between class-conditional Gaussians, with the Calvo-Oller lower bound as a closed-form surrogate, produce features whose QDA accuracy matches or beats LDA…

desk verdict Clean new objective and a useful Hellinger finding, but the Fisher-Rao interpretation rests on a lower bound whose tightness at the optima is unverified; worth reviewing with that fix. read the letter →

arxiv 2502.00168 v5 pith:F2Z2LJWZ submitted 2025-01-31 stat.ML cs.LGmath.DGmath.STstat.TH

Daniel Herrera-Esposito , Johannes Burge This is my paper · ORCID
classification stat.MLcs.LGmath.DGmath.STstat.TH MSC 62H3062B1068T10
keywords Fisher-RaodistanceinformationgeometrysuperviseddimensionalityreductionquadraticdiscriminantanalysisHellingerBhattacharyyaCalvo-Ollerboundlinearfeaturelearning
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

Supervised Quadratic Feature Analysis (SQFA) learns $m$ linear filters by maximizing pairwise Fisher-Rao distances between class-conditional Gaussian distributions in the projected space. Because the Fisher-Rao distance integrates local discriminability along the geodesic between distributions, the authors treat it as a geometrically motivated proxy for class separability. For general Gaussians the true distance has no closed form, so SQFA maximizes the Calvo-Oller lower bound, and a zero-mean variant (smSQFA) uses the exact affine-invariant distance on the manifold of covariance matrices. On SVHN, MNIST, and a naturalistic speed-estimation task, SQFA features support QDA accuracy competitive with or better than LDA, PCA, SPCA, LMNN, and Bhattacharyya-based objectives; the Hellinger variant SQFA-H is consistently the best. The paper's claim is that information geometry offers a practical, efficient objective for supervised linear dimensionality reduction, not just a theoretical framing.

What carries the argument

The central object is the Fisher-Rao distance on the manifold of Gaussian distributions, made computable through two surrogates. For zero-mean Gaussians the manifold is $\mathrm{SPD}(m)$, and the distance is the affine-invariant distance $d_{AI}(A,B)=\sqrt{\sum_{k=1}^m \log^2\lambda_k}$, with $\lambda_k$ the generalized eigenvalues of the pair $(A,B)$; this is the exact objective of smSQFA. For arbitrary Gaussians, the Calvo-Oller embedding $\Omega_i=\begin{pmatrix} \Sigma_i+\mu_i\mu_i^T & \mu_i \\ \mu_i^T & 1 \end{pmatrix}$ maps each class into $\mathrm{SPD}(m+1)$, and the affine-invariant distance in that embedding is a lower bound that is exact when means are equal and, the paper argues, close to the true distance for its datasets. SQFA maximizes the sum of pairwise distances over the linear filter matrix $F\in\mathbb{R}^{n\times m}$, with each gradient step costing $O(c^2m^3+cmn^2)$, which keeps training on a consumer laptop to seconds. The same optimization machinery with the Bhattacharyya and Hellinger distances defines SQFA-B and SQFA-H, isolating the effect of the objective function from the effect of the optimization procedure.

What would settle it

Compute the true Fisher-Rao distance for every class pair of a new dataset together with the Calvo-Oller bound, and check both the bound's tightness and whether SQFA's filters are the ones maximizing the bound. If on a dataset with strong mean-covariance interactions the bound is loose by a wide margin and SQFA's QDA accuracy falls below SQFA-H or LMNN, the claim that the surrogate faithfully represents the Fisher-Rao objective is undermined; a cleaner version would train filters for small $m$ by directly optimizing the numerical Fisher-Rao distance and compare their filters and accuracy with the bound-based SQFA.

Watch

Extended reading notes

Core claim

SQFA's central claim is that maximizing Fisher-Rao distances between class-conditional Gaussians is a workable and competitive objective for supervised linear dimensionality reduction. In the zero-mean case the objective is exact: with class second-moment matrices $\Psi_i$ and $\Psi_j$, the Fisher-Rao distance equals the affine-invariant distance $d_{AI}(\Psi_i,\Psi_j)=\sqrt{\sum_{k=1}^m \log^2 \lambda_k}$, where $\lambda_k$ are the generalized eigenvalues of the matrix pair. For general Gaussians, SQFA maximizes the Calvo-Oller bound $d_{FR}(\theta_i,\theta_j)\ge d_{AI}(\Omega_i,\Omega_j)/\sqrt{2}$ obtained by embedding each Gaussian into an $(m+1)\times(m+1)$ symmetric positive definite matrix via its mean and covariance; the paper verifies empirically that the bound closely tracks the numerically computed Fisher-Rao distance for the class pairs in its three datasets. Using QDA accuracy as the evaluation criterion, SQFA features are competitive with features that maximize Bhattacharyya distance and with LDA, PCA, SPCA, and LMNN, and on the speed-estimation task they come close to AMA-Gauss, a method that directly optimizes quadratic decodability. A second, more specific claim is that the Hellinger variant, SQFA-H, consistently outperforms the Bhattacharyya variant, SQFA-B, in multiclass problems even though the two objectives are monotone-equivalent in the two-class case.

Load-bearing premise

The load-bearing premise is that the Calvo-Oller lower bound stays numerically close to the true Fisher-Rao distance for the class-conditional Gaussians of the data being processed; the paper verifies this on its three datasets but offers no guarantee for a new one.

Editorial extensions

If this is right

  • SQFA makes the Fisher-Rao distance usable in practice: a closed-form, laptop-scale optimization replaces the numerically expensive true distance, giving a linear method whose QDA accuracy matches or exceeds standard supervised and unsupervised baselines on image and naturalistic-video tasks.
  • The Hellinger distance, rarely used for dimensionality reduction, emerges as a consistently strong multiclass objective, outperforming the more common Bhattacharyya distance when classes are unequally separated.
  • Because the Calvo-Oller bound extends to other elliptical distributions (Student-t and Cauchy), the Gaussian assumption can be relaxed without abandoning the information-geometric objective.
  • On the speed-estimation task, SQFA features resemble motion-sensitive receptive fields and perform close to AMA-Gauss, which directly optimizes the quadratic decoder, suggesting that Fisher-Rao maximization is a good proxy for optimal quadratic decoding.
  • LDA is the special case of this framework when class covariances are equal: maximizing pairwise squared Mahalanobis distances is equivalent to the LDA criterion, so SQFA generalizes LDA to heteroscedastic classes.

Reading between the lines

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

  • Editorial extension: the paper does not prove that the Hellinger advantage comes from its boundedness, but if it does, replacing other unbounded pairwise objectives with bounded variants could improve multiclass dimensionality reduction methods beyond SQFA.
  • Editorial extension: the tightness of the Calvo-Oller bound is validated only on the three datasets studied; a direct test would optimize the numerically computed Fisher-Rao distance for small $m$ and compare the resulting filters and QDA accuracy with those produced by the bound-based SQFA.
  • Editorial extension: since the same machinery extends to elliptical distributions, a natural next experiment is to repeat the comparisons on heavy-tailed class-conditional data, which would separate the Gaussian assumption from the information-geometry objective itself.
  • Editorial extension: in neuroscience, SQFA-style objectives could identify axes of neural population activity where condition-dependent covariance differences are largest; the paper frames this as a potential application but does not test it.
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 paper introduces Supervised Quadratic Feature Analysis (SQFA), a linear supervised dimensionality reduction method that learns filters by maximizing Fisher-Rao distances between class-conditional Gaussian distributions in the low-dimensional feature space. For zero-mean Gaussians it uses the exact affine-invariant distance (smSQFA), and for general Gaussians it uses the Calvo-Oller lower bound as a closed-form surrogate. The authors compare SQFA and its Bhattacharyya and Hellinger variants against PCA, LDA, SPCA, and LMNN on SVHN, MNIST, and a naturalistic speed-estimation task, reporting QDA accuracies in which SQFA variants are competitive and SQFA-H is often best. The paper includes several appendices on the distance proxies, computational complexity, invariance properties, and robustness to non-Gaussianity, and it provides a Python implementation.

Significance. If the central claims hold, SQFA is a computationally efficient linear feature learning method that gives the Fisher-Rao geometry a practical role in supervised dimensionality reduction, and the paper provides evidence that the rarely used Hellinger distance is a strong multiclass objective. The manuscript is careful in several respects: the method is precisely specified, the optimization cost is analyzed (Appendix D), the Calvo-Oller bound is numerically compared with the true Fisher-Rao distance (Appendix B), robustness of the evaluation to non-Gaussianity is checked with Gaussian-simulated data and a kNN decoder (Appendix G), and a public implementation is provided. The main unresolved issue is that the validity of the general-Gaussian SQFA objective is established only on raw-data class pairs, not on the feature-space distributions actually optimized, so the geometric interpretation is not yet fully supported at the points where it matters.

major comments (3)
  1. [Section 2.3, Appendix B (Figure 10)] The claim that SQFA maximizes Fisher-Rao distances for general Gaussians rests on the Calvo-Oller bound being a faithful surrogate. Appendix B validates the bound by comparing it with the numerically computed true Fisher-Rao distance for class-conditional Gaussians in the original data space. However, the optimization in Eq. (1) operates on the projected feature-space parameters (µ_i = F^T γ_i, Σ_i = F^T Φ_i F), whose means and covariances are different from the raw-data pairs validated in Figure 10. Maximizing a lower bound need not maximize the true distance if the tightness of the bound varies across projections, and Figure 10 shows nontrivial deviations for MNIST. The geometric motivation in Section 2.2 is therefore load-bearing only if the surrogate tracks the true geodesic distance at the feature-space optima. Please add a validation of the bound at the learned feature-space distributions for each dataset, or provide a theoretical guarantee on the bound's tightness under projection.
  2. [Sections 5.2-5.4, Figures 4-6] The empirical comparison is based on point estimates. The SQFA variants are reported with medians and interquartile ranges over 20 initializations, but the baseline methods have no error bars and no significance testing is reported. The abstract's claim that SQFA-H achieves the best classification accuracy is not supported without uncertainty quantification on the differences. Please report variability over initializations and data resamples, or perform paired significance tests, for all methods, and state explicitly how many repetitions were used for each baseline.
  3. [Section 4, Appendix F.2] The regularization parameter σ² is selected by grid search on a validation split for the digit datasets and fixed to 0.001 for the speed-estimation dataset, but no sensitivity analysis is reported. Since regularization breaks the invariance properties that motivate the objective and can bias filters toward directions with larger second-moment matrices, the comparison could depend on this choice. Please report the sensitivity of the main results to σ² and to the shrinkage parameter used for LDA, or otherwise justify that the chosen values do not drive the conclusions.
minor comments (3)
  1. [Appendix B] The text in Appendix B.1 refers to "Figure 8" when presenting the real-world comparison of the Calvo-Oller bound with the true Fisher-Rao distance, but the actual figure is Figure 10; please correct the cross-references throughout the supplement.
  2. [Appendix G] In the first paragraph of Appendix G, the phrase "which that the data is Gaussian" is ungrammatical and should read "which assumes that the data is Gaussian."
  3. [Appendix B.2] The phrase "alternative expression" should be "alternative expression" (spelling), and the sentences introducing Equations (11) and (12) could be merged for readability.

Circularity Check

0 steps flagged · score 1.0 of 10

No significant circularity: SQFA's derivation is self-contained; the Calvo-Oller bound and benchmarks are external, and the empirical claims are held-out comparisons.

full rationale

The paper's derivation chain is not circular. SQFA's objective is explicitly defined as maximizing the Calvo-Oller lower bound on the Fisher-Rao distance (Section 2.3 and Equation 1), and the bound is an external published result (Calvo and Oller, 1990; Nielsen, 2023), not fitted to the datasets. The paper states that the bound 'closely approximates the true distance for the analyzed datasets' and validates this in Appendix B; this is an empirical claim, not a reduction of the target result to its inputs. Even if the bound were loose in feature space for a new dataset, that would be a correctness or generalization limitation, not circularity, because the paper does not claim to derive the bound's tightness from the datasets it later predicts. The main contributions are evaluated with held-out QDA and kNN accuracy against external baselines (LDA, PCA, SPCA, LMNN, AMA-Gauss), and the speed-estimation dataset and AMA-Gauss method come from prior published work by the same group, but they serve as legitimate external benchmarks rather than as premises in the derivation. The Appendix C proof that LDA maximizes squared Mahalanobis distances is a standard self-contained derivation. No equation in the paper is equivalent to its own inputs by construction, and no fitted parameter is renamed as a prediction. The self-citations present are not load-bearing in deriving SQFA's objective or its geometric interpretation; therefore the appropriate score is low, reflecting only minor non-circular reliance on the authors' earlier benchmark resources.

Assumptions & free parameters 1 free parameters · 5 assumptions · 0 invented entities

The method relies on standard mathematical background (generalized eigenvalue decompositions, affine invariance of the Fisher-Rao metric) and on two domain assumptions: that class-conditional distributions are Gaussian, and that maximizing pairwise Fisher-Rao (or bound) distances is a good proxy for multiclass discriminability. The free parameter is the regularization strength sigma^2, chosen by validation. No invented entities are introduced.

free parameters (1)
  • Regularization parameter sigma^2 = grid-selected for SVHN and MNIST; 0.001 for speed estimation
    Added to class covariance matrices in feature space to avoid ill-conditioning (Section 4). Results may depend on this choice; selected via validation set for digit datasets.
assumptions (5)
  • domain assumption Class-conditional distributions in feature space are Gaussian
    Used throughout Sections 2 and 4 to compute Fisher-Rao, Bhattacharyya and Hellinger objectives, and to justify QDA evaluation.
  • domain assumption The Calvo-Oller lower bound is a close approximation to the true Fisher-Rao distance for the data analyzed
    Section 2.3 and Appendix B; the method optimizes this bound for general Gaussians, so its tightness is required for the geometric interpretation.
  • domain assumption Summing pairwise dissimilarities is a useful proxy for multiclass discriminability
    Equation 1; Discussion acknowledges pairwise errors do not determine multiclass error (Loog et al., 2001; Thangavelu and Raich, 2008).
  • domain assumption Fisher-Rao distance is a good proxy for discriminability
    Section 2.2 argues this via accumulated local discriminability; Appendix A gives supporting evidence only for zero-mean 1D/2D Gaussians.
  • standard math Affine invariance of Fisher-Rao distance and generalized eigenvalue decompositions
    Used in Sections 2.3 and Appendix A/F to justify filter non-uniqueness and distance computations.

how reviews work

0 comments
Cite this review

Pith. "Pith review of Supervised Quadratic Feature Analysis: Information Geometry Approach for Dimensionality Reduction." pith.science (2026). https://pith.science/paper/F2Z2LJWZ

@misc{pith2026250200168,
  author       = {Pith},
  title        = {Pith review of: Supervised Quadratic Feature Analysis: Information Geometry Approach for Dimensionality Reduction},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/F2Z2LJWZ}},
  note         = {Machine review of arXiv:2502.00168}
}
read the original abstract

Supervised dimensionality reduction maps labeled data into a low-dimensional feature space while preserving class separation. A common strategy is to learn features that maximize a measure of statistical dissimilarity between the class-conditional probability distributions. Information geometry, which is rooted in Riemannian geometry, provides an alternative framework for measuring class dissimilarity. It treats probability distributions as points in a statistical manifold and uses the Fisher information metric to define a geodesic distance--the Fisher-Rao distance--between distributions The Fisher-Rao distance is an appealing candidate for measuring class separation because the Fisher information metric is a local measure of discriminability, and because it allows a geometric interpretation. Here, we present Supervised Quadratic Feature Analysis (SQFA), a supervised dimensionality reduction method which learns linear features that maximize Fisher-Rao distances between class-conditional distributions, under Gaussian assumptions. In multiple real world datasets, we find that SQFA features support classification accuracy that is competitive with features that maximize more popular measures of dissimilarity, or that are learned by other state-of-the-art dimensionality reduction methods. Notably, the best classification accuracy is achieved by SQFA-H features, a variant of SQFA that maximizes the Hellinger distance, a rarely used objective for dimensionality reduction. These results demonstrate the potential of information geometry as a tool for supervised dimensionality reduction. We provide a Python implementation of SQFA at https://github.com/dherrera1911/sqfa.

Figures

Figures reproduced from arXiv: 2502.00168 by the authors.

Figure 1
Figure 1. SQFA learns features using information geometry. [PITH_FULL_IMAGE:figures/full_fig_p002_1.png] view at source ↗
Figure 2
Figure 2. SQFA vs. LDA vs. PCA. Ellipses show the probability distributions for three classes (colors) in a 6D toy dataset. Each panel shows two dimensions of the data vector x, where the classes are separated by dif￾ferent statistical properties. Classes are distinguished by large differences in the covariances (dimensions 1- 2), small differences in the means (dimensions 3-4), or neither (dimensions 5-6). Two filters were l… view at source ↗
Figure 3
Figure 3. SQFA vs. smSQFA. Ellipses show the prob [PITH_FULL_IMAGE:figures/full_fig_p006_3.png] view at source ↗
Figures from the paper (12 more)
Figure 4
Figure 4. Figure 4: SQFA extracts useful features using class [PITH_FULL_IMAGE:figures/full_fig_p006_4.png]
Figure 5
Figure 5. Figure 5: SQFA can exploit class-conditional first- and [PITH_FULL_IMAGE:figures/full_fig_p007_5.png]
Figure 6
Figure 6. Figure 6: SQFA features are close to optimal for quadratic discrimination. Top. 4 example videos with the same speed (left) and 4 example videos with dif￾ferent speeds (right). Each video is shown as a 2D space-time plot where the vertical axis is time and the horizontal axis is…
Figure 7
Figure 7. Figure 7: Distances and Bayes error for 1D Gaussian distributions. From left to right, the three panels show, [PITH_FULL_IMAGE:figures/full_fig_p013_7.png]
Figure 8
Figure 8. Figure 8: Distances and Bayes error for 2D Gaussian distributions. From left to right, the panels show for two [PITH_FULL_IMAGE:figures/full_fig_p014_8.png]
Figure 9
Figure 9. Figure 9: Calvo-Oller bound vs. true Fisher-Rao distance for the equal-covariance case. The Fisher-Rao distance [PITH_FULL_IMAGE:figures/full_fig_p015_9.png]
Figure 10
Figure 10. Figure 10: Calvo-Oller bound vs. true Fisher-Rao distance in real-world datasets. Each panel shows the Calvo-Oller [PITH_FULL_IMAGE:figures/full_fig_p016_10.png]
Figure 11
Figure 11. Figure 11: Distances as a function of Mahalanobis distance. The three distances (Fisher-Rao, Hellinger, [PITH_FULL_IMAGE:figures/full_fig_p017_11.png]
Figure 12
Figure 12. Figure 12: Distances as a function of covariance differences. The three distances (Fisher-Rao, Hellinger, [PITH_FULL_IMAGE:figures/full_fig_p018_12.png]
Figure 13
Figure 13. Figure 13: Training times for the different models on the different datasets. Times are in seconds, and indicated [PITH_FULL_IMAGE:figures/full_fig_p020_13.png]
Figure 14
Figure 14. Figure 14: QDA accuracy with simulated Gaussian data. After learning the filters with each method, we simulated [PITH_FULL_IMAGE:figures/full_fig_p023_14.png]
Figure 15
Figure 15. Figure 15: KNN accuracy for each of the datasets. For each dataset, the filters were learned as described in the [PITH_FULL_IMAGE:figures/full_fig_p024_15.png]

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

55 extracted references · 53 canonical work pages

  1. [1]

    Amari, S.-i. (2016). Information Geometry and Its Applications . Springer. Google-Books-ID: UkSFCwAAQBAJ

    work page 2016

  2. [2]

    Arvanitidis, G., González-Duque, M., Pouplin, A., Kalatzis, D., and Hauberg, S. (2022). Pulling back information geometry. In Proceedings of the 25th International Conference on Artificial Intelligence and Statistics , volume 151, Valencia, Spain. PMLR. arXiv:2106.05367 [cs]

    work page Pith review arXiv 2022

  3. [3]

    and Mitchell, A

    Atkinson, C. and Mitchell, A. F. S. (1981). Rao's Distance Measure . Sankhyā: The Indian Journal of Statistics, Series A (1961-2002) , 43(3):345--365. Publisher: Springer

    work page 1981

  4. [4]

    Barshan, E., Ghodsi, A., Azimifar, Z., and Zolghadri Jahromi, M. (2011). Supervised principal component analysis: Visualization , classification and regression on subspaces and submanifolds. Pattern Recognition , 44(7):1357--1371

    work page 2011

  5. [5]

    Bujack, R., Teti, E., Miller, J., Caffrey, E., and Turton, T. L. (2022). The non- Riemannian nature of perceptual color space. Proceedings of the National Academy of Sciences , 119(18):e2119753119. Publisher: Proceedings of the National Academy of Sciences

    work page 2022

  6. [6]

    Burge, J. (2020). Image- Computable Ideal Observers for Tasks with Natural Stimuli . Annual Review of Vision Science , 6(1):491--517. \_eprint: https://doi.org/10.1146/annurev-vision-030320-041134

    work page doi:10.1146/annurev-vision-030320-041134 2020

  7. [7]

    and Geisler, W

    Burge, J. and Geisler, W. S. (2015). Optimal speed estimation in natural image movies predicts human performance. Nature Communications , 6(1):7900. Publisher: Nature Publishing Group

    work page 2015

  8. [8]

    and Oller, J

    Calvo, M. and Oller, J. M. (1990). A distance between multivariate normal distributions based in an embedding into the siegel group. Journal of Multivariate Analysis , 35(2):223--242

    work page 1990

Show all 55 references
  1. [9]

    and Oller, J

    Calvo, M. and Oller, J. M. (2002). A distance between elliptical distributions based in an embedding into the Siegel group. Journal of Computational and Applied Mathematics , 145(2):319--334

    2002

  2. [10]

    M., Raich, R., and Hero, A

    Carter, K. M., Raich, R., and Hero, A. O. (2009). An information geometric approach to supervised dimensionality reduction. In 2009 IEEE International Conference on Acoustics , Speech and Signal Processing , pages 1829--1832, Taipei, Taiwan. IEEE

    2009

  3. [11]

    Chin, B. M. and Burge, J. (2020). Predicting the Partition of Behavioral Variability in Speed Perception with Naturalistic Stimuli . The Journal of Neuroscience , 40(4):864--879

    2020

  4. [12]

    and Lee, C

    Choi, E. and Lee, C. (2003). Feature extraction based on the Bhattacharyya distance. Pattern Recognition , 36(8):1703--1709

    2003

  5. [13]

    and Abbott, L

    Chung, S. and Abbott, L. F. (2021). Neural population geometry: An approach for understanding biological and artificial neural networks. Current Opinion in Neurobiology , 70:137--144

    2021

  6. [14]

    Cunningham, J. P. and Ghahramani, Z. (2015). Linear Dimensionality Reduction : Survey , Insights , and Generalizations . Journal of Machine Learning Research , 16:2859--2900

    2015

  7. [15]

    and Abbott, L

    Dayan, P. and Abbott, L. F. (2005). Theoretical Neuroscience : Computational and Mathematical Modeling of Neural Systems . MIT Press. Google-Books-ID: fLT4DwAAQBAJ

    2005

  8. [16]

    Ding, X., Lee, D., Melander, J., Sivulka, G., Ganguli, S., and Baccus, S. (2023). Information Geometry of the Retinal Representation Manifold . Advances in Neural Information Processing Systems , 36:44310--44322

    2023

  9. [17]

    and Loog, M

    Duin, R. and Loog, M. (2004). Linear dimensionality reduction via a heteroscedastic extension of LDA : the Chernoff criterion. IEEE Transactions on Pattern Analysis and Machine Intelligence , 26(6):732--739. Conference Name: IEEE Transactions on Pattern Analysis and Machine In...

    2004

  10. [18]

    R., Zhou, J., Nassar, J., Berman, J., Olieslagers, J., and Williams, A

    Duong, L. R., Zhou, J., Nassar, J., Berman, J., Olieslagers, J., and Williams, A. H. (2023). Representational dissimilarity metric spaces for stochastic neural networks. arXiv:2211.11665 [cs, q-bio]

    2023 arXiv

  11. [19]

    Dwivedi, A., Wang, S., and Tajer, A. (2022). Discriminant Analysis under f- Divergence Measures . Entropy , 24(2):188. Number: 2 Publisher: Multidisciplinary Digital Publishing Institute

    2022

  12. [20]

    E., Williams, A

    Feather, J., Lipshutz, D., Harvey, S. E., Williams, A. H., and Simoncelli, E. P. (2024). Discriminating image representations with principal distortions. arXiv:2410.15433 [q-bio]

    2024 arXiv

  13. [21]

    Fechner, G. T. (1860). Elemente der psychophysik , volume 2. Breitkopf u. Härtel

  14. [22]

    Flamary, R., Cuturi, M., Courty, N., and Rakotomamonjy, A. (2018). Wasserstein Discriminant Analysis . Machine Learning , 107(12):1923--1945. arXiv:1608.08063 [cs, stat]

    2018 arXiv

  15. [23]

    Fukunaga, K. (1990). Introduction to Statistical Pattern Recognition . Academic Press, second edition. Google-Books-ID: BIJZTGjTxBgC

    1990

  16. [24]

    and Burge, J

    Herrera-Esposito, D. and Burge, J. (2024). Optimal estimation of local motion-in-depth with naturalistic stimuli. Journal of Neuroscience . Publisher: Society for Neuroscience Section: Research Articles

    2024

  17. [25]

    H., and Brainard, D

    Hong, F., Bouhassira, R., Chow, J., Sanders, C., Shvartsman, M., Guan, P., Williams, A. H., and Brainard, D. H. (2025). Comprehensive characterization of human color discrimination thresholds. ISSN: 2692-8205 Pages: 2025.07.16.665219 Section: New Results

    2025

  18. [26]

    and Burge, J

    Jaini, P. and Burge, J. (2017). Linking normative models of natural tasks to descriptive models of neural response. Journal of Vision , 17(12):16

    2017

  19. [27]

    Kailath, T. (1967). The Divergence and Bhattacharyya Distance Measures in Signal Selection . IEEE Transactions on Communications , 15(1):52--60

    1967

  20. [28]

    and Mineiro, P

    Karampatziakis, N. and Mineiro, P. (2014). Discriminative Features via Generalized Eigenvectors . International conference on machine learning

    2014

  21. [29]

    Kohn, A., Coen-Cagli, R., Kanitscheider, I., and Pouget, A. (2016). Correlations and Neuronal Population Information . Annual Review of Neuroscience , 39(1):237--256. \_eprint: https://doi.org/10.1146/annurev-neuro-070815-013851

    2016 doi

  22. [30]

    and Wei, X.-X

    Kriegeskorte, N. and Wei, X.-X. (2021). Neural tuning and representational geometry. Nature Reviews Neuroscience , 22(11):703--718. Publisher: Nature Publishing Group

    2021

  23. [31]

    LeCun, Y., Bengio, Y., and Hinton, G. (2015). Deep learning. Nature , 521(7553):436--444. Publisher: Nature Publishing Group

    2015

  24. [32]

    Lezcano Casado, M. (2019). Trivializations for Gradient - Based Optimization on Manifolds . In Advances in Neural Information Processing Systems , volume 32. Curran Associates, Inc

    2019

  25. [33]

    Loog, M., Duin, R., and Haeb-Umbach, R. (2001). Multiclass linear dimension reduction by weighted pairwise Fisher criteria. IEEE Transactions on Pattern Analysis and Machine Intelligence , 23(7):762--766. Conference Name: IEEE Transactions on Pattern Analysis and Machine Intelligence

    2001

  26. [34]

    T., Tanaka, H., Grant, S., Kastner, D

    Maheswaranathan, N., McIntosh, L. T., Tanaka, H., Grant, S., Kastner, D. B., Melander, J. B., Nayebi, A., Brezovec, L. E., Wang, J. H., Ganguli, S., and Baccus, S. A. (2023). Interpreting the retinal neural code for natural scenes: From computations to neurons. Neuron , 111(17...

    2023

  27. [35]

    K., Meneghetti, F

    Miyamoto, H. K., Meneghetti, F. C. C., Pinele, J., and Costa, S. I. R. (2024). On closed-form expressions for the Fisher – Rao distance. Information Geometry , 7(2):311--354

    2024

  28. [36]

    Moreno-Bote, R., Beck, J., Kanitscheider, I., Pitkow, X., Latham, P., and Pouget, A. (2014). Information-limiting correlations. Nature Neuroscience , 17(10):1410--1417. Publisher: Nature Publishing Group

    2014

  29. [37]

    A., Thompson, I

    Movshon, J. A., Thompson, I. D., and Tolhurst, D. J. (1978). Spatial and temporal contrast sensitivity of neurones in areas 17 and 18 of the cat's visual cortex. The Journal of Physiology , 283(1):101--120

    1978

  30. [38]

    Nejatbakhsh, A., Garon, I., and Williams, A. (2023). Estimating Noise Correlations Across Continuous Conditions With Wishart Processes . Advances in Neural Information Processing Systems , 36:54032--54045

    2023

  31. [39]

    Nielsen, F. (2020). An Elementary Introduction to Information Geometry . Entropy , 22(10):1100. Number: 10 Publisher: Multidisciplinary Digital Publishing Institute

    2020

  32. [40]

    Nielsen, F. (2023). A Simple Approximation Method for the Fisher – Rao Distance between Multivariate Normal Distributions . Entropy , 25(4):654. Number: 4 Publisher: Multidisciplinary Digital Publishing Institute

    2023

  33. [41]

    and Soen, A

    Nielsen, F. and Soen, A. (2024). pyBregMan : A Python library for Bregman Manifolds . arXiv:2408.04175 [cs]

    2024 arXiv

  34. [42]

    Pedregosa, F., Varoquaux, G., Gramfort, A., Michel, V., Thirion, B., Grisel, O., Blondel, M., Prettenhofer, P., Weiss, R., Dubourg, V., Vanderplas, J., Passos, A., and Cournapeau, D. (2011). Scikit-learn: Machine Learning in Python . Journal of Machine Learning Research , 12:2...

    2011

  35. [43]

    J., Lisberger, S

    Priebe, N. J., Lisberger, S. G., and Movshon, J. A. (2006). Tuning for Spatiotemporal Frequency and Speed in Directionally Selective Neurons of Macaque Striate Cortex . Journal of Neuroscience , 26(11):2941--2950. Publisher: Society for Neuroscience Section: Articles

    2006

  36. [44]

    and Herrera, M

    Rueda, L. and Herrera, M. (2008). Linear dimensionality reduction by maximizing the Chernoff distance in the transformed space. Pattern Recognition , 41(10):3138--3152

    2008

  37. [45]

    C., Schwartz, O., Movshon, J

    Rust, N. C., Schwartz, O., Movshon, J. A., and Simoncelli, E. P. (2005). Spatiotemporal Elements of Macaque V1 Receptive Fields . Neuron , 46(6):945--956

    2005

  38. [46]

    Sainburg, T., McInnes, L., and Gentner, T. Q. (2021). Parametric UMAP Embeddings for Representation and Semisupervised Learning . Neural Computation , pages 1--27

    2021

  39. [47]

    Srivastava, A., Jermyn, I., and Joshi, S. (2007). Riemannian Analysis of Probability Density Functions with Applications in Vision . In 2007 IEEE Conference on Computer Vision and Pattern Recognition , pages 1--8, Minneapolis, MN, USA. IEEE

    2007

  40. [48]

    Tao, D., Li, X., Wu, X., and Maybank, S. J. (2007). General Averaged Divergence Analysis . In Seventh IEEE International Conference on Data Mining ( ICDM 2007) , pages 302--311. ISSN: 2374-8486

    2007

  41. [49]

    and Raich, R

    Thangavelu, M. and Raich, R. (2008). Multiclass linear dimension reduction via a generalized Chernoff bound. In 2008 IEEE Workshop on Machine Learning for Signal Processing , pages 350--355. ISSN: 2378-928X

    2008

  42. [50]

    and Mamassian, P

    Vacher, J. and Mamassian, P. (2024). Perceptual Scales Predicted by Fisher Information Metrics . arXiv:2310.11759 [q-bio]

    2024 arXiv

  43. [51]

    d., Carey, C

    Vazelhes, W. d., Carey, C. J., Tang, Y., Vauquier, N., and Bellet, A. (2020). metric-learn: Metric Learning Algorithms in Python . Journal of Machine Learning Research , 21(138):1--6

    2020

  44. [52]

    and Ponce, C

    Wang, B. and Ponce, C. R. (2021). The Geometry of Deep Generative Image Models and its Applications . arXiv:2101.06006 [cs, math]

    2021 arXiv

  45. [53]

    Weinberger, K. Q. and Saul, L. K. (2009). Distance Metric Learning for Large Margin Nearest Neighbor Classification . Journal of Machine Learning Research , 10:207--244

    2009

  46. [54]

    and Wessel, R

    Ye, Z. and Wessel, R. (2024). Speed modulations in grid cell information geometry. Pages: 2024.09.18.613797 Section: New Results

    2024

  47. [55]

    R., and Simoncelli, E

    Zhou, J., Duong, L. R., and Simoncelli, E. P. (2024). A unified framework for perceived magnitude and discriminability of sensory stimuli. Proceedings of the National Academy of Sciences , 121(25):e2312293121. Publisher: Proceedings of the National Academy of Sciences

    2024

Pith tools

Reviewed August 9, 2026 · model on record in the stance chip above.