Pith. sign in

REVIEW 4 major objections 6 minor 43 references

Noise-Aware Ensemble Learning for Efficient Radar Modulation Recognition

T0 review · 4 major / 6 minor · reviewed 2026-08-12 · 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 noise-aware ensemble framework, which inspects the gradient map of a lightweight recognizer and activates a deeper recognizer only when noise is judged to dominate, achieves the highest radar modulation recognition…

desk verdict A genuinely new gradient-map gating idea with real measured USRP data, but the noise-aware gate itself is under-specified and unvalidated, so the headline gain is not yet tied to the mechanism. read the letter →

arxiv 2411.15104 v3 pith:3THT3VFM submitted 2024-11-22 eess.SP

Do-Hyun Park , Min-Wook Jeon , Jinwoo Jeong , Isaac Sim , Sangbom Yun , Junghyun Seo , Hyoung-Nam Kim This is my paper · ORCID
classification eess.SP
keywords radarmodulationrecognitionnoise-awareensemblelearninglowprobabilityofintercepttime-frequencyimageconvolutionalneuralnetworkelectronicwarfaresupportcomputationalefficiencygradientmap
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

Radar electronic support systems must identify modulation schemes from very weak signals, but fixed lightweight CNN models either waste computation when noise is low or lose accuracy when noise is high. The paper proposes a noise-aware ensemble learning (NAEL) framework that first runs a lightweight preliminary recognition network (PRN) and then uses a noise-aware network (NAN) to decide, from where the PRN's attention falls in the time-frequency image, whether that result was corrupted by noise. If the NAN calls the decision unreliable, a deeper advanced recognition network (ARN) is invoked. On a test set of real over-the-air USRP recordings at estimated SNRs of about -4, -15, and -17 dB, the paper reports that NAEL reaches 99.1%, 55.4%, and 38.3% correct classification, the highest among all compared models, with average FLOPs of 402, 606, and 660 mega-FLOPs, below most fixed baselines. The paper's claim is that this adaptive selection makes accurate low-SNR recognition compatible with low-cost high-SNR operation.

What carries the argument

The load-bearing mechanism is the noise gate built from the PRN's gradient map. The importance weight of each feature-map channel is $w_c(k) = \frac{1}{H W} \sum_{t,f} \frac{\partial y_c}{\partial F^k(t,f)}$, and the gradient map is the ReLU-weighted sum $G_c(t,f) = \mathrm{ReLU}\left(\sum_k F^k(t,f) \, w_c(k)\right)$; $f_{\max} = \arg\max_f \sum_t G_c(t,f)$ is the frequency index used to illustrate that correct decisions concentrate activation at center-frequency indices 4 and 5. The NAN takes the whole flattened 8-by-8 map through two fully connected layers (expanding to 256 then 512) and a softmax over two classes, reliable versus unreliable. The PRN and ARN both use FE blocks built from depthwise separable convolutions with channel expansion factor $\alpha$, ReLU6 activations, and skip connections, and the ARN reuses some PRN outputs as inputs; Table IV sets their sizes. The gate converts a fixed architecture into an adaptive one: when the PRN is confident and clean, only the cheap network runs, and when the NAN detects noise-dominated attention, the deeper ARN is called in.

What would settle it

Measure the NAN's hit rate directly on the 1,200 real test signals: compare each reliable/unreliable call with whether the PRN actually got the modulation right. If the NAN is only slightly better than chance, or if its calls do not track the accuracy difference between PRN-only and ARN-refined decisions, the NAEL advantage would be coming from the ARN itself rather than from noise awareness. An ablation that replaces the NAN with an always-on ARN, a never-on ARN, or a random gate, and then compares PCC versus FLOPs, would settle whether adaptive gating is the cause.

Watch

Extended reading notes

Core claim

The central discovery the paper argues for is that the spatial distribution of a CNN's gradient map can serve as a learned indicator of whether noise has corrupted a modulation-recognition decision. The PRN classifies a Choi-Williams time-frequency image; the NAN then flattens the 8-by-8 gradient map $G_c(t,f) = \mathrm{ReLU}\left(\sum_k F^k(t,f) \, w_c(k)\right)$, where $w_c(k)$ is the global-average-pooled gradient of the winning class score with respect to channel $k$, and feeds it through fully connected layers to label the PRN output as reliable or unreliable. Correctly classified simulated signals concentrate their highest gradient activation at frequency indices near the center frequency, while misclassified ones disperse it, and the paper uses this contrast to justify the gating rule. When the gate says unreliable, the ARN, built from the same FE blocks with larger channel expansion and reusing PRN feature maps, recomputes the decision. The paper reports that this three-network ensemble beats every fixed-architecture comparison model in accuracy at all three measured SNR levels while keeping average computational cost at or below most of them.

Load-bearing premise

The load-bearing premise is that the spatial distribution of the PRN's gradient map, specifically whether high activations concentrate near the center frequency, reliably indicates whether the PRN's classification is noise-affected, and that a classifier trained on this signal with simulated data transfers to real USRP measurements. The paper never reports the NAN's own classification accuracy against actual PRN correctness, so the core gating mechanism is unvalidated.

Editorial extensions

If this is right

  • At high SNR, NAEL's average cost is 402 MFLOPs with 99.1% accuracy because the ARN fires only 20 of 1,200 times, so an ES system can get near-perfect recognition without paying for deep feature extraction.
  • At very low SNR, NAEL keeps the best accuracy, 38.3%, by calling the ARN about 57.8% of the time, while averaging 660 MFLOPs, less than every fixed baseline except LPI-Net, which reaches only 23.6% accuracy.
  • Because the three networks are trained separately on simulated data and then applied to real USRP-collected signals, the paper claims the noise gate transfers from simulation to hardware.
  • The ARN is invoked per sample rather than per SNR band, so the model adapts within a scenario to individual signal conditions instead of switching on a fixed noise regime.

Reading between the lines

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

  • The same gate could be lifted onto any base recognizer: any CNN's gradient map could feed a NAN, making the framework a general accuracy-versus-cost dial rather than a single architecture.
  • A graded gate would extend the binary reliable/unreliable choice into a confidence score, letting a deployment pick a FLOPs budget and tolerate a specified drop in accuracy.
  • A direct test of the mechanism would be to compare NAN decisions against the PRN's true correctness on the real test set; the paper does not report that accuracy, so the gating signal's strength remains an open question.
  • Because the USRP SNRs are only estimated, re-evaluating with calibrated noise injection into clean recordings would separate the framework's noise response from hardware distortion and multipath effects.
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

4 major / 6 minor

Summary. The paper proposes a noise-aware ensemble learning (NAEL) framework for radar modulation recognition. The framework consists of a lightweight preliminary recognition network (PRN), a noise-aware network (NAN) that analyzes Grad-CAM-style gradient maps to decide whether the PRN output is noise-affected, and an advanced recognition network (ARN) that is invoked when noise is detected. The authors derive a data-reuse FE-block architecture, train on simulated Choi-Williams distribution images, and evaluate on over-the-air USRP N210 measurements at three approximate SNR levels. They report probability of correct classification (PCC) and FLOPs for NAEL and for eight fixed-architecture baselines, claiming superior accuracy with lower or comparable computational cost.

Significance. The adaptive gating idea is well motivated for real-time electronic warfare support, and evaluating on real USRP measurements rather than only simulated AWGN is a concrete strength. The paper also compares against a broad set of modern lightweight baselines and describes the FE-block architecture in enough detail to be reimplemented. However, the central mechanism is currently not validated: the NAN's training labels are unspecified, its classification accuracy is never reported, and the reported accuracy advantages are small relative to the test set size. If the gating mechanism were validated and the accuracy differences shown to be statistically reliable, the framework would be a useful contribution to low-latency modulation recognition; as it stands, the evidence for the central claim is incomplete.

major comments (4)
  1. [Section III.C and IV.A] The NAN's training labels are never specified. Section IV.A states that 'each network is trained individually whereas keeping the weights of the other networks constant during training,' but it does not define the binary target ('reliable' vs 'unreliable') used to supervise the NAN, nor how those labels are obtained (e.g., from PRN correctness on simulated data at various SNRs, or from a hand-designed SNR threshold). Since the NAN is the sole component that routes inputs to the ARN, this omission makes the core mechanism unreproducible and prevents the reader from assessing whether the reported PCC/FLOPs results in Tables VI–VIII are causally produced by noise-aware gating. Please specify the label generation procedure and report the NAN's own accuracy, precision, recall, and a confusion matrix on held-out simulated and real data.
  2. [Section IV.C, Tables VI–VIII] The claimed accuracy advantage over the best fixed-architecture baselines is small: 0.3 percentage points at high SNR (Table VI), 0.9 points at low SNR (Table VII), and 2.4 points at very low SNR (Table VIII). With 1,200 test signals per scenario, the standard error of a proportion is roughly 1.4 percentage points for PCCs near 50%, so the differences in Tables VI and VII are within the range of sampling variation. No confidence intervals, repeated trials with different weight initializations, or significance tests are reported. Given that the abstract and Section I claim 'superior recognition accuracy,' this evidence is not yet sufficient. I request confidence intervals and/or statistical comparisons (e.g., McNemar's test over paired predictions) and, if feasible, results averaged over several training runs.
  3. [Section IV.C, Tables VII–VIII] The computational-cost claim is not uniformly supported. In the second scenario, NAEL consumes 606 MFLOPs while MobileNetV2 consumes 416 MFLOPs, and LPI-Net consumes 69 MFLOPs; NAEL has higher accuracy than MobileNetV2 (55.4% vs 51.1%) but also higher cost. The statement that NAEL has 'the lowest computational cost among those achieving high recognition accuracy' requires a threshold for what counts as high accuracy, which is never defined. Please either frame the results as a Pareto frontier (accuracy vs. FLOPs), report the accuracy-cost trade-off for each baseline, or refine the claim to avoid giving the impression that NAEL dominates all baselines on both axes.
  4. [Section III.C, Eq. (9), Fig. 8] The load-bearing premise that gradient-map spatial concentration indicates whether the PRN is noise-affected is motivated only by Fig. 8, which shows fmax distributions for 50 simulated signals per class at -15 dB with no quantitative separation statistic (e.g., the fraction of misclassified samples whose fmax falls outside the center-frequency bins, or a related detection metric). Moreover, the NAN's classification accuracy against the actual correctness of the PRN is never reported, so there is no direct evidence that the gating policy generalizes from simulated training data to the real USRP test data. To substantiate the mechanism, add a quantitative evaluation of the NAN (accuracy, precision, recall, or AUC) on both simulated and real data, and include an ablation comparing NAEL with always-activate-PRN and always-activate-ARN policies to confirm that the learned gating, rather than the two recognition architectures alone, produces the reported Pareto improvement.
minor comments (6)
  1. [Section III.C, Eq. (9)] The text says fmax is 'the frequency location where the gradient map Gc(t,f) has the highest activation,' but Eq. (9) sums over time before maximizing; clarify the definition and whether the summation is over the time axis of the 8x8 gradient map.
  2. [Section IV.A] Provide training hyperparameters (learning rate, batch size, data split) and the full architecture of PRN, ARN, and NAN (number of blocks, channel sizes) so that the FLOPs counts can be reproduced.
  3. [Section IV.B] The approximate SNR estimation method is described in only one sentence; specify how the noise-only recording was obtained and how the signal power was estimated, since the reported SNR values (-4, -15, -17 dB) are used to interpret all results.
  4. [Fig. 11] The claim that ARN call counts 'do not exhibit a significant bias' is made without any statistical test; report the distribution or a chi-square test.
  5. [Reference [32]] Reference [32] is labeled as 'An image is worth 16x16 words: Transformers for image recognition at scale' but lists Kolesnikov et al.; the original ViT paper is by Dosovitskiy et al. Verify the citation.
  6. [Table VI] The Vision Transformer FLOPs count of 5,540 MFLOPs should be clarified; it is not clear whether this is for the same input resolution and whether it includes the patch embedding.

Circularity Check

0 steps flagged · score 2.0 of 10

No significant circularity: the NAEL accuracy and FLOPs claims are benchmarked against external models on held-out USRP data, and the only self-citation is a minor, non-load-bearing reference.

full rationale

The derivation chain is not circular in the sense defined by the review criteria. NAEL is a gating ensemble: the PRN produces a recognition result and a gradient map; the NAN classifies the gradient map as reliable or unreliable; the ARN is conditionally invoked. Each subnetwork is trained individually with the others frozen, and the final reported PCCs and MFLOPs (Tables VI-VIII) are measured on real USRP-collected test data against external baselines such as MobileNet, ShuffleNet, MNasNet, LPI-Net, LWCNN, and Vision Transformer. No fitted constant or training-set statistic is relabeled as a prediction of the headline result. The NAN is self-referential in that it predicts PRN reliability using PRN gradient maps, but this is a legitimate auxiliary gating design rather than a forced equivalence: the feature (gradient-map spatial distribution) and the target (correctness/noise-affectedness) are distinct, and the authors provide an empirical motivation in Fig. 8 without assuming the equivalence as a theorem. The absence of reported NAN classification accuracy is a validation gap, not circularity. The only self-citation is reference [15], used to support the incidental statement that 'phase information is highly susceptible to noise-induced distortions'; it does not justify the NAEL architecture or the performance claims. Thus the paper falls in the 0-2 band: essentially self-contained against external benchmarks, with one minor non-load-bearing self-citation.

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

The framework relies on a domain assumption about signal energy at center frequency, a Grad-CAM-style interpretability assumption, and a simulation-to-real transfer assumption. The architecture hyperparameters are hand-tuned and reported without validation details. The NAN training labels are undisclosed, so the axioms that define 'noise-affected' are not fully specified.

free parameters (4)
  • CWD scaling factor sigma = not reported
    Controls time-frequency resolution versus cross-term suppression in Eq. (2); the chosen value is not given, so image generation is not fully reproducible.
  • FE block hyperparameters (alpha, r, Co, s) = values in Table IV
    Section III.B states these were chosen because they yielded the best modulation recognition performance, indicating tuning on data without a documented validation split.
  • NAN hidden layer widths = 256 and 512
    Section III.C defines the two fully connected layers; no ablation or justification is provided.
  • Training SNR range = U(-15, 5) dB in Table V
    The training distribution is chosen by hand; test SNRs extend below the training range (-17 dB), which affects NAN and ARN reliability.
assumptions (4)
  • domain assumption Wireless signals occupy regions near the center frequency, so gradient map activations far from the center frequency indicate noise-affected recognition.
    Section III.C, used to justify the NAN decision criterion and the gradient-map classifier features.
  • domain assumption Gradient maps computed by weighted summation of feature maps (Eqs. 7 and 8) faithfully highlight input regions driving the PRN decision.
    Grad-CAM from reference [38] is adopted without validation on time-frequency images or radar signals.
  • domain assumption A model trained on simulated AWGN-corrupted signals will generalize to real signals with multipath, phase noise, and frequency offset.
    Section IV.A trains on simulated data; Section IV.B tests on USRP-collected signals with no adaptation or calibration.
  • domain assumption The ARN's deeper feature extraction improves accuracy on noise-affected inputs relative to the PRN.
    Underlies the benefit of activating ARN; the paper provides no direct PRN-only versus ARN-only comparison to support this.

how reviews work

0 comments
Cite this review

Pith. "Pith review of Noise-Aware Ensemble Learning for Efficient Radar Modulation Recognition." pith.science (2026). https://pith.science/paper/3THT3VFM

@misc{pith2026241115104,
  author       = {Pith},
  title        = {Pith review of: Noise-Aware Ensemble Learning for Efficient Radar Modulation Recognition},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/3THT3VFM}},
  note         = {Machine review of arXiv:2411.15104}
}
read the original abstract

Electronic warfare support (ES) systems intercept adversary radar signals and estimate various types of signal information, including modulation schemes. The accurate and rapid identification of modulation schemes under conditions of very low signal power remains a significant challenge for ES systems. This paper proposes a recognition model based on a noise-aware ensemble learning (NAEL) framework to efficiently recognize radar modulation schemes in noisy environments. The NAEL framework evaluates the influence of noise on recognition and adaptively selects an appropriate neural network structure, offering significant advantages in terms of computational efficiency and recognition performance. We present the analysis results of the recognition performance of the proposed model based on experimental data. Our recognition model demonstrates superior recognition accuracy with low computational complexity compared to conventional classification models.

Figures

Figures reproduced from arXiv: 2411.15104 by the authors.

Figure 1
Figure 1. Processes of the ES system for emitter identification. [PITH_FULL_IMAGE:figures/full_fig_p002_1.png] view at source ↗
Figure 2
Figure 2. Flow charts of modulation recognition models based on CNN: (a) [PITH_FULL_IMAGE:figures/full_fig_p003_2.png] view at source ↗
Figure 3
Figure 3. A flowchart illustrating the interactions between PRN, NAN, and [PITH_FULL_IMAGE:figures/full_fig_p005_3.png] view at source ↗
Figures from the paper (8 more)
Figure 4
Figure 4. Figure 4: Overall architecture of the NAEL framework. Our proposed recognition model uses SC, FE, and CE blocks to produce feature maps and compute [PITH_FULL_IMAGE:figures/full_fig_p006_4.png]
Figure 5
Figure 5. Figure 5: Block diagram illustrating the structure used in the FE block, excluding [PITH_FULL_IMAGE:figures/full_fig_p006_5.png]
Figure 6
Figure 6. Figure 6: Block diagrams illustrating the structures corresponding to the first [PITH_FULL_IMAGE:figures/full_fig_p006_6.png]
Figure 7
Figure 7. Figure 7: Examples of TFIs and gradient maps for Costas and T2 code simulated signals; (a) and (c) TFIs of transmitted Costas and T2 code signals, (b) and [PITH_FULL_IMAGE:figures/full_fig_p008_7.png]
Figure 8
Figure 8. Figure 8: Distribution of the frequency index fmax in (a) correct and (b) incorrect classifications at an SNR of −15 dB. ber of outliers compared to when classifications are correct. This difference highlights a clear distinction between cor￾rect and incorrect recognition result…
Figure 9
Figure 9. Figure 9: Illustration of measurement environment for test dataset generation [PITH_FULL_IMAGE:figures/full_fig_p010_9.png]
Figure 10
Figure 10. Figure 10: Confusion matrix of the proposed recognition model using the test dataset collected in the (a) first scenario (high-SNR case), (b) second scenario [PITH_FULL_IMAGE:figures/full_fig_p011_10.png]
Figure 11
Figure 11. Figure 11: ARN call count distribution across different modulation schemes in [PITH_FULL_IMAGE:figures/full_fig_p012_11.png]

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

43 extracted references · 42 canonical work pages

  1. [1]

    S. Peng, S. Sun, and Y .-D. Yao, ”A survey of modulation classification using deep learning: Signal representation and data preprocessing,” IEEE Transactions on Neural Networks and Learning Systems , vol. 33, no. 12, pp. 7020-7038, Dec. 2022

    work page 2022

  2. [2]

    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

    work page 2018

  3. [3]

    S. Peng, H. Jiang, H. Wang, H. Alwageed, Y . Zhou, M. M. Sebdani, and Y .-D. Yao, ”Modulation classification based on signal constellation diagrams and deep learning,” IEEE Transactions on Neural Networks and Learning Systems , vol. 30, no. 3, pp. 718-727, 2019

    work page 2019

  4. [4]

    Wiley, ELINT: The interception and analysis of radar signals

    R. Wiley, ELINT: The interception and analysis of radar signals . Artech, 2006

    work page 2006

  5. [5]

    P. E. Pace, Detecting and classifying low probability of intercept radar . Artech House, 2009

    work page 2009

  6. [6]

    Lunden and V

    J. Lunden and V . Koivunen, ”Automatic radar waveform recognition,” IEEE Journal of Selected Topics in Signal Processing , vol. 1, no. 1, pp. 124-136, May 2007

    work page 2007

  7. [7]

    Iglesias, J

    V . Iglesias, J. Grajal, P. Royer, M. A. Sanchez, M. Lopez-Vallejo, and O. A. Yeste-ojeda, ”Real-time low-complexity automatic modulation classifier for pulsed radar signals,” IEEE Transactions on Aerospace and Electronic Systems , vol. 51, no. 1, pp. 108-126, Jan. 2015

    work page 2015

  8. [8]

    Lopez-Risueno, J

    G. Lopez-Risueno, J. Grajal, and A. Sanz-Osorio, ”Digital channelized receiver based on time-frequency analysis for signal interception,” IEEE Transactions on Aerospace and Electronic Systems , vol. 41, no. 3, pp. 879-898, Jul. 2005

    work page 2005

Show all 43 references
  1. [9]

    C. Wang, J. Wang, and X. Zhang, ”Automatic radar waveform recog- nition based on time-frequency analysis and convolutional neural net- work,” in Proc. IEEE Int. Conf. Acoustics, Speech and Signal Processing (ICASSP), New Orleans, LA, USA, Mar. 2017, pp. 2437-2441

    2017

  2. [10]

    E. R. Zilberman and P. E. Pace, ”Autonomous time-frequency morpho- logical feature extraction algorithm for LPI radar modulation classifi- cation,” in Proc. Int. Conf. Image Processing , Atlanta, GA, USA, Oct. 2006, pp. 2321-2324

    2006

  3. [11]

    Huynh-The, V .-S

    T. Huynh-The, V .-S. Doan, C.-H. Hua, Q.-V . Pham, T.-V . Nguyen, and D.-S. Kim, ”Accurate LPI radar waveform recognition with CWD- TFA for deep convolutional network,” IEEE Wireless Communications Letters, vol. 10, no. 8, pp. 1638-1642, Aug. 2021

    2021

  4. [12]

    Zhang, L

    M. Zhang, L. Liu, and M. Diao, ”LPI radar waveform recognition based on time-frequency distribution,” Sensors, vol. 16, no. 10, Article No. 1682, Oct. 2016

    2016

  5. [13]

    Zhang, M

    M. Zhang, M. Diao, and L. Guo, ”Convolutional neural networks for automatic cognitive radio waveform recognition,” IEEE Access , vol. 5, pp. 11074-11082, Jun. 2017

    2017

  6. [14]

    S.-H. Kong, M. Kim, L. M. Hoang, and E. Kim, ”Automatic LPI radar waveform recognition using CNN,” IEEE Access, vol. 6, pp. 4207-4219, Jan. 2018

    2018

  7. [15]

    Park, J.-H

    D.-H. Park, J.-H. Bang, J.-H. Park, and H.-N. Kim, ”A fast and accurate convolutional neural network for LPI radar waveform recognition,” in Proc. 19th European Radar Conference (EuRAD) , Milan, Italy, Sep. 2022, pp. 89-92

    2022

  8. [16]

    E. E. Azzouz and A. K. Nandi, ”Procedure for automatic recognition of analogue and digital modulations,” IEE Proceedings-Communications , vol. 143, no. 5, pp. 259-266, Oct. 1996

    1996

  9. [17]

    Park and D

    C.-S. Park and D. Y . Kim, ”A novel robust feature of modulation classification for reconfigurable software radio,” IEEE Transactions on Consumer Electronics, vol. 52, no. 4, pp. 1193-1200, Nov. 2006

    2006

  10. [18]

    T. S. Tabatabaei, S. Krishnan, and A. Anpalagan, ”SVM-based classifi- cation of digital modulation signals,” in Proc. IEEE Int. Conf. Systems, Man, and Cybernetics , Istanbul, Turkey, Oct. 2010, pp. 277-280

    2010

  11. [19]

    M. D. Zeiler and R. Fergus, ”Visualizing and understanding convolu- tional networks,” in Proc. European Conf. Computer Vision (ECCV) , Zurich, Switzerland, Sep. 2014, pp. 818-833

    2014

  12. [20]

    A. G. Howard, M. Zhu, B. Chen, D. Kalenichenko, W. Wang, T. Weyand, M. Andreetto, and H. Adam, ”MobileNets: Efficient convo- lutional neural networks for mobile vision applications,” arXiv preprint arXiv:1704.04861, 2017. [Online]. Available: https://arxiv.org/abs/1704. 04861

    2017 arXiv

  13. [21]

    Sandler, A

    M. Sandler, A. Howard, M. Zhu, A. Zhmoginov, and L.-C. Chen, ”MobileNetV2: Inverted residuals and linear bottlenecks,” in Proc. IEEE/CVF Conf. Computer Vision and Pattern Recognition (CVPR) , Salt Lake City, UT, USA, Jun. 2018, pp. 4510-4520

    2018

  14. [22]

    C. Xiao, S. Yang, and Z. Feng, ”Complex-valued depthwise separable convolutional neural network for automatic modulation classification,” IEEE Transactions on Instrumentation and Measurement , vol. 72, pp. 1-10, 2023

    2023

  15. [23]

    G. B. Tunze, T. Huynh-The, J.-M. Lee, and D.-S. Kim, ”Sparsely connected CNN for efficient automatic modulation recognition,” IEEE Transactions on V ehicular Technology, vol. 69, no. 12, pp. 15557-15568, 2020

    2020

  16. [24]

    B. Zoph, V . Vasudevan, J. Shlens, and Q. V . Le, ”Learning transferable architectures for scalable image recognition,” in Proc. IEEE/CVF Conf. Computer Vision and Pattern Recognition (CVPR) , Salt Lake City, UT, USA, Jun. 2018, pp. 8697-8710

    2018

  17. [25]

    M. Tan, B. Chen, R. Pang, V . Vasudevan, M. Sandler, A. Howard, and Q. V . Le, ”MnasNet: Platform-aware neural architecture search for mobile,” in Proc. IEEE/CVF Conf. Computer Vision and Pattern Recognition (CVPR), Long Beach, CA, USA, Jun. 2019, pp. 2815-2823

    2019

  18. [26]

    Zhang, H

    X. Zhang, H. Zhao, H. Zhu, B. Adebisi, G. Gui, H. Gacanin, and F. Adachi, ”NAS-AMR: Neural architecture search-based automatic mod- ulation recognition for integrated sensing and communication systems,” IEEE Transactions on Cognitive Communications and Networking , vol. 8, no....

    2022

  19. [27]

    M. Du, X. He, X. Cai, and D. Bi, ”Balanced neural architecture search and its application in specific emitter identification,” IEEE Transactions on Signal Processing , vol. 69, pp. 5051-5065, 2021

    2021

  20. [28]

    Zhang, X

    X. Zhang, X. Zhou, M. Lin, and J. Sun, ”ShuffleNet: An extremely efficient convolutional neural network for mobile devices,” in Proc. IEEE/CVF Conf. Computer Vision and Pattern Recognition (CVPR) , Los Alamitos, CA, USA, Jun. 2018, pp. 6848-6856

    2018

  21. [29]

    N. Ma, X. Zhang, H.-T. Zheng, and J. Sun, ”ShuffleNet V2: Practical guidelines for efficient CNN architecture design,” in Proc. European Conf. Computer Vision (ECCV) , Munich, Germany, Sep. 2018, pp. 116- 131

    2018

  22. [30]

    J. Yin, L. Guo, W. Jiang, S. Hong, and J. Yang, ”ShuffleNet-inspired lightweight neural network design for automatic modulation classifi- cation methods in ubiquitous IoT cyber–physical systems,” Computer Communications, vol. 176, pp. 249-257, 2021

    2021

  23. [31]

    Semisupervised radar intrapulse signal modulation classification with virtual adversarial training,

    J. Cai, M. He, X. Cao, and F. Gan, “Semisupervised radar intrapulse signal modulation classification with virtual adversarial training,” IEEE Internet Things J. , vol. 11, no. 6, pp. 9929–9940, Mar. 2024

    2024

  24. [32]

    Automatic LPI radar waveform recognition using vision transformer,

    J. Kim, S. Cho, S. Hwang, and Y . Choi, “Automatic LPI radar waveform recognition using vision transformer,” in Proc. IEEE Int. Radar Conf. (RADAR), 2023, pp. 1–6

    2023

  25. [33]

    Choi and W

    H.-I. Choi and W. J. Williams, ”Improved time-frequency representation of multicomponent signals using exponential kernels,” IEEE Transac- tions on Acoustics, Speech, and Signal Processing , vol. 37, no. 6, pp. 862-871, Jun. 1989

    1989

  26. [34]

    Sifre, ”Rigid-motion scattering for image classification,” Ph.D

    L. Sifre, ”Rigid-motion scattering for image classification,” Ph.D. dis- sertation, Ecole Polytechnique, CMAP, Oct. 2014

    2014

  27. [35]

    Ioffe and C

    S. Ioffe and C. Szegedy, ”Batch normalization: Accelerating deep network training by reducing internal covariate shift,” in Proc. 32nd Int. Conf. Machine Learning (ICML) , Lille, France, Jul. 2015, pp. 448–456

    2015

  28. [36]

    Nair and G

    V . Nair and G. E. Hinton, ”Rectified linear units improve restricted Boltzmann machines,” in Proc. 27th ICML , Haifa, Israel, Jun. 2010, pp. 807-814

    2010

  29. [37]

    K. He, X. Zhang, S. Ren, and J. Sun, ”Deep residual learning for image recognition,” in Proc. IEEE Conf. Computer Vision and Pattern Recognition (CVPR) , Las Vegas, NV , USA, Jun. 2016, pp. 770-778

    2016

  30. [38]

    R. R. Selvaraju, M. Cogswell, A. Das, R. Vedantam, D. Parikh, and D. Batra, ”Grad-CAM: Visual explanations from deep networks via gradient-based localization,” in Proc. IEEE Int. Conf. Computer Vision (ICCV), Venice, Italy, Oct. 2017, pp. 618-626. JOURNAL OF LATEX CLASS FILES xiii

    2017

  31. [39]

    Removing adversarial noise in class activation feature space,

    D. Zhou et al., “Removing adversarial noise in class activation feature space,” in Proc. IEEE/CVF Int. Conf. Comput. Vis. (ICCV) , 2021, pp. 7858–7867

    2021

  32. [40]

    D. P. Kingma and J. Ba, ”Adam: A method for stochastic optimization,” in Proc. 3rd Int. Conf. Learning Representations (ICLR) , San Diego, CA, USA, May 2015

    2015

  33. [41]

    Ettus Research, Mountain View, CA, USA, 2012

    USRP N200/N210 Networked Series . Ettus Research, Mountain View, CA, USA, 2012. [Online]. Available: https://www.ettus.com/wp-content/ uploads/2019/01/07495 Ettus N200-210 DS Flyer HR 1.pdf

    2012

  34. [42]

    An image is worth 16x16 words: Transformers for image recognition at scale,

    A. Kolesnikov et al., “An image is worth 16x16 words: Transformers for image recognition at scale,” in Proc. 9th Int. Conf. Learn. Represent. (ICLR), 2021, pp. 1–21

    2021

  35. [43]

    Bridging the gap between vision transformers and convolutional neural networks on small datasets,

    Z. Lu, H. Xie, C. Liu, and Y . Zhang, “Bridging the gap between vision transformers and convolutional neural networks on small datasets,” in Proc. 36th Conf. Neural Inf. Process. Syst. (NeurIPS) , 2022, pp. 1–15

    2022

Pith tools

Reviewed August 12, 2026 · model on record in the stance chip above.