Stochastic Configuration Networks Based Adaptive Storage Replica Management for Power Big Data Processing

Huang, Changqin; Huang, Qionghao; Wang, Dianhui

doi:10.1109/tii.2019.2919268

Cited by 58 publications

(16 citation statements)

References 32 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…The SCN has better performance than other randomized neural networks in terms of the fast learning, scope of the random parameters, and the required human intervention. Therefore, it has already been used in many data processing projects, such as [ 5 , 9 , 27 ].…”

Section: Methodsmentioning

confidence: 99%

See 1 more Smart Citation

Optimal Randomness for Stochastic Configuration Network (SCN) with Heavy-Tailed Distributions

Niu

Wei

Chen

2020

Entropy

View full text Add to dashboard Cite

Stochastic Configuration Network (SCN) has a powerful capability for regression and classification analysis. Traditionally, it is quite challenging to correctly determine an appropriate architecture for a neural network so that the trained model can achieve excellent performance for both learning and generalization. Compared with the known randomized learning algorithms for single hidden layer feed-forward neural networks, such as Randomized Radial Basis Function (RBF) Networks and Random Vector Functional-link (RVFL), the SCN randomly assigns the input weights and biases of the hidden nodes in a supervisory mechanism. Since the parameters in the hidden layers are randomly generated in uniform distribution, hypothetically, there is optimal randomness. Heavy-tailed distribution has shown optimal randomness in an unknown environment for finding some targets. Therefore, in this research, the authors used heavy-tailed distributions to randomly initialize weights and biases to see if the new SCN models can achieve better performance than the original SCN. Heavy-tailed distributions, such as Lévy distribution, Cauchy distribution, and Weibull distribution, have been used. Since some mixed distributions show heavy-tailed properties, the mixed Gaussian and Laplace distributions were also studied in this research work. Experimental results showed improved performance for SCN with heavy-tailed distributions. For the regression model, SCN-Lévy, SCN-Mixture, SCN-Cauchy, and SCN-Weibull used less hidden nodes to achieve similar performance with SCN. For the classification model, SCN-Mixture, SCN-Lévy, and SCN-Cauchy have higher test accuracy of 91.5%, 91.7% and 92.4%, respectively. Both are higher than the test accuracy of the original SCN.

show abstract

Section: Methodsmentioning

confidence: 99%

“…SCN algorithms have been commonly studied and used in many areas, such as image data analytics [ 5 , 6 ], prediction of component concentrations in sodium aluminate liquor [ 7 ], and etc. [ 8 , 9 ]. For example, in [ 5 ], Li et al developed a two-dimensional SCNs (2DSCNs) for image data modelling tasks.…”

Section: Introductionmentioning

confidence: 99%

Optimal Randomness for Stochastic Configuration Network (SCN) with Heavy-Tailed Distributions

Niu

Wei

Chen

2020

Entropy

View full text Add to dashboard Cite

show abstract

“…DL (deep learning) has begun to rise. e related algorithms based on this have found a breakthrough for FSL (few-shot learning) [8,9]. We establish a data center configuration management system in line with this practice standard and finely manage various infrastructures, so as to establish a long-term mechanism of production and operation with safety production as the core and effectively improve the service level of the data center.…”

Section: Introductionmentioning

confidence: 99%

Financial Data Center Configuration Management System Based on Random Forest Algorithm and Few-Shot Learning

Wang

2022

Computational Intelligence and Neuroscience

View full text Add to dashboard Cite

To form a unified configuration and information management platform, FCCMS (financial center configuration management system) will integrate and sort information based on various configuration data and relationships as well as integrate processes and permissions. However, the most serious issue that data centers are currently facing is how to effectively manage these infrastructures. For various infrastructures, the data center currently uses a decentralized operation and maintenance management model. When an infrastructure fails due to inexperienced configuration management, this mode is not conducive to quickly locating and resolving the problem. A detection method of RFCO (random forest algorithm based on clustering optimization) is proposed, and an appropriate tree is selected from RF to integrate, so as to achieve the best effect. In this paper, the target matching algorithm based on FSL (few-shot learning) is deeply studied, and the target detection model is applied to the target matching and positioning task by using the ML method. The performance of the algorithm is tested by experiments on relevant datasets to verify the effectiveness of the algorithm in various scenarios.

show abstract

“…The massive quantity of data collected by power IoT systems is not only of large volume but is also more diversified compared with the traditional grid data, and its sources and distribution are more extensive [5]. Advances in Symmetry 2021, 13, 1718 2 of 18 the IoT technologies of the smart grid have led that smart grid data demonstrating typical features of big data such as the abundant amount of data generated, the diversity of data sources and the high velocity of data generation [6,7]. To handle the smart grid with data characterized by big data features, ref.…”

Section: Introductionmentioning

confidence: 99%

A Multi-Source Big Data Security System of Power Monitoring Network Based on Adaptive Combined Public Key Algorithm

et al. 2021

View full text Add to dashboard Cite

The multi-source data collected by the power Internet of Things (IoT) provide the data foundation for the power big data analysis. Due to the limited computational capability and large amount of data collection terminals in power IoT, the traditional security mechanism has to be adapted to such an environment. In order to ensure the security of multi-source data in the power monitoring networks, a security system for multi-source big data in power monitoring networks based on the adaptive combined public key algorithm and an identity-based public key authentication protocol is proposed. Based on elliptic curve cryptography and combined public key authentication, the mapping value of user identification information is used to combine the information in a public and private key factor matrix to obtain the corresponding user key pair. The adaptive key fragment and combination method are designed so that the keys are generated while the status of terminals and key generation service is sensed. An identification-based public key authentication protocol is proposed for the power monitoring system where the authentication process is described step by step. Experiments are established to validate the efficiency and effectiveness of the proposed system. The results show that the proposed model demonstrates satisfying performance in key update rate, key generation quantity, data authentication time, and data security. Finally, the proposed model is experimentally implemented in a substation power IoT environment where the application architecture and security mechanism are described. The security evaluation of the experimental implementation shows that the proposed model can resist a series of attacks such as counterfeiting terminal, data eavesdropping, and tampering.

show abstract

Stochastic Configuration Networks Based Adaptive Storage Replica Management for Power Big Data Processing

Cited by 58 publications

References 32 publications

Optimal Randomness for Stochastic Configuration Network (SCN) with Heavy-Tailed Distributions

Optimal Randomness for Stochastic Configuration Network (SCN) with Heavy-Tailed Distributions

Financial Data Center Configuration Management System Based on Random Forest Algorithm and Few-Shot Learning

A Multi-Source Big Data Security System of Power Monitoring Network Based on Adaptive Combined Public Key Algorithm

Contact Info

Product

Resources

About