Pith. sign in

REVIEW 4 major objections 5 minor 1 cited by

The Exploration of Neural Collapse under Imbalanced Data

T0 review · 4 major / 5 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 Under imbalanced data, the entire neural-collapse geometry of a deep linear model is set by the singular values of a single class-size matrix.

desk verdict The single-layer results and the interlacing theorem are credible, but the central L-layer theorem rests on an unproved SVD-alignment step; the paper deserves a careful referee, not acceptance as-is. read the letter →

arxiv 2411.17278 v1 pith:KLG3S5JY submitted 2024-11-26 cs.LG math.OC

Haixia Liu This is my paper · ORCID
classification cs.LGmath.OC
keywords neuralcollapseimbalanceddataL-extendedunconstrainedfeaturemodelmeansquarederrorsingularvaluedecompositionminorityglobalminimizerclass-meangeometry
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

Neural collapse describes how well-trained classifiers organize their final-layer features, but the classical picture assumes balanced classes. This paper asks what happens when class sizes differ, and claims that in the L-layer unconstrained feature model with bias and squared-error loss, the entire global-minimizer geometry is set by the singular values of a single matrix built from class sizes, $\hat Y = (I_K - \frac{1}{N}\mathbf{n}1_K^\top)D$. If the claim is right, the fate of each class is simple: when its singular value clears a regularization-dependent threshold the classifier keeps an orthogonal, aligned direction for that class, and otherwise the class direction collapses to zero. That matters because imbalanced data are the norm in applications, and minority classes are exactly where this collapse would occur. The paper also computes the singular values of $\hat Y$ explicitly for grouped imbalance patterns, turning the geometric statement into a quantitative sample-size criterion.

What carries the argument

The load-bearing object is the matrix $\hat Y = (I_K - \frac{1}{N} \mathbf{n} 1_K^\top)D$, where $K$ is the number of classes, $\mathbf{n}$ lists per-class sample sizes, $N$ is the total sample count, and $D = \mathrm{diag}(\sqrt{\mathbf{n}})$. Its singular value decomposition $\hat Y = \hat U \hat\kappa \hat V^\top$ gives the coordinates in which the geometry of the global minimizer becomes simple. The argument works by first showing that at any critical point the bias is the class-frequency vector $\mathbf{n}/N$ and the global feature mean is zero; a Jensen-type inequality then forces features within each class to equal their class mean, and the L-layer objective reduces to a nuclear-norm-regularized least-squares problem in the variable $Z = W_L\cdots W_1 \bar H D$. A singular-value-thresholding calculation identifies the optimal $Z$, and a scalar minimization (the lemma on minimizing $1/x^{L+1} + \alpha x$) decides, singular value by singular value, which coordinates of $\hat U^\top W_L\cdots W_1$ and $\bar H D \hat V$ survive as nonzero orthogonal directions.

What would settle it

Solve (4.3) numerically for a small imbalanced case such as $K=3$ classes with sizes $[10,10,1]$, two layers, and small fixed regularization, using many random initializations; if any computed global minimizer shows a nonzero row of $\hat U^\top W_L\cdots W_1$ for a class whose $\hat Y$ singular value is below the paper's threshold, or shows surviving rows that are not mutually orthogonal, Theorem 4.3 would be contradicted. A cheaper check is to audit the unproved inequality in Section B.3 for the specific case of repeated class sizes, where the SVD of $\hat Y$ is non-unique.

Watch

Extended reading notes

Core claim

At the level of the model, the paper's central result is Theorem 4.3. For any global minimizer of the L-extended unconstrained feature model with bias and MSE loss, features within each class collapse to a single class mean, so the effective feature matrix is the class-mean matrix $\bar H$ with entries repeated within each class. Writing $\hat Y = (I_K - \frac{1}{N}\mathbf{n}1_K^\top)D$ and its SVD as $\hat Y = \hat U \hat\kappa \hat V^\top$, the theorem says that $\hat U^\top W_L\cdots W_1$ has rows that are either zero or mutually orthogonal, while $\bar H D \hat V$ has columns of the same kind, and for the surviving coordinates the $i$-th row of the classifier product aligns exactly with the $i$-th column of the scaled class-mean matrix. Which coordinates survive is governed by whether the corresponding singular value of $\hat Y$ exceeds a threshold set by the layer count and the regularization weights. The bias is pinned down as the class-frequency vector and the global feature mean is zero, so the imbalance is carried entirely by $\hat Y$ and not by the features.

Load-bearing premise

The proof that every global minimizer must rotate its singular vectors to match $\hat Y$ is supported by an assertion in Section B.3 that a certain other configuration gives a strictly larger objective value, but the inequality behind that assertion is not demonstrated; if that inequality fails, the orthogonal and aligned structure is not established for every global minimizer.

Editorial extensions

If this is right

  • In this model, minority collapse is not a separate failure mode: a class is represented at the global optimum if and only if the singular value of $\hat Y$ associated with it clears the regularization threshold, and otherwise its classifier row and feature column vanish.
  • For grouped imbalances, the singular values of $\hat Y$ are explicit: within each group of classes sharing a sample size $N_j$, the repeated singular values are $\sqrt{N_j}$, and the remaining singular values come from a small $m\times m$ matrix $G$; in the two-group case the only extra singular value is $\sqrt{K N_1 N_2/N}$.
  • With $L$ layers, the same zero-or-orthogonal geometry holds for the product of $L$ linear classifiers, and the survival threshold depends on the geometric mean of the weight-decay parameters across layers.
  • The bias absorbs the imbalance: the optimal bias is the vector of class frequencies $\mathbf{n}/N$, and the global feature centroid is zero, so the centered label matrix $\hat Y$ is the only place class sizes enter.
  • In the bias-free one-layer case, the survival condition reduces to $n_i > N^2 \lambda_H \lambda_W$, making the collapse threshold a direct inequality on per-class sample count.

Reading between the lines

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

  • The paper does not draw the practical corollary that its threshold is a minimal sample-size rule: the model can only represent a class when the associated singular value of $\hat Y$ clears the regularization threshold, so classes below it are effectively dropped by the global optimum.
  • Because $\hat Y$ is a rank-one perturbation of a diagonal matrix, the same SVD-driven geometry should plausibly appear under other losses whose centered label matrix has the same shape; testing cross-entropy or hinge versions of Theorem 4.3 would be a direct extension the paper does not attempt.
  • Repeated class sizes make the SVD of $\hat Y$ non-unique, so the row-column alignment in Theorem 4.3 is only defined up to rotations inside the repeated eigenspace; a rotation-invariant statement would be needed for datasets with ties, and the paper does not address that case.
  • One could also use the paper's NC2 metrics as a diagnostic on real networks: measure how close the trained product of classifier layers is to the predicted zero/orthogonal pattern, and treat large deviations as evidence that the trained solution is far from the L-EUFM global optimum.
Share X Bluesky LinkedIn Reddit HN

Signed reviews

No signed human review yet.

Editorial analysis

A structured set of objections, weighed in public.

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

Referee Report

4 major / 5 minor

Summary. This paper studies the global minimizers of an L-layer unconstrained feature model with squared loss, weight decay, and a bias term under imbalanced class sizes. The main claim (Theorem 4.3) is that every global minimizer has a geometry determined by the singular value decomposition of \hat Y = (I_K - n 1_K^⊟/N)D: features collapse to class means, the left-transformed classifier product has zero rows for small singular values and orthogonal surviving rows, the right-transformed class-mean matrix has a matching column structure, and surviving rows and columns align. The paper also provides an explicit estimation of the singular values of \hat Y (Theorem 4.4) and numerical experiments on imbalanced CIFAR-10 and EMNIST subsets. Theorems 4.1 and 4.2 cover the single-layer bias-free and biased cases.

Significance. If the results were fully established, they would give a fairly detailed characterization of neural collapse in multilayer linear unconstrained feature models with bias under imbalanced data, extending prior work by Dang et al. and Tirer and Bruna. The single-layer derivations via the nuclear-norm relaxation are useful, and Theorem 4.4's interlacing result and explicit formulas for m=2,3 are concrete and falsifiable. The paper does not provide code, but the experimental setup is described in enough detail to be reproducible in principle. However, the central L-layer theorem is not proved as written, and the SVD-uniqueness problem affects the meaning of the geometric claims, so the paper's main contribution is currently unsupported.

major comments (4)
  1. [Section B.3, proof of Theorem 4.3] The assertion 'Then we can claim that U_WL = \hat U, V_{\bar E} = \hat V' is justified only by the sentence 'Otherwise ... \hat f(W'_L,…,\bar E') > \hat f(W''_L,W'_{L-1},…,W'_1,\bar E''), which conflicts with the definition of a global minimizer.' Replacing the left singular vectors of W_L and the right singular vectors of \bar E by \hat U and \hat V does not preserve the product W_L⋯W_1\bar E, and the Frobenius regularization terms are invariant under the replacement, so the claimed strict inequality does not follow. Since this is the only step that connects every global minimizer to the singular vectors of \hat Y, Theorem 4.3 is not proved as written.
  2. [Theorem 4.3 and Section B.3] When \hat Y has repeated singular values, \hat U and \hat V are not unique; this occurs in the balanced case and more generally under Definition 4.1 whenever any ℓ_j > 1. The proof fixes one SVD and asserts U_WL = \hat U and V_{\bar E} = \hat V for every global minimizer. A common orthogonal rotation within a degenerate singular subspace preserves the objective but changes these singular vectors, so the equality cannot hold for all global minimizers. The statement needs either an explicit gauge fixing or a reformulation in terms of subspaces or invariant quantities, and the 'i-th row aligns with the i-th column' claim requires a definite choice of SVD.
  3. [Theorem 4.2] The statement says κ_1,…,κ_K are the eigenvalues of \tilde Y D and writes \tilde Y D = \tilde U κ \tilde V^⊟. The matrix \tilde Y D = D - n√n^⊟/N is generally not symmetric, so its eigenvalues need not be real or nonnegative; the subsequent soft-thresholding formula (κ - N√(λ_Wλ_H))_+ is valid for singular values, not for eigenvalues. If κ is intended to denote singular values, the wording should be corrected throughout; if eigenvalues are intended, the formula is incorrect.
  4. [Section B.1, proof of Theorem 4.1] After deriving D^{-1}Z^* = VΣ(Σ^2+Nλ_W I)^{-1}Σ V^⊟, the proof concludes V=I because D^{-1}Z^* is diagonal. The correct conclusion is only that V diagonalizes a diagonal matrix, i.e., V is a signed permutation matrix, with arbitrary orthogonal blocks on repeated singular values. Since D and V need not commute, the subsequent formulas W^*(W^*)^⊟ = D^2Σ^2(Σ^2+Nλ_W I)^{-2} and (\bar H^*)^⊟\bar H^* = Σ^2D^{-2} presuppose V=I and do not follow for arbitrary global minimizers.
minor comments (5)
  1. [Section B.4, proof of Theorem 4.4] The proof says there are zeros of f in the intervals (N_i,N_{i+1}) for i=1,…,m-1; since N_1>…>N_m, the intended intervals should be (N_{i+1},N_i).
  2. [Theorem 4.3] The symbol r is defined twice with different arguments, and d_L does not appear as a dimension of any matrix in the theorem (W_L ∈ ℝ^{K×d_{L-1}}); the block sizes in Υ_1 and Υ_2 should be stated consistently with the stated dimensions.
  3. [Appendix C] In the definition of NC_H for the bias-free case, the denominator is written as ‖\bar H^⊟‖_F but should be ‖\bar H^⊟\bar H‖_F to match the numerator.
  4. [Introduction and Appendix C] There are several spacing and typographical errors, including 'F raud detection', 'T ext classification', 'Defect detection' in Section 1, and 'defind' in Appendix C; these should be corrected.
  5. [Section 5] The numerical section reports only plots of the NC metrics; it does not provide a quantitative comparison of the converged values with the singular-value thresholds predicted by Theorem 4.4, so the experimental support is weaker than the text suggests.

Circularity Check

0 steps flagged · score 0.0 of 10

No circularity: the central theorem is derived from the optimization objective via external variational lemmas; the numerical experiments are consistency checks, not fitted predictions.

full rationale

The paper's derivation chain is not circular. Theorem 4.3 is proved from the optimization objective (4.3) by reducing the problem to a nuclear-norm surrogate using Lemma A.1 (cited from ZDZ+21) and a one-variable minimization using Lemma A.2 (cited from DTO+23); both are external, machine-independent results with no author overlap. The singular-value structure of \hat Y is an input of the analysis, and the claimed geometry of global minimizers is a consequence, not a restatement, of that input. No parameter is fitted to the training data and then reported as a prediction: the numerical NC2/NC3 metrics compare trained quantities with the theorem's predicted matrices, which is a consistency check rather than a circular reduction. The only serious weakness is in Section B.3, where the assertion U_{W_L} = \hat U and V_{\bar E} = \hat V rests on an unproved 'Otherwise...conflicts' inequality; however, that is a correctness or proof-completeness concern, not a circularity, because the argument attempts to prove the alignment rather than assuming it. Consequently, the paper does not exhibit any step in which a prediction is equivalent to its inputs by construction.

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

The central claim rests on two standard external lemmas, the UFM modeling assumption, a specific bias-regularization setting, and an unproved singular-vector alignment assertion in the L-layer proof. There are no fitted constants and no new entities.

assumptions (5)
  • standard math Lemma A.1 (nuclear norm factorization): for any Z, ||Z||_* = min_{Z=WE} (||W||_F^2/(2α) + α||E||_F^2/2)
    Invoked in B.1 and B.2 to relax the product W E to a convex nuclear norm problem; cited from ZDZ+21.
  • standard math Lemma A.2 (minimizer of 1/(x^L+1) + αx)
    Used in B.3 to solve the per-coordinate scalar problem that determines which singular directions survive; cited from DTO+23.
  • domain assumption Unconstrained feature model: final-layer features H are free optimization variables
    Definition 3.1; standard in neural collapse theory, but a strong modeling assumption that the paper does not justify beyond citing overparameterization.
  • domain assumption Bias regularization ρ_b=0 and λ_H>0
    Set in Section 3 and used in B.2/B.3 to derive b^*=n/N and global centroid zero; the proofs require λ_H>0, though the setup only states λ_H≥0.
  • ad hoc to paper At a global minimizer of the L-layer objective, the singular vectors of W_L and \bar H D align with \hat U and \hat V of \hat Y
    Asserted in B.3 with an unproved perturbation inequality; this is the load-bearing premise for Theorem 4.3's geometry.

how reviews work

0 comments
Cite this review

Pith. "Pith review of The Exploration of Neural Collapse under Imbalanced Data." pith.science (2026). https://pith.science/paper/KLG3S5JY

@misc{pith2026241117278,
  author       = {Pith},
  title        = {Pith review of: The Exploration of Neural Collapse under Imbalanced Data},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/KLG3S5JY}},
  note         = {Machine review of arXiv:2411.17278}
}
abstract

Neural collapse, a newly identified characteristic, describes a property of solutions during model training. In this paper, we explore neural collapse in the context of imbalanced data. We consider the $L$-extended unconstrained feature model with a bias term and provide a theoretical analysis of global minimizer. Our findings include: (1) Features within the same class converge to their class mean, similar to both the balanced case and the imbalanced case without bias. (2) The geometric structure is mainly on the left orthonormal transformation of the product of $L$ linear classifiers and the right transformation of the class-mean matrix. (3) Some rows of the left orthonormal transformation of the product of $L$ linear classifiers collapse to zeros and others are orthogonal, which relies on the singular values of $\hat Y=(I_K-1/N\mathbf{n}1^\top_K)D$, where $K$ is class size, $\mathbf{n}$ is the vector of sample size for each class, $D$ is the diagonal matrix whose diagonal entries are given by $\sqrt{\mathbf{n}}$. Similar results are for the columns of the right orthonormal transformation of the product of class-mean matrix and $D$. (4) The $i$-th row of the left orthonormal transformation of the product of $L$ linear classifiers aligns with the $i$-th column of the right orthonormal transformation of the product of class-mean matrix and $D$. (5) We provide the estimation of singular values about $\hat Y$. Our numerical experiments support these theoretical findings.

Figures

Figures reproduced from arXiv: 2411.17278 by the authors.

Figure 1
Figure 1. Bias Case: The performance of the NC metrics and training accuracy versus epoch with a [PITH_FULL_IMAGE:figures/full_fig_p009_1.png] view at source ↗
Figure 2
Figure 2. Base case: The performance of the NC metrics and training accuracy versus epoch with a [PITH_FULL_IMAGE:figures/full_fig_p011_2.png] view at source ↗
Figure 3
Figure 3. Bias-free case: The performance of the NC metrics and training accuracy versus epoch [PITH_FULL_IMAGE:figures/full_fig_p025_3.png] view at source ↗
Figures from the paper (1 more)
Figure 4
Figure 4. Figure 4: Bias-free case: The performance of the NC metrics and training accuracy versus epoch [PITH_FULL_IMAGE:figures/full_fig_p026_4.png]

Discussion (0). Continue with ORCID to comment.

Forward citations

Cited by 1 Pith paper

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

  1. Neural Collapse is Globally Optimal in Deep Regularized ResNets and Transformers

    cs.LG 2025-05 conditional novelty 7.0 of 10

    Neural collapse is globally optimal in deep regularized ResNets and transformers, with the approximation improving as depth grows.

Reference graph

Works this paper leans on

31 extracted references · 27 canonical work pages · cited by 1 Pith paper

  1. [1]

    A singular value thresholding algorithm for matrix completion

    Jian-Feng Cai, Emmanuel J Cand \`e s, and Zuowei Shen. A singular value thresholding algorithm for matrix completion. SIAM Journal on optimization , 20(4):1956--1982, 2010

    work page 1956

  2. [2]

    New types of deep neural network learning for speech recognition and related applications: An overview

    Li Deng, Geoffrey Hinton, and Brian Kingsbury. New types of deep neural network learning for speech recognition and related applications: An overview. In 2013 IEEE international conference on acoustics, speech and signal processing , pages 8599--8603. IEEE, 2013

    work page 2013

  3. [3]

    Neural collapse in deep linear networks: from balanced to imbalanced data

    Hien Dang, Tho Tran, Stanley Osher, Hung Tran-The, Nhat Ho, and Tan Nguyen. Neural collapse in deep linear networks: from balanced to imbalanced data. In Proceedings of the 40th International Conference on Machine Learning , pages 6873--6947, 2023

    work page 2023

  4. [4]

    On the emergence of simplex symmetry in the final and penultimate layers of neural network classifiers

    Weinan E and Stephan Wojtowytsch. On the emergence of simplex symmetry in the final and penultimate layers of neural network classifiers. In Mathematical and Scientific Machine Learning , pages 270--290. PMLR, 2022

    work page 2022

  5. [5]

    Cong Fang, Hangfeng He, Qi Long, and Weijie J. Su. Exploring deep neural networks via layer-peeled model: Minority collapse in imbalanced training. Proceedings of the National Academy of Sciences , 118(43):e2103091118, 2021

    work page 2021

  6. [6]

    A review on deep learning techniques applied to semantic segmentation

    Alberto Garcia-Garcia, Sergio Orts-Escolano, Sergiu Oprea, Victor Villena-Martinez, and Jose Garcia-Rodriguez. A review on deep learning techniques applied to semantic segmentation. arXiv preprint arXiv:1704.06857 , 2017

    arXiv 2017

  7. [7]

    Dissecting supervised contrastive learning

    Florian Graf, Christoph Hofer, Marc Niethammer, and Roland Kwitt. Dissecting supervised contrastive learning. In International Conference on Machine Learning , pages 3821--3830. PMLR, 2021

    work page 2021

  8. [8]

    A survey of deep learning techniques for autonomous driving

    Sorin Grigorescu, Bogdan Trasnea, Tiberiu Cocias, and Gigel Macesanu. A survey of deep learning techniques for autonomous driving. Journal of field robotics , 37(3):362--386, 2020

    work page 2020

Show all 31 references
  1. [9]

    G. Gybenko. Approximation by superposition of sigmoidal functions. Mathematics of Control, Signals and Systems , 2(4):303--314, 1989

    1989

  2. [10]

    Evaluation of neural architectures trained with square loss vs cross-entropy in classification tasks

    Like Hui and Mikhail Belkin. Evaluation of neural architectures trained with square loss vs cross-entropy in classification tasks. In International Conference on Learning Representations , 2020

    2020

  3. [11]

    Neural collapse for unconstrained feature model under cross-entropy loss with imbalanced data

    Wanli Hong and Shuyang Ling. Neural collapse for unconstrained feature model under cross-entropy loss with imbalanced data. Journal of Machine Learning Research , 25(192):1--48, 2024

    2024

  4. [12]

    Approximation capabilities of multilayer feedforward networks

    Kurt Hornik. Approximation capabilities of multilayer feedforward networks. Neural networks , 4(2):251--257, 1991

    1991

  5. [13]

    Han, Vardan Papyan, and David L

    X.Y. Han, Vardan Papyan, and David L. Donoho. Neural collapse under mse loss: Proximity to and dynamics on the central path. In International Conference on Learning Representations , 2021

    2021

  6. [14]

    When face recognition meets with deep learning: an evaluation of convolutional neural networks for face recognition

    Guosheng Hu, Yongxin Yang, Dong Yi, Josef Kittler, William Christmas, Stan Z Li, and Timothy Hospedales. When face recognition meets with deep learning: an evaluation of convolutional neural networks for face recognition. In Proceedings of the IEEE international conference on ...

    2015

  7. [15]

    Wenlong Ji, Yiping Lu, Yiliang Zhang, Zhun Deng, and Weijie J. Su. An unconstrained layer-peeled perspective on neural collapse. In International Conference on Learning Representations , 2021

    2021

  8. [16]

    S \'a nchez

    Geert Litjens, Thijs Kooi, Babak Ehteshami Bejnordi, Arnaud Arindra Adiyoso Setio, Francesco Ciompi, Mohsen Ghafoorian, Jeroen Awm Van Der Laak, Bram Van Ginneken, and Clara I. S \'a nchez. A survey on deep learning in medical image analysis. Medical image analysis , 42:60--88, 2017

    2017

  9. [17]

    The expressive power of neural networks: A view from the width

    Zhou Lu, Hongming Pu, Feicheng Wang, Zhiqiang Hu, and Liwei Wang. The expressive power of neural networks: A view from the width. Advances in neural information processing systems , 30, 2017

    2017

  10. [18]

    Neural collapse under cross-entropy loss

    Jianfeng Lu and Stefan Steinerberger. Neural collapse under cross-entropy loss. Applied and Computational Harmonic Analysis , 59:224--241, 2022

    2022

  11. [19]

    Schoenholz

    James Martens, Andy Ballard, Guillaume Desjardins, Grzegorz Swirszcz, Valentin Dalibard, Jascha Sohl-Dickstein, and Samuel S. Schoenholz. Rapid training of deep neural networks without skip connections or normalization layers using deep kernel shaping. arXiv preprint arXiv:211...

    arXiv

  12. [20]

    Towards deep learning models resistant to adversarial attacks

    Aleksander Madry, Aleksandar Makelov, Ludwig Schmidt, Dimitris Tsipras, and Adrian Vladu. Towards deep learning models resistant to adversarial attacks. In International Conference on Learning Representations , 2018

    2018

  13. [21]

    Mixon, Hans Parshall, and Jianzong Pi

    Dustin G. Mixon, Hans Parshall, and Jianzong Pi. Neural collapse with unconstrained features. arXiv preprint arXiv:2011.11619 , 2020

    2011 arXiv

  14. [22]

    Deep double descent: Where bigger models and more data hurt

    Preetum Nakkiran, Gal Kaplun, Yamini Bansal, Tristan Yang, Boaz Barak, and Ilya Sutskever. Deep double descent: Where bigger models and more data hurt. Journal of Statistical Mechanics: Theory and Experiment , 2021(12):124003, 2021

    2021

  15. [23]

    Han, and David L

    Vardan Papyan, X.Y. Han, and David L. Donoho. Prevalence of neural collapse during the terminal phase of deep learning training. Proceedings of the National Academy of Sciences , 117(40):24652--24663, 2020

    2020

  16. [24]

    Deep isometric learning for visual recognition

    Haozhi Qi, Chong You, Xiaolong Wang, Yi Ma, and Jitendra Malik. Deep isometric learning for visual recognition. In International conference on machine learning , pages 7824--7835. PMLR, 2020

    2020

  17. [25]

    Provable approximation properties for deep neural networks

    Uri Shaham, Alexander Cloninger, and Ronald R Coifman. Provable approximation properties for deep neural networks. Applied and Computational Harmonic Analysis , 44(3):537--557, 2018

    2018

  18. [26]

    Extended unconstrained features model for exploring deep neural collapse

    Tom Tirer and Joan Bruna. Extended unconstrained features model for exploring deep neural collapse. In International Conference on Machine Learning , pages 21478--21505. PMLR, 2022

    2022

  19. [27]

    Imbalance trouble: Revisiting neural-collapse geometry

    Christos Thrampoulidis, Ganesh Ramachandra Kini, Vala Vakilian, and Tina Behnia. Imbalance trouble: Revisiting neural-collapse geometry. Advances in Neural Information Processing Systems , 35:27225--27238, 2022

    2022

  20. [28]

    Rethinking bias-variance trade-off for generalization of neural networks

    Zitong Yang, Yaodong Yu, Chong You, Jacob Steinhardt, and Yi Ma. Rethinking bias-variance trade-off for generalization of neural networks. In International Conference on Machine Learning , pages 10767--10777. PMLR, 2020

    2020

  21. [29]

    A geometric analysis of neural collapse with unconstrained features

    Zhihui Zhu, Tianyu Ding, Jinxin Zhou, Xiao Li, Chong You, Jeremias Sulam, and Qing Qu. A geometric analysis of neural collapse with unconstrained features. Advances in Neural Information Processing Systems , 34:29820--29834, 2021

    2021

  22. [30]

    On the optimization landscape of neural collapse under mse loss: Global optimality with unconstrained features

    Jinxin Zhou, Xiao Li, Tianyu Ding, Chong You, Qing Qu, and Zhihui Zhu. On the optimization landscape of neural collapse under mse loss: Global optimality with unconstrained features. In International Conference on Machine Learning , pages 27179--27202. PMLR, 2022

    2022

  23. [31]

    Object detection with deep learning: A review

    Zhong-Qiu Zhao, Peng Zheng, Shou-tao Xu, and Xindong Wu. Object detection with deep learning: A review. IEEE transactions on neural networks and learning systems , 30(11):3212--3232, 2019

    2019

Pith tools

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