Energy-efficient event pattern recognition in wireless sensor networks using multilayer spiking neural networks

Kasi, Shahrukh Khan; Das, Saptarshi; Biswas, Subir

doi:10.1007/s11276-021-02555-9

Cited by 4 publications

(3 citation statements)

References 23 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Although two-layer spiking neural architectures have been utilized in multiple realworld applications [18,19,32] we would like to promote the adaptation of fractional-order derivative-based optimization for more complex state-of-the-art architectures as well. These architectures require further theoretical investigation and a wide range of experiments.…”

Section: Discussionmentioning

confidence: 99%

“…These voltage differences are in the end functions of the weight parameters, and by taking their gradients, classic Gradient Descent (GD) learning can be used. While Tempotron is not the latest architecture, its various applications and enhancements are carried out by numerous research groups even nowadays [18][19][20]. Such reusability of classical methods can be observed in the field of artificial intelligence in general [21][22][23][24].…”

Section: Introductionmentioning

confidence: 99%

See 1 more Smart Citation

Exploring the Effects of Caputo Fractional Derivative in Spiking Neural Network Training

Gyongyossy

Eros²,

Botzheim³

2022

Electronics

View full text Add to dashboard Cite

Fractional calculus is an emerging topic in artificial neural network training, especially when using gradient-based methods. This paper brings the idea of fractional derivatives to spiking neural network training using Caputo derivative-based gradient calculation. We focus on conducting an extensive investigation of performance improvements via a case study of small-scale networks using derivative orders in the unit interval. With particle swarm optimization we provide an example of handling the derivative order as an optimizable hyperparameter to find viable values for it. Using multiple benchmark datasets we empirically show that there is no single generally optimal derivative order, rather this value is data-dependent. However, statistics show that a range of derivative orders can be determined where the Caputo derivative outperforms first-order gradient descent with high confidence. Improvements in convergence speed and training time are also examined and explained by the reformulation of the Caputo derivative-based training as an adaptive weight normalization technique.

show abstract

Section: Discussionmentioning

confidence: 99%

Section: Introductionmentioning

confidence: 99%

Exploring the Effects of Caputo Fractional Derivative in Spiking Neural Network Training

Gyongyossy

Eros²,

Botzheim³

2022

Electronics

View full text Add to dashboard Cite

show abstract

“…Energy-efficient composite event detection is intended to improve energy efficiency while handling massive amounts of data [64,65]. Until recently, the majority of current techniques had been applied to composite event identification [66,67], consuming a significant amount of energy. The effectiveness of WSNs can be measured using a variety of criteria [68].…”

Section: Lifetime Metric and Energy Efficiency In Wsnmentioning

confidence: 99%

A Systematic Review on the Energy Efficiency of Dynamic Clustering in a Heterogeneous Environment of Wireless Sensor Networks (WSNs)

2022

View full text Add to dashboard Cite

There are a variety of applications for wireless sensor networks (WSNs), such as military, health monitoring systems, natural disasters, smartphones, and other surveillance systems. While the primary purpose of sensor nodes is to collect unattended data in hostile environments, many are placed in large numbers and operate independently. Due to limited capabilities, power is often limited. Therefore, these nodes are grouped into clusters to increase communication efficiency. In WSNs, two different routing protocols are possible: apartment and hierarchical or clustering protocols. Due to their significant role in minimizing energy consumption, hierarchical methods have become very popular in clustering. In cluster-based methods, nodes are organized into clusters, and the sensor node with the most resources is appointed as the cluster head (CH). In this paper, we present a Systematic Literature Review (SLR) explaining the difficulties in developing cluster-based methods, critical factors for clustering, and hierarchical clustering protocols. The most important factor of a routing protocol for WSN is the energy consumption and lifetime of a network. Focusing on energy consumption, different cluster-based methods were analyzed to determine which technology should be deployed by analyzing specific criteria to support the selection process. Additionally, the pros and cons of different protocols are listed with their relevance in specific scenarios. To identify these protocols, a systematic literature review was conducted using research studies published from 2010 to 2021, with 30 papers analyzed in the final phase. Based on the results of this SLR, several issues need to be further investigated with respect to the interaction of the potential technology with the Internet of Things (IoT) and Vehicular Ad-Hoc Networks (VANETs).

show abstract

Variance-Constrained Resilient $$H_{\infty }$$ State Estimation for Time-Varying Neural Networks with Random Saturation Observation Under Uncertain Occurrence Probability

Gao

et al. 2023

Neural Process Lett

View full text Add to dashboard Cite

Energy-efficient event pattern recognition in wireless sensor networks using multilayer spiking neural networks

Cited by 4 publications

References 23 publications

Exploring the Effects of Caputo Fractional Derivative in Spiking Neural Network Training

Exploring the Effects of Caputo Fractional Derivative in Spiking Neural Network Training

A Systematic Review on the Energy Efficiency of Dynamic Clustering in a Heterogeneous Environment of Wireless Sensor Networks (WSNs)

Variance-Constrained Resilient $$H_{\infty }$$ State Estimation for Time-Varying Neural Networks with Random Saturation Observation Under Uncertain Occurrence Probability

Contact Info

Product

Resources

About