Pith. sign in

REVIEW 4 major objections 4 minor 38 references

A balanced soft mixture-of-experts model surpasses all baselines for glaucoma detection across three datasets.

Reviewed by Pith at T0; open to challenge. T0 means a machine referee read the full paper against a public rubric. the ladder, T0–T4 →

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

T0 review · deepseek-v4-flash

2026-08-01 02:44 UTC pith:OJFIRYK6

load-bearing objection Solid but incremental SMoE application to multimodal glaucoma detection; the main claimed contribution—the CV-based load-balancing loss—is never ablated and its weight is set so low that the term may be a no-op. the 4 major comments →

arxiv 2607.25324 v1 pith:OJFIRYK6 submitted 2026-07-28 cs.CV cs.AI

Balanced Soft mixture-of-expert model for Glaucoma Detection

Sai Venkatesh Chilukoti , Krishna Rauniyar , Min Shi , Xiali Hei This is my paper
classification cs.CV cs.AI
keywords glaucoma detectionmulti-modal learningmixture of expertsload balancing losscoefficient of variationOCTfundus photographyAUC
verification ladder T0 review T1 audit T2 compute T3 formal T4 reserved

The pith

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

Glaucoma detection from retinal images usually relies on either fundus or OCT scans alone, or naively fuses them. The paper argues that when two modalities are combined, one modality tends to dominate training, leaving the other under-optimized. It proposes a soft mixture-of-experts model with three branches—one for each modality and one for both—whose predictions are blended by a learned gating network. A coefficient-of-variation load-balancing loss keeps the three experts roughly equally used. On three public benchmarks the model reports higher AUC than all uni-modal and multi-modal baselines compared.

Core claim

The central claim is that a soft mixture-of-experts architecture, with separate uni-modal experts and a joint multi-modal expert plus a gating network trained with a coefficient-of-variation load-balancing loss, achieves the highest glaucoma-detection AUC on FairVision (83.29%), FairDomain (84.31%), and HarvardGF (86.33%), outperforming uni-modal EfficientNet, DenseNet, ResNet, ViT-B, and balanced multi-modal methods OPM, OGM, CGGM, and OPM+OGM. The authors interpret this as evidence that adaptive per-sample expert weighting, guided by a balance penalty, mitigates modality imbalance more effectively than forward-pass or gradient-based modulation.

What carries the argument

The load-bearing mechanism is the soft mixture-of-experts (SMoE) module: a lightweight CNN gating network computes a weight per expert (two uni-modal, one multi-modal) that sums to one, and the final prediction is the weighted average of expert outputs. The balance regularizer is the coefficient of variation (CV) of per-expert gating load across a mini-batch, squared and scaled by a hyperparameter; minimizing CV pushes the model toward uniform expert usage, preventing gating collapse.

Load-bearing premise

The paper's central contribution—the coefficient-of-variation load-balancing loss—is never ablated, so the reported AUC gains could be caused by the gating network and three-expert architecture alone, with the balancing loss contributing little or nothing.

What would settle it

Train the same SMoE architecture with the load-balancing loss removed (set the regularization coefficient to 0) and compare AUC on the three datasets. If removing the loss leaves AUC essentially unchanged in all three, the paper's stated contribution is not supported. Alternatively, if the loss is necessary, the AUC should drop measurably when it is disabled.

Watch this falsifier — get emailed when new claim-graph text bears on it.

If this is right

  • If the reported AUCs hold, SMoE provides a drop-in fusion head for two-modality retinal imaging that beats existing balancing tricks without changing the backbone.
  • The method's gain over OPM/OGM/CGGM suggests that per-sample soft weighting, rather than gradient or prediction modulation, is a more effective way to counteract modality dominance in this domain.
  • The approach is modality-agnostic and can be applied to other paired-image diagnostics, such as diabetic retinopathy as the authors note.
  • Stable standard deviations across runs indicate the results are reproducible under the authors' training protocol.

Where Pith is reading between the lines

These are editorial extensions of the paper, not claims the author makes directly.

  • The paper does not ablate the CV loss; the reported gains might come entirely from the three-expert soft routing architecture, with the balancing loss making little difference. A comparison of SMoE with and without the CV term would isolate the contribution.
  • The gating network itself may be learning to ignore noisy samples from one modality, which would make SMoE's robustness a property of weighting rather than balance; this could be tested by checking per-sample gate distributions on misclassified cases.
  • If the CV loss matters, it implies that equal expert usage is a proxy for balanced learning; an alternative test would be to replace CV with entropy of the gate distribution to see if any 'balance' penalty suffices.
  • Because all experts share the same backbone and are trained end-to-end, the multi-modal expert may simply be a fine-tuned copy of the uni-modal experts; inspecting the expert weights on held-out data would reveal whether it actually learns a distinct fusion representation.

Editorial analysis

A structured set of objections, weighed in public.

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

Referee Report

4 major / 4 minor

Summary. The paper proposes a balanced soft mixture-of-experts (SMoE) model for glaucoma detection. The architecture uses three EfficientNet-B1 experts: one for SLO fundus images, one for OCT images, and one that jointly processes both modalities. A lightweight CNN gating network assigns per-sample soft weights to the three experts, and a coefficient-of-variation load-balancing loss (Eq. 1) is added to encourage uniform expert utilization. The model is evaluated on three public datasets (FairVision, FairDomain, HarvardGF) and compared against uni-modal CNN/ViT baselines and balanced multimodal baselines (OPM, OGM, CGGM, OPM+OGM). The authors report the highest overall AUC for SMoE on all three datasets (83.29%, 84.31%, 86.33%, respectively) and claim that the proposed method surpasses all uni-modal, conventional multimodal, and state-of-the-art balanced multimodal baselines.

Significance. If the reported results are statistically robust and the balancing mechanism is actually responsible for the gains, the paper would provide a modest but potentially useful architecture for balanced multimodal medical-image classification, evaluated on three public benchmarks with standard deviations reported. The comparison against OPM/OGM/CGGM is a strength, as is the use of externally published datasets. However, the central contribution—the load-balancing loss—is never isolated experimentally, and the claimed advantage over the strongest baseline is numerically tiny on FairDomain (0.01 AUC) without significance testing. The significance of the paper therefore depends on additional ablation and statistical validation that the current manuscript does not provide.

major comments (4)
  1. The load-balancing loss is the paper's central contribution, but it is never ablated. There is no experiment with λ_dyn=0, no varying λ_dyn, and no analysis of the actual gating-load distribution. Given λ_dyn=0.001 (Section 4.3) and typical cross-entropy loss magnitudes of order 0.5–1, the squared-CV term weighted by 0.001 is likely several orders of magnitude smaller than the classification objective; unless the gating network is trained with separate scaling (which is not described), the balancing term may be numerically negligible. The observed AUC improvements could therefore come entirely from the three-expert architecture and the learned gating network. The Limitations section discusses complexity, missing modalities, and fairness but does not flag this missing ablation. To support the abstract's claim that the balanced design is what makes the method work, the authors must show th
  2. No statistical significance tests are reported. The FairDomain result is particularly problematic: SMoE achieves 84.31±0.0029 AUC while OPM+OGM achieves 84.30±0.0048; the difference is 0.01 percentage points, well within one standard deviation of both runs. Without paired significance tests (e.g., DeLong test or bootstrapped confidence intervals) or multiple-seed comparisons, the claim that SMoE 'surpasses' all balanced multimodal baselines is unsupported on this dataset. The abstract's unconditional superiority claim requires statistical support, especially where the margin is so small.
  3. The architecture is under-specified. The 'Normal' multimodal baseline in Table 2 is not described at all: what fusion does it use, what is the exact architecture, and is it a simple concatenation plus classification head? Similarly, the multi-modal expert is said to 'jointly process both modalities,' but the fusion mechanism inside that expert is not defined (e.g., early concatenation, attention, cross-modal interactions). The gating network is described as processing 'fused multi-modal features,' but the fusion method for the gating input is also omitted. These details are necessary for reproducibility and for interpreting the comparison against conventional multimodal baselines.
  4. The claim that SMoE surpasses 'current state-of-the-art balanced multi-modal models' is broader than the evidence. Table 2 compares against only four balanced strategies (OPM, OGM, CGGM, OPM+OGM). The paper does not compare with other recent balanced multimodal methods or with DynMM, which is cited as inspiration. Unless a more comprehensive comparison is provided, the 'state-of-the-art' claim should be qualified to 'compared methods.'
minor comments (4)
  1. The HarvardGF percentages in Figure 1 sum to 101% (53.4% glaucoma + 47.6% non-glaucoma). Also, the FairVision description says the dataset has 10,000 subjects each for AMD, DR, and glaucoma, but then states 48.7% are diagnosed with glaucoma; these numbers are inconsistent and need clarification.
  2. Many typographical and formatting issues: 'T able 1' and 'T able 2' in table captions, 'F airvision'/'F airdomain' in table rows, 'Comparitive methods' in §4.1, 'In additon' in §3, 'figure [5]' instead of 'Fig. 5', and missing spaces in figure captions.
  3. The sentence 'CGGM performs less poorly than the normal model' should be 'CGGM performs worse than the normal model' (or similar).
  4. Some references are incompletely formatted (e.g., [8] 'Hwang, D.K., al.', [25] 'Cai, Y., et al.'); full author lists or consistent et al. formatting should be used.

Circularity Check

0 steps flagged

No significant circularity: empirical benchmark comparison, no prediction reduces to a fitted parameter or self-citation.

full rationale

The paper is an empirical architecture study. Its central claim is that the proposed soft mixture-of-experts with a CV load-balancing loss obtains the highest AUC on three external benchmarks. The only closed-form equation (Eq. 1) defines the auxiliary loss, not the evaluation metric; no test AUC is constructed from the fitted loss or from gating weights. The hyperparameter λ_dyn=0.001 is a training configuration, not a parameter fitted to the reported AUC. The datasets FairVision, FairDomain, and HarvardGF were co-created by an author, but they are published benchmarks with independent ground-truth labels; citing them does not make the comparison definitionally equivalent to the paper's own method. The absence of an ablation of the load-balancing loss (and the possibility that λ_dyn=0.001 makes the term weak) is a correctness/attribution risk, not circularity: even a no-op balancing loss would not make the AUC numbers equal to the paper's inputs by construction. The stated limitations (no explicit fairness optimization, EfficientNet-only backbone, added complexity, assumption of complete modalities) are caveats, not admissions of circularity. No uniqueness theorem or ansatz from prior self-cited work is invoked to force the design choice. Therefore score 0.

Axiom & Free-Parameter Ledger

1 free parameters · 4 axioms · 0 invented entities

The paper introduces one free hyperparameter (λ_dyn=0.001) and relies on several domain assumptions about label quality, metric choice, backbone selection, and the benefit of balanced expert usage. No new physical or conceptual entities are postulated.

free parameters (1)
  • λ_dyn (load balancing loss coefficient) = 0.001
    Chosen by hand (Section 4.3); controls strength of the CV-based balancing loss. No sensitivity analysis is reported.
axioms (4)
  • domain assumption Visual-field-based glaucoma labels in FairVision, FairDomain, and HarvardGF are accurate.
    Ground truth for all AUC computations; any label noise biases the comparison (Section 2.1).
  • domain assumption AUC is the appropriate performance metric for glaucoma screening.
    Used throughout; no discussion of calibration or clinical operating points.
  • ad hoc to paper EfficientNet-B1 is the best backbone for all experts.
    Chosen because it 'achieves the best overall performance' among uni-modal baselines (Section 4.2); this post-hoc choice could favor the proposed model.
  • ad hoc to paper Balanced expert utilization improves generalization.
    Core motivation for the load-balancing loss; not tested with an ablation.

pith-pipeline@v1.3.0-alltime-deepseek · 12061 in / 10023 out tokens · 85083 ms · 2026-08-01T02:44:51.404457+00:00 · methodology

0 comments
read the original abstract

Glaucoma is a group of eye diseases that damage the optic nerve, often caused by elevated intraocular pressure. It is a leading cause of irreversible vision loss and is typically developed slowly and painlessly, making it difficult to notice until significant damage has occurred. Therefore, early detection is crucial to prevent or slow the progression of vision loss. In recent years, deep learning based uni-modal models have improved the accuracy and efficiency of glaucoma detection, empowering doctors with tools for earlier diagnosis, better monitoring, and timely treatment. Building on this, multi-modal models have emerged, leveraging the strengths of different imaging modalities to learn richer and more robust representations, further enhancing glaucoma detection accuracy. However, multi-modal learning faces challenges such as imbalanced and under-optimized uni-modal representations due to joint learning objectives. To address this, we propose a balanced soft mixture-experts model with three experts and load balancing loss. The performance is measured by AUC, our proposed method surpasses the performance of all uni-modal baselines, conventional multi-modal models, and current stateof- the-art balanced multi-modal models. The proposed model can be generalized to other disease detections such as diabetic retinopathy.

discussion (0)

Sign in with ORCID, Apple, or X to comment. Anyone can read and Pith papers without signing in.

Reference graph

Works this paper leans on

38 extracted references · 8 canonical work pages

  1. [1]

    https://glaucoma

    Glaucoma Research Foundation: Glaucoma Facts and Stats. https://glaucoma. 14 org/articles/glaucoma-facts-and-stats. Accessed: 2025-07-09; Last updated March 12, 2025 (2025)

    2025

  2. [2]

    https://www.mayoclinic.org/ diseases-conditions/glaucoma/symptoms-causes/syc-20372839

    Mayo Clinic: Glaucoma: Symptoms and Causes. https://www.mayoclinic.org/ diseases-conditions/glaucoma/symptoms-causes/syc-20372839. Accessed: 2025- 07-09 (2025)

    2025

  3. [3]

    Survey of Ophthalmology53(Suppl 1), 17–36 (2008) https://doi.org/10.1016/j.survophthal.2007.11.008

    Sharma, P., Sample, P.A., Zangwill, L.M., Schuman, J.S.: Diagnostic tools for glaucoma detection and management. Survey of Ophthalmology53(Suppl 1), 17–36 (2008) https://doi.org/10.1016/j.survophthal.2007.11.008 . Review article; accessed via PubMed Central

    work page doi:10.1016/j.survophthal.2007.11.008 2008

  4. [4]

    Ophthalmology126(12), 1627–1639 (2019) https: //doi.org/10.1016/j.ophtha.2019.07.024

    Phene, S., Dunn, R.C., Hammel, N., Liu, Y., Krause, J., Kitade, N., Schaek- ermann, M., Sayres, R., Singh, A., Easterbrook, E.B., Wu, D.J., Bora, A., Semturs, C., Misra, A., Huang, A.E., Spitze, A., Medeiros, F.A., Maa, A.Y., Gandhi, M., Corrado, G.S., Peng, L., Webster, D.R.: Deep learning and glaucoma specialists: The relative importance of optic disc f...

    work page doi:10.1016/j.ophtha.2019.07.02 2019

  5. [5]

    Scientific Reports8(1), 16685 (2018) https://doi.org/10.1038/s41598-018-35044-9

    Christopher, M., Belghith, A., Bowd, C., Proudfoot, J.A., Goldbaum, M.H., Weinreb, R.N., Girkin, C.A., Liebmann, J.M., Zangwill, L.M.: Performance of deep learning architectures and transfer learning for detecting glaucomatous optic neuropathy in fundus photographs. Scientific Reports8(1), 16685 (2018) https://doi.org/10.1038/s41598-018-35044-9

    work page doi:10.1038/s41598-018-35044-9 2018

  6. [6]

    PLOS ONE13(12), 0207982 (2018) https://doi.org/10.1371/journal.pone.0207982

    Mahmood, F., Sandhu, H., Garrido, I., Plant, G.T., Gupta, V., Teasdale, S., Lawrenson, J.G., Biousse, V., Newman, J.: Automated glaucoma diagnosis using multi-modal imaging and deep learning. PLOS ONE13(12), 0207982 (2018) https://doi.org/10.1371/journal.pone.0207982

    work page doi:10.1371/journal.pone.0207982 2018

  7. [7]

    IEEE Transactions on Medical Imaging39(2), 413–424 (2020) https://doi.org/ 10.1109/TMI.2019.2927226

    Li, L., Xu, M., Liu, H., Li, Y., Wang, X., Jiang, L., Wang, Z., Fan, X., Wang, N.: A large-scale database and a cnn model for attention-based glaucoma detection. IEEE Transactions on Medical Imaging39(2), 413–424 (2020) https://doi.org/ 10.1109/TMI.2019.2927226

    arXiv 2020

  8. [8]

    American Journal of Ophthalmology (2025)

    Hwang, D.K., al.: Multimodal neural networks for glaucoma detection using fun- dus, oct, and visual fields. American Journal of Ophthalmology (2025). Published online ahead of print via PubMed

    2025

  9. [9]

    In: Proc

    Du, J., Cang, Y., Zhou, T., Hu, J., He, W.: Deep learning with hm-vgg: Ai strategies for multi-modal image analysis. In: Proc. 3rd International Sympo- sium on Sensor Technology and Control (ISSTC) (2024). https://doi.org/10.1109/ ISSTC63573.2024.10824093

    arXiv 2024

  10. [10]

    arXiv preprintarXiv:2310.02492(2023)

    Luo, Y., Khan, M.O., Tian, Y., Shi, M., Dou, Z., Elze, T., Fang, Y., Wang, M.: Fairvision: Equitable deep learning for eye disease screening via fair identity 15 scaling. arXiv preprintarXiv:2310.02492(2023). Version v3 (revised 12 April 2024), cs.CV

    Pith/arXiv arXiv 2023

  11. [11]

    arXiv preprint arXiv:2407.08813 (2024) [eess.IV]

    Tian, Y., Wen, C., Shi, M., Afzal, M.M., Huang, H., Khan, M.O., Luo, Y., Fang, Y., Wang, M.: Fairdomain: Achieving fairness in cross-domain medical image segmentation and classification. arXiv preprint arXiv:2407.08813 (2024) [eess.IV]. Version v2, last revised 18 July 2024

    Pith/arXiv arXiv 2024

  12. [12]

    IEEE Transactions on Medical Imaging 43(7), 2623–2633 (2024) https://doi.org/10.1109/TMI.2024.3377552

    Luo, Y., Tian, Y., Shi, M., Pasquale, L.R., Shen, L.Q., Zebardast, N., Elze, T., Wang, M.: Harvard glaucoma fairness: A retinal nerve disease dataset for fairness learning and fair identity normalization. IEEE Transactions on Medical Imaging 43(7), 2623–2633 (2024) https://doi.org/10.1109/TMI.2024.3377552

    arXiv 2024

  13. [13]

    In: Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition (CVPR) (2017)

    Huang, G., Liu, Z., Maaten, L., Weinberger, K.Q.: Densely connected convolu- tional networks. In: Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition (CVPR) (2017)

    2017

  14. [14]

    In: Proceedings of the 36th International Conference on Machine Learning (ICML), pp

    Tan, M., Le, Q.V.: Efficientnet: Rethinking model scaling for convolutional neu- ral networks. In: Proceedings of the 36th International Conference on Machine Learning (ICML), pp. 6105–6114. PMLR, ??? (2019)

    2019

  15. [15]

    In: Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition (CVPR), pp

    He, K., Zhang, X., Ren, S., Sun, J.: Deep residual learning for image recogni- tion. In: Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition (CVPR), pp. 770–778 (2016)

    2016

  16. [16]

    In: International Conference on Learning Representations (ICLR) (2021)

    Dosovitskiy, A., Beyer, L., Kolesnikov, A., Weissenborn, D., Zhai, X., Unterthiner, T., Dehghani, M., Minderer, M., Heigold, G., Gelly, S., Uszkoreit, J., Houlsby, N.: An image is worth 16x16 words: Transformers for image recognition at scale. In: International Conference on Learning Representations (ICLR) (2021). arXiv:2010.11929

    Pith/arXiv arXiv 2021

  17. [17]

    IEEE Transactions on Pattern Analysis and Machine Intelligence (2024) https://doi.org/10.1109/TPAMI.2024.3468315

    Wei, Y., Hu, D., Du, H., Wen, J.-R.: On-the-fly modulation for balanced multi- modal learning. IEEE Transactions on Pattern Analysis and Machine Intelligence (2024) https://doi.org/10.1109/TPAMI.2024.3468315 . Early access

    arXiv 2024

  18. [18]

    In: Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR), pp

    Peng, X., Wei, Y., Deng, A., Wang, D., Hu, D.: Balanced multimodal learning via on-the-fly gradient modulation. In: Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR), pp. 8238–8247 (2022)

    2022

  19. [19]

    In: Proceedings of the Thirty-Eighth Annual Conference on Neural Information Processing Systems (NeurIPS) (2024)

    Guo, Z., Jin, T., Chen, J., Zhao, Z.: Classifier-guided gradient modulation for enhanced multimodal learning. In: Proceedings of the Thirty-Eighth Annual Conference on Neural Information Processing Systems (NeurIPS) (2024)

    2024

  20. [20]

    In: Proceedings of the Multi-Modal Learning and Applications Workshop (MULA), CVPR (2023)

    Xue, Z., Marculescu, R.: Dynamic multimodal fusion. In: Proceedings of the Multi-Modal Learning and Applications Workshop (MULA), CVPR (2023)

    2023

  21. [21]

    Translational Vision Science & Technology9(2), 12 (2020) https://doi.org/10.1167/tvst.9.2.12

    Russakoff, D.B., Mannil, S.S., Oakley, J.D.,et al.: A 3d deep learning system 16 for detecting referable glaucoma using full oct macular cube scans. Translational Vision Science & Technology9(2), 12 (2020) https://doi.org/10.1167/tvst.9.2.12

    work page doi:10.1167/tvst.9.2.12 2020

  22. [22]

    Ophthalmology127(3), 346–356 (2020) https://doi.org/10.1016/j.ophtha.2019.09.036

    Christopher, M., Bowd, C., Belghith, A.,et al.: Deep learning approaches predict glaucomatous visual field damage from oct optic nerve head en face images and retinal nerve fiber layer thickness maps. Ophthalmology127(3), 346–356 (2020) https://doi.org/10.1016/j.ophtha.2019.09.036

    work page doi:10.1016/j.ophtha.2019.09.036 2020

  23. [23]

    Frontiers in Neuroscience16, 939472 (2022) https://doi.org/10.3389/fnins.2022.939472

    Huang, L.,et al.: Deep learning with fundus and visual field fusion. Frontiers in Neuroscience16, 939472 (2022) https://doi.org/10.3389/fnins.2022.939472

    arXiv 2022

  24. [24]

    arXiv preprint arXiv:2311.08032 (2023)

    Li, J., Pun, C.M.: ELF: End-to-end local-global fusion of OCT and fundus for glaucoma grading. arXiv preprint arXiv:2311.08032 (2023). Version 1, cs.CV

    Pith/arXiv arXiv 2023

  25. [25]

    arXiv preprint arXiv:2201.03795 (2022)

    Cai, Y., et al.: COROLLA: Contrastive learning for multimodal glaucoma detection. arXiv preprint arXiv:2201.03795 (2022)

    Pith/arXiv arXiv 2022

  26. [26]

    Scientific Reports13, 2254 (2023) https://doi.org/10.1038/s41598-022-27045-6

    Li, Y.,et al.: GMNNnet: A metadata-aware glaucoma classification framework. Scientific Reports13, 2254 (2023) https://doi.org/10.1038/s41598-022-27045-6

    work page doi:10.1038/s41598-022-27045-6 2023

  27. [27]

    In: Advances in Neural Information Processing Systems

    Nguyen, D.M.H., Nguyen, H., Diep, N.T., Pham, T.N., Cao, T., Nguyen, B.T., Swoboda, P., Ho, N., Albarqouni, S., Xie, P., Sonntag, D., Niepert, M.: L VM-Med: Learning large-scale self-supervised vision models for medical imaging via second- order graph matching. In: Advances in Neural Information Processing Systems. NeurIPS, vol. 36, pp. 27922–27950 (2023)

    2023

  28. [28]

    Computation13(4), 88 (2025) https://doi.org/10.3390/computation13040088

    Gupta, A., Osman, I., Shehata, M.S., Braun, W..J., Feldman, R.E.: MedMAE: A self-supervised backbone for medical imaging tasks. Computation13(4), 88 (2025) https://doi.org/10.3390/computation13040088

    work page doi:10.3390/computation13040088 2025

  29. [29]

    In: Medical Imaging with Deep Learning (MIDL), pp

    Sowrirajan, H., Yang, J., Ng, A.Y., Rajpurkar, P.: MoCo-CXR: Moco pretraining improves representation and transferability of chest x-ray models. In: Medical Imaging with Deep Learning (MIDL), pp. 728–744 (2021)

    2021

  30. [30]

    Nature622, 156– 163 (2023) https://doi.org/10.1038/s41586-023-06555-x

    Zhou, Y., Chia, M.A., Wagner, S.K., Ayhan, M.S., Williamson, D.J., Struyven, R.R., Liu, T., Xu, M., Lozano, M.G., Woodward-Court, P.,et al.: A foundation model for generalizable disease detection from retinal images. Nature622, 156– 163 (2023) https://doi.org/10.1038/s41586-023-06555-x

    doi:10.1038/s41586-023-06555-x 2023

  31. [31]

    In: Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR), pp

    Zanella, M., Ben Ayed, I.: Low-rank few-shot adaptation of vision-language mod- els. In: Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR), pp. 1593–1603 (2024)

    2024

  32. [32]

    arXiv preprint arXiv:1701.06538 (2017) 17

    Shazeer, N., Mirhoseini, A., Maziarz, P., Davis, A., Le, Q., Hinton, G., Dean, J.: Outrageously large neural networks: The sparsely-gated mixture-of-experts layer. arXiv preprint arXiv:1701.06538 (2017) 17

    Pith/arXiv arXiv 2017

  33. [33]

    IEEE Transactions on Pattern Analysis and Machine Intelligence45(9), 10795–10814 (2023) https://doi.org/10.1109/TPAMI.2022.3201094

    Yan, J., Xie, J., He, X., Li, X., Zha, H.: Mixture of experts in deep learning: A survey. IEEE Transactions on Pattern Analysis and Machine Intelligence45(9), 10795–10814 (2023) https://doi.org/10.1109/TPAMI.2022.3201094

    arXiv 2023

  34. [34]

    IEEE Transactions on Pattern Analysis and Machine Intel- ligence41(2), 423–443 (2019) https://doi.org/10.1109/TPAMI.2018.2798607

    Baltruˇ saitis, T., Ahuja, C., Morency, L.-P.: Multimodal machine learning: A survey and taxonomy. IEEE Transactions on Pattern Analysis and Machine Intel- ligence41(2), 423–443 (2019) https://doi.org/10.1109/TPAMI.2018.2798607

    arXiv 2019

  35. [35]

    In: Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR), pp

    Tan, M., Le, Q.V.: MnasNet: Platform-aware neural architecture search for mobile. In: Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR), pp. 2820–2828 (2019)

    2019

  36. [36]

    In: Proceedings of the IEEE/CVF Con- ference on Computer Vision and Pattern Recognition (CVPR), pp

    Sandler, M., Howard, A., Zhu, M., Zhmoginov, A., Chen, L.-C.: MobileNetV2: Inverted residuals and linear bottlenecks. In: Proceedings of the IEEE/CVF Con- ference on Computer Vision and Pattern Recognition (CVPR), pp. 4510–4520 (2018)

    2018

  37. [37]

    arXiv preprint arXiv:1704.04861 (2017)

    Howard, A.G., Zhu, M., Chen, B., Kalenichenko, D., Wang, W., Weyand, T., Andreetto, M., Adam, H.: MobileNets: Efficient convolutional neural networks for mobile vision applications. arXiv preprint arXiv:1704.04861 (2017)

    Pith/arXiv arXiv 2017

  38. [38]

    In: Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR), pp

    Hu, J., Shen, L., Sun, G.: Squeeze-and-excitation networks. In: Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR), pp. 7132–7141 (2018) 18

    2018