Pith. sign in

REVIEW 3 major objections 4 minor 31 references

Enhancing Automatic Modulation Recognition With a Reconstruction-Driven Vision Transformer Under Limited Labels

T0 review · 3 major / 4 minor · reviewed 2026-08-05 · deepseek-v4-flash

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

Pith's one-line read A reconstruction-only pretraining objective, rather than contrastive or joint objectives, gives a Vision Transformer the best accuracy on automatic modulation recognition with 10–20% labeled data, approaching fully supervised ResNet perform

desk verdict Plausible idea and a legitimate empirical question, but the dataset construction is arithmetically impossible and the reported numbers don't hang together; the evidence as presented is not usable. read the letter →

arxiv 2508.20193 v2 pith:6MXELFQS submitted 2025-08-27 cs.CV eess.SP

Hossein Ahmadi , Banafsheh Saffari , Sajjad Emdadi Mahdimahalleh , Mohammad Esmaeil Safari , Aria Ahmadi This is my paper · ORCID
classification cs.CVeess.SP
keywords automaticmodulationrecognitionVisionTransformersemi-supervisedlearningself-supervisedpretrainingreconstructionlossI/QsignalslabelefficiencySNRrobustness
verification ladder T0 review T1 audit T2 compute T3 formal

The pith

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

The reading

This paper tries to show that a Vision Transformer can do automatic modulation recognition with very few labels if its pretraining objective is reconstruction rather than contrastive invariance. The central question is whether forcing the encoder to map augmented I/Q signals back to their original form teaches features that transfer to classification under data scarcity. On the RML2018.01A benchmark, the paper finds that reconstruction-only pretraining beats both contrastive-only and joint reconstruction+contrastive objectives at 10%, 15%, and 20% labels, and that a semi-supervised ViT with 15% labels approaches a fully supervised ResNet. If true, this gives wireless systems a single-stage recipe for label-efficient spectrum monitoring.

What carries the argument

The reconstruction-to-original branch: a lightweight convolutional decoder maps augmented I/Q patches back to the original signal, while the ViT encoder's patch embeddings are averaged (rather than using a class token) for both reconstruction projection and classification. This forces the encoder to preserve fine phase and amplitude structure instead of learning only contrastive invariance, and the reconstruction loss can be trained jointly with or without a contrastive loss, letting the paper isolate the contribution of reconstruction.

What would settle it

Re-run Scenario 1 (reconstruction-only) and the baselines on a subset with an explicitly documented sample count, for instance 500 samples per class per SNR for a stated set of integer SNR levels, and check whether Table 4's ordering and the 71.0% figure at 15% labels reproduce. A second check is to zero out the reconstruction loss while keeping the decoder and augmentations, which should lower low-label accuracy if reconstruction is the active ingredient.

Watch

Extended reading notes

Core claim

On RML2018.01A with 16 modulation classes and SNR −2 to +20 dB, a ViT pretrained by reconstructing augmented I/Q signals back to their originals, then fine-tuned on 10–20% of labels, reaches 68.2–73.7% accuracy. At 15% labels it reaches 71.0%, beating a fully supervised CNN (61.8%) and a fully supervised ViT (68.2%) and approaching a fully supervised ResNet (78.5%). Reconstruction-only pretraining outperforms reconstruction+contrastive and contrastive-only across every tested label ratio, with the largest gains on high-order QAM and APSK classes, where contrastive-only learning largely collapses.

Load-bearing premise

The load-bearing premise is that the RML2018.01A subset is built exactly as described; Section 5.2 claims 1,000 samples per class per SNR across 16 classes and a total of 220,000, which no integer number of SNR levels in the stated −2 to +21 dB range can produce, so the true subsampling is the assumption on which all reported accuracies rest.

Editorial extensions

If this is right

  • Practitioners can drop the contrastive branch: reconstruction-only pretraining outperforms reconstruction+contrastive and contrastive-only at every label ratio (68.2/71.0/73.7% vs 66.4/70.2/71.4% and 53.4/54.4/56.4%).
  • Label budgets for AMR can shrink to 10–20%: with 15% labels the semi-supervised ViT reaches 71.0%, above the fully supervised CNN (61.8%) and ViT (68.2%) and within about 7.5 points of the fully supervised ResNet (78.5%).
  • High-order constellations are where reconstruction helps most: at 15% labels, reconstruction-only beats contrastive-only by large margins on 32APSK, 64APSK, 128APSK, 16QAM, 32QAM, and 128QAM.
  • The single-stage pipeline avoids teacher–student architectures and multi-stage pretraining, simplifying deployment for cognitive radio and spectrum monitoring.
  • SNR robustness is preserved: the framework stays above roughly 70% accuracy for SNR ≥ 5 dB, so the low-label gains are not limited to clean signals.

Reading between the lines

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

  • The reconstruction-to-original objective can be read as denoising in I/Q space: it likely forces the encoder to keep phase and amplitude detail that rotation and flip contrastive views discard. A direct test would be ablating each augmentation family to see which one drives the low-label gain.
  • The same pretraining should transfer across I/Q datasets better than contrastive invariance; evaluating on a different modulation benchmark or on hardware-varied channel data would test this directly.
  • A masked-autoencoder variant that reconstructs only masked I/Q patches, rather than full augmented signals, is a natural next step that could push the 10%-label regime lower.
  • The absolute accuracy figures are likely tied to the exact class/SNR balance, so comparing them across papers requires reproducing that balance; the paper's own stated counts are internally inconsistent on this point.
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

3 major / 4 minor

Summary. The paper proposes a unified ViT-based framework for automatic modulation recognition (AMR) that combines supervised classification with self-supervised reconstruction and, optionally, contrastive learning. A ViT encoder processes I/Q signals as patches; a lightweight convolutional decoder reconstructs augmented signals back to their original forms; a linear head on mean-pooled patch embeddings performs classification. The authors claim that reconstruction-only pretraining gives the strongest low-label performance, that the semi-supervised ViT with 15% labels outperforms fully supervised CNN and ViT baselines and approaches ResNet accuracy, and that the method is robust across SNR levels on RML2018.01A.

Significance. If the empirical claims were reliable, the paper would make a useful, simple contribution to low-label AMR: the architecture is clearly described, avoids teacher-student complexity, and the reconstruction-to-original anchoring is a sensible pretext task. The pseudo-label self-training loop does not touch the held-out test evaluation, so there is no derivation-level circularity. Unfortunately, the experimental validation is undermined by multiple internal inconsistencies in the dataset construction and in the reported numbers; no code or data-release link is provided. As a result, the central empirical claims cannot currently be accepted.

major comments (3)
  1. [Section 5.2, Table 4 caption, Section 5.3.5, Figure 4 caption] The dataset construction is arithmetically impossible as stated. Section 5.2 says 16 classes, 1,000 samples per class per SNR, and a total of 220,000 signals. That would require 13.75 SNR levels. The SNR range is given inconsistently as -2 to +21 dB (Section 5.2), -2 to +20 dB (Table 4 caption and Section 5.3), and -2 to +22 dB (Figure 4 caption); no integer number of SNR levels in any of these ranges gives 220,000 total samples (e.g., 24 levels would give 384,000). The 70/10/20 split implies a test set of about 44,000 samples, but Section 5.3.5 reports about 55,000 test samples, and Section 5.3.6 reports 700 test samples per SNR. These numbers are mutually incompatible. Because every reported accuracy depends on the actual subsample and split, this is a load-bearing error: the results are attached to an undefined dataset.
  2. [Tables 4 and 5] The value 68.21 appears twice: as Scenario 1 (Reconstruction Only) with 10% labels in Table 4, and as the fully supervised ViT with 100% labels in Table 5. The paper's central claim is label efficiency, so a fully supervised model using ten times as many labels should not produce an identical overall accuracy to the 10%-label model. This strongly suggests a table or computation mix-up and further erodes confidence in the other reported numbers.
  3. [Section 5.3, Tables 4 and 5] The claim that the proposed model 'outperforms supervised CNN and ViT baselines in low-label regimes' is not supported by the reported experiments. Table 5 gives supervised CNN, ResNet, and ViT results only at 100% labels. Table 4 compares three pretraining scenarios against each other at 10%, 15%, and 20% labels but contains no CNN, ResNet, or supervised ViT baselines at those same label fractions. Without matching-label-fraction baseline comparisons, the headline low-label superiority claim is not demonstrated.
minor comments (4)
  1. [Figure 3] The t-SNE visualization is explicitly for 10 modulation classes, while the experiments use 16 classes. The claim of well-separated clusters under low-label regimes is therefore only verified on a subset; please provide the full 16-class visualization or qualify the claim.
  2. [Section 5.1 / 5.3] The SNR range is stated inconsistently across the paper. Please unify: Section 5.1 says -2 to +21 dB, Section 5.3 and Table 4 say -2 to +20 dB, and Figure 4 says -2 to +22 dB.
  3. [Section 5.2 / Related Work] No code or data-release link is provided, so the reported numbers cannot be externally audited. Also, reference [22] is incompletely formatted ('A. names not provided in abstract'), and Section 2.4 includes several citations that are not clearly connected to the AMR problem; please tighten.
  4. [Section 5.3.2 / Table 5] The text 'our ViT-based framework outperforms both CNN and ResNet baselines across all labeling ratios' is stronger than the data shown; Table 5 has only 100%-label baselines, and the semi-supervised ViT (71.01%) is below supervised ResNet (78.50%). Please align the wording with what is actually tabulated.

Circularity Check

0 steps flagged · score 0.0 of 10

No significant circularity: the central claims are empirical and evaluated on a held-out test set; the dataset arithmetic inconsistency is a reporting/correctness issue, not a circularity.

full rationale

Walking the paper's derivation chain, the central claims are empirical comparisons on the external RML2018.01A benchmark with a 70/10/20 train/validation/test split and five repeated shuffles. The reconstruction-only pretraining result is an experimental observation, not a definitional consequence of the loss formulation. No fitted parameter is renamed as a prediction, and no uniqueness theorem or prior self-citation is invoked to force the architectural choice. The pseudo-label self-training loop uses the model's own confident predictions on unlabeled training data, but final accuracies are computed on a disjoint test set, so the evaluation is not circular. Self-citations [14] and [16] appear only in related-work context and are not load-bearing for the main claims. The arithmetic inconsistency in §5.2 (16 classes × 1,000 samples/class/SNR cannot total 220,000 for the stated SNR range) and the duplicated 68.21% in Tables 4 and 5 are serious correctness/auditability concerns, but they do not constitute circularity: they do not make any reported result equivalent to its input by construction. The paper is self-contained against an external benchmark, so no circular step can be identified.

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

The framework rests on several hand-chosen hyperparameters (pseudo-label threshold/weight, loss weights) and on the unverified assumption that reconstruction of augmentations is a good pretext task. The data sampling itself is inconsistently described, making the numerical basis fragile.

free parameters (4)
  • Pseudo-label confidence threshold = 0.8
    Used to decide which unlabeled samples enter the classification loss; no ablation reported.
  • Pseudo-label weight = 0.5
    Weight for pseudo-labeled loss term; hand-chosen, no sensitivity analysis.
  • Loss balancing weights (alpha, beta in Eq. 5)
    Scalar weights for reconstruction, contrastive, and classification losses are mentioned but not specified; presumably tuned.
  • Data subsampling (samples per class per SNR) = 1,000; total 220,000
    The stated selection is arithmetically inconsistent with 16 classes and SNR range -2 to +20 dB, leaving the exact data construction unknown.
assumptions (4)
  • domain assumption Deep neural networks can extract discriminative features from I/Q time series
    Standard premise in AMR deep learning, not verified in this paper.
  • domain assumption The selected augmentations (rotation, flip, noise, time/magnitude warping) preserve modulation class labels
    Plausible but not verified; aggressive warping may distort constellations.
  • ad hoc to paper Reconstruction of augmented signals to originals is a useful pretext task for classification
    This is the core design assumption that the paper does not justify theoretically.
  • domain assumption RML2018.01A is a valid benchmark for AMR
    Widely used dataset, but the paper restricts a subset without clear justification.

how reviews work

0 comments
Cite this review

Pith. "Pith review of Enhancing Automatic Modulation Recognition With a Reconstruction-Driven Vision Transformer Under Limited Labels." pith.science (2026). https://pith.science/paper/6MXELFQS

@misc{pith2026250820193,
  author       = {Pith},
  title        = {Pith review of: Enhancing Automatic Modulation Recognition With a Reconstruction-Driven Vision Transformer Under Limited Labels},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/6MXELFQS}},
  note         = {Machine review of arXiv:2508.20193}
}
read the original abstract

Automatic modulation recognition (AMR) is critical for cognitive radio, spectrum monitoring, and secure wireless communication. However, existing solutions often rely on large labeled datasets or multi-stage training pipelines, which limit scalability and generalization in practice. We propose a unified Vision Transformer (ViT) framework that integrates supervised, self-supervised, and reconstruction objectives. The model combines a ViT encoder, a lightweight convolutional decoder, and a linear classifier; the reconstruction branch maps augmented signals back to their originals, anchoring the encoder to fine-grained I/Q structure. This strategy promotes robust, discriminative feature learning during pretraining, while partial label supervision in fine-tuning enables effective classification with limited labels. On the RML2018.01A dataset, our approach outperforms supervised CNN and ViT baselines in low-label regimes, approaches ResNet-level accuracy with only 15-20% labeled data, and maintains strong performance across varying SNR levels. Overall, the framework provides a simple, generalizable, and label-efficient solution for AMR.

Figures

Figures reproduced from arXiv: 2508.20193 by the authors.

Figure 1
Figure 1. Overview of the main model architecture. [PITH_FULL_IMAGE:figures/full_fig_p009_1.png] view at source ↗
Figure 2
Figure 2. Building on these classical methods, Pi et al. [31] proposed augmenta￾tion techniques that specifically target temporal signal dynamics. Scaling adjusts the amplitude of the signal using a Gaussian-distributed scalar mul￾tiplier, while magnitude warping applies smooth nonlinear distortions gener￾ated through cubic spline interpolation. Time warping, on the other hand, compresses or stretches the signal along the tem… view at source ↗
Figure 2
Figure 2. Examples of original and augmented I/Q signals: (a) Original, (b) [PITH_FULL_IMAGE:figures/full_fig_p011_2.png] view at source ↗
Figures from the paper (2 more)
Figure 3
Figure 3. Figure 3: t-SNE visualization of Scenario 1 (reconstruction-only pretraining) [PITH_FULL_IMAGE:figures/full_fig_p017_3.png]
Figure 4
Figure 4. Figure 4: Classification accuracy of the proposed model across different SNR [PITH_FULL_IMAGE:figures/full_fig_p019_4.png]

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

31 extracted references · 30 canonical work pages

  1. [1]

    A deep learning method based on convolutional neural network for automatic modulation classi- fication of wireless signals,

    Y. Xu, D. Li, Z. Wang, Q. Guo, and W. Xiang, “A deep learning method based on convolutional neural network for automatic modulation classi- fication of wireless signals,”Wireless Networks, vol. 25, no. 7, pp. 3735– 3746, 2019. 20

    work page 2019

  2. [2]

    Automatic mod- ulation classification using cnn with features fusion of spwvd and bjd,

    Z. Zhang, C. Wang, C. Gan, S. Sun, and M. Wang, “Automatic mod- ulation classification using cnn with features fusion of spwvd and bjd,” IEEE Transactions on Signal and Information Processing over Net- works, vol. 5, no. 3, pp. 469–481, 2019

    work page 2019

  3. [3]

    A transformer-based contrastive semi-supervised learning framework for automatic modula- tionrecognition,

    W. Kong, X. Jiao, Y. Xu, B. Zhang, and Q. Yang, “A transformer-based contrastive semi-supervised learning framework for automatic modula- tionrecognition,”IEEE Transactions on Cognitive Communications and Networking, vol. 9, no. 4, pp. 950–963, 2023

    work page 2023

  4. [4]

    Over-the-air deep learning based radio signal classification,

    T. J. O’Shea, T. Roy, and T. C. Clancy, “Over-the-air deep learning based radio signal classification,”IEEE Journal of Selected Topics in Signal Processing, vol. 12, no. 1, pp. 168–179, 2018

    2018

  5. [5]

    Adversarial transfer learning for deep learning based automatic modulation classification,

    K. Bu, Y. He, X. Jing, and J. Han, “Adversarial transfer learning for deep learning based automatic modulation classification,”IEEE Signal Processing Letters, vol. 27, pp. 880–884, 2020

    work page 2020

  6. [6]

    Transfer learning for semi-supervised amc in zf-mimo systems,

    Y. Wang, G. Gui, H. Gacanin, T. Ohtsuki, H. Sari, and F. Adachi, “Transfer learning for semi-supervised amc in zf-mimo systems,”IEEE Journal on Emerging and Selected Topics in Circuits and Systems, vol. 10, no. 2, pp. 231–243, 2020

    work page 2020

  7. [7]

    Self-contrastive learn- ing based semi-supervised radio modulation classification,

    D. Liu, P. Wang, T. Wang, and T. Abdelzaher, “Self-contrastive learn- ing based semi-supervised radio modulation classification,” inMILCOM 2021 - Special Session on Internet of Battlefield Things, IEEE, 2021

    work page 2021

  8. [8]

    Semi-supervised modula- tion classification via an ensemble sigmatch method,

    H. Wang, S. Yang, Z. Feng, and B. Huang, “Semi-supervised modula- tion classification via an ensemble sigmatch method,”IEEE Internet of Things Journal, vol. 11, no. 20, pp. 32985–32997, 2024

    work page 2024

Show all 31 references
  1. [9]

    Sscl-amc: A self-supervised automatic modulation classification method via dynamic augmentation and ensemble learning,

    Y. Cai, D. Li, S. Wu, M. Shao, S. Hong, and H. Sun, “Sscl-amc: A self-supervised automatic modulation classification method via dynamic augmentation and ensemble learning,” inICASSP 2025 - IEEE Inter- national Conference on Acoustics, Speech and Signal Processing, IEEE, 2025

    2025

  2. [10]

    Modulation classification with data augmentation based on a semi-supervised generative model,

    L. Yin, X. Xiang, Y. Liang, and K. Liu, “Modulation classification with data augmentation based on a semi-supervised generative model,”Wire- less Networks, vol. 30, pp. 5683–5696, 2024

    2024

  3. [11]

    Contrastive semi- supervised learning with pseudo-label for radar signal automatic mod- 21 ulation recognition,

    D. Wu, J. Shi, Z. Li, M. Du, F. Liu, and F. Zeng, “Contrastive semi- supervised learning with pseudo-label for radar signal automatic mod- 21 ulation recognition,”IEEE Sensors Journal, vol. 24, no. 19, pp. 30399– 30410, 2024

    2024

  4. [12]

    Augmented semi-supervised learning for cnn based automatic modulation classification,

    H. Liu and Z. Zhu, “Augmented semi-supervised learning for cnn based automatic modulation classification,” inIEEE ICCC, IEEE, 2022

    2022

  5. [13]

    Meta supervised contrastive learning for few-shot open-set modulation classification with signal con- stellation,

    J. Zhao, H. Wang, S. Peng, and Y.-D. Yao, “Meta supervised contrastive learning for few-shot open-set modulation classification with signal con- stellation,”IEEE Communications Letters, vol. 28, no. 4, pp. 837–841, 2024

    2024

  6. [14]

    Unsuper- visedtime-seriessignalanalysiswithautoencodersandvisiontransform- ers: A review of architectures and applications,

    H. Ahmadi, S. E. Mahdimahalleh, A. Farahat, and B. Saffari, “Unsuper- visedtime-seriessignalanalysiswithautoencodersandvisiontransform- ers: A review of architectures and applications,”Journal of Intelligent Learning Systems and Applications, vol. 17, no. 2, pp. –, 2025

    2025

  7. [15]

    Gaf-mae: A self-supervised automatic modulation classification method based on gramian angular field and masked autoencoder,

    Y. Shi, H. Xu, Y. Zhang, Z. Qi, and D. Wang, “Gaf-mae: A self-supervised automatic modulation classification method based on gramian angular field and masked autoencoder,”IEEE Transactions on Cognitive Communications and Networking, vol. 10, no. 1, pp. 94–106, 2024

    2024

  8. [16]

    Multiheart: Secure and ro- bust heartbeat pattern recognition in multimodal cardiac monitoring system,

    H. Ahmadi, Y. Zhang, and N. H. Tran, “Multiheart: Secure and ro- bust heartbeat pattern recognition in multimodal cardiac monitoring system,”Electronics, vol. 14, no. 15, p. 3149, 2025

    2025

  9. [17]

    Spectrum interference-based two-level data augmentation method in deep learning for automatic modulation classification,

    Q. Zheng, P. Zhao, Y. Li, H. Wang, and Y. Yang, “Spectrum interference-based two-level data augmentation method in deep learning for automatic modulation classification,”Neural Computing and Appli- cations, vol. 33, no. 15, pp. 7723–7745, 2021

    2021

  10. [18]

    Modulation classifier: A few-shot learning semi-supervised method based on multimodal infor- mation and domain adversarial network,

    W. Deng, X. Wang, Z. Huang, and Q. Xu, “Modulation classifier: A few-shot learning semi-supervised method based on multimodal infor- mation and domain adversarial network,”IEEE Communications Let- ters, vol. 27, no. 2, pp. 576–580, 2023

    2023

  11. [19]

    Soamc: A semi-supervised open-set recognition algorithm for automatic modulation classification,

    C. Di, J. Ji, C. Sun, and L. Liang, “Soamc: A semi-supervised open-set recognition algorithm for automatic modulation classification,”Elec- tronics, vol. 13, no. 21, p. 4196, 2024

    2024

  12. [20]

    Sanitizing manufacturing dataset labels using vision-language models,

    N. Mahjourian and V. Nguyen, “Sanitizing manufacturing dataset labels using vision-language models,”arXiv preprint arXiv:2506.23465, 2025. 22

    2025 arXiv

  13. [21]

    Classimbalance-awareactivelearningwithvisiontransformersinfeder- ated histopathological imaging,

    M. A. L. Khankiki, M. Mirzaeibonehkhatker, and S. Esfandiari Fard, “Classimbalance-awareactivelearningwithvisiontransformersinfeder- ated histopathological imaging,”Journal of Medicine and Medical Stud- ies, vol. 1, no. 2, pp. 65–73, 2025. Published: 07 May, 2025

    2025

  14. [22]

    Real-time wearable sensor compli- ance monitoring using a two-stage gradient-boosting classifier,

    A. names not provided in abstract, “Real-time wearable sensor compli- ance monitoring using a two-stage gradient-boosting classifier,” in2023 International Conference on Electrical, Computer and Energy Technolo- gies (ICECET), (Cape Town, South Africa), pp. 1–6, IEEE, Nov 2023...

    2023

  15. [23]

    Fine-grained open set signal modulation classification via self-supervised pre-training,

    Z. Feng, H. Pei, S. Yang, and C. Yang, “Fine-grained open set signal modulation classification via self-supervised pre-training,”IEEE Trans- actions on Cognitive Communications and Networking, 2025

    2025

  16. [24]

    A graph- based semi-supervised approach for few-shot class-incremental modula- tion classification,

    X. Zhou, P. Qi, Q. Liu, Y. Ding, S. Zheng, and Z. Li, “A graph- based semi-supervised approach for few-shot class-incremental modula- tion classification,”China Communications, vol. 21, no. 11, pp. 88–103, 2024

    2024

  17. [25]

    Network anomaly detection for iot using hyperdimensional computing on nsl-kdd,

    G. Ghajari, A. Ghimire, E. Ghajari, and F. Amsaad, “Network anomaly detection for iot using hyperdimensional computing on nsl-kdd,”arXiv preprint arXiv:2503.03031, 2025

    2025

  18. [26]

    Msmcnet: A modular few-shot learning framework for signal modulation classification,

    Y. Wang, J. Bai, Z. Xiao, H. Zhou, and L. Jiao, “Msmcnet: A modular few-shot learning framework for signal modulation classification,”IEEE Transactions on Signal Processing, vol. 70, pp. 3789–3804, 2022

    2022

  19. [27]

    A novel automatic modulation classification scheme based on multi-scale networks,

    H. Zhang, F. Zhou, Q. Wu, W. Wu, and R. Q. Hu, “A novel automatic modulation classification scheme based on multi-scale networks,”IEEE Transactions on Cognitive Communications and Networking, vol. 8, no. 1, pp. 97–111, 2022

    2022

  20. [28]

    Supervised ml method based modu- lation formats classification for wdm systems,

    R. Kebaili, S. Driz, and B. Fassi, “Supervised ml method based modu- lation formats classification for wdm systems,” inISPA, 2024

    2024

  21. [29]

    Modulation format identification using su- pervised learning and high-dimensional features,

    A. K. Ali and E. Ercelebi, “Modulation format identification using su- pervised learning and high-dimensional features,”Arabian Journal for Science and Engineering, vol. 48, no. 4, pp. 1461–1486, 2023

    2023

  22. [30]

    Data augmentation for deep learning-based radio modulation classification,

    L. Huang, W. Pan, Y. Zhang, L. Qian, N. Gao, and Y. Wu, “Data augmentation for deep learning-based radio modulation classification,” IEEE Access, vol. 8, pp. 1498–1506, 2020. 23

    2020

  23. [31]

    Improving modulation recognition using time series data augmentation via a spatiotemporal multi-channel framework,

    S. Pi, S. Zhang, S. Wang, B. Guo, and W. Yan, “Improving modulation recognition using time series data augmentation via a spatiotemporal multi-channel framework,”Electronics, vol. 12, no. 1, p. 96, 2023. 24

    2023

Pith tools

Reviewed August 5, 2026 · model on record in the stance chip above.