Pith. sign in

REVIEW 4 major objections 6 minor 1 cited by

Adaptive Regularization of Labels

T0 review · 4 major / 6 minor · reviewed 2026-08-14 · deepseek-v4-flash

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

Pith's one-line read A self-learned regularizer that recycles a network's own mistakes improves classification accuracy across image and text benchmarks.

desk verdict The paper's own ImageNet table undercuts its 'all settings' claim, but the residual-label idea is a real variant on label smoothing and worth refereeing after the overstatement is fixed. read the letter →

arxiv 1908.05474 v1 pith:UEUE4YCD submitted 2019-08-15 cs.LG stat.ML

Qianggang Ding , Sifan Wu , Hao Sun , Jiadong Guo , Shu-Tao Xia This is my paper · ORCID
classification cs.LGstat.ML
keywords labelregularizationresidualcorrelationmatrixlabelssoftknowledgedistillationimageclassificationtext
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

This paper tries to establish that a classifier can regularize itself by recycling its own mistakes. The authors define a residual label—a probability distribution over the classes a sample does not belong to—and add a residual loss that pushes the network's erroneous outputs toward this learned distribution, with a total loss of $L_{\mathrm{tot}} = L_{\mathrm{hard}} + (1 - \mathrm{acc}_{\mathrm{train}})L_{\mathrm{res}} + L_{\mathrm{upd}}$. Unlike knowledge distillation, no separate teacher network is needed; the same network learns inter-class correlations online through a tiny embedding layer. If the claim is right, any one-hot classification task can be regularized nearly for free, and the reported experiments on CIFAR, ImageNet, and text benchmarks all show accuracy gains.

What carries the argument

The load-bearing object is the residual correlation matrix: a $K \times (K-1)$ matrix whose $i$-th row, after softmax, gives the residual label, i.e. the probabilities that a sample of class $i$ is misclassified into each of the other $K-1$ classes. It is stored as the weight of an embedding layer, so a label can be converted to a residual label by lookup, and it is updated online by $L_{\mathrm{upd}}$ while $L_{\mathrm{res}}$ transfers the same distribution back to the backbone network. This two-way coupling is what lets the network learn from its own erroneous experience and is also what makes the method teacher-free.

What would settle it

Run the same training recipe on CIFAR-10 with a fraction of labels randomly flipped, or with one class heavily overrepresented; if the residual labels $q^{(\mathrm{res})}$ concentrate on the noisy or dominant classes and the final test accuracy falls below the plain baseline, the central assumption fails.

Watch

Extended reading notes

Core claim

The central claim is that a network's own confusion pattern—what it tends to wrongly predict for each class—is useful training signal, not just noise. The method erases the ground-truth position from the network's output logits, builds a residual probability vector $p^{(\mathrm{res})}$ by softmax, and matches it to residual labels $q^{(\mathrm{res})}$ stored in an embedding layer. The update loss $L_{\mathrm{upd}}$ adapts $q^{(\mathrm{res})}$ to the current network, while the residual loss $L_{\mathrm{res}}$ pushes the network toward that same distribution, with a weight $(1 - \mathrm{acc}_{\mathrm{train}})$ that fades the correction as training accuracy rises. The paper reports that this adaptive label regularization improves top-1 accuracy over standard training under all tested image and text classification settings, and that combining it with label smoothing gives further gains.

Load-bearing premise

The method assumes that the patterns of a network's own mistakes on the training set are stable and trustworthy inter-class knowledge, so that updating residual labels online cannot amplify the model's errors.

Editorial extensions

If this is right

  • A standard one-hot classifier can be augmented with a residual-label embedding of size roughly $K \times (K-1)$, adding only a few thousand parameters even for 100-class tasks.
  • The method is architecture-agnostic in the reported experiments, improving ResNet, WideResNet, FastText, TextRNN, CharCNN, and Transformer baselines.
  • Combining adaptive label regularization with label smoothing is complementary; the largest reported gain on CIFAR-100 reaches 1.54% over the baseline.
  • On ImageNet, top-1 accuracy improves while top-5 accuracy can be slightly lower, indicating the regularization changes the structure of errors rather than only their rate.
  • The method reaches accuracy comparable to deep mutual learning while using only about half the parameters.

Reading between the lines

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

  • Because residual labels are learned online from the network's own confusion, they may implicitly encode a data-driven taxonomy; a trained residual correlation matrix could be reused as a soft-label prior for transfer or semi-supervised learning.
  • The method's safety likely depends on the training distribution being balanced and clean; a natural test is to run it under label noise or class imbalance, where the online-updated residual labels could drift toward the noise and turn $L_{\mathrm{res}}$ from a regularizer into an error amplifier.
  • The $(1 - \mathrm{acc}_{\mathrm{train}})$ weighting suggests a testable variation: replace batch accuracy with per-class confidence or uncertainty estimates to see whether the correction schedule can be made more local.
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 / 6 minor

Summary. The paper proposes Adaptive Label Regularization (ALR), a teacher-free method that maintains a residual correlation matrix as an embedding layer. Given a ground-truth label, this layer produces a residual label q^(res) over the K-1 non-ground-truth classes. The total loss is Ltot = Lhard + (1 - acctrain)Lres + Lupd, where Lupd fits q^(res) to the network's current residual prediction distribution p^(res) (Eq. 4) and Lres trains the network's residual predictions to match q^(res) (Eq. 5). The method is evaluated on CIFAR-10, CIFAR-100, ImageNet, and three text classification datasets, and is compared with label smoothing and Deep Mutual Learning. The authors claim significant improvement under all experimental settings and provide a qualitative visualization analysis of the learned residual labels on CIFAR-10.

Significance. If the claims are correct, ALR is an attractive low-overhead regularizer: it needs no teacher network, adds only a small embedding layer, can be combined with label smoothing, and shows gains across image and text classification. The paper's strengths are the simple formulation, the parameter efficiency compared with mutual-distillation baselines (Table 4), and the breadth of the evaluation. However, the headline claim is overstated by the paper's own ImageNet top-5 results, the gradient derivation contains a missing factor, statistical support is weak, and the stability of the self-referential residual-label update is not established. These issues are central to the paper's contribution and require substantial revision.

major comments (4)
  1. [Abstract; Section 4.1, Table 2] The abstract claims "significant improvement under all experimental settings," but Table 2 shows that ALR decreases ImageNet top-5 accuracy on all three backbones (ResNet-50: 93.03 to 92.73; ResNet-101: 93.82 to 93.63; ResNet-152: 93.89 to 93.80), and ALR-S decreases top-5 accuracy on ResNet-50 and ResNet-101. The authors acknowledge this in Section 4.1: "the top-5 accuracy of our methods is slightly lower than it of the baselines." This contradiction is load-bearing because the central empirical claim is the universality of the improvement. The authors must either restrict the claim to top-1 accuracy, temper the abstract, or provide a convincing explanation and analysis of the top-5 trade-off.
  2. [Section 3.2, Eqs. (4)-(8)] The gradient formula in Eq. (8) omits the 1/(K-1) factor that is present in the definition of Lres in Eq. (5). With Lres = -1/(K-1) * sum_i q_i^(res) log p_i^(res), direct differentiation over the softmax of the erased logits gives dLres/dz_i = (p_i^(res) - q_i^(res))/(K-1), not (p_i^(res) - q_i^(res)). In addition, Eq. (7) is presented as the gradient of Lb = Lhard + Lres, but the final training objective is Ltot = Lhard + (1 - acctrain)Lres + Lupd, so the displayed gradient does not correspond to the loss actually minimized. These are technical errors in the core derivation and should be corrected.
  3. [Section 4, Tables 1-3] The word "significant" is not supported by the reported statistics. On CIFAR-10, the ResNet-18 baseline is 95.28±0.21 and ALR is 95.42±0.16, a difference well within one standard deviation; on CIFAR-100 with WideResNet-28-10, baseline 81.20±0.08 vs ALR 81.31±0.14 likewise overlaps. Table 3 reports text-classification accuracies as single numbers with no standard deviations, repeated runs, or significance tests, and the ImageNet results appear to be single runs with no error bars. The authors should either provide error bars across multiple seeded runs and significance tests, or avoid claiming statistical significance.
  4. [Section 3.2, Eqs. (4)-(5); Section 5] The residual-label update is self-referential: Lupd moves q^(res) toward the network's own current p^(res), while Lres moves the network's p^(res) toward q^(res). At a fixed point p^(res) is approximately q^(res), so the regularizer ultimately targets the model's own error distribution. The paper offers no fixed-point analysis, no convergence argument, and no experiments under label noise or class imbalance; Section 5 provides only qualitative visualization on CIFAR-10. This leaves open the possibility that the method amplifies systematic errors or collapses to a degenerate distribution on harder or noisier problems, and the mechanism behind the reported gains remains underdetermined.
minor comments (6)
  1. [Section 1] In the discussion of knowledge distillation, "the hard loss Lsoft" should read "the hard loss Lhard"; this appears to be a typo.
  2. [Section 3.2, Eq. (3) and footnote 2] The notation defining the residual correlation matrix is garbled: the sentence "each row vector si = (ri)i 2" is incomplete and hard to parse. Please rewrite the definition of S and the erasing operation clearly.
  3. [Section 3.2] The text first defines Ltot = Lhard + Lres + Lupd (Eq. 6), then later writes Ltot = Lhard + (1 - acctrain)Lres + Lupd without renumbering, and Figure 2 does not show the (1 - acctrain) coefficient. Please reconcile the notation and the figure.
  4. [Section 3.2] The quantity acctrain is used as an adaptive weight but is never formally defined, for example whether it is a running average over the current epoch or over the whole training history.
  5. [Section 4.1] The paper states that ImageNet-12 consists of "more than14 million samples in 1,000 classes"; the ILSVRC2012 training set actually has roughly 1.28 million images. Please correct this factual detail.
  6. [Section 3.2] The claim that the method "has no hyper-parameter that needed to be manually adjusted" is overstated, because (1 - acctrain) is a hyperparameter of the loss and ALR-S additionally uses a label-smoothing coefficient.

Circularity Check

0 steps flagged · score 0.0 of 10

No circularity: the learned residual labels are self-referential by design, but test-time evaluation is independent and no load-bearing derivation reduces to its inputs.

full rationale

The paper's adaptive label regularization defines residual labels q^(res) that are updated online by Lupd (Eq. 4) to match the network's own non-ground-truth probabilities p^(res), while Lres (Eq. 5) trains the backbone toward q^(res). The text makes this explicit: 'Then we use the cross entropy loss between p^(res) and q^(res) as update loss Lupd to update residual labels. To regularize the output logits, we use the reverse cross entropy loss between p^(res) and q^(res) as residual loss Lres.' This is a self-referential feedback loop, but it is the method's stated design ('learn from erroneous experience'), not a hidden reduction of a predicted quantity to a fitted input. The residual labels are not presented as externally derived ground truth; they are explicitly learned from the network itself, and the paper's claims are evaluated by held-out test accuracy on CIFAR, ImageNet, and text datasets, which are independent of the training-time residual labels. There are no self-citations carrying a load-bearing argument, no imported uniqueness theorem, and no fitted parameter relabeled as a prediction. The abstract's 'significant improvement under all experimental settings' is contradicted by the ImageNet top-5 numbers in Table 2 (e.g., ResNet-50 top-5 93.03 baseline vs 92.73 ALR), but that is a correctness/support concern, not circularity. Under the circularity pass, no step reduces by construction to its own input.

Assumptions & free parameters 3 free parameters · 4 assumptions · 2 invented entities

The method introduces a learned residual correlation matrix and a training-dependent loss weight, both free parameters. The central assumption is that the network's own confusion distribution is a stable source of inter-class knowledge; this is asserted from visualization rather than proved. The residual labels and matrix are invented entities with no external handle.

free parameters (3)
  • Residual correlation matrix S = learned online, not reported
    The central learned component of the method: a K x (K-1) matrix whose softmax rows define residual labels. It is fit to the network's own predictions via Lupd (Eq. 4), making it a free parameter introduced by the paper.
  • Residual loss weight (1 - acctrain) = depends on running training accuracy
    A heuristic, training-dependent scaling factor applied to Lres in Section 3.2. It is a design choice that controls the strength of residual regularization and is not a standard hyperparameter of the base models.
  • Label smoothing coefficient in ALR-S = not reported
    The combination ALR-S adds label smoothing, but its epsilon value is not stated in Section 4, leaving the configuration underspecified.
assumptions (4)
  • domain assumption The softmax of a row of the residual correlation matrix is a valid and useful target distribution for regularization.
    Section 3.2 defines residual labels this way and the method assumes they carry inter-class knowledge worth learning.
  • ad hoc to paper The model's training-set confusion distribution is stable and transferable enough that the online residual labels do not amplify errors.
    Section 5 supports this only with a qualitative visualization on CIFAR-10; there is no proof or test on larger or noisier datasets.
  • domain assumption Uniform initialization of residual labels maximizes entropy and is a safe starting point.
    Section 3.2 sets all residual logits to zero based on the principle of maximum entropy.
  • domain assumption The modified ResNet and training schedules used in experiments are appropriate baselines.
    Standard practice, but the empirical claims depend on these settings.
invented entities (2)
  • Residual correlation matrix S
    purpose: Stores a soft label per class representing the probability of being misclassified into each other class; updated online and used as a regularization target.
    This matrix is internal to the training loop and has no falsifiable handle outside the paper; it is fit to the model's own outputs.
  • Residual label q^(res)
    purpose: The softmax-normalized row of S, used as the target in the residual loss Lres.
    Derived entirely from S and the network's predictions; no independent evidence.

how reviews work

0 comments
Cite this review

Pith. "Pith review of Adaptive Regularization of Labels." pith.science (2026). https://pith.science/paper/UEUE4YCD

@misc{pith2026190805474,
  author       = {Pith},
  title        = {Pith review of: Adaptive Regularization of Labels},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/UEUE4YCD}},
  note         = {Machine review of arXiv:1908.05474}
}
read the original abstract

Recently, a variety of regularization techniques have been widely applied in deep neural networks, such as dropout, batch normalization, data augmentation, and so on. These methods mainly focus on the regularization of weight parameters to prevent overfitting effectively. In addition, label regularization techniques such as label smoothing and label disturbance have also been proposed with the motivation of adding a stochastic perturbation to labels. In this paper, we propose a novel adaptive label regularization method, which enables the neural network to learn from the erroneous experience and update the optimal label representation online. On the other hand, compared with knowledge distillation, which learns the correlation of categories using teacher network, our proposed method requires only a minuscule increase in parameters without cumbersome teacher network. Furthermore, we evaluate our method on CIFAR-10/CIFAR-100/ImageNet datasets for image recognition tasks and AGNews/Yahoo/Yelp-Full datasets for text classification tasks. The empirical results show significant improvement under all experimental settings.

Figures

Figures reproduced from arXiv: 1908.05474 by the authors.

Figure 1
Figure 1. The relation among correlation matrix, residual correlation matrix, and residual labels: (1) [PITH_FULL_IMAGE:figures/full_fig_p004_1.png] view at source ↗
Figure 2
Figure 2. Our proposed adaptive label regularization pipeline: the part in blue is backbone network [PITH_FULL_IMAGE:figures/full_fig_p005_2.png] view at source ↗
Figure 3
Figure 3. The softmax output of residual correlation matrix on CIFAR-10: each row of them is the [PITH_FULL_IMAGE:figures/full_fig_p008_3.png] view at source ↗

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. Dynamic Frequency-Adaptive Knowledge Distillation for Speech Enhancement

    cs.SD 2025-02 conditional novelty 5.0 of 10

    A dynamic frequency-adaptive knowledge distillation method, using the steepest point in the running maximum of the teacher spectrum as a crossover, improves speech enhancement student models by small PESQ margins over...

Reference graph

Works this paper leans on

32 extracted references · 12 canonical work pages · cited by 1 Pith paper

  1. [1]

    La- bel refinery: Improving imagenet classification through label progression

    Hessam Bagherinezhad, Maxwell Horton, Mohammad Rastegari, and Ali Farhadi. La- bel refinery: Improving imagenet classification through label progression. arXiv preprint arXiv:1805.02641, 2018

    arXiv 2018

  2. [2]

    Recur- rent batch normalization

    Tim Cooijmans, Nicolas Ballas, César Laurent, Ça˘glar Gülçehre, and Aaron Courville. Recur- rent batch normalization. arXiv preprint arXiv:1603.09025, 2016

    arXiv 2016

  3. [3]

    Ranking a stream of news

    Gianna M Del Corso, Antonio Gulli, and Francesco Romani. Ranking a stream of news. In Proceedings of the 14th international conference on World Wide Web , pages 97–106. ACM, 2005

    work page 2005

  4. [4]

    Maxout networks

    Ian J Goodfellow, David Warde-Farley, Mehdi Mirza, Aaron Courville, and Yoshua Bengio. Maxout networks. arXiv preprint arXiv:1302.4389, 2013

    arXiv 2013

  5. [5]

    Deep residual learning for image recognition

    Kaiming He, Xiangyu Zhang, Shaoqing Ren, and Jian Sun. Deep residual learning for image recognition. In Proceedings of the IEEE conference on computer vision and pattern recognition , pages 770–778, 2016

    2016

  6. [6]

    Identity mappings in deep residual networks

    Kaiming He, Xiangyu Zhang, Shaoqing Ren, and Jian Sun. Identity mappings in deep residual networks. In European conference on computer vision, pages 630–645. Springer, 2016

    2016

  7. [7]

    Distilling the knowledge in a neural network

    Geoffrey Hinton, Oriol Vinyals, and Jeff Dean. Distilling the knowledge in a neural network. arXiv preprint arXiv:1503.02531, 2015

    arXiv 2015

  8. [8]

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

    Sergey Ioffe and Christian Szegedy. Batch normalization: Accelerating deep network training by reducing internal covariate shift. international conference on machine learning , pages 448–456, 2015

    work page 2015

Show all 32 references
  1. [9]

    Information theory and statistical mechanics

    E T Jaynes. Information theory and statistical mechanics. Physical Review, 106(2):620–630, 1957

    1957

  2. [10]

    Bag of tricks for efficient text classification

    Armand Joulin, Edouard Grave, Piotr Bojanowski, and Tomas Mikolov. Bag of tricks for efficient text classification. conference of the european chapter of the association for computational linguistics, 2:427–431, 2017

    2017

  3. [11]

    Learning multiple layers of features from tiny images

    Alex Krizhevsky and Geoffrey Hinton. Learning multiple layers of features from tiny images. Technical report, Citeseer, 2009

    2009

  4. [12]

    Imagenet classification with deep convolutional neural networks

    Alex Krizhevsky, Ilya Sutskever, and Geoffrey E Hinton. Imagenet classification with deep convolutional neural networks. In Advances in neural information processing systems , pages 1097–1105, 2012

    2012

  5. [13]

    Recurrent convolutional neural networks for text classification

    Siwei Lai, Liheng Xu, Kang Liu, and Jun Zhao. Recurrent convolutional neural networks for text classification. pages 2267–2273, 2015

    2015

  6. [14]

    Deeply- supervised nets

    Chen-Yu Lee, Saining Xie, Patrick Gallagher, Zhengyou Zhang, and Zhuowen Tu. Deeply- supervised nets. arXiv preprint arXiv:1409.5185, 2014

    2014 arXiv

  7. [15]

    Unifying distillation and privileged information

    David Lopez-Paz, Léon Bottou, Bernhard Schölkopf, and Vladimir Vapnik. Unifying distillation and privileged information. arXiv preprint arXiv:1511.03643, 2015

    2015 arXiv

  8. [16]

    Mixed precision training

    Paulius Micikevicius, Sharan Narang, Jonah Alben, Gregory Diamos, Erich Elsen, David Garcia, Boris Ginsburg, Michael Houston, Oleksii Kuchaiev, Ganesh Venkatesh, et al. Mixed precision training. arXiv preprint arXiv:1710.03740, 2017. 9

    arXiv

  9. [17]

    Reg- ularizing neural networks by penalizing confident output distributions

    Gabriel Pereyra, George Tucker, Jan Chorowski, Łukasz Kaiser, and Geoffrey Hinton. Reg- ularizing neural networks by penalizing confident output distributions. arXiv preprint arXiv:1701.06548, 2017

    2017 arXiv

  10. [18]

    Imagenet large scale visual recognition challenge

    Olga Russakovsky, Jia Deng, Hao Su, Jonathan Krause, Sanjeev Satheesh, Sean Ma, Zhiheng Huang, Andrej Karpathy, Aditya Khosla, Michael Bernstein, et al. Imagenet large scale visual recognition challenge. International journal of computer vision , 115(3):211–252, 2015

    2015

  11. [19]

    Dropout: a simple way to prevent neural networks from overfitting

    Nitish Srivastava, Geoffrey Hinton, Alex Krizhevsky, Ilya Sutskever, and Ruslan Salakhutdinov. Dropout: a simple way to prevent neural networks from overfitting. The Journal of Machine Learning Research, 15(1):1929–1958, 2014

    1929

  12. [20]

    Highway networks

    Rupesh Kumar Srivastava, Klaus Greff, and Jürgen Schmidhuber. Highway networks. arXiv preprint arXiv:1505.00387, 2015

    2015 arXiv

  13. [21]

    On the importance of initialization and momentum in deep learning

    Ilya Sutskever, James Martens, George Dahl, and Geoffrey Hinton. On the importance of initialization and momentum in deep learning. In International conference on machine learning, pages 1139–1147, 2013

    2013

  14. [22]

    Re- thinking the inception architecture for computer vision

    Christian Szegedy, Vincent Vanhoucke, Sergey Ioffe, Jon Shlens, and Zbigniew Wojna. Re- thinking the inception architecture for computer vision. In Proceedings of the IEEE conference on computer vision and pattern recognition , pages 2818–2826, 2016

    2016

  15. [23]

    Attention is all you need

    Ashish Vaswani, Noam Shazeer, Niki Parmar, Jakob Uszkoreit, Llion Jones, Aidan N Gomez, Lukasz Kaiser, and Illia Polosukhin. Attention is all you need. neural information processing systems, pages 5998–6008, 2017

    2017

  16. [24]

    Regularization of neural networks using dropconnect

    Li Wan, Matthew Zeiler, Sixin Zhang, Yann Le Cun, and Rob Fergus. Regularization of neural networks using dropconnect. In International conference on machine learning , pages 1058–1066, 2013

    2013

  17. [25]

    Group normalization

    Yuxin Wu and Kaiming He. Group normalization. In Proceedings of the European Conference on Computer Vision (ECCV), pages 3–19, 2018

    2018

  18. [26]

    Disturblabel: Regularizing cnn on the loss layer

    Lingxi Xie, Jingdong Wang, Zhen Wei, Meng Wang, and Qi Tian. Disturblabel: Regularizing cnn on the loss layer. In Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition, pages 4753–4762, 2016

    2016

  19. [27]

    Shakedrop regulariza- tion for deep residual learning

    Yoshihiro Yamada, Masakazu Iwamura, Takuya Akiba, and Koichi Kise. Shakedrop regulariza- tion for deep residual learning. arXiv preprint arXiv:1802.02375, 2018

    2018 arXiv

  20. [28]

    Wide residual networks

    Sergey Zagoruyko and Nikos Komodakis. Wide residual networks. arXiv preprint arXiv:1605.07146, 2016

    2016 arXiv

  21. [29]

    Character-level convolutional networks for text classification

    Xiang Zhang, Junbo Jake Zhao, and Yann Lecun. Character-level convolutional networks for text classification. neural information processing systems, pages 649–657, 2015

    2015

  22. [30]

    Deep mutual learning

    Ying Zhang, Tao Xiang, Timothy M Hospedales, and Huchuan Lu. Deep mutual learning. In Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition , pages 4320–4328, 2018

    2018

  23. [31]

    Knowledge distillation by on-the-fly native ensemble

    Xiatian Zhu, Shaogang Gong, et al. Knowledge distillation by on-the-fly native ensemble. In Advances in Neural Information Processing Systems , pages 7517–7527, 2018

    2018

  24. [32]

    Learning transferable architectures for scalable image recognition

    Barret Zoph, Vijay Vasudevan, Jonathon Shlens, and Quoc V Le. Learning transferable architectures for scalable image recognition. InProceedings of the IEEE conference on computer vision and pattern recognition, pages 8697–8710, 2018. 10

    2018

Pith tools

Reviewed August 14, 2026 · model on record in the stance chip above.