Pith. sign in

REVIEW 3 major objections 6 minor 1 cited by

Seeing the Whole in the Parts in Self-Supervised Representation Learning

T0 review · 3 major / 6 minor · reviewed 2026-08-10 · deepseek-v4-flash

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

Pith's one-line read A local-global alignment objective improves self-supervised category learning and robustness.

desk verdict A solid, honest empirical SSL paper on local-global alignment; the headline ImageNet gain is plausible but the 1.4-point gap over their own BYOL baseline is not yet pinned to the CO loss (single runs, withheld configs, selection asymmetry), so treat the headline numbers as provisional. read the letter →

arxiv 2501.02860 v1 pith:PQVTUL6P submitted 2025-01-06 cs.LG cs.CV

Arthur Aubret , Céline Teulière , Jochen Triesch This is my paper · ORCID
classification cs.LGcs.CV
keywords self-supervisedlearningrepresentationspatialco-occurrencelocal-globalalignmentreceptivefieldrobustnesslinearprobingImageNet-1K
verification ladder T0 review T1 audit T2 compute T3 formal

The pith

A machine-rendered reading of the paper's core claim, the machinery that carries it, and where it could break.

The reading

The paper tries to establish that modeling spatial co-occurrences, by aligning the local representations of an image with its global representation, improves self-supervised visual learning. If true, this gives a simple, general principle: make co-occurring visual features similar, and category recognition becomes less sensitive to which features are present. The paper implements this as CO-SSL, instantiates CO-BYOL and CO-DINO, and reports that CO-BYOL reaches 71.5% top-1 linear-probe accuracy on ImageNet-1K after 100 pretraining epochs, 1.4 points above their BYOL reproduction. It also reports improved robustness to ImageNet-C noise, internal masking, and small PGD attacks, explained by the high redundancy of the learned local representations.

What carries the argument

The load-bearing object is the local-global alignment loss: the global representation of an image is used as a target for the local representations of the same image taken before pooling, with the SSL loss averaged over local positions. RF-ResNet is the complementary architectural mechanism: it bounds the receptive field of local representations (for example, 99 by 99 pixels on 224 by 224 images) so that aligning them with the global view requires learning co-occurrence statistics rather than merging features that already cover the image. The loss weight $w_s$ trades off the standard global loss against the local-global loss.

What would settle it

A controlled run that adds CO-BYOL's local projection heads and loss weight to BYOL but aligns each local representation with the global representation of a randomly sampled different image should show no gain over BYOL; if it still shows the ImageNet-1K gain, then the improvement is not caused by learning co-occurrence statistics within an image.

Watch

Extended reading notes

Core claim

The central claim is that aligning local representations of an image with its global representation teaches a model the spatial co-occurrence statistics of visual features, and that this improves category recognition and robustness. The authors introduce CO-SSL, which adds a self-supervised loss between each local representation taken right before average pooling and the global embedding of a differently augmented view. They also introduce RF-ResNet, a convolutional architecture that bounds the receptive field of local representations so that alignment requires learning true co-occurrence statistics. With 100 epochs of pretraining, CO-BYOL on the RF99-ResNet50 reaches 71.5% top-1 on ImageNet-1K; even with a standard ResNet50 it reaches 71.4%, and an averaged bag of patch representations alone reaches 71.2%. Robustness experiments show CO-BYOL with a ResNet50 consistently surpassing BYOL under ImageNet-C noise, internal masking, and PGD attacks, and the paper attributes this to the high similarity among local representations within an image.

Load-bearing premise

The results depend on attributing the accuracy and robustness gains to the local-global alignment objective itself, rather than to the changed receptive-field size, the extra projection heads, or the hyperparameters that were tuned before the main comparison.

Editorial extensions

If this is right

  • If the central claim holds, any instance-discrimination SSL method can be upgraded with a local-global alignment term, and the CO-MoCoV3 and CO-DINO results suggest the gain is not specific to BYOL.
  • Category recognition improves with modest receptive fields: the paper finds RF sizes between 67 and 163 pixels work best on 224 by 224 images, so bounded receptive fields are part of the recipe.
  • Robustness to masking follows directly from redundancy: because many local representations are pulled toward the same global representation, corrupting or removing some of them leaves the global feature mostly intact.
  • The approach is sample-efficient compared with multicrop methods: CO-BYOL processes two images per iteration yet matches or beats multicrop baselines that consume more images.
  • A simple averaged bag of small patch representations, with no global integration beyond pooling, already produces strong category accuracy, suggesting that co-occurrence statistics can be learned locally.

Reading between the lines

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

  • Our inference: if the co-occurrence principle is general, the same local-global alignment could be applied to masked-image modeling and vision transformers, potentially improving their robustness at small patch sizes.
  • Our inference: the redundancy explanation suggests a testable prediction that adversarial robustness should scale with the number of local representations aligned per image, up to the point where receptive fields become too large.
  • Our inference: the biological motivation points to a concrete experiment, measuring whether representations trained with local-global alignment show the same tolerance to partial occlusion as human object recognition.
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 / 6 minor

Summary. The paper proposes CO-SSL, a family of self-supervised representation learning methods that add a local-to-global alignment loss to existing SSL objectives such as BYOL, DINO, and MoCoV3. The local representations are taken before the final pooling layer and are aligned with the global image representation through a separate projection head. The paper also introduces RF-ResNet, a ResNet variant with controlled receptive field sizes, and reports that CO-BYOL with RF99-ResNet50 reaches 71.5% top-1 linear-probe accuracy on ImageNet-1K after 100 pre-training epochs, compared with 70.1% for the authors' BYOL reproduction. Additional experiments address robustness to ImageNet-C corruptions, internal masking, and small PGD attacks, together with ablations on ImageNet-100 and analyses of effective receptive fields.

Significance. If the central attribution holds, the contribution is significant: a simple and general local-global alignment term improves both accuracy and robustness of convolutional SSL, and the RF-ResNet construction provides a useful tool for studying receptive field effects. The paper has clear strengths: it tests three SSL bases, includes ablation studies, analyzes receptive fields explicitly, and offers a mechanistic explanation in terms of redundant local representations. However, the evidence is currently single-run, partly confounded by hyperparameter selection, and the robustness claims are contradicted for the default RF99 architecture in several tables. The significance is therefore conditional on resolving these empirical issues.

major comments (3)
  1. [Section 4.2, Tables 3 and 4, Abstract] The abstract and introduction state that CO-SSL is more robust to noise corruption, internal corruption, and small adversarial attacks, but this is not supported for the default RF99-ResNet50 model. In Table 3, CO-BYOL (RF99-R50) obtains 19.6/18.9/8.0 on Gaussian, Shot, and Impulse noise versus 28.5/26.7/19.2 for BYOL, and is also worse on Defocus, Glass, Motion, and Zoom blur; it is better only on the Mask corruption. In Table 4, CO-BYOL (RF99-R50) is substantially worse than BYOL under several PGD settings, e.g. 18.6 versus 51.7 at epsilon=0.003 with gamma=epsilon/10 and 3.8 versus 27.4 at epsilon=0.01. Similarly, CO-DINO in Table 12 is worse than DINO on internal masking (37.6 versus 30.4). The robustness claims should be restricted to the specific variants and corruption types for which they hold, or the paper should provide an explanation for the RF99-R50 failures; as written, the headline robustness claim is overclaimed.
  2. [Section 3.3, Tables 1 and 6] The attribution of the ImageNet-1K gains to the co-occurrence loss is not fully established. Section 3.3 states that the RF size was hyperparameterized on ImageNet-100, that ws was selected in {0.2, 0.5}, that cmin was set to 0.2 for CO-SSL, that a two-hidden-layer projection head was adopted because it 'works better' for BYOL-based methods, and that the initial learning rate was fine-tuned per method; the precise configuration files are withheld until acceptance. The baseline BYOL was not re-tuned on ImageNet-100 in the same way, and the headline 71.5 versus 70.1 comparison includes architecture changes (RF99-R50 versus R50) as well as the extra local heads. The CO-BYOL (R50) result of 71.4 partially addresses the architecture confound, but the selection-asymmetry confound remains. The paper should release the exact configurations, run a baseline tuned under the same protocol, and report multiple seeds with error bars so the 1.3-1.4 point differences can be assessed.
  3. [Tables 1, 2 and 7; general protocol] All headline results are single runs without error bars or seed information, and several comparisons mix published numbers with the authors' reproductions under different epochs and training protocols. For example, Table 1 compares 'BYOL (pub)' at 69.3, 'BYOL+' at 70.1, and 'BYOL (ours)' at 70.1, while Table 7 lists 'BYOL (ours)' at 72.4 and 'BYOL (Grill et al., 2020)' at 72.5; the differences among these baselines are not explained beyond the '+ denotes improved reproduction' note. Given that the central claim rests on margins of about one point, the paper should report at least three seeds for the main comparisons and specify the exact protocol for each reproduced baseline. This is a reproducibility and statistical-support issue, not an allegation of fabrication.
minor comments (6)
  1. [Section 3.3] The sentence 'For ImageNet-1K, Tiny-ImageNet, ImageNet-100, we apply hyperparameters provided in the original papers for ImageNet-1K' is ambiguous because it lists ImageNet-100 twice and does not clarify which original paper provides the hyperparameters for each dataset.
  2. [Table 5] The table footnote says 'CO-BYOL uses a RF99-ResNet5°'; the degree sign should be a zero, i.e. RF99-ResNet50.
  3. [Table 6] The column header 'RF99-50 (1 head)' should read 'RF99-R50 (1 head)' for consistency with the rest of the paper.
  4. [Appendix A.4] The description of memory calculation is internally inconsistent: it says memory is 'the number of parameters multiplied by the number of input images and four' but then states 'a float32 is encoded with eight memory bytes.' A float32 is four bytes, so the accounting should be corrected and clarified.
  5. [Section 4.3 and Figure 3A] The abbreviation 'C/R' is used without definition; it should be defined as crop/resize at first use. Also, there is a missing space in 'cmin' in the Figure 3A caption.
  6. [Appendix A.4 and Table 11] The claim that CO-BYOL is 'three times more sample-efficient' than BYOL-mc is misleading: BYOL-mc uses six augmented views but those views come from the same two original images, whereas CO-BYOL uses two original images. Sample efficiency is typically measured per unique training image, not per augmented view, and the table does not report unique-image counts.

Circularity Check

0 steps flagged · score 0.0 of 10

No significant circularity: CO-SSL is an empirical method paper whose claims are supported by controlled ablations and external benchmarks, not by construction or self-citation.

full rationale

The paper is an empirical method paper, and no load-bearing step in its derivation chain reduces by construction to its own inputs. The CO-BYOL loss is explicitly defined as the standard BYOL global loss plus a weighted local-global loss, and Table 6 ablates the local loss weight at ws=0, so the reported gains are not a fitted parameter renamed as a prediction. The RF-ResNet ablations (RF99-R50v0 versus RF99-R50, and evaluation before versus after the MLP) provide controlled architecture comparisons. Selecting hyperparameters on ImageNet-100 is ordinary tuning, not circularity, and the authors state that precise configuration files will be released with the code. The only self-citation, Aubret et al. 2023, appears in a conclusion sentence about temporal co-occurrences and is not load-bearing. Robustness is measured independently against ImageNet-C corruptions and PGD attacks; the redundancy explanation is an interpretive analysis of observed representations, not a derivation equivalent to the objective. Accordingly, the central claim retains independent empirical content.

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

The central claim rests on standard SSL assumptions plus several tuned hyperparameters. The most important free parameter is the receptive-field size of local representations, which was selected on ImageNet-100 before the ImageNet-1K runs. No new physical or metaphysical entities are introduced; CO-SSL and RF-ResNet are a loss family and an architecture.

free parameters (5)
  • ws (weight of local-global loss Ll) = 0.2 or 0.5, best on ImageNet-100
    Defined in the L-CO-BYOL loss in Section 3.1; Table 5 sweeps ws over {0.1, 0.2, 0.5, 1} and the best values are selected on ImageNet-100.
  • cmin (minimum crop ratio) = 0.2 default
    Section 3.3 sets cmin=0.2; Figure 3A shows accuracy depends strongly on cmin, so this choice influences all headline results.
  • RF size of local representations = 99x99 for RF99-ResNet50 on 224x224 inputs
    RF sizes were hyperparameterized on ImageNet-100 in Section 4.3 and transferred to ImageNet-1K; this architectural choice is central to CO-SSL.
  • initial learning rate = one of {0.2, 0.4, 0.8, 1.6}
    Section 3.3 says the initial learning rate was finetuned in this range for all methods.
  • n2 (number of local representations used in Ll) = 196 by default
    Table 5 varies n2 from 0 to 196; the default uses the maximum, and accuracy changes by about 3.4 points across the range.
assumptions (5)
  • domain assumption BYOL-style EMA target, stop-gradient, and predictor head prevent representation collapse
    Section 3.1 inherits the standard BYOL machinery; without this assumption the local-global loss could be trivially minimized by collapse.
  • domain assumption Top-1 linear probe accuracy on frozen features is a valid measure of representation quality
    Section 3.3 evaluates all models with an online linear probe; this standard protocol assumes linear separability approximates downstream usefulness.
  • domain assumption Pre-pooling local representations with bounded theoretical receptive fields correspond to localized visual features
    Section 4.4 uses effective-receptive-field saliency maps; RF-ResNet bounds the theoretical RF, but semantic locality of the features is not guaranteed.
  • domain assumption Crop and resize augmentation produces useful spatial co-occurrences for category learning
    Sections 4.3 and A.3 interpret varying cmin as changing the amount of co-occurrence signal available to the model.
  • standard math The Araujo et al. receptive-field formula correctly predicts the RF sizes used to configure RF-ResNet
    Appendix B.1 applies this formula and validates it with saliency maps of a random network.

how reviews work

0 comments
Cite this review

Pith. "Pith review of Seeing the Whole in the Parts in Self-Supervised Representation Learning." pith.science (2026). https://pith.science/paper/PQVTUL6P

@misc{pith2026250102860,
  author       = {Pith},
  title        = {Pith review of: Seeing the Whole in the Parts in Self-Supervised Representation Learning},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/PQVTUL6P}},
  note         = {Machine review of arXiv:2501.02860}
}
read the original abstract

Recent successes in self-supervised learning (SSL) model spatial co-occurrences of visual features either by masking portions of an image or by aggressively cropping it. Here, we propose a new way to model spatial co-occurrences by aligning local representations (before pooling) with a global image representation. We present CO-SSL, a family of instance discrimination methods and show that it outperforms previous methods on several datasets, including ImageNet-1K where it achieves 71.5% of Top-1 accuracy with 100 pre-training epochs. CO-SSL is also more robust to noise corruption, internal corruption, small adversarial attacks, and large training crop sizes. Our analysis further indicates that CO-SSL learns highly redundant local representations, which offers an explanation for its robustness. Overall, our work suggests that aligning local and global representations may be a powerful principle of unsupervised category learning.

Figures

Figures reproduced from arXiv: 2501.02860 by the authors.

Figure 1
Figure 1. Architecture of CO-BYOL. We augment an image through two augmentations pipelines and forward one to the visual encoder fθ and one to the momentum encoder fξ. Both output a set of local representations. Then, we spatially average these local representations into two global representation, which are fed into MLPs computing projections and predictions for the standard loss function Lg of BYOL. In CO-BYOL, we also indiv… view at source ↗
Figure 2
Figure 2. RF-ResNet architecture. We omit residual connections for better readability. [PITH_FULL_IMAGE:figures/full_fig_p005_2.png] view at source ↗
Figure 3
Figure 3. A) Top-1 ImageNet-100 validation accuracy. We train different RF-ResNet50 and different [PITH_FULL_IMAGE:figures/full_fig_p008_3.png] view at source ↗
Figures from the paper (3 more)
Figure 4
Figure 4. Figure 4: Visualization of effective receptive fields of 4 local representations computed on two [PITH_FULL_IMAGE:figures/full_fig_p009_4.png]
Figure 5
Figure 5. Figure 5: Top-1 ImageNet-100 validation accuracy. This is the same data shown in Figure 3, but we [PITH_FULL_IMAGE:figures/full_fig_p017_5.png]
Figure 6
Figure 6. Figure 6: Effective receptive fields of local representations computed on ImageNet-1k validation set [PITH_FULL_IMAGE:figures/full_fig_p019_6.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. Self-Supervised Learning with a Multi-Task Latent Space Objective

    cs.CV 2026-02 conditional novelty 6.0 of 10

    Assigning a dedicated predictor to each view type stabilizes multi-crop Siamese SSL and, combined with asymmetric cutout views, yields consistent ImageNet gains over BYOL, SimSiam, and MoCo v3.

Reference graph

Works this paper leans on

77 extracted references · 49 canonical work pages · cited by 1 Pith paper

  1. [1]

    Self-supervised classification network

    Elad Amrani, Leonid Karlinsky, and Alex Bronstein. Self-supervised classification network. In European Conference on Computer Vision, pp.\ 116--132. Springer, 2022

    work page 2022

  2. [2]

    Unsupervised state representation learning in atari

    Ankesh Anand, Evan Racah, Sherjil Ozair, Yoshua Bengio, Marc-Alexandre C \^o t \'e , and R Devon Hjelm. Unsupervised state representation learning in atari. In Proceedings of the 33rd International Conference on Neural Information Processing Systems, pp.\ 8769--8782, 2019

    work page 2019

  3. [3]

    Computing receptive fields of convolutional neural networks

    Andr \'e Araujo, Wade Norris, and Jack Sim. Computing receptive fields of convolutional neural networks. Distill, 4 0 (11): 0 e21, 2019

    2019

  4. [4]

    Self-supervised learning from images with a joint-embedding predictive architecture

    Mahmoud Assran, Quentin Duval, Ishan Misra, Piotr Bojanowski, Pascal Vincent, Michael Rabbat, Yann LeCun, and Nicolas Ballas. Self-supervised learning from images with a joint-embedding predictive architecture. In Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, pp.\ 15619--15629, 2023

    work page 2023

  5. [5]

    Ernst, C \'e line Teuli \`e re, and Jochen Triesch

    Arthur Aubret, Markus R. Ernst, C \'e line Teuli \`e re, and Jochen Triesch. Time to augment self-supervised visual representation learning. In The Eleventh International Conference on Learning Representations, 2023. URL https://openreview.net/forum?id=o8xdgmwCP8l

    work page 2023

  6. [6]

    Does the brain's ventral visual pathway compute object shape? Trends in Cognitive Sciences, 26 0 (12): 0 1119--1132, 2022

    Vladislav Ayzenberg and Marlene Behrmann. Does the brain's ventral visual pathway compute object shape? Trends in Cognitive Sciences, 26 0 (12): 0 1119--1132, 2022

    work page 2022

  7. [7]

    Learning representations by maximizing mutual information across views

    Philip Bachman, R Devon Hjelm, and William Buchwalter. Learning representations by maximizing mutual information across views. Advances in neural information processing systems, 32, 2019

    2019

  8. [8]

    Deep learning models fail to capture the configural nature of human shape perception

    Nicholas Baker and James H Elder. Deep learning models fail to capture the configural nature of human shape perception. Iscience, 25 0 (9), 2022

    work page 2022

Show all 77 references
  1. [9]

    Deep convolutional networks do not classify based on global object shape

    Nicholas Baker, Hongjing Lu, Gennady Erlikhman, and Philip J Kellman. Deep convolutional networks do not classify based on global object shape. PLoS computational biology, 14 0 (12): 0 e1006613, 2018

    2018

  2. [10]

    Beit: Bert pre-training of image transformers

    Hangbo Bao, Li Dong, Songhao Piao, and Furu Wei. Beit: Bert pre-training of image transformers. In International Conference on Learning Representations, 2021

    2021

  3. [11]

    Vicreg: Variance-invariance-covariance regularization for self-supervised learning

    Adrien Bardes, Jean Ponce, and Yann Lecun. Vicreg: Variance-invariance-covariance regularization for self-supervised learning. In ICLR 2022-10th International Conference on Learning Representations, 2022 a

    2022

  4. [12]

    Vicregl: Self-supervised learning of local visual features

    Adrien Bardes, Jean Ponce, and Yann LeCun. Vicregl: Self-supervised learning of local visual features. Advances in Neural Information Processing Systems, 35: 0 8799--8810, 2022 b

    2022

  5. [13]

    Towards democratizing joint-embedding self-supervised learning

    Florian Bordes, Randall Balestriero, and Pascal Vincent. Towards democratizing joint-embedding self-supervised learning. arXiv preprint arXiv:2303.01986, 2023

    2023 arXiv

  6. [14]

    Approximating cnns with bag-of-local-features models works surprisingly well on imagenet

    Wieland Brendel and Matthias Bethge. Approximating cnns with bag-of-local-features models works surprisingly well on imagenet. In International Conference on Learning Representations, 2018

    2018

  7. [15]

    Unsupervised learning of visual features by contrasting cluster assignments

    Mathilde Caron, Ishan Misra, Julien Mairal, Priya Goyal, Piotr Bojanowski, and Armand Joulin. Unsupervised learning of visual features by contrasting cluster assignments. Advances in neural information processing systems, 33: 0 9912--9924, 2020

    2020

  8. [16]

    Emerging properties in self-supervised vision transformers

    Mathilde Caron, Hugo Touvron, Ishan Misra, Herv \'e J \'e gou, Julien Mairal, Piotr Bojanowski, and Armand Joulin. Emerging properties in self-supervised vision transformers. In Proceedings of the IEEE/CVF international conference on computer vision, pp.\ 9650--9660, 2021

    2021

  9. [17]

    A simple framework for contrastive learning of visual representations

    Ting Chen, Simon Kornblith, Mohammad Norouzi, and Geoffrey Hinton. A simple framework for contrastive learning of visual representations. In International conference on machine learning, pp.\ 1597--1607. PMLR, 2020 a

    2020

  10. [18]

    Exploring simple siamese representation learning

    Xinlei Chen and Kaiming He. Exploring simple siamese representation learning. In Proceedings of the IEEE/CVF conference on computer vision and pattern recognition, pp.\ 15750--15758, 2021

    2021

  11. [19]

    Improved baselines with momentum contrastive learning

    Xinlei Chen, Haoqi Fan, Ross Girshick, and Kaiming He. Improved baselines with momentum contrastive learning. arXiv preprint arXiv:2003.04297, 2020 b

    2003 arXiv

  12. [20]

    An empirical study of training self-supervised vision transformers

    Xinlei Chen, Saining Xie, and Kaiming He. An empirical study of training self-supervised vision transformers. In Proceedings of the IEEE/CVF international conference on computer vision, pp.\ 9640--9649, 2021

    2021

  13. [21]

    Intra-instance vicreg: Bag of self-supervised image patch embedding explains the performance

    Yubei Chen, Adrien Bardes, ZENGYI LI, and Yann LeCun. Intra-instance vicreg: Bag of self-supervised image patch embedding explains the performance. 2022

    2022

  14. [22]

    solo-learn: A library of self-supervised methods for visual representation learning

    Victor Guilherme Turrisi Da Costa, Enrico Fini, Moin Nabi, Nicu Sebe, and Elisa Ricci. solo-learn: A library of self-supervised methods for visual representation learning. Journal of Machine Learning Research, 23 0 (56): 0 1--6, 2022

    2022

  15. [23]

    Scaling up your kernels to 31x31: Revisiting large kernel design in cnns

    Xiaohan Ding, Xiangyu Zhang, Jungong Han, and Guiguang Ding. Scaling up your kernels to 31x31: Revisiting large kernel design in cnns. In Proceedings of the IEEE/CVF conference on computer vision and pattern recognition, pp.\ 11963--11975, 2022

    2022

  16. [24]

    With a little help from my friends: Nearest-neighbor contrastive learning of visual representations

    Debidatta Dwibedi, Yusuf Aytar, Jonathan Tompson, Pierre Sermanet, and Andrew Zisserman. With a little help from my friends: Nearest-neighbor contrastive learning of visual representations. In Proceedings of the IEEE/CVF International Conference on Computer Vision, pp.\ 9588--...

    2021

  17. [25]

    All4one: Symbiotic neighbour contrastive learning via self-attention and redundancy reduction

    Imanol G Estepa, Ignacio Saras \'u a, Bhalaji Nagarajan, and Petia Radeva. All4one: Symbiotic neighbour contrastive learning via self-attention and redundancy reduction. In Proceedings of the IEEE/CVF International Conference on Computer Vision, pp.\ 16243--16253, 2023

    2023

  18. [26]

    Unsupervised statistical learning of higher-order spatial structures from visual scenes

    J \'o zsef Fiser and Richard N Aslin. Unsupervised statistical learning of higher-order spatial structures from visual scenes. Psychological science, 12 0 (6): 0 499--504, 2001

    2001

  19. [27]

    Statistical learning of higher-order temporal structure from visual shape sequences

    J \'o zsef Fiser and Richard N Aslin. Statistical learning of higher-order temporal structure from visual shape sequences. Journal of Experimental Psychology: Learning, Memory, and Cognition, 28 0 (3): 0 458, 2002

    2002

  20. [28]

    Encoding multielement scenes: statistical learning of visual feature hierarchies

    J \'o zsef Fiser and Richard N Aslin. Encoding multielement scenes: statistical learning of visual feature hierarchies. Journal of Experimental Psychology: General, 134 0 (4): 0 521, 2005

    2005

  21. [29]

    A review on discriminative self-supervised learning methods

    Nikolaos Giakoumoglou and Tania Stathaki. A review on discriminative self-supervised learning methods. arXiv preprint arXiv:2405.04969, 2024

    2024 arXiv

  22. [30]

    Learning representations by predicting bags of visual words

    Spyros Gidaris, Andrei Bursuc, Nikos Komodakis, Patrick P \'e rez, and Matthieu Cord. Learning representations by predicting bags of visual words. In Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, pp.\ 6928--6938, 2020

    2020

  23. [31]

    Bootstrap your own latent: A new approach to self-supervised learning

    Jean-Bastien Grill, Florian Strub, Florent Altch \'e , Corentin Tallec, Pierre Richemond, Elena Buchatskaya, Carl Doersch, Bernardo Pires, Zhaohan Guo, Mohammad Azar, et al. Bootstrap your own latent: A new approach to self-supervised learning. In Neural Information Processing...

    2020

  24. [32]

    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, pp.\ 770--778, 2016

    2016

  25. [33]

    Momentum contrast for unsupervised visual representation learning

    Kaiming He, Haoqi Fan, Yuxin Wu, Saining Xie, and Ross Girshick. Momentum contrast for unsupervised visual representation learning. In Proceedings of the IEEE/CVF conference on computer vision and pattern recognition, pp.\ 9729--9738, 2020

    2020

  26. [34]

    Masked autoencoders are scalable vision learners

    Kaiming He, Xinlei Chen, Saining Xie, Yanghao Li, Piotr Doll \'a r, and Ross Girshick. Masked autoencoders are scalable vision learners. In Proceedings of the IEEE/CVF conference on computer vision and pattern recognition, pp.\ 16000--16009, 2022

    2022

  27. [35]

    Benchmarking neural network robustness to common corruptions and perturbations

    Dan Hendrycks and Thomas Dietterich. Benchmarking neural network robustness to common corruptions and perturbations. In International Conference on Learning Representations, 2018

    2018

  28. [36]

    Data augmentation instead of explicit regularization

    Alex Hern \'a ndez-Garc \' a and Peter K \"o nig. Data augmentation instead of explicit regularization. arXiv preprint arXiv:1806.03852, 2018

    2018 arXiv

  29. [37]

    Learning deep representations by mutual information estimation and maximization

    R Devon Hjelm, Alex Fedorov, Samuel Lavoie-Marchildon, Karan Grewal, Phil Bachman, Adam Trischler, and Yoshua Bengio. Learning deep representations by mutual information estimation and maximization. In International Conference on Learning Representations, 2018

    2018

  30. [38]

    Adco: Adversarial contrast for efficient learning of unsupervised representations from self-trained negative adversaries

    Qianjiang Hu, Xiao Wang, Wei Hu, and Guo-Jun Qi. Adco: Adversarial contrast for efficient learning of unsupervised representations from self-trained negative adversaries. In Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, pp.\ 1074--1083, 2021

    2021

  31. [39]

    Texture-like representation of objects in human visual cortex

    Akshay V Jagadeesh and Justin L Gardner. Texture-like representation of objects in human visual cortex. Proceedings of the National Academy of Sciences, 119 0 (17): 0 e2115302119, 2022

    2022

  32. [40]

    Self-distilled self-supervised representation learning

    Jiho Jang, Seonhoon Kim, Kiyoon Yoo, Chaerin Kong, Jangho Kim, and Nojun Kwak. Self-distilled self-supervised representation learning. In Proceedings of the IEEE/CVF Winter Conference on Applications of Computer Vision, pp.\ 2829--2839, 2023

    2023

  33. [41]

    Shape-selective processing in deep networks: integrating the evidence on perceptual integration

    Christian Jarvers and Heiko Neumann. Shape-selective processing in deep networks: integrating the evidence on perceptual integration. Frontiers in Computer Science, 5: 0 1113609, 2023

    2023

  34. [42]

    Transformers in vision: A survey

    Salman Khan, Muzammal Naseer, Munawar Hayat, Syed Waqas Zamir, Fahad Shahbaz Khan, and Mubarak Shah. Transformers in vision: A survey. ACM computing surveys (CSUR), 54 0 (10s): 0 1--41, 2022

    2022

  35. [43]

    Adversarial examples in the physical world

    Alexey Kurakin, Ian J Goodfellow, and Samy Bengio. Adversarial examples in the physical world. In Artificial intelligence safety and security, pp.\ 99--112. Chapman and Hall/CRC, 2018

    2018

  36. [44]

    Are we ready for a new paradigm shift? a survey on visual deep mlp

    Ruiyang Liu, Yinghui Li, Linmi Tao, Dun Liang, and Hai-Tao Zheng. Are we ready for a new paradigm shift? a survey on visual deep mlp. Patterns, 3 0 (7), 2022 a

    2022

  37. [45]

    Self-supervised learning via maximum entropy coding

    Xin Liu, Zhongdao Wang, Ya-Li Li, and Shengjin Wang. Self-supervised learning via maximum entropy coding. Advances in Neural Information Processing Systems, 35: 0 34091--34105, 2022 b

    2022

  38. [46]

    A convnet for the 2020s

    Zhuang Liu, Hanzi Mao, Chao-Yuan Wu, Christoph Feichtenhofer, Trevor Darrell, and Saining Xie. A convnet for the 2020s. In Proceedings of the IEEE/CVF conference on computer vision and pattern recognition, pp.\ 11976--11986, 2022 c

    2022

  39. [47]

    Understanding the effective receptive field in deep convolutional neural networks

    Wenjie Luo, Yujia Li, Raquel Urtasun, and Richard Zemel. Understanding the effective receptive field in deep convolutional neural networks. Advances in neural information processing systems, 29, 2016

    2016

  40. [48]

    Deep reinforcement and infomax learning

    Bogdan Mazoure, Remi Tachet des Combes, Thang Long Doan, Philip Bachman, and R Devon Hjelm. Deep reinforcement and infomax learning. Advances in Neural Information Processing Systems, 33: 0 3686--3698, 2020

    2020

  41. [49]

    Unsupervised learning of visual representations by solving jigsaw puzzles

    Mehdi Noroozi and Paolo Favaro. Unsupervised learning of visual representations by solving jigsaw puzzles. In European conference on computer vision, pp.\ 69--84. Springer, 2016

    2016

  42. [50]

    Self-supervised learning with an information maximization criterion

    Serdar Ozsoy, Shadi Hamdan, Sercan Arik, Deniz Yuret, and Alper Erdogan. Self-supervised learning with an information maximization criterion. Advances in Neural Information Processing Systems, 35: 0 35240--35253, 2022

    2022

  43. [51]

    On the integration of self-attention and convolution

    Xuran Pan, Chunjiang Ge, Rui Lu, Shiji Song, Guanfu Chen, Zeyi Huang, and Gao Huang. On the integration of self-attention and convolution. In Proceedings of the IEEE/CVF conference on computer vision and pattern recognition, pp.\ 815--825, 2022

    2022

  44. [52]

    Unsupervised visual representation learning by synchronous momentum grouping

    Bo Pang, Yifan Zhang, Yaoyi Li, Jia Cai, and Cewu Lu. Unsupervised visual representation learning by synchronous momentum grouping. In European Conference on Computer Vision, pp.\ 265--282. Springer, 2022

    2022

  45. [53]

    Self-supervised video pretraining yields robust and more human-aligned visual representations

    Nikhil Parthasarathy, SM Ali Eslami, Joao Carreira, and Olivier J Henaff. Self-supervised video pretraining yields robust and more human-aligned visual representations. In Thirty-seventh Conference on Neural Information Processing Systems, 2023

    2023

  46. [54]

    Learning transferable visual models from natural language supervision

    Alec Radford, Jong Wook Kim, Chris Hallacy, Aditya Ramesh, Gabriel Goh, Sandhini Agarwal, Girish Sastry, Amanda Askell, Pamela Mishkin, Jack Clark, et al. Learning transferable visual models from natural language supervision. In International conference on machine learning, pp...

    2021

  47. [55]

    Foolbox: A python toolbox to benchmark the robustness of machine learning models

    Jonas Rauber, Wieland Brendel, and Matthias Bethge. Foolbox: A python toolbox to benchmark the robustness of machine learning models. arXiv preprint arXiv:1707.04131, 2017

    2017 arXiv

  48. [56]

    Mobilenetv2: Inverted residuals and linear bottlenecks

    Mark Sandler, Andrew Howard, Menglong Zhu, Andrey Zhmoginov, and Liang-Chieh Chen. Mobilenetv2: Inverted residuals and linear bottlenecks. In Proceedings of the IEEE conference on computer vision and pattern recognition, pp.\ 4510--4520, 2018

    2018

  49. [57]

    Efficientnet: Rethinking model scaling for convolutional neural networks

    Mingxing Tan and Quoc Le. Efficientnet: Rethinking model scaling for convolutional neural networks. In International conference on machine learning, pp.\ 6105--6114. PMLR, 2019

    2019

  50. [58]

    Understanding self-supervised learning dynamics without contrastive pairs

    Yuandong Tian, Xinlei Chen, and Surya Ganguli. Understanding self-supervised learning dynamics without contrastive pairs. In International Conference on Machine Learning, pp.\ 10268--10278. PMLR, 2021

    2021

  51. [59]

    Emp-ssl: Towards self-supervised learning in one training epoch

    Shengbang Tong, Yubei Chen, Yi Ma, and Yann Lecun. Emp-ssl: Towards self-supervised learning in one training epoch. arXiv preprint arXiv:2304.03977, 2023

    2023 arXiv

  52. [60]

    Patches are all you need? Transactions on Machine Learning Research, 2023

    Asher Trockman and J Zico Kolter. Patches are all you need? Transactions on Machine Learning Research, 2023

    2023

  53. [61]

    The automaticity of visual statistical learning

    Nicholas B Turk-Browne, Justin A Jung \'e , and Brian J Scholl. The automaticity of visual statistical learning. Journal of Experimental Psychology: General, 134 0 (4): 0 552, 2005

    2005

  54. [62]

    Self-supervised learning by estimating twin class distribution

    Feng Wang, Tao Kong, Rufeng Zhang, Huaping Liu, and Hang Li. Self-supervised learning by estimating twin class distribution. IEEE Transactions on Image Processing, 2023 a

    2023

  55. [63]

    Pose-aware self-supervised learning with viewpoint trajectory regularization

    Jiayun Wang, Yubei Chen, and Stella X Yu. Pose-aware self-supervised learning with viewpoint trajectory regularization. In European Conference on Computer Vision, pp.\ 19--37. Springer, 2025

    2025

  56. [64]

    Contrastive learning with stronger augmentations

    Xiao Wang and Guo-Jun Qi. Contrastive learning with stronger augmentations. IEEE transactions on pattern analysis and machine intelligence, 45 0 (5): 0 5549--5560, 2022

    2022

  57. [65]

    Caco: Both positive and negative samples are directly learnable via cooperative-adversarial contrastive learning

    Xiao Wang, Yuhang Huang, Dan Zeng, and Guo-Jun Qi. Caco: Both positive and negative samples are directly learnable via cooperative-adversarial contrastive learning. IEEE Transactions on Pattern Analysis and Machine Intelligence, 2023 b

    2023

  58. [66]

    Dense contrastive learning for self-supervised visual pre-training

    Xinlong Wang, Rufeng Zhang, Chunhua Shen, Tao Kong, and Lei Li. Dense contrastive learning for self-supervised visual pre-training. In Proceedings of the IEEE/CVF conference on computer vision and pattern recognition, pp.\ 3024--3033, 2021

    2021

  59. [67]

    Detco: Unsupervised contrastive learning for object detection

    Enze Xie, Jian Ding, Wenhai Wang, Xiaohang Zhan, Hang Xu, Peize Sun, Zhenguo Li, and Ping Luo. Detco: Unsupervised contrastive learning for object detection. In Proceedings of the IEEE/CVF international conference on computer vision, pp.\ 8392--8401, 2021 a

    2021

  60. [68]

    Propagate yourself: Exploring pixel-level consistency for unsupervised visual representation learning

    Zhenda Xie, Yutong Lin, Zheng Zhang, Yue Cao, Stephen Lin, and Han Hu. Propagate yourself: Exploring pixel-level consistency for unsupervised visual representation learning. In Proceedings of the IEEE/CVF conference on computer vision and pattern recognition, pp.\ 16684--16693, 2021 b

    2021

  61. [69]

    Patch-level representation learning for self-supervised vision transformers

    Sukmin Yun, Hankook Lee, Jaehyung Kim, and Jinwoo Shin. Patch-level representation learning for self-supervised vision transformers. In Proceedings of the IEEE/CVF conference on computer vision and pattern recognition, pp.\ 8354--8363, 2022

    2022

  62. [70]

    Barlow twins: Self-supervised learning via redundancy reduction

    Jure Zbontar, Li Jing, Ishan Misra, Yann LeCun, and St \'e phane Deny. Barlow twins: Self-supervised learning via redundancy reduction. In International conference on machine learning, pp.\ 12310--12320. PMLR, 2021

    2021

  63. [71]

    Matrix information theory for self-supervised learning

    Yifan Zhang, Zhiquan Tan, Jingqin Yang, Weiran Huang, and Yang Yuan. Matrix information theory for self-supervised learning. In Forty-first International Conference on Machine Learning

  64. [72]

    Places: A 10 million image database for scene recognition

    Bolei Zhou, Agata Lapedriza, Aditya Khosla, Aude Oliva, and Antonio Torralba. Places: A 10 million image database for scene recognition. IEEE Transactions on Pattern Analysis and Machine Intelligence, 2017

    2017

  65. [73]

    Self-supervised learning of object parts for semantic segmentation

    Adrian Ziegler and Yuki M Asano. Self-supervised learning of object parts for semantic segmentation. In Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, pp.\ 14502--14511, 2022

    2022

  66. [74]

    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 gl...

  67. [75]

    @esa (Ref

    \@ifxundefined[1] #1\@undefined \@firstoftwo \@secondoftwo \@ifnum[1] #1 \@firstoftwo \@secondoftwo \@ifx[1] #1 \@firstoftwo \@secondoftwo [2] @ #1 \@temptokena #2 #1 @ \@temptokena \@ifclassloaded agu2001 natbib The agu2001 class already includes natbib coding, so you should ...

  68. [76]

    \@lbibitem[] @bibitem@first@sw\@secondoftwo \@lbibitem[#1]#2 \@extra@b@citeb \@ifundefined br@#2\@extra@b@citeb \@namedef br@#2 \@nameuse br@#2\@extra@b@citeb \@ifundefined b@#2\@extra@b@citeb @num @parse #2 @tmp #1 NAT@b@open@#2 NAT@b@shut@#2 \@ifnum @merge>\@ne @bibitem@firs...

  69. [77]

    u ckstiess, Thomas and Sehnke, Frank and Schaul, Tom and Wierstra, Daan and Sun, Yi and Schmidhuber, J \

    @open @close @open @close and [1] URL: #1 \@ifundefined chapter * \@mkboth \@ifxundefined @sectionbib * \@mkboth * \@mkboth\@gobbletwo \@ifclassloaded amsart * \@ifclassloaded amsbook * \@ifxundefined @heading @heading NAT@ctr thebibliography [1] @ \@biblabel @NAT@ctr \@bibset...

    2013 arXiv

Pith tools

Reviewed August 10, 2026 · model on record in the stance chip above.