Pith. sign in

REVIEW 3 major objections 5 minor 3 cited by

Pseudo-Labeling and Confirmation Bias in Deep Semi-Supervised Learning

T0 review · 3 major / 5 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 Semi-supervised image classifiers can learn from their own predictions if confirmation bias is controlled.

desk verdict A genuinely simple pseudo-labeling recipe that performs surprisingly well, but the state-of-the-art comparison is confounded by differences in augmentation and dropout between the proposed method and the cited baselines. read the letter →

arxiv 1908.02983 v5 pith:FXU4USJG submitted 2019-08-08 cs.CV

Eric Arazo , Diego Ortego , Paul Albert , Noel E. O'Connor , Kevin McGuinness This is my paper · ORCID
classification cs.CV
keywords semi-supervisedlearningpseudo-labelingconfirmationbiasmixupaugmentationconsistencyregularizationimageclassificationsoftlabelslabeloversampling
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 claims that pseudo-labeling, training a network on its own predictions for unlabeled images, has been prematurely written off, and that its main failure mode, confirmation bias, can be controlled with two simple regularizers. The authors show that a naive pseudo-labeling scheme overfits incorrect predictions, then demonstrate that mixup augmentation plus a minimum number of labeled samples per mini-batch substantially reduces that bias. With these additions, their purely pseudo-labeling method reaches state-of-the-art test error on CIFAR-10/100, SVHN, and Mini-ImageNet, beating consistency-regularization methods such as Mean Teacher and MixMatch. If correct, this reframes semi-supervised learning: consistency between perturbed views of unlabeled data is not required for top performance.

What carries the argument

The central mechanism is a loss decomposition: the semi-supervised objective splits into a labeled-data term and an unlabeled pseudo-label term, and with few labels the pseudo-label term vastly outweighs the labeled term, so the model fits its own errors. The paper's two fixes target this imbalance and its confidence dynamics. Mixup, defined as training on convex combinations $\delta x_p+(1-\delta)x_q$ with labels $\delta y_p+(1-\delta)y_q$ for $\delta$ drawn from a $\beta$ distribution, acts as label smoothing that prevents overconfident wrong pseudo-labels. A minimum of $k$ labeled samples per mini-batch, with $k=16$ used in most experiments, restores the labeled term's influence by oversampling the clean labels. Around this core sit an all-classes regularization $R_A$ that prevents collapse to one class and an entropy regularization $R_H$ that concentrates soft pseudo-labels, both adapted from the noisy-label method [24], plus dropout and data augmentation that are disabled in the second forward pass so soft pseudo-labels are computed cleanly.

What would settle it

Retrain the strongest consistency baselines with the same 13-CNN, data augmentation, dropout, weight decay, learning-rate schedule, warm-up, and labeled/unlabeled splits used here, and compare CIFAR-10 test errors at 500, 1000, and 4000 labels; if any baseline meets or beats 8.80, 6.85, and 5.97 percent, the paper's central comparison is refuted.

Watch

Extended reading notes

Core claim

The paper's central claim is that pseudo-labeling alone, without consistency regularization, can be state of the art in deep semi-supervised image classification, contrary to the prevailing assumption in the literature. The confirmation-bias diagnosis is that when a network's own sometimes wrong predictions are reused as labels, the model grows more confident in those errors and resists correction; the authors quantify this with a per-epoch measure of the average certainty of misclassified unlabeled samples. They show that naively adapting the soft-relabeling scheme of [24] fails, but two modifications, training on mixup interpolations and enforcing at least $k$ labeled samples in every mini-batch, reduce the certainty of incorrect predictions and let the model converge. In head-to-head comparisons, the approach reports lower test error than published consistency-regularization baselines on CIFAR-10/100, SVHN, and Mini-ImageNet, including a 10-point margin on Mini-ImageNet and a large gap at 500 CIFAR-10 labels (8.80% versus 27.45% for Mean Teacher).

Load-bearing premise

The comparison with state-of-the-art assumes the published baseline numbers come from settings comparable to the authors' training protocol; if those baselines were not tuned under the same augmentation, schedule, and tuning budget, the reported margin over consistency regularization could shrink or disappear.

Editorial extensions

If this is right

  • At 500 labeled CIFAR-10 images, the method's 8.80% test error is over 18 points better than the Mean Teacher baseline (27.45%) under the paper's comparison.
  • On Mini-ImageNet with 4,000 labels, the method reaches 56.49% error versus 70.29% for label-propagation pseudo-labeling, a margin of more than 10 points.
  • Soft pseudo-labels are reported to work better than hard one-hot pseudo-labels, confirming the noisy-label observation the method adapts.
  • Purely pseudo-labeling approaches, not just consistency regularization, belong in future semi-supervised learning benchmarks as first-class contenders.
  • The approach needs one network and a single extra forward pass per sample, unlike multi-network or many-pass consistency methods.

Reading between the lines

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

  • If the diagnosis is right, any self-training pipeline that reuses model predictions as targets should inherit the same bias, so minimum-$k$ oversampling and mixup-style label smoothing may transfer to text, audio, and time-series settings where consistency regularization is less natural.
  • The observation that the 13-CNN is more robust at 250 labels than a Wide ResNet or PreAct ResNet suggests that architecture capacity interacts with pseudo-label noise; a testable extension would sweep depth and width at fixed parameter count to separate capacity effects from skip-connection effects.
  • Since stronger mixup ($\alpha=4$ or $8$) helped in the ablation but was not additive with dropout and extra augmentation, an adaptive $\alpha$ schedule that starts strong to suppress early confirmation bias and anneals later is a plausible extension the paper does not test.
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

3 major / 5 minor

Summary. The paper proposes a semi-supervised image classification method based on pseudo-labeling. The network's softmax predictions on unlabeled samples serve as soft pseudo-labels in a cross-entropy loss, augmented by an all-class regularization term and an entropy regularization term. To counter confirmation bias, the method uses mixup, a minimum number of labeled samples per mini-batch, dropout, and color-jitter augmentation. Experiments on CIFAR-10/100, SVHN, and Mini-ImageNet with the 13-CNN architecture and several other networks report state-of-the-art results and the claim that pseudo-labeling alone can outperform consistency-regularization methods.

Significance. If the comparative claims hold, the paper provides a valuable and conceptually simple counterpoint to the consistency-regularization paradigm that has dominated recent deep semi-supervised learning. The strengths of the work include careful ablation studies (Tables I-III), an explicit quantitative measure of confirmation bias (rt in Figure 2), averaged results with standard deviations over three runs, and released source code. However, as detailed below, the headline comparison against consistency-regularization baselines is weakened by a protocol mismatch, and the paper's central claim needs to be either supported by same-protocol baseline runs or carefully qualified.

major comments (3)
  1. [Section IV-E, Tables IV and V] The headline comparison is not apples-to-apples. Ours* uses the final configuration developed in Section IV-D (dropout p=0.1, color-jitter augmentation, and a second forward pass for pseudo-labels), while the consistency-regularization entries (MT, ICT, Pi-model, etc.) are unmarked numbers taken from prior papers. Section IV-A states only that supervised baselines use the same augmentation and dropout, not that the cited SSL baselines do. Table III shows that these additions are not negligible: on CIFAR-10 with 500 labels, M* improves from 13.68 to 9.16 when dropout and color jitter are added. The claimed margin over consistency regularization, e.g. 8.80 vs 27.45 (MT) in Table V, could therefore be explained by the additional regularization rather than by pseudo-labeling itself. I ask the authors to rerun the key consistency baselines under their exact training protocol, or to clearly restrict the claim so that it does not depend on unmarked baseline numbers.
  2. [Section IV-A] The validation set is used for hyperparameter selection in Subsections IV-B through IV-D and is then folded back into the training set for the final comparisons in Subsection IV-E. The paper states that the 5K validation samples are added back for the results in IV-E, where test results are reported from the model at the best epoch. This means the final test numbers are produced by a pipeline whose hyperparameters and epoch selection were partly determined on data that later becomes training data. While this is not unusual in the literature, it removes a layer of independence and can inflate the reported margins; the authors should either report results under a held-out validation protocol or explicitly discuss the potential optimistic bias.
  3. [Section IV-E, Table VI] The only consistency-regularization baseline that the authors rerun under their own protocol, ICT* on WR-28, is not used in the main comparison of Tables IV and V. Moreover, for 250 labels, Ours* on WR-28 (24.81) is much worse than the cited MixMatch result on the same architecture (11.08), and on PR-18 it also degrades sharply (23.86). This does not invalidate the method on the 13-CNN architecture, but it directly contradicts the abstract's general claim that 'pseudo-labeling alone can outperform consistency regularization methods.' The claim should be qualified by architecture and label regime, and the same-protocol baseline comparisons should be extended to the settings where the claim is made.
minor comments (5)
  1. [Equations (5)-(6)] The notation δ∈{0,1} is incorrect; since δ is sampled from a beta distribution, it should be δ∈[0,1] (or the text should say that δ is a continuous value in this interval).
  2. [Section IV-A] There is a typo: 'respectivelly' should be 'respectively'.
  3. [Tables IV and V captions] The caption notes that '*' denotes runs by the authors, but it does not indicate which of the unmarked baselines used the same data augmentation, dropout, or learning-rate schedule as Ours*. A clear statement is needed to interpret the comparison.
  4. [Equation (8)] The notation in the decomposition 𝓁* = Nl𝓁l + Nu𝓁u is slightly confusing because Nl and Nu are counts while 𝓁l and 𝓁u are averages; explicitly defining the per-sample averaged losses would remove ambiguity.
  5. [Section III-A] The phrase 'as well as other purely pseudo-labeling approaches and their combination with consistency regularization methods' in Section IV-E is broad; consider listing the specific methods or grouping them more precisely, since the table includes several hybrid methods.

Circularity Check

0 steps flagged · score 0.0 of 10

No circularity: the paper is an empirical pseudo-labeling method evaluated on external benchmarks, with no prediction that reduces to a fitted input or to a self-citation chain.

full rationale

No circular step is present. The paper proposes an empirical training procedure (soft pseudo-labels generated from network predictions, mixup, a minimum number of labeled samples per mini-batch, dropout, and extra data augmentation) and evaluates it on CIFAR-10/100, SVHN, and Mini-ImageNet. The loss equations (1)-(4), (7), and (8) define the training objective; they do not encode the final test-error comparisons. The RA and RH regularizers are taken from Tanaka et al. [24] by external citation, with lambda_A=0.8 and lambda_H=0.4 fixed, and the paper reports an ablation (Table II) rather than claiming these values are derived. Hyperparameters are chosen on a validation split, and the validation samples are added back to training for the final comparisons, which is ordinary model selection rather than circularity. The only self-citation, Arazo et al. [9], appears in the related-work discussion of label noise and is not load-bearing for the central claim. The skeptic's concern that unmarked baseline numbers in Tables IV and V may not share the authors' augmentation and dropout protocol is a question of experimental comparability and confound control, not circularity: the headline claim is an empirical benchmark claim, and the paper does not construct those baseline numbers from its own fitted parameters or equations. Accordingly, the derivation chain is self-contained with respect to the circularity criteria, and the correct score is 0.

Assumptions & free parameters 6 free parameters · 4 assumptions · 0 invented entities

The method is empirical and does not introduce new theoretical entities. Its load-bearing choices are hyperparameters and domain assumptions about class balance and the SSL evaluation protocol.

free parameters (6)
  • alpha (mixup beta distribution parameter) = 1 (default; studied 0.1, 1, 4, 8)
    Controls mixup strength; chosen as default with marginal differences in the ablation (Sec IV-C).
  • lambda_A (all-class regularization weight) = 0.8
    Taken from Tanaka et al. [24]; ablation in Sec IV-C shows near-optimal.
  • lambda_H (entropy regularization weight) = 0.4
    Taken from Tanaka et al. [24]; ablation in Sec IV-C shows near-optimal.
  • k (minimum number of labeled samples per mini-batch) = 16
    Selected from {8, 16, 32, 64} on the validation set (Table I); strongly affects results, especially for 500 labels.
  • dropout probability p = 0.1 (for 13-CNN/WR-28), 0.3 studied
    Chosen on validation set per architecture (Table III).
  • warm-up epochs = 10 (CIFAR, Mini-ImageNet), 150 (SVHN)
    Warm-up phase for initial pseudo-labels; the SVHN value is unusually long.
assumptions (4)
  • domain assumption The training and evaluation are restricted to class-balanced scenarios.
    Stated in Sec I: 'we focus the study on class-balanced scenarios.' The method may not extend to class-imbalanced SSL.
  • domain assumption The validation set (5K samples) is representative of the test distribution and is used for hyperparameter selection.
    Sec IV-A: hyperparameters are studied on the validation set in IV-B/IV-D and then the validation set is added back for final training.
  • domain assumption Soft pseudo-labels from the network's softmax outputs are suitable targets for unlabeled samples.
    Sec III: the approach stores softmax predictions as soft pseudo-labels; this assumes the network's confidence is informative.
  • domain assumption Mixup's label smoothing reduces overconfidence and thus confirmation bias.
    Sec III-A: relies on the mixup calibration effect [26] to justify the regularization.

how reviews work

0 comments
Cite this review

Pith. "Pith review of Pseudo-Labeling and Confirmation Bias in Deep Semi-Supervised Learning." pith.science (2026). https://pith.science/paper/FXU4USJG

@misc{pith2026190802983,
  author       = {Pith},
  title        = {Pith review of: Pseudo-Labeling and Confirmation Bias in Deep Semi-Supervised Learning},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/FXU4USJG}},
  note         = {Machine review of arXiv:1908.02983}
}
read the original abstract

Semi-supervised learning, i.e. jointly learning from labeled and unlabeled samples, is an active research topic due to its key role on relaxing human supervision. In the context of image classification, recent advances to learn from unlabeled samples are mainly focused on consistency regularization methods that encourage invariant predictions for different perturbations of unlabeled samples. We, conversely, propose to learn from unlabeled data by generating soft pseudo-labels using the network predictions. We show that a naive pseudo-labeling overfits to incorrect pseudo-labels due to the so-called confirmation bias and demonstrate that mixup augmentation and setting a minimum number of labeled samples per mini-batch are effective regularization techniques for reducing it. The proposed approach achieves state-of-the-art results in CIFAR-10/100, SVHN, and Mini-ImageNet despite being much simpler than other methods. These results demonstrate that pseudo-labeling alone can outperform consistency regularization methods, while the opposite was supposed in previous work. Source code is available at https://git.io/fjQsC.

Figures

Figures reproduced from arXiv: 1908.02983 by the authors.

Figure 1
Figure 1. Pseudo-labeling in the “two moons” data (4 labels/class) for 1000 samples. From left to right: no mixup, mixup, and mixup with a minimum number [PITH_FULL_IMAGE:figures/full_fig_p004_1.png] view at source ↗
Figure 2
Figure 2. Example of certainty of incorrect predictions [PITH_FULL_IMAGE:figures/full_fig_p005_2.png] view at source ↗

Discussion (0). Continue with ORCID to comment.

Forward citations

Cited by 3 Pith papers

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

  1. Heuristic-Free Multi-Teacher Learning

    cs.LG 2024-11 conditional novelty 6.0 of 10

    By turning each teacher's labels into a separate confidence-prediction task conditioned on a teacher token, Teacher2Task claims to outperform the best individual teacher on open-vocabulary tagging without manual aggre...

  2. Denoising Mutual Knowledge Distillation in Bi-Directional Multiple Instance Learning

    cs.CV 2025-05 conditional novelty 4.0 of 10

    A two-branch MIL training method with self-confidence losses and attention calibration improves bag- and instance-level whole-slide image classification on CAMELYON16 and TCGA-NSCLC over WENO and MHIM-MIL baselines.

  3. Optimising Language Models for Downstream Tasks: A Post-Training Perspective

    cs.CL 2025-06 conditional novelty 3.0 of 10

    A dissertation that repackages the author's previously published papers on continued pre-training, prompt tuning, and instruction modelling into a single narrative.

Reference graph

Works this paper leans on

48 extracted references · 43 canonical work pages · cited by 3 Pith papers

  1. [1]

    Focal Loss for Dense Object Detection,

    T. Lin, P. Goyal, R. Girshick, K. He, and P. Dollar, “Focal Loss for Dense Object Detection,” in IEEE International Conference on Computer Vision (ICCV), 2017

    work page 2017

  2. [2]

    Learning towards Minimum Hyperspherical Energy,

    W. Liu, R. Lin, Z. Liu, L. Liu, Z. Yu, B. Dai, and L. Song, “Learning towards Minimum Hyperspherical Energy,” in Advances in Neural Information Processing Systems (NeurIPS) , 2018

    work page 2018

  3. [3]

    Multi-object Tracking with Neural Gating Using Bilinear LSTM,

    C. Kim, F. Li, and J. Rehg, “Multi-object Tracking with Neural Gating Using Bilinear LSTM,” in European Conference on Computer Vision (ECCV), 2018

    work page 2018

  4. [4]

    Rethinking Spatiotempo- ral Feature Learning: Speed-Accuracy Trade-offs in Video Classification,

    S. Xie, C. Sun, J. Huang, Z. Tu, and K. Murphy, “Rethinking Spatiotempo- ral Feature Learning: Speed-Accuracy Trade-offs in Video Classification,” in European Conference on Computer Vision (ECCV) , September 2018

    work page 2018

  5. [5]

    WebVision Database: Visual Learning and Understanding from Web Data,

    W. Li, L. Wang, W. Li, E. Agustsson, and L. Van Gool, “WebVision Database: Visual Learning and Understanding from Web Data,” arXiv: 1708.02862, 2017

    arXiv 2017

  6. [6]

    Realistic Evaluation of Deep Semi-Supervised Learning Algorithms,

    A. Oliver, A. Odena, C. Raffel, E. Cubuk, and I. Goodfellow, “Realistic Evaluation of Deep Semi-Supervised Learning Algorithms,” in Advances in Neural Information Processing Systems (NeurIPS) , 2018

    work page 2018

  7. [7]

    Exploiting Unlabeled Data in CNNs by Self-supervised Learning to Rank,

    X. Liu, J. Van De Weijer, and A. D. Bagdanov, “Exploiting Unlabeled Data in CNNs by Self-supervised Learning to Rank,” IEEE Transactions on Pattern Analysis and Machine Intelligence , 2019

    work page 2019

  8. [8]

    Deep visual domain adaptation: A survey,

    M. Wang and W. Deng, “Deep visual domain adaptation: A survey,” Neurocomputing, vol. 312, pp. 135–153, 2018

    2018

Show all 48 references
  1. [9]

    Un- supervised Label Noise Modeling and Loss Correction,

    E. Arazo, D. Ortego, P. Albert, N. O’Connor, and K. McGuinness, “Un- supervised Label Noise Modeling and Loss Correction,” in International Conference on Machine Learning (ICML) , 2019

    2019

  2. [10]

    Unsupervised Representation Learning by Predicting Image Rotations,

    S. Gidaris, P. Singh, and N. Komodakis, “Unsupervised Representation Learning by Predicting Image Rotations,” in International Conference on Learning Representations (ICLR) , 2018

    2018

  3. [11]

    Decoupled Certainty-Driven Consistency Loss for Semi-supervised Learning,

    Y . Li, L. Liu, and R. Tan, “Decoupled Certainty-Driven Consistency Loss for Semi-supervised Learning,” arXiv: 1901.05657, 2019

    1901 arXiv

  4. [12]

    Enhanced semi-supervised learning for multimodal emotion recognition,

    Z. Zhang, F. Ringeval, B. Dong, E. Coutinho, E. Marchi, and B. Sch ¨uller, “Enhanced semi-supervised learning for multimodal emotion recognition,” in IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP), 2016

    2016

  5. [13]

    Self-labeling techniques for semi-supervised time series classification: an empirical study,

    M. Gonz ´alez, C. Bergmeir, I. Triguero, Y . Rodr ´ıguez, and J. Ben ´ıtez, “Self-labeling techniques for semi-supervised time series classification: an empirical study,” Knowledge and Information Systems , vol. 55, no. 2, pp. 493–528, 2018

    2018

  6. [14]

    Adversarial Training Methods for Semi-Supervised Text Classification,

    T. Miyato, A. Dai, and I. Goodfellow, “Adversarial Training Methods for Semi-Supervised Text Classification,” arXiv: 1605.07725, 2016

    2016 arXiv

  7. [15]

    Regularization With Stochastic Transformations and Perturbations for Deep Semi-Supervised Learning,

    M. Sajjadi, M. Javanmardi, and T. Tasdizen, “Regularization With Stochastic Transformations and Perturbations for Deep Semi-Supervised Learning,” in Advances in Neural Information Processing Systems (NeurIPS), 2016

    2016

  8. [16]

    Pseudo-label: The simple and efficient semi-supervised learning method for deep neural networks,

    D. Lee, “Pseudo-label: The simple and efficient semi-supervised learning method for deep neural networks,” in International Conference on Machine Learning Workshops (ICMLW) , 2013

    2013

  9. [17]

    Label Propagation for Deep Semi-supervised Learning,

    A. Iscen, G. Tolias, Y . Avrithis, and O. Chum, “Label Propagation for Deep Semi-supervised Learning,” in IEEE Conference on Computer Vision and Pattern Recognition (CVPR) , 2019

    2019

  10. [18]

    Mean teachers are better role models: Weight-averaged consistency targets improve semi-supervised deep learning results,

    A. Tarvainen and H. Valpola, “Mean teachers are better role models: Weight-averaged consistency targets improve semi-supervised deep learning results,” in Advances in Neural Information Processing Systems (NeurIPS), 2017

    2017

  11. [19]

    Transductive Semi-Supervised Deep Learning using Min-Max Features,

    W. Shi, Y . Gong, C. Ding, Z. Ma, T. Xiaoyu, and N. Zheng, “Transductive Semi-Supervised Deep Learning using Min-Max Features,” in European Conference on Computer Vision (ECCV) , 2018

    2018

  12. [20]

    Virtual Adversarial Training: A Regularization Method for Supervised and Semi-Supervised Learning,

    T. Miyato, S. Maeda, S. Ishii, and M. Koyama, “Virtual Adversarial Training: A Regularization Method for Supervised and Semi-Supervised Learning,” IEEE Transactions on Pattern Analysis and Machine Intelli- gence, 2018

    2018

  13. [21]

    Deep Co-Training for Semi-Supervised Image Recognition,

    S. Qiao, W. Shen, Z. Zhang, B. Wang, and A. Yuille, “Deep Co-Training for Semi-Supervised Image Recognition,” in European Conference on Computer Vision (ECCV) , 2018

    2018

  14. [22]

    Smooth Neighbors on Teacher Graphs for Semi-Supervised Learning,

    Y . Luo, J. Zhu, M. Li, Y . Ren, and B. Zhang, “Smooth Neighbors on Teacher Graphs for Semi-Supervised Learning,” in IEEE Conference on Computer Vision and Pattern Recognition (CVPR) , 2018

    2018

  15. [23]

    MixMatch: A Holistic Approach to Semi-Supervised Learn- ing,

    D. Berthelot, N. Carlini, I. Goodfellow, N. Papernot, A. Oliver, and C. Raffel, “MixMatch: A Holistic Approach to Semi-Supervised Learn- ing,” in Advances in Neural Information Processing Systems (NeurIPS) , 2019

    2019

  16. [24]

    Joint Optimization Framework for Learning with Noisy Labels,

    D. Tanaka, D. Ikami, T. Yamasaki, and K. Aizawa, “Joint Optimization Framework for Learning with Noisy Labels,” in IEEE Conference on Computer Vision and Pattern Recognition (CVPR) , 2018

    2018

  17. [25]

    mixup: Beyond Empirical Risk Minimization,

    H. Zhang, M. Cisse, Y . Dauphin, and D. Lopez-Paz, “mixup: Beyond Empirical Risk Minimization,” in International Conference on Learning Representations (ICLR), 2018

    2018

  18. [26]

    On Mixup Training: Improved Calibration and Predictive Uncertainty for Deep Neural Networks,

    S. Thulasidasan, G. Chennupati, J. Bilmes, T. Bhattacharya, and S. Michalak, “On Mixup Training: Improved Calibration and Predictive Uncertainty for Deep Neural Networks,” arXiv: 1905.11001, 2019

    1905 arXiv

  19. [27]

    Dropout: A Simple Way to Prevent Neural Networks from Overfitting,

    N. Srivastava, G. Hinton, A. Krizhevsky, I. Sutskever, and R. Salakhut- dinov, “Dropout: A Simple Way to Prevent Neural Networks from Overfitting,” Journal of Machine Learning Research , vol. 15, pp. 1929– 1958, 2014

    1929

  20. [28]

    Interpola- tion Consistency Training for Semi-Supervised Learning,

    V . Verma, A. Lamb, J. Kannala, Y . Bengio, and D. Lopez-Paz, “Interpola- tion Consistency Training for Semi-Supervised Learning,” inInternational Joint Conference on Artificial Intelligence (IJCAI) , 2019

    2019

  21. [29]

    There Are Many Consistent Explanations of Unlabeled Data: Why You Should Average,

    B. Athiwaratkun, M. Finzi, P. Izmailov, and A. Wilson, “There Are Many Consistent Explanations of Unlabeled Data: Why You Should Average,” in International Conference on Learning Representations (ICLR) , 2019

    2019

  22. [30]

    Temporal Ensembling for Semi-Supervised Learning,

    S. Laine and T. Aila, “Temporal Ensembling for Semi-Supervised Learning,” in International Conference on Learning Representations (ICLR), 2017

    2017

  23. [31]

    Semi-Supervised Deep Learning with Memory,

    Y . Chen, X. Zhu, and S. Gong, “Semi-Supervised Deep Learning with Memory,” in European Conference on Computer Vision (ECCV) , 2018

    2018

  24. [32]

    Semi-supervised Learning by Entropy Minimization,

    Y . Grandvalet and Y . Bengio, “Semi-supervised Learning by Entropy Minimization,” in International Conference on Neural Information Processing Systems (NIPS) , 2004

    2004

  25. [33]

    Learning multiple layers of features from tiny images,

    A. Krizhevsky and G. Hinton, “Learning multiple layers of features from tiny images,” University of Toronto, Tech. Rep., 2009

    2009

  26. [34]

    Reading digits in natural images with unsupervised feature learning,

    Y . Netzer, T. Wang, A. Coates, A. Bissacco, . Wu, B, and A. Ng, “Reading digits in natural images with unsupervised feature learning,” in Advances in Neural Information Processing Systems (NeurIPS) , 2011

    2011

  27. [35]

    Matching Networks for One Shot Learning,

    O. Vinyals, C. Blundell, T. Lillicrap, K. Kavukcuoglu, and D. Wierstra, “Matching Networks for One Shot Learning,” in Advances in Neural Information Processing Systems (NeurIPS) , 2016

    2016

  28. [36]

    ImageNet: A large-scale hierarchical image database,

    J. Deng, W. Dong, R. Socher, L. Li, K. Li, and L. Fei-Fei, “ImageNet: A large-scale hierarchical image database,” in IEEE Conference on Computer Vision and Pattern Recognition (CVPR) , 2009

    2009

  29. [37]

    Optimization as a model for few-shot learning,

    S. Ravi and H. Larochelle, “Optimization as a model for few-shot learning,” in International Conference on Learning Representations (ICLR), 2017

    2017

  30. [38]

    Deep Residual Learning for Image Recognition,

    K. He, X. Zhang, S. Ren, and J. Sun, “Deep Residual Learning for Image Recognition,” in IEEE Conference on Computer Vision and Pattern Recognition (CVPR), 2016

    2016

  31. [39]

    Surprising Effectiveness of Few- Image Unsupervised Feature Learning,

    Y . Asano, C. Rupprecht, and A. Vedaldi, “Surprising Effectiveness of Few- Image Unsupervised Feature Learning,” in IEEE International Conference on Computer Vision (ICCV) , 2019

    2019

  32. [40]

    Weight Normalization: A Simple Repa- rameterization to Accelerate Training of Deep Neural Networks,

    T. Salimans and D. Kingma, “Weight Normalization: A Simple Repa- rameterization to Accelerate Training of Deep Neural Networks,” in Advances in Neural Information Processing Systems (NeurIPS) , 2016

    2016

  33. [41]

    DisturbLabel: Regularizing CNN on the Loss Layer,

    L. Xie, J. Wang, Z. Wei, M. Wang, and Q. Tian, “DisturbLabel: Regularizing CNN on the Loss Layer,” in IEEE Conference on Computer Vision and Pattern Recognition (CVPR) , 2016

    2016

  34. [42]

    Revisiting Self-Supervised Visual Representation Learning,

    A. Kolesnikov, X. Zhai, and L. Beyer, “Revisiting Self-Supervised Visual Representation Learning,” in IEEE Conference on Computer Vision and Pattern Recognition (CVPR), 2019

    2019

  35. [43]

    ImageNet Classification with Deep Convolu- tional Neural Networks,

    I. S. G. H. A. Krizhevsky, “ImageNet Classification with Deep Convolu- tional Neural Networks,” in Advances in Neural Information Processing Systems (NeurIPS), 2012

    2012

  36. [44]

    Deep Image Prior,

    D. Ulyanov, A. Vedaldi, and V . Lempitsky, “Deep Image Prior,” in IEEE Conference on Computer Vision and Pattern Recognition (CVPR) , 2018

    2018

  37. [45]

    Wide Residual Networks,

    N. K. S. Zagoruyko, “Wide Residual Networks,” in British Machine Vision Conference (BMVC), 2016

    2016

  38. [46]

    Identity Mappings in Deep Residual Networks,

    K. He, X. Zhang, S. Ren, and J. Sun, “Identity Mappings in Deep Residual Networks,” in European Conference on Computer Vision (ECCV) , 2016

    2016

  39. [47]

    Population Based Augmentation: Efficient Learning of Augmentation Policy Schedules,

    D. Ho, E. Liang, X. Chen, I. Stoica, and P. Abbeel, “Population Based Augmentation: Efficient Learning of Augmentation Policy Schedules,” in International Conference on Machine Learning (ICML) , 2019

    2019

  40. [48]

    Semi- Supervised Learning with Scarce Annotations,

    S.-A. Rebuffi, S. Ehrhardt, K. Han, A. Vedaldi, and A. Zisserman, “Semi- Supervised Learning with Scarce Annotations,” in IEEE International Conference on Computer Vision (ICCV) , 2019

    2019

Pith tools

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