Pith. sign in

REVIEW 5 major objections 8 minor 71 references

Biologically-inspired Semi-supervised Semantic Segmentation for Biomedical Imaging

T0 review · 5 major / 8 minor · reviewed 2026-08-11 · deepseek-v4-flash

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

Pith's one-line read The paper claims that a two-stage pipeline—unsupervised Hebbian pretraining of both the downsampling and upsampling paths, followed by a short backpropagation fine-tuning on a few labels—beats current semi-supervised segmentation methods…

desk verdict Worth a look if you work on semi-supervised medical segmentation or Hebbian learning; the new TSA rule for transpose-conv layers is a genuine contribution, but the abstract overclaims what the paper's own tables show. read the letter →

arxiv 2412.03192 v2 pith:2T6TWR24 submitted 2024-12-04 cs.CV

Luca Ciampi , Gabriele Lagani , Giuseppe Amato , Fabrizio Falchi This is my paper · ORCID
classification cs.CV
keywords Hebbianlearningsemi-supervisedsemanticsegmentationbiomedicalimagingtranspose-convolutionallayersSWTAHPCAencoder-decoderpretraining
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 show that an encoder-decoder segmentation network can be pretrained entirely without labels using Hebbian “fire together, wire together” updates, and that this pretraining makes the subsequent supervised fine-tuning data-efficient enough for biomedical imaging. The first stage applies local Hebbian rules to the ordinary convolutional layers and to the transpose-convolutional (upsampling) layers, for which the paper derives new SWTA-TSA and HPCA-TSA update rules. The second stage fine-tunes the pretrained weights with standard backpropagation on a small labeled subset. On four biomedical datasets across 1% to 20% label regimes, the method reports higher Dice overlap scores than state-of-the-art pseudo-labeling, consistency-training, variational-autoencoder, and self-supervised baselines, and it improves those baselines when used as their initialization. The practical upshot, if the claim holds, is that accurate pixel-level segmentation needs far fewer annotated images.

What carries the argument

The load-bearing machinery is the generalized Hebbian update $\Delta w_{i,j} = \eta\, g_j\,(s_i - s^*_{i,j})$, built from three blocks: a gating block $\gamma(y)$ that sets the update size, a reconstruction block $\rho(y,w)$ that reproduces the layer input from its outputs, and the plasticity function that moves the weight along the difference between a target signal $s_i$ and a reconstruction $s^*_{i,j}$. For transpose-convolutional layers the paper adds the named Transposed-Structure-Aware reconstruction block $\rho^*(y,w)$: it takes patches from the upsampled feature map at different offsets and combines them with the weight matrix through an ordinary convolution, fixing the shape mismatch while preserving the SWTA/HPCA pattern-discovery reading. This block is what allows the upsampling path to be pretrained without backpropagation or labels.

What would settle it

Fine-tune the full model from the SWTA-TSA initialization and from the same initialization with only the transpose-convolutional layers reset to random weights; if the Dice scores are statistically indistinguishable across several seeds, the new T-Conv rule is not carrying the claimed benefit.

Watch

Extended reading notes

Core claim

The central discovery the authors aim to establish is that a bio-inspired local learning rule, applied layer-by-layer and including the upsampling path, produces a useful unsupervised initialization for semi-supervised semantic segmentation. The main technical step is the definition of Transposed-Structure-Aware (TSA) Hebbian rules for transpose-convolutional layers, where a new reconstruction block $\rho^*$ treats the downsampled feature map as the target and derives gate and reconstruction signals from the upsampled feature map, so that the clustering (SWTA) or principal-component (HPCA) interpretation survives the reversed input-output geometry. In their comparisons, SWTA-TSA is the best of the four derived rules, and supplying its pretrained weights as initialization improves most single-stage semi-supervised competitors on most datasets and label regimes. The authors also report that the same two-stage pipeline carries over to 3D volumetric MRI segmentation, where it finishes best or second best in most settings.

Load-bearing premise

The load-bearing premise is that the new TSA rule for upsampling layers still discovers useful patterns even though the usual input and output roles are reversed, a property the paper proves only for the forward convolutional HPCA case and not for the TSA variant.

Editorial extensions

If this is right

  • If the central claim holds, an encoder-decoder segmentation model can be initialized from unlabeled images alone, so the number of pixel-level annotations needed for a usable biomedical segmentation system drops substantially.
  • The new TSA Hebbian rules mean the upsampling path is no longer a leftover that needs backpropagation; the entire architecture can be pretrained with local, biologically plausible updates.
  • Since the Hebbian initialization improves most pseudo-labeling and consistency-training baselines, the benefit compounds with existing semi-supervised losses rather than replacing them.
  • The same pipeline transfers from 2D histology, dermoscopy, and pupil images to 3D volumetric MRI, where it reaches best or second-best results at most label fractions.
  • The method's gains are concentrated in low-label regimes; the authors observe a smaller unsupervised-stage benefit on HMEPS at 1% labels, where the labeled set is already relatively large.

Reading between the lines

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

  • Beyond the paper's experiments, the TSA shape-mismatch argument applies to any upsampling layer, so SWTA-TSA could be tested as a label-free pretraining step for natural-image super-resolution and image synthesis.
  • Because the first stage uses only local updates and no backpropagation, it may run on neuromorphic or energy-constrained hardware, but the paper does not measure energy consumption, so that benefit remains an open empirical question.
  • The paper's ablation shows SWTA-TSA beats HPCA-TSA on GlaS, which suggests cluster discovery is a better inductive bias than PCA-style decorrelation for pixel-wise classification; a cross-task study would show whether that ranking is general.
  • The reported small performance drops when the Hebbian initialization is applied to some consistency methods indicate that the pretraining is not universally beneficial, so mapping which methods and regimes tolerate it would sharpen the recipe.
Share X Bluesky LinkedIn Reddit HN

Editorial analysis

A structured set of objections, weighed in public.

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

Referee Report

5 major / 8 minor

Summary. The paper proposes a two-stage semi-supervised semantic segmentation pipeline for biomedical imaging. In the first stage, an encoder-decoder (UNet-like) architecture is pretrained without labels using Hebbian learning rules (SWTA and HPCA), including newly formulated rules for transpose-convolutional layers (SWTA-TSA and HPCA-TSA). In the second stage, the pretrained weights are fine-tuned with standard backpropagation on a small labeled subset. The method is evaluated on three 2D datasets (GlaS, PH2, HMEPS) and one 3D dataset (LA), across label regimes from 1% to 20%. The authors report Dice/Jaccard/boundary metrics, compare with several single- and two-stage semi-supervised baselines, and show that initializing those baselines with their Hebbian pretraining often improves performance. The central claim is that the proposed method outperforms SOTA across different levels of label availability.

Significance. If the claims are substantiated, the paper makes a useful contribution by showing that biologically plausible Hebbian updates can pretrain upsampling layers without labels, complementing prior work that focused on convolutional or fully connected layers. The release of code and the inclusion of a linear-probing ablation are strengths, as is the evaluation across multiple imaging modalities and a volumetric dataset. However, the manuscript's headline superiority claim is not yet supported by the reported statistics or by its own 3D results, and several protocol details (temperature selection, baseline tuning, significance testing) need to be clarified before the empirical contribution can be fully assessed.

major comments (5)
  1. [§5.2, Tables 2–4] The claim that the method outperforms SOTA is not supported by the reported statistics in the 2D experiments. For example, on GlaS at 1% labeled data, Ours achieves DC 69.95 ± 1.09 versus CPS at 69.32 ± 0.59; on PH2 at 10%, Ours achieves 88.26 ± 0.51 versus URPC at 88.06 ± 0.40. In nearly all regimes the 90% confidence intervals overlap substantially, and no paired significance tests over the ten runs are reported. Without paired tests (e.g., Wilcoxon signed-rank or paired t-test), the numerical advantages of 0.2–1.3 DC points do not establish a reliable improvement. Please add appropriate statistical tests or temper the abstract and conclusions to 'numerically higher in most 2D regimes' where significance is not demonstrated.
  2. [§5.5, Table 8, and Abstract] The abstract's claim of outperforming SOTA 'across different levels of label availability' is directly contradicted by the paper's own LA results. At 10% and 20% labels, UAMT exceeds Ours by 1.84 and 1.74 DC points (88.95 vs. 87.11 and 90.91 vs. 89.17, respectively), and at 2% and 5% DTC or UAMT are best. Similarly, on HMEPS at 1% (Table 4), CCT achieves 91.09 vs. Ours 90.75. The paper itself acknowledges the HMEPS 1% case but the abstract remains overbroad. Please restrict the claims to the regimes and datasets where the method is actually best, or provide additional experiments/analysis to support a global claim.
  3. [§5.4, Table 5, and Appendix B] The per-dataset SWTA temperature (100 for GlaS, 20 for PH2 and HMEPS) appears to be selected on the test set: no validation protocol is described for this choice. Since the main results in Tables 2–4 use these selected temperatures, any selection based on test performance would invalidate the comparison with baselines that used default or differently tuned hyperparameters. Please document explicitly how the temperature (and the learning rates for competitor methods mentioned in Appendix B) were chosen on a validation split, and report the sensitivity of the final results to these choices.
  4. [§4.2 and Appendix A] The newly proposed TSA Hebbian rules for transpose-convolutional layers are the main theoretical novelty, but the manuscript does not provide a convergence or pattern-discovery analysis for them. Appendix A proves the HPCA equilibrium only for the forward convolutional case; the TSA variant reverses the roles of input and output and introduces a new reconstruction block ρ*, so the claim that SWTA-TSA/HPCA-TSA inherit the clustering/PCA behavior of the forward rules is an unproven inductive assumption. Please either provide an equilibrium/convergence analysis for the TSA rules or explicitly frame them as heuristics whose usefulness is established empirically (e.g., via the linear probing and initialization results).
  5. [§5.2 and Appendix B] The description of baseline hyperparameter tuning is not sufficiently specified. The statement that for some competitor techniques the authors 'opted for a lower initial learning rate, as we observed better results empirically' does not constitute a reproducible protocol. It is unclear whether the same validation-based selection procedure was applied to the proposed method and to all baselines, or whether the baselines were tuned with equal effort. Please specify the exact selection procedure for all hyperparameters of all methods, or the fairness of the comparison cannot be assessed.
minor comments (8)
  1. [§4.2] There is a typo: 'Straightfarward' should be 'Straightforward' in the description of the first T-Conv strategy.
  2. [§5.4, Table 5] For GlaS, the best temperature is the maximum tested (100), so the statement that performance 'reaches a plateau at that point' is not warranted; higher values should be tested or the wording should be softened.
  3. [§5.4, Table 7] The linear probing results are reported without confidence intervals or a description of the probing protocol (which features are used, classifier details, number of runs). Please add these details for consistency with the rest of the evaluation.
  4. [Figure 4] The y-axis ranges are described as 'embedded in the most convenient range for best readability,' which can visually exaggerate small differences. Please use fixed axis ranges or add error bars so that the improvements/worsenings are not misleading.
  5. [Appendix B] The phrase '10-fold cross-validation protocol (5-fold for the LA dataset), varying the training and test splits' is ambiguous. Please clarify whether this is standard k-fold cross-validation or k independent random splits, and how the validation split for model selection was obtained.
  6. [Abstract and §1] The wording 'substantially improves performance' is stronger than the quantitative evidence in the tables. Please align the abstract and introduction with the actual effect sizes and statistical support.
  7. [References] Reference [54] is listed as 'to appear'; if the ECCV 2024 workshop paper has been published, please update the citation.
  8. [Throughout] The text contains typographical artifacts such as 'di fferent' and 'e fficient' that should be corrected in the final version.

Circularity Check

0 steps flagged · score 0.0 of 10

No circularity: Hebbian pretraining is evaluated against independent SOTA baselines and test Dice is not used to fit the unsupervised weights.

full rationale

The first stage uses standard unsupervised Hebbian updates (SWTA/HPCA) extended to transpose-convolutional layers; the updates depend only on unlabeled inputs and current weights, not on labels or test metrics. The second-stage fine-tuning uses standard backpropagation on labeled data. Test-set Dice, JI, 95HD, and ASD are computed after training and are not fed back into the Hebbian weights, so the empirical comparison is not circular. The TSA reconstruction block is defined by the paper's own equations (Sec. 4.2, Eq. 4) and implements a transposed-structure-aware reconstruction as an ordinary convolution; it does not assume the target result. There are self-citations ([54] for the preliminary T-Conv rule, [21,24,53] for related Hebbian work), but they are used for context and novelty claims, not as the sole justification for the central empirical claim, which is benchmarked against independently reimplemented SOTA methods. The absence of a convergence proof for the TSA variant and the per-dataset temperature selection are correctness/protocol concerns, not circularity.

Assumptions & free parameters 4 free parameters · 5 assumptions · 0 invented entities

The paper introduces no new physical or ontological entities. Its load-bearing assumptions are standard Hebbian learning theory, a domain assumption about clustering as an inductive bias, and an unproven structural assumption about the reversed-role TSA update. The main free parameters are the SWTA temperature, the Hebbian learning rate (not reported), the stage-1 epoch count, and the baseline loss weight lambda_max.

free parameters (4)
  • SWTA softmax temperature t = 20 (PH2, HMEPS), 100 (GlaS), swept 1-100 in Table 5
    Hyperparameter controlling softmax sharpness in Eq. 2; chosen per dataset based on DC at the 20% labeled regime, affecting all stage-1 Hebbian updates.
  • Hebbian learning rate eta = not reported in text
    The Hebbian update rules (Eqs. 1-4) require a learning rate value; the paper reports optimizer learning rates for stage 2 but not the Hebbian stage rate, so replication depends on the repo.
  • Stage-1 training epochs = 200
    Chosen empirically ('we empirically set the end of the training to 200 epochs'); the stopping point affects which features the unsupervised stage discovers.
  • Consistency loss lambda_max for baselines = 5
    For competitor pseudo-labeling/consistency methods, the loss weight increases linearly to lambda_max=5; this tuning choice affects the strength of the baselines being compared against.
assumptions (5)
  • standard math SWTA weight updates converge to cluster centroids under competitive learning.
    Invoked in Section 3 and Figure 7, citing Grossberg and Rumelhart & Zipser; the paper relies on this convergence to claim the unsupervised stage discovers useful clusters.
  • standard math HPCA learning rule aligns weight vectors with principal components of the data.
    Derived in Appendix A using the covariance eigen-equation; the derivation assumes the standard linear neuron model y = Wx and standard linear algebra.
  • domain assumption Clustering in feature space is a useful inductive bias for semantic segmentation pretraining.
    Stated at the end of Section 4.2; supported indirectly by the linear-probing ablation (Table 7) but not proven theoretically.
  • domain assumption U-Net-like downsampling-upsampling architectures are an appropriate backbone for biomedical segmentation.
    The entire method is built on UNet-style architectures, citing standard medical segmentation literature; the claim does not generalize beyond this architecture family.
  • ad hoc to paper The TSA reconstruction block rho* preserves the intended Hebbian pattern-discovery behavior when input and output roles are reversed.
    Introduced in Section 4.2 (Figure 3c); the paper does not provide a convergence proof for this reversed-role update, so the empirical usefulness of the TSA rule is an assumed inductive property.

how reviews work

0 comments
Cite this review

Pith. "Pith review of Biologically-inspired Semi-supervised Semantic Segmentation for Biomedical Imaging." pith.science (2026). https://pith.science/paper/2T6TWR24

@misc{pith2026241203192,
  author       = {Pith},
  title        = {Pith review of: Biologically-inspired Semi-supervised Semantic Segmentation for Biomedical Imaging},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/2T6TWR24}},
  note         = {Machine review of arXiv:2412.03192}
}
read the original abstract

We propose a novel bio-inspired semi-supervised learning approach for training downsampling-upsampling semantic segmentation architectures. The first stage does not use backpropagation. Rather, it exploits the Hebbian principle ``fire together, wire together'' as a local learning rule for updating the weights of both convolutional and transpose-convolutional layers, allowing unsupervised discovery of data features. In the second stage, the model is fine-tuned with standard backpropagation on a small subset of labeled data. We evaluate our methodology through experiments conducted on several widely used biomedical datasets, deeming that this domain is paramount in computer vision and is notably impacted by data scarcity. Results show that our proposed method outperforms SOTA approaches across different levels of label availability. Furthermore, we show that using our unsupervised stage to initialize the SOTA approaches leads to performance improvements. The code to replicate our experiments can be found at https://github.com/ciampluca/hebbian-bootstraping-semi-supervised-medical-imaging

Figures

Figures reproduced from arXiv: 2412.03192 by the authors.

Figure 1
Figure 1. Our semi-supervised, bio-inspired approach for semantic segmentation architectures with downsampling and upsampling paths. The [PITH_FULL_IMAGE:figures/full_fig_p002_1.png] view at source ↗
Figure 2
Figure 2. Building blocks for Hebbian learning. Hebbian updates are [PITH_FULL_IMAGE:figures/full_fig_p004_2.png] view at source ↗
Figure 3
Figure 3. Hebbian learning in convolutional and transpose-convolutional layers. [PITH_FULL_IMAGE:figures/full_fig_p005_3.png] view at source ↗
Figures from the paper (4 more)
Figure 4
Figure 4. Figure 4: Performance changes (with green and red arrows) obtained with single-stage SOTA semi-supervised approaches initialized with our [PITH_FULL_IMAGE:figures/full_fig_p009_4.png]
Figure 5
Figure 5. Figure 5: Qualitative results from our semi-supervised approach based on Hebbian SWTA-TSA. Each row corresponds to a di [PITH_FULL_IMAGE:figures/full_fig_p010_5.png]
Figure 6
Figure 6. Figure 6: Qualitative results with 3D images on the LA dataset [28]. We provide two views for better readability – a slide view and a 3D view. [PITH_FULL_IMAGE:figures/full_fig_p010_6.png]
Figure 7
Figure 7. Figure 7: The illustration represents weight vectors (arrows) and data points (dots) within a data space. Fig. 7a shows how a weight vector updates [PITH_FULL_IMAGE:figures/full_fig_p011_7.png]

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

71 extracted references · 37 canonical work pages

  1. [1]

    Ronneberger, P

    O. Ronneberger, P. Fischer, T. Brox, U-net: Convolutional net- works for biomedical image segmentation, in: Medical Im- age Computing and Computer-Assisted Intervention – MICCAI 2015, Springer International Publishing, Cham, 2015, pp. 234–

    work page 2015

  2. [2]

    L. Chen, G. Papandreou, F. Schro ff, H. Adam, Rethinking atrous convolution for semantic image segmentation, CoRR abs/1706.05587 (2017). arXiv:1706.05587

    arXiv 2017

  3. [3]

    J. Long, E. Shelhamer, T. Darrell, Fully convolutional net- works for semantic segmentation, in: IEEE Conference on Com- puter Vision and Pattern Recognition, CVPR 2015, Boston, MA, USA, June 7-12, 2015, IEEE Computer Society, 2015, pp. 3431–3440. doi:10.1109/CVPR.2015.7298965

    arXiv 2015

  4. [4]

    Milletari, N

    F. Milletari, N. Navab, S. Ahmadi, V-net: Fully convolutional neural networks for volumetric medical image segmentation, in: 2016 Fourth International Conference on 3D Vision (3DV), IEEE Computer Society, Los Alamitos, CA, USA, 2016, pp. 565–571. doi:10.1109/3DV .2016.79

    doi:10.1109/3dv 2016

  5. [5]

    Y . Xie, J. Zhang, C. Shen, Y . Xia, Cotr: Efficiently bridging cnn and transformer for 3d medical image segmentation, in: Medical Image Computing and Computer Assisted Intervention – MIC- CAI 2021, Springer International Publishing, Cham, 2021, pp. 171–180. doi:10.1007/978-3-030-87199-4 16

    doi:10.1007/978-3-030-87199-4 2021

  6. [6]

    Hatamizadeh, Y

    A. Hatamizadeh, Y . Tang, V . Nath, D. Yang, A. Myronenko, B. Landman, H. R. Roth, D. Xu, Unetr: Transformers for 3d medical image segmentation, in: 2022 IEEE /CVF Winter Con- ference on Applications of Computer Vision (W ACV), 2022, pp. 1748–1758. doi:10.1109/W ACV51458.2022.00181

    arXiv 2022

  7. [7]

    H. Cao, Y . Wang, J. Chen, D. Jiang, X. Zhang, Q. Tian, M. Wang, Swin-unet: Unet-like pure transformer for medical image segmentation, in: Computer Vision – ECCV 2022 Work- shops, Springer Nature Switzerland, Cham, 2023, pp. 205–218. doi:10.1007/978-3-031-25066-8 9

    doi:10.1007/978-3-031-25066-8 2022

  8. [8]

    J. M. J. Valanarasu, V . A. Sindagi, I. Hacihaliloglu, V . M. Patel, Kiu-net: Overcomplete convolutional architectures for biomedical image and volumetric segmentation, IEEE Transactions on Medical Imaging 41 (4) (2022) 965–976. doi:10.1109/TMI.2021.3130469

    arXiv 2022

Show all 71 references
  1. [9]

    Isensee, P

    F. Isensee, P. F. Jaeger, S. A. A. Kohl, J. Petersen, K. H. Maier- Hein, nnu-net: a self-configuring method for deep learning- based biomedical image segmentation, Nature Methods 18 (2) (2020) 203–211. doi:10.1038/s41592-020-01008-z

    2020 doi

  2. [10]

    C ¸ ic ¸ek, A

    ¨O. C ¸ ic ¸ek, A. Abdulkadir, S. S. Lienkamp, T. Brox, O. Ron- neberger, 3d u-net: Learning dense volumetric segmenta- tion from sparse annotation, in: Medical Image Comput- ing and Computer-Assisted Intervention – MICCAI 2016, Springer International Publishing, Cham, 2016, p...

    2016 doi

  3. [11]

    J. M. J. Valanarasu, P. Oza, I. Hacihaliloglu, V . M. Patel, Med- ical transformer: Gated axial-attention for medical image seg- mentation, in: Medical Image Computing and Computer As- sisted Intervention – MICCAI 2021, Springer International Pub- lishing, Cham, 2021, pp. 36–4...

    2021 doi

  4. [12]

    Basak, R

    H. Basak, R. Kundu, R. Sarkar, Mfsnet: A multi focus segmen- tation network for skin lesion segmentation, Pattern Recogn. 128 (C) (aug 2022). doi:10.1016/j.patcog.2022.108673

    2022

  5. [13]

    Karimi, K

    A. Karimi, K. Faez, S. Nazari, Deu-net: Dual-encoder u-net for automated skin lesion segmentation, IEEE Access 11 (2023) 134804–134821. doi:10.1109/ACCESS.2023.3337528

    2023

  6. [14]

    C. D. Schuman, S. R. Kulkarni, M. Parsa, J. P. Mitchell, P. Date, B. Kay, Opportunities for neuromorphic computing algorithms and applications, Nature Computational Science 2 (1) (2022) 10–19

    2022

  7. [15]

    Shrestha, H

    A. Shrestha, H. Fang, Z. Mei, D. P. Rider, Q. Wu, Q. Qiu, A sur- vey on neuromorphic computing: Models and hardware, IEEE Circuits and Systems Magazine 22 (2) (2022) 6–35

    2022

  8. [16]

    Haykin, Neural networks and learning machines, 3rd Edition, Pearson, 2009

    S. Haykin, Neural networks and learning machines, 3rd Edition, Pearson, 2009

    2009

  9. [17]

    Gerstner, W

    W. Gerstner, W. M. Kistler, Spiking neuron models: Single neu- rons, populations, plasticity, Cambridge university press, 2002

    2002

  10. [18]

    Lagani, F

    G. Lagani, F. Falchi, C. Gennaro, G. Amato, Synaptic plas- ticity models and bio-inspired unsupervised deep learning: A survey, CoRR abs /2307.16236 (2023). arXiv:2307.16236, doi:10.48550/ARXIV .2307.16236

    2023 doi arXiv

  11. [19]

    Krotov, J

    D. Krotov, J. J. Hopfield, Unsupervised learning by competing hidden units, Proceedings of the National Academy of Sciences 116 (16) (2019) 7723–7731. doi:10.1073/pnas.1820458116

    2019 doi

  12. [20]

    Moraitis, D

    T. Moraitis, D. Toichkin, A. Journ ´e, Y . Chua, Q. Guo, Softhebb: Bayesian inference in unsupervised hebbian soft winner-take- all networks, Neuromorphic Computing and Engineering 2 (4) (2022) 044017. doi:10.1088/2634-4386/aca710

    2022 doi

  13. [21]

    Lagani, F

    G. Lagani, F. Falchi, C. Gennaro, G. Amato, Comparing the performance of hebbian against backpropagation learning us- ing convolutional neural networks, Neural Computing and Ap- plications 34 (8) (2022) 6503–6519. doi:10.1007 /s00521-021- 06701-4

    2022

  14. [22]

    Pehlevan, T

    C. Pehlevan, T. Hu, D. B. Chklovskii, A Hebbian /Anti- Hebbian Neural Network for Linear Subspace Learning: A Derivation from Multidimensional Scaling of Stream- ing Data, Neural Computation 27 (7) (2015) 1461–1495. doi:10.1162/NECO a 00745

    2015 doi

  15. [23]

    Bahroun, A

    Y . Bahroun, A. Soltoggio, Online representation learning with single and multi-layer hebbian networks for image classifica- tion, in: Artificial Neural Networks and Machine Learning – ICANN 2017, Springer International Publishing, Cham, 2017, pp. 354–363. doi:10.1007/978-3-31...

    2017 doi

  16. [24]

    Lagani, F

    G. Lagani, F. Falchi, C. Gennaro, G. Amato, Heb- bian semi-supervised learning in a sample e ffi- ciency setting, Neural Networks 143 (2021) 719–731. doi:https://doi.org/10.1016/j.neunet.2021.08.003

    2021 doi

  17. [25]

    Sirinukunwattana, J

    K. Sirinukunwattana, J. P. Pluim, H. Chen, X. Qi, P.-A. Heng, Y . B. Guo, L. Y . Wang, B. J. Matuszewski, E. Bruni, U. Sanchez, A. B ¨ohm, O. Ronneberger, B. B. Cheikh, D. Racoceanu, P. Kainz, M. Pfei ffer, M. Urschler, D. R. Snead, N. M. Ra- jpoot, Gland segmentation in colon...

    2017 doi

  18. [26]

    Mendonc ¸a, P

    T. Mendonc ¸a, P. M. Ferreira, J. S. Marques, A. R. S. Marcal, J. Rozeira, Ph2 - a dermoscopic image database 13 for research and benchmarking, in: 2013 35th An- nual International Conference of the IEEE Engineering in Medicine and Biology Society (EMBC), 2013, pp. 5437–5440. ...

    2013

  19. [27]

    Mazziotti, F

    R. Mazziotti, F. Carrara, A. Viglione, L. Leonardo, L. V . Luca, B. Alessandro, R. Giulia, S. Giulia, A. Giuseppe, P. Tommaso, Human and Mouse Eyes for Pupil Semantic Segmentation (Feb. 2021). doi:10.5281/zenodo.4488164

    2021 doi

  20. [28]

    Xiong, Q

    Z. Xiong, Q. Xia, Z. Hu, N. Huang, C. Bian, Y . Zheng, S. Vesal, N. Ravikumar, A. Maier, X. Yang, P.-A. Heng, D. Ni, C. Li, Q. Tong, W. Si, E. Puybareau, Y . Khoudli, T. G´eraud, C. Chen, W. Bai, D. Rueckert, L. Xu, X. Zhuang, X. Luo, S. Jia, M. Sermesant, Y . Liu, K. Wang, D....

    2021

  21. [29]

    T. Vu, H. Jain, M. Bucher, M. Cord, P. P ´erez, ADVENT: adver- sarial entropy minimization for domain adaptation in semantic segmentation, in: IEEE Conference on Computer Vision and Pattern Recognition, CVPR 2019, Long Beach, CA, USA, June 16-20, 2019, Computer Vision Foundati...

    2019

  22. [30]

    L. Yu, S. Wang, X. Li, C.-W. Fu, P.-A. Heng, Uncertainty-aware self-ensembling model for semi-supervised 3d left atrium seg- mentation, in: Medical Image Computing and Computer As- sisted Intervention – MICCAI 2019, Springer International Pub- lishing, Cham, 2019, pp. 605–613

    2019

  23. [31]

    Ouali, C

    Y . Ouali, C. Hudelot, M. Tami, Semi-supervised seman- tic segmentation with cross-consistency training, in: 2020 IEEE/CVF Conference on Computer Vision and Pattern Recog- nition, CVPR 2020, Seattle, W A, USA, June 13-19, 2020, Computer Vision Foundation / IEEE, 2020, pp. 1267...

    2020

  24. [32]

    X. Chen, Y . Yuan, G. Zeng, J. Wang, Semi-supervised semantic segmentation with cross pseudo supervision, in: IEEE Confer- ence on Computer Vision and Pattern Recognition, CVPR 2021, virtual, June 19-25, 2021, Computer Vision Foundation / IEEE, 2021, pp. 2613–2622. doi:10.1109...

    2021

  25. [33]

    X. Luo, G. Wang, W. Liao, J. Chen, T. Song, Y . Chen, S. Zhang, D. N. Metaxas, S. Zhang, Semi-supervised med- ical image segmentation via uncertainty rectified pyramid consistency, Medical Image Analysis 80 (2022) 102517. doi:https://doi.org/10.1016/j.media.2022.102517

    2022

  26. [34]

    X. Luo, J. Chen, T. Song, G. Wang, Semi-supervised medi- cal image segmentation through dual-task consistency, Proceed- ings of the AAAI Conference on Artificial Intelligence 35 (10) (2021) 8801–8809. doi:10.1609/aaai.v35i10.17066

    2021 doi

  27. [35]

    D. P. Kingma, M. Welling, Auto-encoding variational bayes, in: Y . Bengio, Y . LeCun (Eds.), 2nd International Conference on Learning Representations, ICLR 2014, Ban ff, AB, Canada, April 14-16, 2014, Conference Track Proceedings, 2014

    2014

  28. [36]

    Ouyang, C

    C. Ouyang, C. Bi ffi, C. Chen, T. Kart, H. Qiu, D. Rueckert, Self-supervision with superpixels: Training few-shot medical image segmentation without annotation, in: Computer Vision - ECCV 2020 - 16th European Conference, Glasgow, UK, Au- gust 23-28, 2020, Proceedings, Part XXI...

    2020 doi

  29. [37]

    Devlin, M.-W

    J. Devlin, M.-W. Chang, K. Lee, K. Toutanova, BERT: Pre- training of deep bidirectional transformers for language under- standing, in: Proceedings of the 2019 Conference of the North American Chapter of the Association for Computational Lin- guistics: Human Language Technologi...

    2019

  30. [38]

    Ciampi, F

    L. Ciampi, F. Carrara, V . Totaro, R. Mazziotti, L. Lu- pori, C. Santiago, G. Amato, T. Pizzorusso, C. Gen- naro, Learning to count biological structures with raters’ uncertainty, Medical Image Analysis 80 (2022) 102500. doi:https://doi.org/10.1016/j.media.2022.102500

    2022

  31. [39]

    Dosovitskiy, L

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

    2020 arXiv

  32. [40]

    Y . Roh, G. Heo, S. E. Whang, A survey on data collection for machine learning: A big data - ai integration perspective, IEEE Transactions on Knowledge and Data Engineering 33 (4) (2021) 1328–1347. doi:10.1109/TKDE.2019.2946162

    2021

  33. [41]

    Badar, A

    A. Badar, A. Varma, A. Staniec, M. Gamal, O. Magdy, H. Iqbal, E. Arani, B. Zonooz, Highlighting the importance of reducing research bias and carbon emissions in cnns, in: International Conference of the Italian Association for Artificial Intelligence, Springer, 2021, pp. 515–531

    2021

  34. [42]

    B. M. Lake, S. T. Piantadosi, People infer recursive visual con- cepts from just a few examples, Computational Brain & Behav- ior 3 (2020) 54–65

    2020

  35. [43]

    Javed, Q

    F. Javed, Q. He, L. E. Davidson, J. C. Thornton, J. Albu, L. Boxt, N. Krasnow, M. Elia, P. Kang, S. Heshka, et al., Brain and high metabolic rate organ mass: contributions to resting energy ex- penditure beyond fat-free mass, The American journal of clini- cal nutrition 91 (4)...

    2010

  36. [44]

    Hassabis, D

    D. Hassabis, D. Kumaran, C. Summerfield, M. Botvinick, Neuroscience-inspired artificial intelligence, Neuron 95 (2) (2017) 245–258

    2017

  37. [45]

    B. M. Lake, T. D. Ullman, J. B. Tenenbaum, S. J. Gershman, Building machines that learn and think like people, Behavioral and brain sciences 40 (2017)

    2017

  38. [46]

    Y . Wu, L. Deng, G. Li, J. Zhu, Y . Xie, L. Shi, Direct training for spiking neural networks: Faster, larger, better, in: Proceed- ings of the AAAI Conference on Artificial Intelligence, V ol. 33, 2019, pp. 1311–1318

    2019

  39. [47]

    C. Lee, S. S. Sarwar, P. Panda, G. Srinivasan, K. Roy, Enabling spike-based backpropagation for training deep neural network architectures, Frontiers in neuroscience 14 (2020) 119

    2020

  40. [48]

    S. Zhou, X. Li, Y . Chen, S. T. Chandrasekaran, A. Sanyal, Temporal-coded deep spiking neural network with easy training and robust performance, in: Proceedings of the AAAI Confer- ence on Artificial Intelligence, V ol. 35, 2021, pp. 11143–11151

    2021

  41. [49]

    G ¨oltz, L

    J. G ¨oltz, L. Kriener, A. Baumbach, S. Billaudelle, O. Bre- itwieser, B. Cramer, D. Dold, A. F. Kungl, W. Senn, J. Schem- mel, et al., Fast and energy-efficient neuromorphic deep learning with first-spike times, Nature machine intelligence 3 (9) (2021) 823–835

    2021

  42. [50]

    H. Wang, Z. He, T. Wang, J. He, X. Zhou, Y . Wang, L. Liu, N. Wu, M. Tian, C. Shi, Triplebrain: A compact neuromorphic hardware core with fast on-chip self-organizing and reinforce- ment spike-timing dependent plasticity, IEEE Transactions on Biomedical Circuits and Systems 16...

    2022

  43. [51]

    R. C. O’Reilly, Y . Munakata, Computational explorations in cognitive neuroscience: Understanding the mind by simulating the brain, MIT press, 2000

    2000

  44. [52]

    Journ ´e, H

    A. Journ ´e, H. G. Rodriguez, Q. Guo, T. Moraitis, Hebbian deep learning without feedback, in: The Eleventh International Con- ference on Learning Representations, 2023

    2023

  45. [53]

    Lagani, C

    G. Lagani, C. Gennaro, H. Fassold, G. Amato, Fasthebb: Scal- 14 ing hebbian training of deep neural networks to imagenet level, in: Similarity Search and Applications: 15th International Con- ference, SISAP 2022, Bologna, Italy, October 5–7, 2022, Pro- ceedings, Springer, 2022...

    2022

  46. [54]

    Ciampi, G

    L. Ciampi, G. Lagani, G. Amato, F. Falchi, A biologically- inspired approach to biomedical image segmentation, in: Com- puter Vision - ECCV 2024 Workshops - Milan, Italy, September 29-October 4, 2024 (Accepted), 2024, to appear

    2024

  47. [55]

    Badrinarayanan, A

    V . Badrinarayanan, A. Kendall, R. Cipolla, Segnet: A deep convolutional encoder-decoder architecture for image segmen- tation, IEEE Trans. Pattern Anal. Mach. Intell. 39 (12) (2017) 2481–2495. doi:10.1109/TPAMI.2016.2644615

    2017

  48. [56]

    Bengio, P

    Y . Bengio, P. Lamblin, D. Popovici, H. Larochelle, Greedy layer-wise training of deep networks, in: B. Sch ¨olkopf, J. Platt, T. Hoffman (Eds.), Advances in Neural Information Processing Systems, V ol. 19, MIT Press, 2006

    2006

  49. [57]

    Larochelle, Y

    H. Larochelle, Y . Bengio, J. Louradour, P. Lamblin, Exploring strategies for training deep neural networks, J. Mach. Learn. Res. 10 (2009) 1–40

    2009

  50. [58]

    D. P. Kingma, D. J. Rezende, S. Mohamed, M. Welling, Semi- supervised learning with deep generative models, in: Proceed- ings of the 27th International Conference on Neural Informa- tion Processing Systems - V olume 2, NIPS’14, MIT Press, Cam- bridge, MA, USA, 2014, p. 3581–3589

    2014

  51. [59]

    Zhang, K

    Y . Zhang, K. Lee, H. Lee, Augmenting supervised neural net- works with unsupervised objectives for large-scale image clas- sification, in: Proceedings of the 33rd International Conference on International Conference on Machine Learning - V olume 48, ICML’16, JMLR.org, 2016, p...

    2016

  52. [60]

    Grossberg, Adaptive Pattern Classification and Univer- sal Recoding, I: Parallel Development and Coding of Neu- ral Feature Detectors, The MIT Press, 1991, p

    S. Grossberg, Adaptive Pattern Classification and Univer- sal Recoding, I: Parallel Development and Coding of Neu- ral Feature Detectors, The MIT Press, 1991, p. 203–232. doi:10.7551/mitpress/5271.003.0008

    1991 doi

  53. [61]

    D. E. Rumelhart, D. Zipser, Feature discovery by com- petitive learning*, Cognitive Science 9 (1) (1985) 75–112. doi:10.1207/s15516709cog0901 5

    1985 doi

  54. [62]

    T. D. Sanger, Optimal unsupervised learning in a single-layer linear feedforward neural network, Neural Networks 2 (6) (1989) 459–473. doi:10.1016/0893-6080(89)90044-0

    1989 doi

  55. [63]

    Karhunen, J

    J. Karhunen, J. Joutsensalo, Generalizations of principal com- ponent analysis, optimization problems, and neural networks, Neural Networks 8 (4) (1995) 549–562. doi:10.1016 /0893- 6080(94)00098-7

    1995

  56. [64]

    Becker, M

    S. Becker, M. Plumbley, Unsupervised neural network learning procedures for feature extraction and classification, Applied In- telligence 6 (3) (1996) 185–203. doi:10.1007/bf00126625

    1996 doi

  57. [65]

    Luo, Architectures of neuronal circuits, Science 373 (6559) (2021) eabg7285

    L. Luo, Architectures of neuronal circuits, Science 373 (6559) (2021) eabg7285. doi:10.1126/science.abg7285

    2021 doi

  58. [66]

    Dumoulin, F

    V . Dumoulin, F. Visin, A guide to convolution arith- metic for deep learning, CoRR abs /1603.07285 (2016). arXiv:1603.07285

    2016 arXiv

  59. [67]

    Lagani, F

    G. Lagani, F. Falchi, C. Gennaro, H. Fassold, G. Amato, Scal- able bio-inspired training of deep neural networks with fasthebb, Neurocomputing (2024) 127867

    2024

  60. [68]

    Journ ´e, H

    A. Journ ´e, H. G. Rodriguez, Q. Guo, T. Moraitis, Hebbian deep learning without feedback, arXiv preprint arXiv:2209.11883 (2022)

    2022 arXiv

  61. [69]

    W. V . Gansbeke, S. Vandenhende, S. Georgoulis, L. V . Gool, Unsupervised semantic segmentation by contrasting ob- ject mask proposals, in: 2021 IEEE /CVF International Con- ference on Computer Vision, ICCV 2021, Montreal, QC, Canada, October 10-17, 2021, IEEE, 2021, pp. 10032...

    2021

  62. [70]

    X. Ji, A. Vedaldi, J. F. Henriques, Invariant information clustering for unsupervised image classification and segmen- tation, in: 2019 IEEE /CVF International Conference on Computer Vision, ICCV 2019, Seoul, Korea (South), Octo- ber 27 - November 2, 2019, IEEE, 2019, pp. 9864...

    2019

  63. [241]

    doi:10.1007/978-3-319-24574-4 28

    doi

Pith tools

Reviewed August 11, 2026 · model on record in the stance chip above.