Pith. sign in

REVIEW 4 major objections 4 minor 50 references

Eigen Neural Network: Unlocking Generalizable Vision with Eigenbasis

T0 review · 4 major / 4 minor · reviewed 2026-08-06 · deepseek-v4-flash

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

Pith's one-line read The paper claims that rewriting every layer's weights in a single learned orthonormal basis forces the weights to be decorrelated and aligned by construction, improving large-scale image classification, cross-modal retrieval, and…

desk verdict A bold architectural idea with strong empirical claims; the abstract alone cannot carry them, and the shared-basis expressivity question is the first thing a referee must press. read the letter →

arxiv 2508.01219 v1 pith:L2ZMD6VH submitted 2025-08-02 cs.CV cs.LG

Anzhe Cheng , Chenzhong Yin , Mingxi Cheng , Shukai Duan , Shahin Nazarian , Paul Bogdan This is my paper
classification cs.CVcs.LG MSC 68T0715A18
keywords EigenNeuralNetworkorthonormaleigenbasisbackpropagation-freelearninglocalweightdecorrelationImageNetclassificationcross-modalretrievalrepresentation
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

Deep neural networks trained by backpropagation tend to develop disordered, correlated weight matrices, and the paper attributes part of their training difficulty and feature noise to that disorder. The Eigen Neural Network (ENN) addresses this by writing every layer's weight matrix in one layer-shared, learned orthonormal eigenbasis, so decorrelation and alignment of weights hold by construction rather than by regularization. The paper reports that ENN with standard backpropagation outperforms existing methods on large-scale image classification, including ImageNet, and that its representations set a new state-of-the-art result in cross-modal image-text retrieval. It also introduces ENN-$\ell$, a backpropagation-free local learning variant, which it reports trains over twice as fast through parallelism while exceeding the accuracy of end-to-end backpropagation. If these results hold, architectural structure alone can replace weight regularization and make local learning competitive with global training.

What carries the argument

The central object is the layer-shared, learned orthonormal eigenbasis. An orthonormal basis is a set of mutually perpendicular unit directions; calling it an eigenbasis signals that each weight matrix is diagonal in this basis. All layers use the same learned basis, while each layer has its own coefficients, so the network's inductive bias is that useful weight directions align across layers. This object carries the argument because it transforms decorrelation from a regularized soft constraint into a hard structural fact: the basis is orthogonal by construction, and weights expressed in it cannot be arbitrarily aligned. The basis is learned during training, so the network can adapt the shared directions to the data.

What would settle it

Run ENN and a matched standard network of the same width and depth on a task engineered so that the optimal first-layer and last-layer weight matrices have incompatible singular vectors; if ENN's shared basis cannot match the standard network's accuracy, the expressiveness assumption fails.

Watch

Extended reading notes

Core claim

On the paper's own terms, the central discovery is that the disordered weight structure produced by gradient-based optimization is not an unavoidable byproduct of training but a representational choice that can be redesigned. By reparameterizing each layer's weights in a shared orthonormal basis, ENN guarantees that weight directions are mutually orthogonal and aligned, and only layer-specific coefficients vary. This structural guarantee produces more discriminative features, which the paper demonstrates by improved ImageNet accuracy and a new state-of-the-art cross-modal retrieval result. The same structure removes the need for end-to-end gradient flow: in ENN-$\ell$, each layer can be trained locally in parallel, giving over 2x training speedup and, the paper reports, accuracy above standard backpropagation. The intended lesson is that the geometry of the parameter space, not just the optimization algorithm, determines how well and how fast deep networks learn.

Load-bearing premise

The load-bearing premise is that one learned orthonormal basis shared by every layer is expressive enough to represent the useful weight matrices of all layers; if different layers need fundamentally different directions, the architecture cannot match an unconstrained network.

Editorial extensions

If this is right

  • Standard backpropagation on ENN outperforms comparable baselines on large-scale image classification, including ImageNet.
  • ENN representations transfer to cross-modal image-text retrieval and set a new state-of-the-art result on that task.
  • The backpropagation-free local learning variant ENN-$\ell$ trains more than twice as fast through layer parallelism.
  • ENN-$\ell$ reaches higher accuracy than end-to-end backpropagation on the tested tasks.
  • Weight decorrelation and alignment, normally encouraged by regularization, become structural properties in ENN rather than training targets.

Reading between the lines

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

  • A direct extension the paper leaves untested: if the shared basis is the active ingredient, the learned basis itself should transfer across datasets as a fixed representation, so one could test whether the basis improves generalization even when the classifier head is retrained.
  • The same reparameterization could in principle be applied to attention or recurrent layers, where weight correlation also affects training, though the paper does not report such experiments.
  • The reported speedup suggests that local layer-wise training can be competitive when weights are preconditioned by structure, so ENN-$\ell$ could be explored as a pretraining strategy for very deep networks.
  • One could check whether the structural decorrelation guarantee flattens the loss landscape enough to reduce sensitivity to initialization and learning rate, which would be a separate claim from the accuracy results.
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 / 4 minor

Summary. The manuscript proposes the Eigen Neural Network (ENN), which reparameterizes each layer's weight matrix in a shared learned orthonormal eigenbasis, i.e., weights of the form U diag(s_l) V^T with common U and V across layers. It claims that this structure axiomatically enforces decorrelated, well-aligned weight dynamics and yields three headline results: consistent state-of-the-art accuracy on ImageNet, a new state-of-the-art on a cross-modal image-text retrieval benchmark, and a backpropagation-free variant (ENN-ell) that trains more than 2x faster than standard backpropagation while surpassing its final accuracy. The submitted full text is heavily corrupted and largely illegible; the readable portions contain no experimental tables, architecture equations, hyperparameters, or derivations that can be independently checked.

Significance. If the claims held, the paper would be significant: it would show that a purely architectural parameterization can replace regularization, that a low-dimensional shared eigenbasis retains enough expressivity for ImageNet-scale models, and that a local learning rule can beat end-to-end backpropagation in both speed and accuracy. The core idea is clearly stated and falsifiable in principle, and the abstract promises concrete, testable comparisons. However, the submitted evidence is far below the bar for such claims: no reproducible protocol, code, numerical results, or verifiable derivations are available in the reviewed text, so the significance cannot be assessed at this stage.

major comments (4)
  1. [Abstract] The abstract states that ENN 'consistently outperforms state-of-the-art methods on large-scale image classification benchmarks, including ImageNet' and sets 'a new benchmark in cross-modal image-text retrieval,' but the submitted text contains no results tables, model sizes, training schedules, baselines, or dataset names. Without these details, the central empirical claims are unverifiable.
  2. [Abstract] The parameterization W_l = U diag(s_l) V^T with a single learned pair U,V shared across all layers restricts the set of representable weight matrices to a low-dimensional family (about O(n^2 + Ln) parameters rather than O(Ln^2), and it is ill-defined for layers of different widths without padding). The manuscript provides no expressivity theorem or approximation bound showing that this family can realize ImageNet-competitive functions, so the capacity premise underlying the claimed gains is unsupported.
  3. [Abstract] The claim that ENN-ell surpasses end-to-end backpropagation requires a precise specification of the local update rule, the per-layer loss or target information, and the parallelization scheme. None of these ingredients is readable in the submitted text, so the 2x speedup and accuracy comparisons cannot be checked or reproduced.
  4. [Full text (as received)] Most of the body is corrupted or illegible, including the equations and tables. Because the core derivations and experimental evidence cannot be inspected, I cannot verify any of the paper's formal or empirical statements; this is a blocking issue independent of the scientific content.
minor comments (4)
  1. [Abstract] The phrase 'Deep Neural Networks(DNN)' should be 'Deep Neural Networks (DNN)' with a space; please run a formatting pass over the text.
  2. [Abstract] The abstract uses 'we introduced' where 'we introduce' would be more consistent with the submission's timeframe.
  3. [Abstract] The term 'new benchmark' is ambiguous; specify whether it is a new dataset, a new evaluation protocol, or simply a new state-of-the-art result on an existing benchmark.
  4. [Full text (as received)] Many section headers and table captions appear garbled; please regenerate the manuscript from the LaTeX source so that all symbols and tables render correctly.

Circularity Check

0 steps flagged · score 0.0 of 10

No significant circularity: the architecture claim is empirical and the decorrelation property is a construction-level consequence, not a derived prediction.

full rationale

The paper's central claim is an empirical architecture proposal: weights are reparameterized in a shared, learned orthonormal eigenbasis, and ImageNet classification and cross-modal retrieval results are reported for that architecture. I find no step where a predicted quantity is equal by construction to a fitted input. The abstract's phrase 'enforces decorrelated, well-aligned weight dynamics axiomatically' describes a design property: by construction, layer weight matrices share the same left and right singular vectors, so their Gram matrices are simultaneously diagonalizable. That is a property of the parameterization, not a circular derivation of the benchmark outcomes. The ENN-ell backpropagation-free variant's speedup and accuracy comparisons are external benchmark statements, not reductions to the training loss. No load-bearing self-citation could be identified in the legible text. The expressivity question raised by the layer-shared basis (whether the restriction W_l = U diag(s_l) V^T preserves the function class) is a correctness or approximation concern, not circularity: if false, the empirical claims would fail, but the paper does not assume the conclusion it is trying to demonstrate. The manuscript is therefore self-contained as an empirical claim, and the appropriate finding is no significant circularity.

Assumptions & free parameters 1 free parameters · 3 assumptions · 0 invented entities

No new physical or geometric entities are introduced. The learned eigenbasis is a parameterization of existing weights, not an invented mechanism. The free parameter list is minimal because the abstract does not provide enough detail to identify additional tuned constants. The axioms are the unstated and stated premises that the architecture's success relies on.

free parameters (1)
  • learned eigenbasis matrices U and V = not stated
    The abstract says the orthonormal eigenbasis is 'learned'. The entries of the shared basis matrices are fitted to data during training, and the central claim depends on the particular values learned. No parameter count or initialization is given in the abstract.
assumptions (3)
  • domain assumption Disordered weight structures harm feature clarity and degrade learning dynamics.
    This is the motivating premise of the paper, stated in the first sentence of the abstract. If disorder does not actually hurt learning, the architecture's benefit would be unclear.
  • domain assumption A layer-shared orthonormal eigenbasis preserves the expressive capacity of standard dense layers.
    Implicit in the claim that ENN outperforms standard networks. If the shared basis restricts the function class, the network could not achieve the same accuracy, so the architecture would be limited.
  • domain assumption The local learning rule used by ENN-ell provides sufficiently accurate gradient information to match or beat backpropagation.
    The abstract claims ENN-ell surpasses end-to-end backpropagation accuracy. This is possible only if the local learning rule's updates do not diverge from useful gradient directions, which is a strong assumption about the optimization landscape.

how reviews work

0 comments
Cite this review

Pith. "Pith review of Eigen Neural Network: Unlocking Generalizable Vision with Eigenbasis." pith.science (2026). https://pith.science/paper/L2ZMD6VH

@misc{pith2026250801219,
  author       = {Pith},
  title        = {Pith review of: Eigen Neural Network: Unlocking Generalizable Vision with Eigenbasis},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/L2ZMD6VH}},
  note         = {Machine review of arXiv:2508.01219}
}
abstract

The remarkable success of Deep Neural Networks(DNN) is driven by gradient-based optimization, yet this process is often undermined by its tendency to produce disordered weight structures, which harms feature clarity and degrades learning dynamics. To address this fundamental representational flaw, we introduced the Eigen Neural Network (ENN), a novel architecture that reparameterizes each layer's weights in a layer-shared, learned orthonormal eigenbasis. This design enforces decorrelated, well-aligned weight dynamics axiomatically, rather than through regularization, leading to more structured and discriminative feature representations. When integrated with standard BP, ENN consistently outperforms state-of-the-art methods on large-scale image classification benchmarks, including ImageNet, and its superior representations generalize to set a new benchmark in cross-modal image-text retrieval. Furthermore, ENN's principled structure enables a highly efficient, backpropagation-free(BP-free) local learning variant, ENN-$\ell$. This variant not only resolves BP's procedural bottlenecks to achieve over 2$\times$ training speedup via parallelism, but also, remarkably, surpasses the accuracy of end-to-end backpropagation. ENN thus presents a new architectural paradigm that directly remedies the representational deficiencies of BP, leading to enhanced performance and enabling a more efficient, parallelizable training regime.

Discussion (0). Sign in to comment.

Reference graph

Works this paper leans on

50 extracted references · 31 canonical work pages

  1. [1]

    , " * write output.state after.block = add.period write newline

    ENTRY address archivePrefix author booktitle chapter edition editor eid eprint howpublished institution isbn journal key month note number organization pages publisher school series title type volume year label extra.label sort.label short.list INTEGERS output.state before.all mid.sentence after.sentence after.block FUNCTION init.state.consts #0 'before.a...

  2. [2]

    write newline

    " write newline "" before.all 'output.state := FUNCTION n.dashify 't := "" t empty not t #1 #1 substring "-" = t #1 #2 substring "--" = not "--" * t #2 global.max substring 't := t #1 #1 substring "-" = "-" * t #2 global.max substring 't := while if t #1 #1 substring * t #2 global.max substring 't := if while FUNCTION word.in bbl.in capitalize " " * FUNCT...

  3. [3]

    Bansal, N.; Chen, X.; and Wang, Z. 2018. Can we gain more from orthogonality regularizations in training deep cnns? arxiv 2018. arXiv preprint arXiv:1810.09102

    arXiv 2018

  4. [4]

    Bardes, A.; Ponce, J.; and LeCun, Y. 2021. Vicreg: Variance-invariance-covariance regularization for self-supervised learning. arXiv preprint arXiv:2105.04906

    arXiv 2021

  5. [5]

    Bau, D.; Zhou, B.; Khosla, A.; Oliva, A.; and Torralba, A. 2017. Network dissection: Quantifying interpretability of deep visual representations. In Proceedings of the IEEE conference on computer vision and pattern recognition, 6541--6549

    2017

  6. [6]

    Belilovsky, E.; Eickenberg, M.; and Oyallon, E. 2020. Decoupled greedy learning of cnns. In International Conference on Machine Learning, 736--745. PMLR

    work page 2020

  7. [7]

    B \"o ttcher, A.; and Wenzel, D. 2008. The Frobenius norm and the commutator. Linear algebra and its applications, 429(8-9): 1864--1885

    work page 2008

  8. [8]

    Chen, Y.; Yuille, A.; and Zhou, Z. 2023. Which layer is learning faster? a systematic exploration of layer-wise convergence rate for deep neural networks. In The Eleventh International Conference on Learning Representations

    work page 2023

Show all 50 references
  1. [9]

    Cheng, A.; Ping, H.; Wang, Z.; Xiao, X.; Yin, C.; Nazarian, S.; Cheng, M.; and Bogdan, P. 2024. Unlocking Deep Learning: A BP-Free Approach for Parallel Block-Wise Training of Neural Networks. In ICASSP 2024-2024 IEEE International Conference on Acoustics, Speech and Signal Pr...

    2024

  2. [10]

    Cheon, J.; Bae, J.; and Paik, S.-B. 2025. One-Time Soft Alignment Enables Resilient Learning without Weight Transport. arXiv preprint arXiv:2505.20892

    2025 arXiv

  3. [11]

    Deng, J.; Dong, W.; Socher, R.; Li, L.-J.; Li, K.; and Fei-Fei, L. 2009. Imagenet: A large-scale hierarchical image database. In 2009 IEEE conference on computer vision and pattern recognition, 248--255. Ieee

    2009

  4. [12]

    Denil, M.; Shakibi, B.; Dinh, L.; Ranzato, M.; and De Freitas, N. 2013. Predicting parameters in deep learning. Advances in neural information processing systems, 26

    2013

  5. [13]

    Dosovitskiy, A.; Beyer, L.; Kolesnikov, A.; Weissenborn, D.; Zhai, X.; Unterthiner, T.; Dehghani, M.; Minderer, M.; Heigold, G.; Gelly, S.; et al. 2020. An image is worth 16x16 words: Transformers for image recognition at scale. arXiv preprint arXiv:2010.11929

    2020 arXiv

  6. [14]

    u gel, K.; Coquelin, D.; Weiel, M.; Debus, C.; Streit, A.; and G \

    Fl \"u gel, K.; Coquelin, D.; Weiel, M.; Debus, C.; Streit, A.; and G \"o tz, M. 2024. Feed-Forward Optimization With Delayed Feedback for Neural Network Training. In International Conference on Neural Information Processing, 74--88. Springer

    2024

  7. [15]

    T.; Nazeer, K

    Fokam, C. T.; Nazeer, K. K.; Mayr, C.; Subramoney, A.; and Kappel, D. 2024. A variational framework for local learning with probabilistic latent representations. The Twelfth International Conference on Learning Representations

    2024

  8. [16]

    Gong, J.; Li, B.; and Abdulla, W. 2025. Mono-Forward: Backpropagation-Free Algorithm for Efficient Neural Network Training Harnessing Local Errors. arXiv preprint arXiv:2501.09238

    2025 arXiv

  9. [17]

    Goupy, G.; Tirilly, P.; and Bilasco, I. M. 2025. Frozen Backpropagation: Relaxing Weight Symmetry in Temporally-Coded Deep Spiking Neural Networks. arXiv preprint arXiv:2505.13741

    2025 arXiv

  10. [18]

    He, J.; Du, J.; and Ma, W. 2024. Preventing dimensional collapse in self-supervised learning via orthogonality regularization. Advances in Neural Information Processing Systems, 37: 95579--95606

    2024

  11. [19]

    He, K.; Zhang, X.; Ren, S.; and Sun, J. 2016. Deep residual learning for image recognition. In Proceedings of the IEEE conference on computer vision and pattern recognition, 770--778

    2016

  12. [20]

    Hinton, G. 2022. The forward-forward algorithm: Some preliminary investigations. arXiv preprint arXiv:2212.13345, 2(3): 5

    2022 arXiv

  13. [21]

    Huang, G.; Liu, Z.; Van Der Maaten, L.; and Weinberger, K. Q. 2017. Densely connected convolutional networks. In Proceedings of the IEEE conference on computer vision and pattern recognition, 4700--4708

    2017

  14. [22]

    Huo, Z.; Gu, B.; and Huang, H. 2018. Training neural networks using features replay. Advances in Neural Information Processing Systems, 31

    2018

  15. [23]

    M.; Osindero, S.; Vinyals, O.; Graves, A.; Silver, D.; and Kavukcuoglu, K

    Jaderberg, M.; Czarnecki, W. M.; Osindero, S.; Vinyals, O.; Graves, A.; Silver, D.; and Kavukcuoglu, K. 2017. Decoupled neural interfaces using synthetic gradients. In International conference on machine learning, 1627--1635. PMLR

    2017

  16. [24]

    Ji-An, L.; and Benna, M. K. 2024. Deep learning without weight symmetry. ArXiv, arXiv--2405

    2024

  17. [25]

    Krizhevsky, A.; Hinton, G.; et al. 2009. Learning multiple layers of features from tiny images

    2009

  18. [26]

    Le, Y.; and Yang, X. 2015. Tiny imagenet visual recognition challenge. CS 231N, 7(7): 3

    2015

  19. [27]

    Lee, C.-Y.; Xie, S.; Gallagher, P.; Zhang, Z.; and Tu, Z. 2015. Deeply-supervised nets. In Artificial intelligence and statistics, 562--570. Pmlr

    2015

  20. [28]

    W.; and Pascanu, R

    Li, Q.; Teh, Y. W.; and Pascanu, R. 2025. NoProp: Training Neural Networks without Back-propagation or Forward-propagation. arXiv preprint arXiv:2503.24322

    2025 arXiv

  21. [29]

    Liao, Q.; Leibo, J.; and Poggio, T. 2016. How important is weight symmetry in backpropagation? In Proceedings of the AAAI Conference on Artificial Intelligence, volume 30

    2016

  22. [30]

    P.; Santoro, A.; Marris, L.; Akerman, C

    Lillicrap, T. P.; Santoro, A.; Marris, L.; Akerman, C. J.; and Hinton, G. 2020. Backpropagation and the brain. Nature Reviews Neuroscience, 21(6): 335--346

    2020

  23. [31]

    Lin, T.-Y.; Maire, M.; Belongie, S.; Hays, J.; Perona, P.; Ramanan, D.; Doll \'a r, P.; and Zitnick, C. L. 2014. Microsoft coco: Common objects in context. In European conference on computer vision, 740--755. Springer

    2014

  24. [32]

    Mostafa, H.; Ramesh, V.; and Cauwenberghs, G. 2018. Deep supervised learning using local errors. Frontiers in neuroscience, 12: 608

    2018

  25. [33]

    N kland, A. 2016. Direct feedback alignment provides learning in deep neural networks. Advances in neural information processing systems, 29

    2016

  26. [34]

    N kland, A.; and Eidnes, L. H. 2019. Training neural networks with local error signals. In International conference on machine learning, 4839--4850. PMLR

    2019

  27. [35]

    A.; Wang, L.; Cervantes, C

    Plummer, B. A.; Wang, L.; Cervantes, C. M.; Caicedo, J. C.; Hockenmaier, J.; and Lazebnik, S. 2015. Flickr30k entities: Collecting region-to-phrase correspondences for richer image-to-sentence models. In Proceedings of the IEEE international conference on computer vision, 2641--2649

    2015

  28. [36]

    Pyeon, M.; Moon, J.; Hahn, T.; and Kim, G. 2020. Sedona: Search for decoupled neural networks toward greedy block-wise learning. In International conference on learning representations

    2020

  29. [37]

    W.; Hallacy, C.; Ramesh, A.; Goh, G.; Agarwal, S.; Sastry, G.; Askell, A.; Mishkin, P.; Clark, J.; et al

    Radford, A.; Kim, J. W.; Hallacy, C.; Ramesh, A.; Goh, G.; Agarwal, S.; Sastry, G.; Askell, A.; Mishkin, P.; Clark, J.; et al. 2021. Learning transferable visual models from natural language supervision. In International conference on machine learning, 8748--8763. PmLR

    2021

  30. [38]

    Shervani-Tabar, N.; and Rosenbaum, R. 2023. Meta-learning biologically plausible plasticity rules with random feedback pathways. Nature Communications, 14(1): 1805

    2023

  31. [39]

    Sun, Q.; Fang, Y.; Wu, L.; Wang, X.; and Cao, Y. 2023. Eva-clip: Improved training techniques for clip at scale. arXiv preprint arXiv:2303.15389

    2023 arXiv

  32. [40]

    Szegedy, C.; Liu, W.; Jia, Y.; Sermanet, P.; Reed, S.; Anguelov, D.; Erhan, D.; Vanhoucke, V.; and Rabinovich, A. 2015. Going deeper with convolutions. In Proceedings of the IEEE conference on computer vision and pattern recognition, 1--9

    2015

  33. [41]

    Szegedy, C.; Zaremba, W.; Sutskever, I.; Bruna, J.; Erhan, D.; Goodfellow, I.; and Fergus, R. 2013. Intriguing properties of neural networks. arXiv preprint arXiv:1312.6199

    2013 arXiv

  34. [42]

    Wang, B.; Ning, Z.; Ding, J.; Gao, X.; Li, Y.; Jiang, D.; Yang, J.; and Liu, W. 2025. FIX-CLIP: Dual-Branch Hierarchical Contrastive Learning via Synthetic Captions for Better Understanding of Long Text. arXiv preprint arXiv:2507.10095

    2025 arXiv

  35. [43]

    Wang, J.; Chen, Y.; Chakraborty, R.; and Yu, S. X. 2020. Orthogonal convolutional neural networks. In Proceedings of the IEEE/CVF conference on computer vision and pattern recognition, 11505--11515

    2020

  36. [44]

    Wang, Q.; Gao, M.; Zhang, Z.; Xie, J.; Li, P.; and Hu, Q. 2022. DropCov: A simple yet effective method for improving deep architectures. Advances in Neural Information Processing Systems, 35: 33576--33588

    2022

  37. [45]

    F.; Massett, R

    Yin, C.; Imms, P.; Cheng, M.; Amgalan, A.; Chowdhury, N. F.; Massett, R. J.; Chaudhari, N. N.; Chen, X.; Thompson, P. M.; Bogdan, P.; et al. 2023. Anatomically interpretable deep learning of brain age captures domain-specific cognitive impairment. Proceedings of the National A...

    2023

  38. [46]

    Yoshida, Y.; and Miyato, T. 2017. Spectral norm regularization for improving the generalizability of deep learning. arXiv preprint arXiv:1705.10941

    2017 arXiv

  39. [47]

    Yu, Y.; Scheidegger, S.; and McKelvey, T. 2024. Building efficient CNNs using Depthwise Convolutional Eigen-Filters (DeCEF). Neurocomputing, 609: 128461

    2024

  40. [48]

    D.; and Fergus, R

    Zeiler, M. D.; and Fergus, R. 2014. Visualizing and understanding convolutional networks. In European conference on computer vision, 818--833. Springer

    2014

  41. [49]

    Zhang, A.; Yang, Z.; Wang, N.; Qi, Y.; Xin, J.; Li, X.; and Yin, P. 2025. Comq: A backpropagation-free algorithm for post-training quantization. IEEE Access

    2025

  42. [50]

    Zhang, B.; Zhang, P.; Dong, X.; Zang, Y.; and Wang, J. 2024. Long-clip: Unlocking the long-text capability of clip. In European conference on computer vision, 310--325. Springer

    2024

Pith tools

Reviewed August 6, 2026 · model on record in the stance chip above.