Automatic Modulation Recognition Using Neural Architecture Search

Wei, Shengyun; Zou, Shun; Liao, Feifan; Lang, Weimin; Wan, Wenhui

doi:10.1109/hpbdis.2019.8735458

Cited by 11 publications

(10 citation statements)

References 14 publications

Supporting

Mentioning

Contrasting

Unclassified

Order By: Relevance

“…In the context of modulation classification, GA methods have been employed to extract and optimize classification features [29]- [32] or to optimize the DNN architecture [14,15]. However, none of these prior work considers the joint optimization of both the connections and the hyperparameters of the AMC's DNN architecture.…”

Section: Related Workmentioning

confidence: 99%

“…However, none of these prior work considers the joint optimization of both the connections and the hyperparameters of the AMC's DNN architecture. Moreover, only the simple CNN or FFNN network architectures have been considered [14,15]. In this paper, we aim to jointly optimize both the connections and hyperparameters of the AMC's DNN architecture by a novel memetic algorithm that addresses the aforementioned drawbacks of GA methods in image processing.…”

Section: Related Workmentioning

confidence: 99%

“…In contrast, GAs are highly-efficient for MOOPs [12]. NAS has seldom been considered for AMC [14,15]. Dataset-centric solutions: Most existing humancrafted AMC DNN architectures were optimized for a single set of modulations [1,16].…”

Section: Introductionmentioning

confidence: 99%

“…However, such the human-crafted DNN architectures might not be optimized in terms of the connections and hyperparameters values. As already mentioned, GA provides a few techniques to solve MOOP, but GAs applied for AMC's NAS are still single-objective driven [14,15].…”

Section: Introductionmentioning

confidence: 99%

See 3 more Smart Citations

Evolutionary Optimization of Residual Neural Network Architectures for Modulation Classification

Perenda¹,

Rajendran²,

Bovet³

et al. 2021

Preprint

View full text Add to dashboard Cite

Automatic Modulation Classification (AMC) receives significant interest in the context of current and future wireless communication systems. Deep learning emerged as a powerful AMC tool, as it allows for the joint learning of discriminative features, and signal classification. However, the optimization of Deep Neural Network (DNN) architectures for AMC is a manual and time-consuming process that requires profound domain knowledge and much effort. Moreover, most proposed solutions focus mainly on classification accuracy, while optimization of network complexity is neglected. In this paper, we propose a novel bi-objective memetic algorithm, BO-NSMA, to search optimal DNN architectures for AMC to maximize classification accuracy and minimize network complexity. We show that BO-NSMA, with a small initial population of six individuals and only ten generations, finds a DNN architecture that outperforms all human-crafted State-of-the-Art (SoA) models. BO-NSMA discovered the first low-complexity Convolutional Neural Network (CNN)-based model, which achieves slightly better performance than costly Recurrent Neural Network (RNN)-based approaches, allowing a 2.8-fold reduction in network complexity with 0.7% performance improvement. Compared to its counterparts from Network Architecture Search (NAS), BO-NSMA finds the best architecture, which achieves up to 18.24% accuracy gain and up to a 78.71-fold reduction in network complexity.

show abstract

Section: Related Workmentioning

confidence: 99%

Section: Related Workmentioning

confidence: 99%

Section: Introductionmentioning

confidence: 99%

Section: Introductionmentioning

confidence: 99%

See 2 more Smart Citations

Evolutionary Optimization of Residual Neural Network Architectures for Modulation Classification

Perenda¹,

Rajendran²,

Bovet³

et al. 2021

Preprint

View full text Add to dashboard Cite

show abstract

“…Dentre os métodos de classificação mais utilizados na literatura atualmente estão as redes neurais artificiais (ANN, do inglês antificial neural network) [7], redes neurais convolucionais (CNN, do inglês convolutional neural network) [8] [9] e SVM (do inglês support vector machine) [1]. O método de classificação proposto neste trabalho é uma ANN com RB e algoritmo de retropropagação (do inglês backpropagation) de LM.…”

Section: Introductionunclassified

Classificação Automática de Modulações utilizando Redes Neurais Artificiais com regularização Bayesiana e algoritmo de retropropagação de Levenberg-Marquardt

Valadao¹,

Silva²,

Costa³

et al. 2020

Anais De XXXVIII Simpósio Brasileiro De Telecomunicações E Processamento De Sinais

View full text Add to dashboard Cite

Resumo-Com o aumento da demanda por espectro de frequência, foi necessário buscar maneiras mais eficientes de alocar usuários no espectro. O rádio cognitivo sensoreia o espectro e aloca dinamicamente usuários em espaços inutilizados. As técnicas de classificação automática de modulação vieram para fornecer informações que auxiliam no sensoriamento do espectro. Neste trabalho, foram extraídas características de sinais modulados em banda passante em diferentes SNR (do inglês signal noise ratio). O classificador utilizado foi uma rede perceptron com Regularização Bayesiana (RB) e algoritmo de retropropagação de Levenberg-Marquardt (LM). Para o método proposto, obteve-se acurácias que variam de 74, 8 % a 95, 5 %.

show abstract

Survey of Research on Application of Deep Learning in Modulation Recognition

Sun,

2023

Wireless Pers Commun

View full text Add to dashboard Cite

Automatic Modulation Recognition Using Neural Architecture Search

Cited by 11 publications

References 14 publications

Evolutionary Optimization of Residual Neural Network Architectures for Modulation Classification

Evolutionary Optimization of Residual Neural Network Architectures for Modulation Classification

Classificação Automática de Modulações utilizando Redes Neurais Artificiais com regularização Bayesiana e algoritmo de retropropagação de Levenberg-Marquardt

Survey of Research on Application of Deep Learning in Modulation Recognition

Contact Info

Product

Resources

About