A Distributed Learning Automata Scheme for Spectrum Management in Self-Organized Cognitive Radio Network

Fahimi, Mina; Ghasemi, Abdorasoul

doi:10.1109/tmc.2016.2601926

Cited by 27 publications

(25 citation statements)

References 34 publications

Supporting

Mentioning

Contrasting

Unclassified

Order By: Relevance

“…The primary task of supervised learning is to train a "black-box" model from limited data samples, where the training process is generally in a single machine. However, this centralized training manner may not be suitable for those large-scale or privacy-concerned problems, such as big data applications [5] and recommendation systems [6], which could only be or better addressed in a distributed manner [7], [8].…”

Section: Introductionmentioning

confidence: 99%

Distributed Training for Multi-Layer Neural Networks by Consensus

Liu

Ding

2020

IEEE Trans. Neural Netw. Learning Syst.

View full text Add to dashboard Cite

Over the past decade, there has been a growing interest in large-scale and privacy-concerned machine learning, especially in the situation where the data cannot be shared due to privacy protection or cannot be centralized due to computational limitations. Parallel computation has been proposed to circumvent these limitations, usually based on the master-slave and decentralized topologies, and the comparison study shows that a decentralized graph could avoid the possible communication jam on the central agent but incur extra communication cost. In this brief, a consensus algorithm is designed to allow all agents over the decentralized graph to converge to each other, and the distributed neural networks with enough consensus steps could have nearly the same performance as the centralized training model. Through the analysis of convergence, it is proved that all agents over an undirected graph could converge to the same optimal model even with only a single consensus step, and this can significantly reduce the communication cost. Simulation studies demonstrate that the proposed distributed training algorithm for multi-layer neural networks without data exchange could exhibit comparable or even better performance than the centralized training model.

show abstract

Section: Introductionmentioning

confidence: 99%

Distributed Training for Multi-Layer Neural Networks by Consensus

Liu

Ding

2020

IEEE Trans. Neural Netw. Learning Syst.

View full text Add to dashboard Cite

show abstract

“…Clustering is a method of partitioning items into clusters, in such a way that items in the same cluster are more similar to each other and items in different clusters are dissimilar (Aggarwal & Reddy, ; Bezdek, ; Everitt, ; Hansen & Jaumard, ; Jain, Murty, & Flynn, ). Clustering is a useful technique which is incorporated in a variety of fields, such as engineering (e.g., machine learning, artificial intelligence, and pattern recognition), computer sciences (e.g., web mining and image segmentation), and medical sciences (e.g., genetics, biology, microbiology, and pathology) (Abbas & Fahmy, ; Aggarwal & Reddy, ; Anari, Torkestani, & Rahmani, ; Andrews, ; Bali & Kumar, ; Das, Abraham, & Konar, ; Díaz‐Cortés, Cuevas, & Rojas, ; Evangelou, Hadjimitsis, Lazakidou, & Clayton, ; Everitt, ; Fahimi & Ghasemi, ; Hansen & Jaumard, ; Judd, McKinley, & Jain, ; Shariat, Movaghar, & Hoseinzadeh, ).…”

Section: Introductionmentioning

confidence: 99%

“…LA have been used as optimization tools in complex and dynamic environments. In recent years, they have successfully been applied to a wide range of applications, such as information retrieval (Torkestani, 2012), pattern recognition (Maravall & de Lope, 2011), dynamic optimization (Rezvanian & Meybodi, 2010), power systems (Vlachogiannis, 2009), clustering (Anari et al, 2017;Hasanzadeh-Mofrad & Rezvanian, 2017), segmentation (Díaz-Cortés et al, 2017), routing protocol (Fahimi & Ghasemi, 2017), vehicular system (Shariat et al, 2017), and cloud computing (Morshedlou & Meybodi, 2014). Each LA has a number of actions, and the goal of LA is to find the optimal action from its action set.…”

mentioning

confidence: 99%

A learning automata‐based clustering algorithm using ant swarm intelligence

2018

View full text Add to dashboard Cite

Ant‐based clustering algorithms are inspired by the behaviour of real ants. In most ant‐based clustering algorithms, each ant lacks global visibility and only uses local search to find the best place on a grid to drop its data item. This paper presents a new clustering method called learning automata‐based clustering algorithm using learning automata (LA) and ant swarm intelligence. In this paper, the problem of finding the best location of grid for dropping a data item is solved using LA. We introduced a new drop operation using the global search capability of LA that can increase the quality of clustering. In the proposed method, each ant is equipped with an LA and a two‐dimensional grid is partitioned into a number of clusters. To drop a data item, the LA of each ant will be responsible for finding one of the best clusters on a grid unlike other methods that use the random search of a grid. To evaluate the results, some clustering evaluation criteria, such as the number of obtained clusters, f‐measure, Rand index, intracluster variance, Dunn index, error rate, and time cost were employed. Additionally, the proposed algorithms were compared with k‐means, Lumer and Faieta, ant clustering algorithm, modified version of the ant‐clustering algorithm, Chaotic Ant clustering (CAC), chaotic ant clustering algorithm, adaptive ant‐based clustering algorithm, adaptive artificial ant clustering algorithm, and ant clustering algorithm with information theoretic learning algorithms on both real and synthetic datasets. Experimental results show that the proposed method can increase the efficiency of the real dataset by 11.13% and that of the synthetic dataset by 16.11%. It also reduces the number of neighbouring function calls by 16% for the real dataset and by 28% for the synthetic dataset.

show abstract

“…Thus, repeating of the process increases the possibility of selecting the ideal action. LA is used in several areas including wireless sensor networks, online social networks, resources allocation, pattern classification, signal processing, and in some other troublesome areas of wireless communication such as those in the literatures …”

Section: Introductionmentioning

confidence: 99%

“…LA is used in several areas including wireless sensor networks, 24 online social networks, 25 resources allocation, 26 pattern classification, 27 signal processing, 28 and in some other troublesome areas of wireless communication such as those in the literatures. [29][30][31][32] Optimization problems become increasingly complex in the practical scenarios and real communication systems. On the other hand, time and energy limitations demand better optimization algorithms.…”

Section: Introductionmentioning

confidence: 99%

Channel estimation based on learning automata for OFDM systems

Salehi

Majidi

Neda

2018

Int J Communication

View full text Add to dashboard Cite

Summary The abundant benefits of Orthogonal Frequency‐Division Multiplexing and its high flexibility have resulted in its widespread applications in many telecommunication standards. One important parameter for improving wireless system's efficiency is the accurate estimation of channel state information. In the literatures, many techniques have been studied in order to estimate the channel state information. Nowadays, the techniques based on intelligent algorithms such as genetic algorithm and particle swarm optimization (PSO) have attracted attention of researchers. With a very low pilot overhead, these techniques are able to estimate the channel frequency response properly only using the received signals. Unfortunately, each of these techniques suffers a common weakness: they have a slow convergence rate. In this paper, a new intelligent and different method has been presented for channel estimation using learning automata, entitled LA estimator, where the learning automata are search agents, and each pair is responsible for searching 1 complex coefficient of the channel frequency response. This method can achieve an accurate channel estimation with a moderate computational complexity in comparison with GA and PSO estimators. Furthermore, with higher convergence rate, our proposed method is capable of providing the same performance as GA and PSO. For a 2‐path fast fading channel, simulation results demonstrate the robustness of our proposed scheme according to the bit error rate and the mean square error.

show abstract

A Distributed Learning Automata Scheme for Spectrum Management in Self-Organized Cognitive Radio Network

Cited by 27 publications

References 34 publications

Distributed Training for Multi-Layer Neural Networks by Consensus

Distributed Training for Multi-Layer Neural Networks by Consensus

A learning automata‐based clustering algorithm using ant swarm intelligence

Channel estimation based on learning automata for OFDM systems

Contact Info

Product

Resources

About