Pith. sign in

REVIEW 4 major objections 6 minor 40 references

Regularization via f-Divergence: An Application to Multi-Oxide Spectroscopic Analysis

T0 review · 4 major / 6 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 This paper proposes an f-divergence regularizer that penalizes the distributional gap between a network's predictions and its noisy targets, and reports that it matches or beats L1, L2, and dropout on Mars-rock LIBS spectra while further…

desk verdict Honest, careful experiments on a useful problem, but the paper's central f-divergence claim is not supported by the math and the estimator can violate its own [0,1] bound. read the letter →

arxiv 2502.03755 v1 pith:GKXRPLDQ submitted 2025-02-06 cs.LG

Weizhi Li , Natalie Klein , Brendan Gifford , Elizabeth Sklute , Carey Legett , Samuel Clegg This is my paper · ORCID
classification cs.LG
keywords oxide-weightsmulti-responseregressionneuralnetworkregularizationf-divergenceLIBSspectroscopyoverfittingmulti-target
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 tries to establish that adding an f-divergence penalty between a neural network's predictions and its noisy regression targets reduces overfitting on small spectroscopic datasets. On multi-oxide weight prediction from LIBS spectra collected under Mars-like conditions, the proposed regularizer reportedly achieves smaller or comparable test root mean squared errors than L1, L2, and dropout, and combining it with those regularizers further improves performance. The motivating intuition is that a good regression model keeps an intermediate level of divergence between predictions and noisy targets: too little divergence means overfitting, too much means underfitting. If the claim holds, this gives planetary scientists a practical way to train convolutional networks on the small LIBS datasets that rover instruments produce.

What carries the argument

The load-bearing mechanism is the differentiable f-divergence estimator: a fully connected Euclidean graph over each batch of predictions and targets, with edge weights $e^{-w(i,j)/\lambda}$ normalized by softmax, whose cross-type weight sum $\hat{t}_n$ in Eq. (5) approximates the cut-edge count of a nearest-neighbor graph. Theorem 1, assembled from graph-based divergence results, states that the true cut-edge ratio converges to $2\alpha(1-\alpha)(1-D_f)$, so Eq. (6) turns the differentiable sum into a surrogate divergence bounded in $[0,1]$. This surrogate feeds the squared penalty in the loss, making the regularizer trainable by backpropagation and giving the hyperparameter $\gamma$ a bounded search range.

What would settle it

Compare the differentiable estimate from Eqs. (5)-(6) with the true nearest-neighbor cut-edge ratio on the same prediction-target batches used in the LIBS experiments reported in the paper, or on synthetic samples with known f-divergence; if the smooth estimate deviates from the true ratio or leaves $[0,1]$ at batch size 16, the divergence interpretation and the gamma range are not supported. A second check is to train the same CNN with the true but non-differentiable cut-edge statistic using a surrogate gradient and see whether the RMSE gains in Tables II and III persist.

Watch

Extended reading notes

Core claim

The central claim is that a regularization term of the form $w(\hat{D}_f(p_{\hat{Y}}\|p_Y;\Theta)-\gamma)^2$, added to the mean squared error in Eq. (4), constrains how far the distribution of network outputs can drift from the distribution of noisy targets. Here $\hat{D}_f$ is a differentiable estimate of an f-divergence, a measure of the difference between two probability distributions, computed from a softmax-weighted sum over a fully connected graph of predictions and targets, following the graph-based cut-edge estimator whose asymptotic limit is stated in Theorem 1 as $\lim T_n/n = 2\alpha(1-\alpha)(1-D_f)$. With $\alpha=1/2$ for equal batch sizes, the estimate is $\hat{D}_f = 1 - 2\hat{t}_n/n$ and is claimed to inherit the $[0,1]$ bound of the true divergence. The paper reports that this regularizer alone outperforms or matches L1, L2, and dropout on multi-oxide LIBS prediction tasks, and that combining it with those standard regularizers lowers RMSE further in most comparisons.

Load-bearing premise

The method's name and its $[0,1]$ bound for $\gamma$ depend on an unproven assumption: that the smooth softmax-weighted graph sum used in training behaves like the true nearest-neighbor cut-edge ratio when batches are small, so the estimated quantity really is an f-divergence.

Editorial extensions

If this is right

  • Training a 1D-CNN with f-divergence regularization yields test RMSE on multi-oxide LIBS prediction that is significantly smaller than or comparable to L1, L2, and dropout, averaged across all oxides.
  • Appending the f-divergence term to L1, L2, or dropout improves or maintains average all-oxide RMSE in most strength settings, so the regularizer is complementary rather than redundant.
  • Because the estimated divergence is bounded between 0 and 1, the target divergence $\gamma$ lies in a small range, simplifying hyperparameter search across tasks.
  • The regularizer acts as both a stabilizer, pushing predictions away from the noisy target distribution when divergence is too small, and an auxiliary loss, pulling them back when divergence is too large.
  • When a specific oxide matters, training a single-oxide network with the combined regularizer is more effective than training one multi-oxide network.

Reading between the lines

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

  • The same output-space penalty could transfer to other small-data multi-target regression problems, but the reported gains are established only on LIBS spectra, so any broader claim is an extrapolation.
  • The paper gives no proof that the softmax-weighted sum in Eq. (5) tracks the true nearest-neighbor cut-edge ratio at batch size 16, so the practical benefit may come from an ad hoc output-space smoothing effect rather than from f-divergence theory; an ablation replacing the smooth estimator with the true graph statistic would separate these explanations.
  • The method is essentially label smoothing transplanted to regression, suggesting it could be compared against other output-regularization losses such as adding noise to targets or penalizing output variance.
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

4 major / 6 minor

Summary. The paper proposes an output-space regularization method for multi-target regression. The loss is L(Θ) = MSE + w( D̂f − γ )², where D̂f is a differentiable estimate of an f-divergence between the distribution of network predictions and noisy targets. The differentiable estimate is obtained by replacing the non-differentiable nearest-neighbor cut-edge count of Theorem 1 with a softmax-weighted complete-graph statistic (Eq. (5)), then applying the asymptotic inversion D̂f = 1 − 2t̂n/n (Eq. (6)). The method is evaluated on LIBS spectra from ChemCam and SuperCam testbed data, comparing against no regularization, L1, L2, and dropout, with paired t-tests and an ablation study.

Significance. The empirical study is carefully executed: 15 runs, validation-based hyperparameter selection, paired t-tests, public benchmarks, and an ablation study addressing an observed failure mode (ChemCam dropout-0.06 row). The proposal to regularize the output distribution in multi-target regression is reasonable and the combination results in Tables III(a)–(c) show a fairly consistent improvement in the averaged-over-oxides RMSE. If the method were simply presented as a heuristically motivated softmax-graph output regularizer, the empirical content could support a moderate claim. However, the paper's central theoretical packaging — that Eq. (6) estimates the Df of Theorem 1 and is bounded in [0,1] — is not established, and the specific statements made are contradicted by a direct computation. As it stands, the theoretical framing is a load-bearing part of the paper's contribution claim.

major comments (4)
  1. [§III-D, Fig. 5, Eqs. (5)–(6)] No argument is provided that the softmax-weighted complete-graph statistic t̂n in Eq. (5) estimates the nearest-neighbor cut-edge ratio Tn/n of Theorem 1. The cited reference [34] concerns a different object and a different purpose (two-sample testing for implicit generative models), and the present paper offers no consistency statement for fixed λ=2 and batch size b=16. The claim that 'Eq. (6) estimates Df' is therefore unsupported; at most this is a heuristic graph-based loss.
  2. [§III-D, Eq. (6), and §III-B] Eq. (6) violates the claimed bound 0 ≤ D̂f ≤ 1. Setting ŷi = yi for all i (identical distributions), all 2b nodes coincide, so each node's softmax cross-class weight is b/(2b−1); with b=16 this gives t̂n/n = 16/31 and D̂f = 1 − 32/31 = −1/31 < 0. Thus the estimator is negative for identical distributions, contradicting the assertion that 'γ is also bounded between 0 and 1' used to justify the hyperparameter grid in Table I. The sentence in §III-B invoking [0,1] to reduce the search space is therefore not supported.
  3. [§II-B, Theorem 1 and §III-C] The application of Theorem 1 presupposes i.i.d. samples from each distribution, but the construction in §III-D pairs each prediction ŷi with its target yi for the same input xi. The pairs are dependent, and the induced empirical distributions of {yi} and {ŷi} are not independent draws; moreover, at the start of training the prediction distribution is typically a near-deterministic function of the inputs rather than a sample from a fixed marginal pŶ. The paper does not address how Theorem 1's asymptotic applies to this paired, deterministic-conditional setting at b=16.
  4. [§V-B and §V-C, Tables II and III] The abstract and conclusion claim that f-divergence regularization 'performs better than or comparable to' standard regularizers and that combinations 'further enhance performance.' The t-test results in Table II for individual oxides are mixed (e.g., L1 vs. ours shows ✗ for MnO, MgO, and K2O on ChemCam), and the ChemCam rows of Table III(c) show the combined method significantly worse than dropout alone for 5 of 9 oxides at rate 0.06, which the paper itself acknowledges. The averaged-over-oxides improvements are often modest, and the 'all oxides' row is not a single measured quantity but an average over 9 (or 8) separately evaluated regression tasks. The global claim should be tempered accordingly, and the multiple-comparison issue (dozens of t-tests at α=0.1) is not discussed.
minor comments (6)
  1. [§V-C, Table II and text] The text says 'all nine oxide weights' for both datasets, but the SuperCam results table lists eight oxides (SiO2, TiO2, Al2O3, FeO, MgO, CaO, Na2O, K2O). Please correct the count and the corresponding description.
  2. [§III-D, Eq. (5)] The notation is overloaded: wi denotes a vector of edge weights incident to vi in one place and the scalar w(i,j) in another, and the paper alternates between 'ˆtn' and 'tn' for the same quantity. Please clarify.
  3. [Fig. 5 and §III-D] The passage is typeset as a figure with the definition of Eq. (5); it would be clearer as a normal numbered display equation.
  4. [§V-C, Table II] The row 'f-divergence (proposed)' is repeated after each baseline; presenting the proposed method once, with baselines aligned, would improve readability.
  5. [§V-D, Table III(a)] For the L1 ChemCam strength-0.0001 row, the combined method has a larger all-oxides RMSE (2.36 vs. 2.30) and 5 ✗ marks; describing this as a success supports the need for a more careful reading of the tables, and the summary text should reflect the mixed nature of the results.
  6. [General] The learning rate is reported as r = 1 with Adadelta; Adadelta is scale-invariant, but reporting the default epsilon and rho would make the training setup reproducible. Also, the spectral preprocessing (e.g., continuum removal, normalization) is not described.

Circularity Check

0 steps flagged · score 0.0 of 10

No significant circularity: the empirical comparisons are held-out and the f-divergence estimator is imported from external references, not built from the paper's own conclusions.

full rationale

The paper's central empirical claim—that f-divergence regularization matches or improves on L1, L2, and dropout—is evaluated on held-out test sets after choosing (w, gamma) on a validation set, and it is compared against external baselines. No fitted parameter is renamed as a prediction, and the target RMSE values are not forced by construction of the loss. The differentiable estimator in Eqs. (5)-(6) is attributed to external works [31]-[34], not derived from the paper's own conclusion; the theoretical gap that Eq. (6) is not proven consistent for Theorem 1's D_f, and may even leave the claimed [0,1] bound (e.g., negative for identical predictions), is a soundness/correctness concern rather than circularity. Self-citations are also not load-bearing: [20] is a related label-smoothing citation, and the divergence properties are attributed to [23] and [31], which are not authored by the present paper's authors. Accordingly, no circular step can be exhibited under the rules requiring a specific equation-level reduction, so the circularity score is 0.

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

The central claim rests on the f-divergence theorem for graph cut-edges, on an unproven smooth approximation of those cut-edges, and on validation-tuned hyperparameters. No new physical entities are introduced.

free parameters (3)
  • gamma = 0.001 to 0.03, selected by validation (Table I)
    Target divergence between the prediction and target distributions in Eq. (4). It is tuned per experiment rather than derived from the noise model, despite the paper's statement that ideally it is proportional to epsilon.
  • w = 0.00005 to 0.01 (Table I)
    Regularization strength balancing the MSE term and the divergence term in Eq. (4); selected by validation.
  • lambda = 2 (Section V-B)
    Scaling parameter in the softmax edge weights in Eq. (5). It is fixed by hand with no sensitivity analysis.
assumptions (4)
  • standard math The cut-edge ratio asymptotic in Theorem 1 holds for the distributions of predictions and targets, including required regularity conditions.
    The paper relies on [31,32,33] to equate the normalized cut-edge number with an f-divergence and uses this as the foundation for Eq. (6).
  • ad hoc to paper The softmax-weighted complete-graph sum in Eq. (5) faithfully approximates the nearest-neighbor cut-edge number in Theorem 1.
    No theorem or experiment in the paper establishes this. The approximation is borrowed from [34] but is not shown to inherit the f-divergence interpretation or the stated [0,1] bound.
  • domain assumption The data model Y = g(X) + epsilon with i.i.d. samples and noise independent of the spectrum is adequate for LIBS oxide measurements.
    This model is stated in Section II-A and motivates the regularizer, but the paper does not validate whether oxide-weight measurement noise is additive, homoscedastic, or independent.
  • domain assumption The per-mini-batch divergence estimate, with batch size 16, is representative of the full training distribution.
    The regularizer is computed on small batches and the asymptotic theorem is applied to n=32 points, with no finite-sample correction or bias analysis.

how reviews work

0 comments
Cite this review

Pith. "Pith review of Regularization via f-Divergence: An Application to Multi-Oxide Spectroscopic Analysis." pith.science (2026). https://pith.science/paper/GKXRPLDQ

@misc{pith2026250203755,
  author       = {Pith},
  title        = {Pith review of: Regularization via f-Divergence: An Application to Multi-Oxide Spectroscopic Analysis},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/GKXRPLDQ}},
  note         = {Machine review of arXiv:2502.03755}
}
abstract

In this paper, we address the task of characterizing the chemical composition of planetary surfaces using convolutional neural networks (CNNs). Specifically, we seek to predict the multi-oxide weights of rock samples based on spectroscopic data collected under Martian conditions. We frame this problem as a multi-target regression task and propose a novel regularization method based on f-divergence. The f-divergence regularization is designed to constrain the distributional discrepancy between predictions and noisy targets. This regularizer serves a dual purpose: on the one hand, it mitigates overfitting by enforcing a constraint on the distributional difference between predictions and noisy targets. On the other hand, it acts as an auxiliary loss function, penalizing the neural network when the divergence between the predicted and target distributions becomes too large. To enable backpropagation during neural network training, we develop a differentiable f-divergence and incorporate it into the f-divergence regularization, making the network training feasible. We conduct experiments using spectra collected in a Mars-like environment by the remote-sensing instruments aboard the Curiosity and Perseverance rovers. Experimental results on multi-oxide weight prediction demonstrate that the proposed $f$-divergence regularization performs better than or comparable to standard regularization methods including $L_1$, $L_2$, and dropout. Notably, combining the $f$-divergence regularization with these standard regularization further enhances performance, outperforming each regularization method used independently.

Figures

Figures reproduced from arXiv: 2502.03755 by the authors.

Figure 1
Figure 1. True function and various approximation functions are compared, along with their training and test mean squared errors (MSE). The training MSE [PITH_FULL_IMAGE:figures/full_fig_p002_1.png] view at source ↗
Figure 2
Figure 2. (a) and (b) illustrate two scenarios of samples generated from [PITH_FULL_IMAGE:figures/full_fig_p003_2.png] view at source ↗
Figure 3
Figure 3. Candidates for L2 regularized curves are highlighted in blue. These curves avoid overfitting to the noisy data. By accounting for the presence of f-divergence between noisy data and predictions made by the target function, candidate curves with divergence exceeding a specified threshold (γ in (4)) are eliminated. This yields a final selected curve that maintains an appropriate f-divergence and closely approximates t… view at source ↗
Figures from the paper (4 more)
Figure 4
Figure 4. Figure 4: (a) and (b) illustrate two scenarios of predictions (red nodes) [PITH_FULL_IMAGE:figures/full_fig_p004_4.png]
Figure 5
Figure 5. Figure 5: Differentiable approximation of the cut-edge ratio [PITH_FULL_IMAGE:figures/full_fig_p005_5.png]
Figure 7
Figure 7. Figure 7: The f-divergence between predictions and targets in (b) appears appropriate even the approximation and true functions are distant, due to the loss of pairing knowledge between predictions (or targets) and predictors in (a). V. EXPERIMENTAL RESULTS FOR MULTI-OXIDE SPECT…
Figure 8
Figure 8. Figure 8: The testbed ChemCam (a) and SuperCam (b), a chamber (c) for [PITH_FULL_IMAGE:figures/full_fig_p007_8.png]

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

40 extracted references · 35 canonical work pages

  1. [34]

    Learning Implicit Generative Models Using Differentiable Graph Tests

    J. Djolonga and A. Krause, “Learning implicit generative models using differentiable graph tests,” arXiv preprint arXiv:1709.01006 , 2017

    work page Pith review arXiv 2017

  2. [1]

    Mars’ surface radiation environment measured with the mars science laboratory’s curiosity rover,

    D. M. Hassler, C. Zeitlin, R. F. Wimmer-Schweingruber, B. Ehresmann, S. Rafkin, J. L. Eigenbrode, D. E. Brinza, G. Weigle, S. B ¨ottcher, E. B¨ohm et al., “Mars’ surface radiation environment measured with the mars science laboratory’s curiosity rover,” science, vol. 343, no. 6169, p. 1244797, 2014

    work page 2014

  3. [2]

    The collinearity problem in linear regression. the partial least squares (pls) approach to generalized inverses,

    S. Wold, A. Ruhe, H. Wold, and W. Dunn, Iii, “The collinearity problem in linear regression. the partial least squares (pls) approach to generalized inverses,” SIAM Journal on Scientific and Statistical Computing , vol. 5, no. 3, pp. 735–743, 1984

    work page 1984

  4. [3]

    Partial least squares regression (plsr) applied to nir and hsi spectral data modeling to predict chemical properties of fish muscle,

    J.-H. Cheng and D.-W. Sun, “Partial least squares regression (plsr) applied to nir and hsi spectral data modeling to predict chemical properties of fish muscle,” Food engineering reviews , vol. 9, pp. 36– 49, 2017

    work page 2017

  5. [4]

    The prediction of soil chemical and physical properties from mid-infrared spectroscopy and combined partial least-squares regression and neural networks (pls-nn) analysis,

    L. Janik, S. Forrester, and A. Rawson, “The prediction of soil chemical and physical properties from mid-infrared spectroscopy and combined partial least-squares regression and neural networks (pls-nn) analysis,” Chemometrics and Intelligent Laboratory Systems , vol. 97, no. 2, pp. 179–188, 2009

    work page 2009

  6. [5]

    Prediction of aged red wine aroma properties from aroma chemical composition. partial least squares regression models,

    M. Aznar, R. L ´opez, J. Cacho, and V . Ferreira, “Prediction of aged red wine aroma properties from aroma chemical composition. partial least squares regression models,” Journal of Agricultural and Food Chemistry, vol. 51, no. 9, pp. 2700–2707, 2003

    work page 2003

  7. [6]

    Fractional-order derivative spectral transformations improved partial least squares regression estimation of photosynthetic capacity from hyperspectral reflectance,

    G. Song, Q. Wang, and J. Jin, “Fractional-order derivative spectral transformations improved partial least squares regression estimation of photosynthetic capacity from hyperspectral reflectance,” IEEE Transac- tions on Geoscience and Remote Sensing , vol. 61, pp. 1–10, 2023

    work page 2023

  8. [7]

    Using partial least squares-artificial neural network for inversion of inland water chlorophyll-a,

    K. Song, L. Li, S. Li, L. Tedesco, H. Duan, Z. Li, K. Shi, J. Du, Y . Zhao, and T. Shao, “Using partial least squares-artificial neural network for inversion of inland water chlorophyll-a,” IEEE Transactions on Geoscience and Remote Sensing, vol. 52, no. 2, pp. 1502–1517, 2013

    work page 2013

Show all 40 references
  1. [8]

    Multivariate analysis of remote laser-induced breakdown spectroscopy spectra using partial least squares, principal component analysis, and related techniques,

    S. M. Clegg, E. Sklute, M. D. Dyar, J. E. Barefield, and R. C. Wiens, “Multivariate analysis of remote laser-induced breakdown spectroscopy spectra using partial least squares, principal component analysis, and related techniques,” Spectrochimica Acta Part B: Atomic Spectrosco...

    2009

  2. [9]

    Overview of the planetary data system,

    S. K. McMahon, “Overview of the planetary data system,” Planetary and Space Science , vol. 44, no. 1, pp. 3–12, 1996

    1996

  3. [10]

    A review on object detection based on deep convolutional neural networks for autonomous driving,

    J. Lu, S. Tang, J. Wang, H. Zhu, and Y . Wang, “A review on object detection based on deep convolutional neural networks for autonomous driving,” in 2019 Chinese Control And Decision Conference (CCDC) . IEEE, 2019, pp. 5301–5308

    2019

  4. [11]

    Face recognition based on convolutional neural network,

    M. Cos ¸kun, A. Uc ¸ar,¨O. Yildirim, and Y . Demir, “Face recognition based on convolutional neural network,” in 2017 international conference on modern electrical and energy systems (MEES) . IEEE, 2017, pp. 376– 379

    2017

  5. [12]

    Medical image analysis using convolutional neural net- works: a review,

    S. M. Anwar, M. Majid, A. Qayyum, M. Awais, M. Alnowami, and M. K. Khan, “Medical image analysis using convolutional neural net- works: a review,” Journal of medical systems , vol. 42, pp. 1–13, 2018

    2018

  6. [13]

    Convolutional neural networks for vibrational spec- troscopic data analysis,

    J. Acquarelli, T. van Laarhoven, J. Gerretzen, T. N. Tran, L. M. Buydens, and E. Marchiori, “Convolutional neural networks for vibrational spec- troscopic data analysis,” Analytica chimica acta , vol. 954, pp. 22–31, 2017

    2017

  7. [14]

    Automated spectroscopic mod- elling with optimised convolutional neural networks,

    Z. Shen and R. Viscarra Rossel, “Automated spectroscopic mod- elling with optimised convolutional neural networks,” Scientific Reports, vol. 11, no. 1, p. 208, 2021

    2021

  8. [15]

    Deep learning spectroscopy: Neural networks for molecular excitation spectra,

    K. Ghosh, A. Stuke, M. Todorovi ´c, P. B. Jørgensen, M. N. Schmidt, A. Vehtari, and P. Rinke, “Deep learning spectroscopy: Neural networks for molecular excitation spectra,” Advanced science , vol. 6, no. 9, p. 1801367, 2019

    2019

  9. [16]

    Character-level convolutional net- works for text classification,

    X. Zhang, J. Zhao, and Y . LeCun, “Character-level convolutional net- works for text classification,” Advances in neural information processing systems, vol. 28, 2015

    2015

  10. [17]

    On measures of entropy and information,

    A. R ´enyi, “On measures of entropy and information,” in Proceedings of the fourth Berkeley symposium on mathematical statistics and probabil- ity, volume 1: contributions to the theory of statistics, vol. 4. University of California Press, 1961, pp. 547–562

    1961

  11. [18]

    On the solution of ill-posed problems and the method of regularization,

    A. N. Tikhonov, “On the solution of ill-posed problems and the method of regularization,” in Doklady akademii nauk , vol. 151, no. 3. Russian Academy of Sciences, 1963, pp. 501–504

    1963

  12. [19]

    Dropout training as adaptive regularization,

    S. Wager, S. Wang, and P. S. Liang, “Dropout training as adaptive regularization,” Advances in neural information processing systems , vol. 26, 2013

    2013

  13. [20]

    Regularization via structural label smoothing,

    W. Li, G. Dasarathy, and V . Berisha, “Regularization via structural label smoothing,” in International Conference on Artificial Intelligence and Statistics. PMLR, 2020, pp. 1453–1463

    2020

  14. [21]

    Rethinking the inception architecture for computer vision,

    C. Szegedy, V . Vanhoucke, S. Ioffe, J. Shlens, and Z. Wojna, “Rethinking the inception architecture for computer vision,” in Proceedings of the IEEE conference on computer vision and pattern recognition , 2016, pp. 2818–2826

    2016

  15. [22]

    Learning repre- sentations by back-propagating errors,

    D. E. Rumelhart, G. E. Hinton, and R. J. Williams, “Learning repre- sentations by back-propagating errors,” nature, vol. 323, no. 6088, pp. 533–536, 1986

    1986

  16. [23]

    Empirically estimable classification bounds based on a nonparametric divergence measure,

    V . Berisha, A. Wisler, A. O. Hero, and A. Spanias, “Empirically estimable classification bounds based on a nonparametric divergence measure,” IEEE Transactions on Signal Processing , vol. 64, no. 3, pp. 580–591, 2015

    2015

  17. [24]

    Re- calibration of the mars science laboratory chemcam instrument with an 11 expanded geochemical database,

    S. M. Clegg, R. C. Wiens, R. Anderson, O. Forni, J. Frydenvang, J. Lasue, A. Cousin, V . Payre, T. Boucher, M. D. Dyar et al. , “Re- calibration of the mars science laboratory chemcam instrument with an 11 expanded geochemical database,” Spectrochimica Acta Part B: Atomic Spec...

    2017

  18. [25]

    Empirical risk minimization with f-divergence regularization in statistical learning,

    F. Daunas, I. Esnaola, S. M. Perlaza, and H. V . Poor, “Empirical risk minimization with f-divergence regularization in statistical learning,” Ph.D. dissertation, Inria, 2023

    2023

  19. [26]

    Posterior differential regularization with f-divergence for improving model robustness,

    H. Cheng, X. Liu, L. Pereira, Y . Yu, and J. Gao, “Posterior differential regularization with f-divergence for improving model robustness,” arXiv preprint arXiv:2010.12638, 2020

    2010 arXiv

  20. [27]

    Learning fair classifiers via min-max f- divergence regularization,

    M. Zhong and R. Tandon, “Learning fair classifiers via min-max f- divergence regularization,” in 2023 59th Annual Allerton Conference on Communication, Control, and Computing (Allerton) . IEEE, 2023, pp. 1–8

    2023

  21. [28]

    Multivariate f-divergence estimation with confidence,

    K. Moon and A. Hero, “Multivariate f-divergence estimation with confidence,” Advances in neural information processing systems, vol. 27, 2014

    2014

  22. [29]

    Practical and consistent estimation of f-divergences,

    P. Rubenstein, O. Bousquet, J. Djolonga, C. Riquelme, and I. O. Tol- stikhin, “Practical and consistent estimation of f-divergences,” Advances in Neural Information Processing Systems , vol. 32, 2019

    2019

  23. [30]

    f-divergence estimation and two-sample homogeneity test under semiparametric density-ratio models,

    T. Kanamori, T. Suzuki, and M. Sugiyama, “ f-divergence estimation and two-sample homogeneity test under semiparametric density-ratio models,” IEEE Transactions on Information Theory , vol. 58, no. 2, pp. 708–720, 2011

    2011

  24. [31]

    Empirical non-parametric estimation of the fisher information,

    V . Berisha and A. O. Hero, “Empirical non-parametric estimation of the fisher information,” IEEE Signal Processing Letters , vol. 22, no. 7, pp. 988–992, 2014

    2014

  25. [32]

    A multivariate two-sample test based on the number of nearest neighbor type coincidences,

    N. Henze, “A multivariate two-sample test based on the number of nearest neighbor type coincidences,” The Annals of Statistics , vol. 16, no. 2, pp. 772–783, 1988

    1988

  26. [33]

    On the multivariate runs test,

    N. Henze and M. D. Penrose, “On the multivariate runs test,” Annals of statistics, pp. 290–298, 1999

    1999

  27. [35]

    Systems engineering the curiosity rover: A retrospective,

    R. Welch, D. Limonadi, and R. Manning, “Systems engineering the curiosity rover: A retrospective,” in 2013 8th international conference on system of systems engineering . IEEE, 2013, pp. 70–75

    2013

  28. [36]

    The mars 2020 engineering cameras and microphone on the perseverance rover: A next-generation imaging system for mars exploration,

    J. Maki, D. Gruel, C. McKinney, M. Ravine, M. Morales, D. Lee, R. Willson, D. Copley-Woods, M. Valvo, T. Goodsall et al., “The mars 2020 engineering cameras and microphone on the perseverance rover: A next-generation imaging system for mars exploration,” Space science reviews,...

    2020

  29. [37]

    The chemcam instrument suite on the mars science laboratory (msl) rover: Science objectives and mast unit description,

    S. Maurice, R. Wiens, M. Saccoccio, B. Barraclough, O. Gasnault, O. Forni, N. Mangold, D. Baratoux, S. Bender, G. Berger et al. , “The chemcam instrument suite on the mars science laboratory (msl) rover: Science objectives and mast unit description,” Space science reviews , vo...

    2012

  30. [38]

    The supercam instrument suite on the mars 2020 rover: Science objectives and mast- unit description,

    S. Maurice, R. C. Wiens, P. Bernardi, P. Ca ¨ıs, S. Robinson, T. Nelson, O. Gasnault, J.-M. Reess, M. Deleuze, F. Rull et al. , “The supercam instrument suite on the mars 2020 rover: Science objectives and mast- unit description,” Space Science Reviews , vol. 217, pp. 1–108, 2021

    2020

  31. [39]

    Adadelta: An adaptive learning rate method,

    M. Zeiler, “Adadelta: An adaptive learning rate method,” arXiv preprint arXiv:1212.5701, 2012

    2012 arXiv

  32. [40]

    Dropout: a simple way to prevent neural networks from over- fitting,

    N. Srivastava, G. Hinton, A. Krizhevsky, I. Sutskever, and R. Salakhut- dinov, “Dropout: a simple way to prevent neural networks from over- fitting,” The journal of machine learning research , vol. 15, no. 1, pp. 1929–1958, 2014

    1929

Pith tools

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