Pith. sign in

REVIEW 4 major objections 6 minor 2 cited by

RadioLLM: Introducing Large Language Model into Cognitive Radio via Hybrid Prompt and Token Reprogrammings

T0 review · 4 major / 6 minor · reviewed 2026-08-10 · deepseek-v4-flash

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

Pith's one-line read A single pretrained language model, adapted with hybrid prompts and frequency fusion, beats task-specific radio classifiers and denoisers on most benchmarks.

desk verdict Plausible but not yet proven: the first real application of LLM token reprogramming to raw I/Q radio signals, undermined by asymmetric pretraining and an ablation inconsistency. read the letter →

arxiv 2501.17888 v3 pith:H3CICUQL submitted 2025-01-28 eess.SP cs.AIcs.LG

Shuai Chen , Yong Zu , Zhixi Feng , Shuyuan Yang , Mengchang Li This is my paper · ORCID
classification eess.SPcs.AIcs.LG
keywords cognitiveradiolargelanguagemodelssignalclassificationtokenreprogramminghybridpromptingfrequency-attunedfusiondenoisingfew-shotlearning
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 establish that a large language model can act as a single universal engine for cognitive radio tasks. It proposes RadioLLM, which feeds raw I/Q radio samples to a GPT-2 backbone after reprogramming them into token embeddings and prefixing them with hybrid prompts that compress expert knowledge into a few retrieved semantic anchors. A frequency-attuned fusion module injects CNN-extracted high-frequency features so the transformer does not lose transient and phase-detail information. Across seven classification benchmarks and three denoising benchmarks, the paper reports the highest overall accuracy and Kappa on six of the seven classification sets and the highest SSIM on all three denoising sets. If correct, radio-signal applications would no longer need a separate task-specific network for every modulation family and noise regime.

What carries the argument

The load-bearing machinery is the pair of modules that make an LLM accept radio signals. Hybrid Prompt and Token Reprogramming (HPTR) does two things: it selects the top-K embeddings from a pretrained word-token space that are most similar to a concise hardware prompt, forming a short hybrid prefix, and it uses a multi-head cross-attention layer, with raw I/Q patches as queries and the anchor embeddings as keys and values, to turn signal patches into LLM-compatible tokens without a natural-language intermediate. Frequency-Attuned Fusion (FAF) runs the raw signal through three convolutional high-frequency extraction layers and fuses those local features with the reprogrammed tokens before the LLM, compensating for the attention mechanism's bias toward low-frequency global structure. The output side is a lightweight decoder that can either reconstruct the denoised I/Q signal or feed pooled features to a linear classification head.

What would settle it

Retrain every baseline under the same multi-dataset pretraining protocol that RadioLLM receives and compare on the held-out test splits; if the baselines match or exceed RadioLLM's reported OA, Kappa, and SSIM numbers, the claimed advantage of HPTR and FAF over existing methods is not established.

Watch

Extended reading notes

Core claim

The central claim is that a GPT-2 model, mostly frozen and adapted with LoRA, can master radio signal classification and denoising when its input is reprogrammed and its attention is supplemented with high-frequency information. Specifically, RadioLLM obtains the best overall accuracy on RML16A (58.10), RML16B (58.35), RML16C (68.19), RML22 (59.39), RML18A (52.03), and Wi-Fi (35.41) under a 100-shot labeling regime, and the best SSIM on all three RML denoising tasks (0.838, 0.893, and 0.846). The authors attribute the gains to pretraining across multiple radio datasets, the hybrid prompt's replacement of verbose templates with top-K semantic anchors, and the FAF module's correction of the transformer's low-frequency bias. They also report that on ADS-B the model is not the top performer, and that the remaining confusions are between similar modulation families such as 16QAM/64QAM and AM-DSB/WBFM.

Load-bearing premise

The evaluation assumes that pretraining RadioLLM on the same datasets on which it is later benchmarked (RML16A/B/C, RML18A, Wi-Fi, and ADS-B) does not give it an unfair advantage, since most baselines receive only the 100 labeled examples and only TcssAMR and SemiAMC receive a multi-domain pretraining comparison on three of the datasets.

Editorial extensions

If this is right

  • If correct, a single LLM backbone with lightweight task heads can replace separate radio classification and denoising networks, reducing deployment complexity in cognitive radio systems.
  • The hybrid prompt's top-K retrieval shortens the prompt prefix, and the paper reports 31.85% faster inference with a 0.85% accuracy gain, implying prompt compression matters for latency-sensitive radio settings.
  • FAF's success implies transformer-based radio models should not rely on attention alone; injecting convolutional high-frequency features is a transferable recipe for low-SNR robustness.
  • The gains grow as labeled samples increase, so the model is most useful in semi-supervised regimes where a large unlabeled signal pool can be pretrained on, rather than in extreme 1-shot settings.

Reading between the lines

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

  • The reported margins may largely reflect in-distribution pretraining rather than HPTR and FAF, because RadioLLM is pretrained on the same benchmark datasets it is later tested on; an out-of-distribution holdout would settle this.
  • The same hybrid-prompt and token-reprogramming design could be applied to other physical-layer signals such as radar, sonar, or biomedical I/Q streams by swapping the anchor embedding space, though the paper does not test these.
  • If the semantic anchors are the main source of the improvement, a smaller non-LLM encoder conditioned on the same anchors might reproduce most of the gain, which would weaken the case that a full language model is necessary; this is an untested hypothesis.
  • The FAF design suggests a direct modification of transformer attention to be frequency-aware could yield similar or better low-SNR performance without a separate CNN branch, which the paper does not explore.
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 RadioLLM, a framework that reprograms GPT-2 to process raw I/Q radio signals via Hybrid Prompt and Token Reprogramming (HPTR) and a Frequency-Attuned Fusion (FAF) module. The model is pretrained on six public radio datasets and then evaluated on 100-shot classification over seven datasets and on denoising over three datasets. The paper claims that RadioLLM achieves superior performance over task-specific RSC and denoising baselines in the majority of testing scenarios, with Table I reporting the highest OA/Kappa on six of seven classification benchmarks and Table II reporting the highest SSIM on all three denoising benchmarks.

Significance. If the result holds, RadioLLM would be a meaningful demonstration that a relatively small LLM backbone can be reprogrammed into a unified radio-signal processing front-end, which is a plausible and timely direction for cognitive radio. The claim is made more credible by the deliberate holdout of RML22 from pretraining, by the equalization experiment in Section IV-G showing that two pretrained baselines still trail on three RML datasets, and by the inclusion of an ablation table. However, the fairness of the benchmark comparison is the central load-bearing issue: most evaluation datasets are also pretraining datasets for RadioLLM but not for the baselines. The paper does not yet provide the evidence needed to attribute the RML18A and Wi-Fi wins to HPTR and FAF rather than to asymmetric data exposure.

major comments (4)
  1. [IV-A, IV-F] Pretraining in Section IV-A uses RML16A, RML16B, RML16C, RML18A, ADS-B, and Wi-Fi, and classification results in Section IV-F are then reported on those same datasets, while the baselines receive only 100 labeled samples. The reported gains on RML18A (52.03 vs. 47.66) and Wi-Fi (35.41 vs. 34.59) may therefore reflect the pretraining exposure rather than the efficacy of HPTR and FAF. Section IV-G equalizes only TcssAMR and SemiAMC, and only on RML16A/B/C; the equalized baselines on those three datasets still trail RadioLLM, which is evidence in the paper's favor, but no equalized comparison is provided for RML18A or Wi-Fi. An equalized comparison on those datasets, or removal of those wins from the central claim, is required. The RML22 result, which is not pretrained, is a fair test and supports the method.
  2. [IV-I, Table III] The ablation contradicts the text's claim of consistent improvements when HTRP and FAF are combined: the SSIM of the full HTRP+FAF configuration (0.838) is lower than that of FAF-only (0.857). The paragraph in Section IV-I states that the joint configuration 'delivers consistent improvements across all evaluation metrics,' which is directly contradicted by the same table. This needs an explanation, for example if SSIM is evaluated only on the denoising task while the primary target is classification, and the wording should be corrected to match the numeric results.
  3. [IV-F, IV-H, Tables I and II] The paper reports a single run for each configuration without variance, confidence intervals, or significance testing. Several reported margins are small (for example, 58.10 vs. 57.60 for the equalized TcssAMR on RML16A, and 35.41 vs. 34.59 on Wi-Fi), so the 'superior performance' claim is not yet quantitatively grounded. At least a small number of random seeds and a statement of variance are needed to establish that the observed differences are not noise.
  4. [IV-H, Table II] The denoising evaluation on RML16A, RML16B, and RML16C is subject to the same asymmetry as the classification comparison: RadioLLM was pretrained on these datasets, while the denoising baselines SGFilter and DNCNet were not. The qualitative and quantitative superiority claimed in Section IV-H would be more convincing if the pretrained-equivalent baselines from Section IV-G were also evaluated for denoising, or if denoising were reported on a dataset not used in pretraining.
minor comments (6)
  1. [I, IV-I] The abbreviation 'HTRP' appears in several places (the contributions bullet in Section I and the Table III caption and surrounding text) but the paper consistently defines the module as HPTR; please unify the abbreviation.
  2. [Index Terms] The Index Terms contain the typo 'Technolog'; it should read 'Technology'.
  3. [III-E] In Section III-E, 'fine-tune GPT-2 using the LoRA technique [24]' is correct in the reference list, but the later sentence 'with only a subset updated via LoRA [7]' cites the network optimization reference [7] instead of the LoRA reference; please fix the citation.
  4. [IV-B] For RML18A, the text states 2,555,904 total samples, 24 modulation types, and 26 SNR levels with 4096 samples per class per SNR; these numbers imply 24*26*4096 = 2,555,904, which is consistent, but the phrase '26 SNR levels from -20 dB to 30 dB in 2 dB increments' actually lists 26 values, so the count is consistent; please confirm the SNR range wording is intended.
  5. [IV-F, Fig. 4] Figure 4 has panels labeled (a)-(e), (f)-(j), and (k)-(o), but the text refers to 'Fig. 4 (o)' both for the RML18A confusion matrix and for the misclassification discussion; please renumber or refer to the specific panel of the confusion matrix more clearly.
  6. [IV-G, Fig. 5] The text in Section IV-G discusses the results shown in Fig. 5 but does not explicitly cite the figure number in the paragraph; please add a citation to Fig. 5.

Circularity Check

0 steps flagged · score 0.0 of 10

No derivation-level circularity: HPTR and FAF are trained components evaluated against external baselines, and self-citations are not load-bearing.

full rationale

The paper's claimed derivation chain is self-contained in the sense required for circularity analysis. HPTR is a differentiable reprogramming layer whose top-K anchor selection is a fixed cosine-similarity retrieval from pretrained token embeddings, not a fit to the target accuracy metric; FAF is a CNN branch trained jointly with the LLM. The downstream classification and denoising numbers in Tables I and II are obtained by training these components and comparing them with external baselines, so no reported quantity is equivalent to an input by construction. The authors' self-citations ([4], [22]) appear only as background for prior DL-based CRT work and do not supply the central result or forbid alternatives. The more serious concern is evaluation fairness: Section IV-A pretrains on RML16A/B/C, RML18A, ADS-B, and Wi-Fi, and Section IV-F evaluates 100-shot classification on the same datasets, with equalization of pretraining reported only for TcssAMR and SemiAMC on RML16A/B/C. This could explain some margins, especially on RML18A and Wi-Fi, but it is a benchmark-fairness/validity issue rather than a circular derivation. Additionally, the ablation in Table III shows the full HTRP+FAF SSIM (0.838) is lower than FAF-only (0.857), contradicting the text's claim of consistent improvements across all metrics; this is an internal consistency problem, not circularity. Overall circularity score is 0.

Assumptions & free parameters 6 free parameters · 5 assumptions · 0 invented entities

The framework rests on several unproved modeling assumptions rather than derived theorems. The main nonstandard assumptions are that GPT-2 word embeddings form a useful semantic space for radio tokens, that transformer attention is low-pass while CNN is high-pass, and that pretraining on the evaluation datasets is a fair comparison. No new physical entities are introduced; the semantic anchors E' are trained artifacts, not independent existence claims.

free parameters (6)
  • Top-K anchor count = 7
    Selected via sensitivity analysis on RML16A in Section IV-J.1 and used across all datasets and tasks, so the headline results are tied to this hand-tuned value.
  • Pretraining SNR threshold = 14 dB
    High-quality samples for the RML series were selected with SNR >= 14 dB in Section IV-A; no ablation on this threshold is reported.
  • Loss balancing factors b_i = not reported
    Algorithm 1 uses per-dataset loss balancing factors b_i, but their values are never given; they affect the pretraining objective.
  • Signal patch size and embedding dimension = not reported
    Section III-B splits the input into patches to form X_s, but the patch length and projection dimensions are not specified.
  • LoRA rank and alpha = not reported
    LoRA is used to fine-tune GPT-2 in Section IV-C, but rank and alpha values are absent.
  • Number of HFE layers = 3
    FAF uses three high-frequency extraction layers in Section III-C; no ablation on depth is reported.
assumptions (5)
  • domain assumption Pretrained LLM word embeddings form a semantic space in which top-K cosine-similarity anchors are more effective prompts than full text templates.
    Used in Equations (3)-(5) for hybrid prompt and token reprogramming; the paper provides an ablation but no mechanism or external evidence that word embeddings are semantically aligned with radio signals.
  • domain assumption Transformer attention is biased toward low-frequency signal content and CNN convolution toward high-frequency content, so fusing the two improves radio classification.
    Statement in Section III-C and the introduction; the cited references [13]-[16] are about vision transformers and CNNs, not radio signals.
  • domain assumption Unlabeled pretraining on clean samples from the same datasets used for downstream evaluation is fair and does not constitute leakage.
    Section IV-A selects high-SNR RML samples plus ADS-B and Wi-Fi for pretraining; Section IV-F benchmarks on those same datasets, with only partial baseline control in Section IV-G.
  • domain assumption MSE reconstruction pretraining on clean I/Q signals transfers to both classification and denoising downstream tasks.
    The pretraining objective in Section III-D minimizes MSE; no analysis separates the contribution of pretraining from the architecture.
  • standard math Standard deep learning assumptions such as backpropagation, AdamW optimization, and LoRA fine-tuning hold for this setup.
    Used throughout training in Section IV-C and Algorithm 1; not contested by the paper.

how reviews work

0 comments
Cite this review

Pith. "Pith review of RadioLLM: Introducing Large Language Model into Cognitive Radio via Hybrid Prompt and Token Reprogrammings." pith.science (2026). https://pith.science/paper/H3CICUQL

@misc{pith2026250117888,
  author       = {Pith},
  title        = {Pith review of: RadioLLM: Introducing Large Language Model into Cognitive Radio via Hybrid Prompt and Token Reprogrammings},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/H3CICUQL}},
  note         = {Machine review of arXiv:2501.17888}
}
read the original abstract

The growing scarcity of spectrum resources and rapid proliferation of wireless devices make efficient radio network management critical. While deep learning-enhanced Cognitive Radio Technology (CRT) provides promising solutions for tasks such as radio signal classification (RSC), denoising, and spectrum allocation, existing DL-based CRT frameworks are typically task-specific and lack scalability in diverse real-world applications. This limitation naturally leads to the exploration of Large Language Models (LLMs), whose exceptional cross-domain generalization capabilities offer new potential for advancing CRT. To bridge this gap, we propose RadioLLM, a novel framework that integrates Hybrid Prompt and Token Reprogramming (HPTR) for combining radio signal features with expert knowledge, and a Frequency-Attuned Fusion (FAF) module for enhanced high-frequency feature modeling. Extensive evaluations on multiple benchmark datasets demonstrate that RadioLLM achieves superior performance compared to existing baselines in the majority of testing scenarios.

Figures

Figures reproduced from arXiv: 2501.17888 by the authors.

Figure 1
Figure 1. Comparison of existing CRT frameworks with our proposed approach. [PITH_FULL_IMAGE:figures/full_fig_p001_1.png] view at source ↗
Figure 2
Figure 2. The model framework of RadioLLM. The input radio signal is preprocessed to generate signal embeddings [PITH_FULL_IMAGE:figures/full_fig_p003_2.png] view at source ↗
Figure 3
Figure 3. An illustrative example of the prompt template used in this study. [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 evaluation of RadioLLM across multiple datasets and SNR levels. Each column corresponds to one dataset: RML2016a, RML16B, [PITH_FULL_IMAGE:figures/full_fig_p008_4.png]
Figure 5
Figure 5. Figure 5: Comparison of OA for different methods on RML16A, RML16B, and [PITH_FULL_IMAGE:figures/full_fig_p009_5.png]
Figure 6
Figure 6. Figure 6: Comparison of denoising results on randomly selected high-SNR samples from the RML16A, RML16B, and RML16C datasets. For each dataset, the [PITH_FULL_IMAGE:figures/full_fig_p010_6.png]
Figure 7
Figure 7. Figure 7: Performance Comparison Across Different K Values. 2) Decoder Usage: To further investigate the impact of the decoder module on the overall model performance, we conducted an ablation study in which the Transformer decoder was replaced with a Linear decoder. Experimenta…

Discussion (0). Continue with ORCID to comment.

Forward citations

Cited by 2 Pith papers

Reviewed papers in the Pith corpus that reference this work. Sorted by Pith novelty score. Full citation record

  1. EMind: A Foundation Model for Multi-task Electromagnetic Signals Understanding

    eess.SP 2025-08 conditional novelty 5.0 of 10

    EMind reports that one masked-autoencoder transformer, pretrained on 81 million heterogeneous IQ samples, transfers to seven electromagnetic signal tasks with strong accuracy, but post-hoc checkpoint selection and mis...

  2. The Hitchhiker's Guide to Agentic AI: From Foundations to Systems

    cs.AI 2026-06 unverdicted novelty 2.0 of 10

    A survey-style reference book mapping the full agentic-AI stack from transformer internals to production deployment, with no new research result.

Reference graph

Works this paper leans on

42 extracted references · 32 canonical work pages · cited by 2 Pith papers

  1. [1]

    Intent-aware radio re- source scheduling in a ran slicing scenario using reinforcement learning,

    C. V . Nahum, V . H. L. Lopes, R. M. Dreifuerst, P. Batista, I. Correa, K. V . Cardoso, A. Klautau, and R. W. Heath, “Intent-aware radio re- source scheduling in a ran slicing scenario using reinforcement learning,” IEEE Trans. Wireless Commun. , vol. 23, no. 3, pp. 2253–2267, 2023

    work page 2023

  2. [2]

    Gnn-based power allocation and user association in digital twin network for the terahertz band,

    H. Zhang, X. Ma, X. Liu, L. Li, and K. Sun, “Gnn-based power allocation and user association in digital twin network for the terahertz band,” IEEE J. Sel. Area Commun. , vol. 41, no. 10, pp. 3111–3121, 2023

    work page 2023

  3. [3]

    Spectral and energy efficiency analysis for cognitive radio networks,

    F. Haider, C.-X. Wang, H. Haas, E. Hepsaydir, X. Ge, and D. Yuan, “Spectral and energy efficiency analysis for cognitive radio networks,” IEEE Trans. Wireless Commun. , vol. 14, no. 6, pp. 2969–2980, 2015

    work page 2015

  4. [4]

    Learning tempo- ral–spectral feature fusion representation for radio signal classification,

    Z. Feng, S. Chen, Y . Ma, Y . Gao, and S. Yang, “Learning tempo- ral–spectral feature fusion representation for radio signal classification,” IEEE Trans. Ind. Inf. , pp. 1–10, 2024

    work page 2024

  5. [5]

    Exploring llm- based multi-agent situation awareness for zero-trust space-air-ground integrated network,

    X. Cao, G. Nan, H. Guo, H. Mu, L. Wang, Y . Lin, Q. Zhou, J. Li, B. Qin, Q. Cui, X. Tao, H. Fang, H. Du, and T. Q. Quek, “Exploring llm- based multi-agent situation awareness for zero-trust space-air-ground integrated network,” IEEE J. Sel. Area Commun. , pp. 1–1, 2025

    work page 2025

  6. [6]

    Toward intelligent communications: Large model empowered semantic communications,

    H. Xie, Z. Qin, X. Tao, and Z. Han, “Toward intelligent communications: Large model empowered semantic communications,” IEEE Commun. Mag., pp. 1–7, 2024

    work page 2024

  7. [7]

    Adapting foundation models for information syn- thesis of wireless communication specifications,

    M. Kotaru, “Adapting foundation models for information syn- thesis of wireless communication specifications,” arXiv preprint arXiv:2308.04033, 2023

    arXiv 2023

  8. [8]

    Wirelessllm: Empowering large language models towards wireless intelligence,

    J. Shao, J. Tong, Q. Wu, W. Guo, Z. Li, Z. Lin, and J. Zhang, “Wirelessllm: Empowering large language models towards wireless intelligence,” J. Commun. Inf. Netw. , vol. 9, no. 2, pp. 99–112, 2024

    work page 2024

Show all 42 references
  1. [9]

    Llm4wm: Adapting llm for wireless multi-tasking,

    X. Liu, S. Gao, B. Liu, X. Cheng, and L. Yang, “Llm4wm: Adapting llm for wireless multi-tasking,” arXiv preprint arXiv:2501.12983, 2025

    2025 arXiv

  2. [10]

    Wirelessagent: Large language model agents for intelligent wireless networks,

    J. Tong, J. Shao, Q. Wu, W. Guo, Z. Li, Z. Lin, and J. Zhang, “Wirelessagent: Large language model agents for intelligent wireless networks,” arXiv preprint arXiv:2409.07964 , 2024

    2024 arXiv

  3. [11]

    Time-LLM: Time series forecasting by reprogramming large language models,

    M. Jin, S. Wang, L. Ma, Z. Chu, J. Y . Zhang, X. Shi, P.-Y . Chen, Y . Liang, Y .-F. Li, S. Pan, and Q. Wen, “Time-LLM: Time series forecasting by reprogramming large language models,” in The Twelfth International Conference on Learning Representations , 2024. [Online]. Availa...

    2024

  4. [12]

    TEMPO: Prompt-based generative pre-trained transformer for time series forecasting,

    D. Cao, F. Jia, S. O. Arik, T. Pfister, Y . Zheng, W. Ye, and Y . Liu, “TEMPO: Prompt-based generative pre-trained transformer for time series forecasting,” in The Twelfth International Conference on Learning Representations , 2024. [Online]. Available: https: //openreview.net...

    2024

  5. [13]

    Inception transformer,

    C. Si, W. Yu, P. Zhou, Y . Zhou, X. Wang, and S. Yan, “Inception transformer,” Proc. Adv. Neural Inf. Process. Syst. (NeurIPS) , vol. 35, pp. 23 495–23 509, 2022

    2022

  6. [14]

    Exploring frequency-inspired optimization in transformer for efficient single image super-resolution,

    A. Li, L. Zhang, Y . Liu, and C. Zhu, “Exploring frequency-inspired optimization in transformer for efficient single image super-resolution,” IEEE Trans. Pattern Anal. Mach. Intell. , 2025

    2025

  7. [15]

    Improving vision transformers by revisiting high-frequency components,

    J. Bai, L. Yuan, S.-T. Xia, S. Yan, Z. Li, and W. Liu, “Improving vision transformers by revisiting high-frequency components,” pp. 1–18, 2022

    2022

  8. [16]

    High-frequency component helps explain the generalization of convolutional neural networks,

    H. Wang, X. Wu, Z. Huang, and E. P. Xing, “High-frequency component helps explain the generalization of convolutional neural networks,” in Proceedings of the IEEE/CVF conference on computer vision and pattern recognition, 2020, pp. 8684–8694

    2020

  9. [18]

    Real-time radio technology and modulation classification via an lstm auto-encoder,

    Z. Ke and H. Vikalo, “Real-time radio technology and modulation classification via an lstm auto-encoder,” IEEE Trans. Wireless Commun., vol. 21, no. 1, pp. 370–382, 2021

    2021

  10. [20]

    Mclhn: Towards automatic modulation classification via masked contrastive learning with hard negatives,

    C. Xiao, S. Yang, Z. Feng, and L. Jiao, “Mclhn: Towards automatic modulation classification via masked contrastive learning with hard negatives,” IEEE Trans. Wireless Commun. , 2024

    2024

  11. [21]

    Better approach for denoising eeg signals,

    G. Kaushal, A. Singh, and V . Jain, “Better approach for denoising eeg signals,” in 2016 5th International Conference on Wireless Networks and Embedded Systems (WECON) . IEEE, 2016, pp. 1–3

    2016

  12. [22]

    A generative self-supervised framework for cognitive radio leveraging time-frequency features and attention-based fusion,

    S. Chen, Z. Feng, S. Yang, Y . Ma, J. Liu, and Z. Qi, “A generative self-supervised framework for cognitive radio leveraging time-frequency features and attention-based fusion,” IEEE Trans. Wireless Commun. , 2024

    2024

  13. [23]

    S2ip-llm: Semantic space informed prompt learning with llm for time series forecasting,

    Z. Pan, Y . Jiang, S. Garg, A. Schneider, Y . Nevmyvaka, and D. Song, “S2ip-llm: Semantic space informed prompt learning with llm for time series forecasting,” in Forty-first International Conference on Machine Learning, 2024

    2024

  14. [24]

    Lora: Low-rank adaptation of large language models,

    E. J. Hu, Y . Shen, P. Wallis, Z. Allen-Zhu, Y . Li, S. Wang, L. Wang, and W. Chen, “Lora: Low-rank adaptation of large language models,” arXiv preprint arXiv:2106.09685 , 2021

    2021 arXiv

  15. [25]

    Radio machine learning dataset generation with gnu radio,

    T. J. O’shea and N. West, “Radio machine learning dataset generation with gnu radio,” in Proceedings of the GNU radio conference , vol. 1, no. 1, 2016

    2016

  16. [26]

    Convolutional radio mod- ulation recognition networks,

    T. J. O’Shea, J. Corgan, and T. C. Clancy, “Convolutional radio mod- ulation recognition networks,” in Engineering Applications of Neural Networks: 17th International Conference, EANN 2016, Aberdeen, UK, September 2-5, 2016, Proceedings 17 . Springer, 2016, pp. 213–226

    2016

  17. [27]

    Rml22: Realistic dataset generation for wireless modulation classification,

    V . Sathyanarayanan, P. Gerstoft, and A. E. Gamal, “Rml22: Realistic dataset generation for wireless modulation classification,” IEEE Trans. Wireless Commun., vol. 22, no. 11, pp. 7663–7675, 2023

    2023

  18. [28]

    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 J. Sel. Top. Signal Process. , vol. 12, no. 1, pp. 168–179, 2018

    2018

  19. [29]

    Large-scale real-world radio signal recognition with deep learning,

    T. Ya, L. Yun, Z. Haoran, J. Zhang, W. Yu, G. Guan, and M. Shiwen, “Large-scale real-world radio signal recognition with deep learning,” Chin. J. Aeronaut. , vol. 35, no. 9, pp. 35–48, 2022

    2022

  20. [30]

    Oracle: Optimized radio classification through convo- lutional neural networks,

    K. Sankhe, M. Belgiovine, F. Zhou, S. Riyaz, S. Ioannidis, and K. Chowdhury, “Oracle: Optimized radio classification through convo- lutional neural networks,” in IEEE INFOCOM 2019-IEEE Conference on Computer Communications . IEEE, 2019, pp. 370–378

    2019

  21. [31]

    A hierarchical classification head based convolutional gated deep neural network for automatic modulation classification,

    S. Chang, R. Zhang, K. Ji, S. Huang, and Z. Feng, “A hierarchical classification head based convolutional gated deep neural network for automatic modulation classification,” IEEE Trans. Wireless Commun. , vol. 21, no. 10, pp. 8713–8728, 2022

    2022

  22. [32]

    An efficient deep learning model for automatic modulation recognition based on parameter estimation and transformation,

    F. Zhang, C. Luo, J. Xu, and Y . Luo, “An efficient deep learning model for automatic modulation recognition based on parameter estimation and transformation,” IEEE Commun. Lett. , vol. 25, no. 10, pp. 3287–3290, 2021

    2021

  23. [33]

    A spatiotemporal multi-channel learning framework for automatic modulation recognition,

    J. Xu, C. Luo, G. Parr, and Y . Luo, “A spatiotemporal multi-channel learning framework for automatic modulation recognition,” IEEE Wire- less Commun. Lett. , vol. 9, no. 10, pp. 1629–1632, 2020

    2020

  24. [34]

    An efficient specific emitter identification method based on complex- valued neural networks and network compression,

    Y . Wang, G. Gui, H. Gacanin, T. Ohtsuki, O. A. Dobre, and H. V . Poor, “An efficient specific emitter identification method based on complex- valued neural networks and network compression,” IEEE J. Sel. Areas Commun., vol. 39, no. 8, pp. 2305–2317, 2021

    2021

  25. [35]

    Resource- constrained specific emitter identification using end-to-end sparse feature selection,

    M. Tao, X. Fu, Y . Lin, Y . Wang, Z. Yao, S. Shi, and G. Gui, “Resource- constrained specific emitter identification using end-to-end sparse feature selection,” in GLOBECOM 2023 - 2023 IEEE Global Communications Conference, 2023, pp. 6067–6072

    2023

  26. [36]

    Cnn-based automatic modulation classification for beyond 5g communications,

    A. P. Hermawan, R. R. Ginanjar, D.-S. Kim, and J.-M. Lee, “Cnn-based automatic modulation classification for beyond 5g communications,” IEEE Commun. Lett. , vol. 24, no. 5, pp. 1038–1041, 2020

    2020

  27. [37]

    A two-stage model based on a complex-valued separate residual network for cross-domain iiot devices identification,

    G. Han, Z. Xu, H. Zhu, Y . Ge, and J. Peng, “A two-stage model based on a complex-valued separate residual network for cross-domain iiot devices identification,” IEEE Trans. Ind. Inf. , vol. 20, no. 2, pp. 2589– 2599, 2024

    2024

  28. [38]

    A transformer- based contrastive semi-supervised learning framework for automatic modulation recognition,

    W. Kong, X. Jiao, Y . Xu, B. Zhang, and Q. Yang, “A transformer- based contrastive semi-supervised learning framework for automatic modulation recognition,” IEEE Trans. Cognit. Commun. Networking , vol. 9, no. 4, pp. 950–962, 2023

    2023

  29. [39]

    Self-contrastive learning based semi-supervised radio modulation classification,

    D. Liu, P. Wang, T. Wang, and T. Abdelzaher, “Self-contrastive learning based semi-supervised radio modulation classification,” in MILCOM 2021 - 2021 IEEE Military Communications Conference (MILCOM) , 2021, pp. 777–782

    2021

  30. [40]

    What is a savitzky-golay filter?[lecture notes],

    R. W. Schafer, “What is a savitzky-golay filter?[lecture notes],” IEEE Signal Process Mag. , vol. 28, no. 4, pp. 111–117, 2011

    2011

  31. [41]

    Dncnet: Deep radar signal denoising and recognition,

    M. Du, P. Zhong, X. Cai, and D. Bi, “Dncnet: Deep radar signal denoising and recognition,” IEEE Trans. Aerosp. Electron. Syst., vol. 58, no. 4, pp. 3549–3562, 2022

    2022

  32. [42]

    Bert: Pre-training of deep bidirectional transformers for language understanding,

    J. Devlin, M.-W. Chang, K. Lee, and K. Toutanova, “Bert: Pre-training of deep bidirectional transformers for language understanding,” in Pro- ceedings of the 2019 conference of the North American chapter of the association for computational linguistics: human language technolo...

    2019

  33. [43]

    Language models are unsupervised multitask learners,

    A. Radford, J. Wu, R. Child, D. Luan, D. Amodei, I. Sutskever et al., “Language models are unsupervised multitask learners,” OpenAI blog , vol. 1, no. 8, p. 9, 2019

    2019

  34. [44]

    The llama 3 herd of models,

    A. Grattafiori, A. Dubey, A. Jauhri, A. Pandey, A. Kadian, A. Al-Dahle, A. Letman, A. Mathur, A. Schelten, A. Vaughan et al. , “The llama 3 herd of models,” arXiv preprint arXiv:2407.21783 , 2024

    2024 arXiv

Pith tools

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