Home / Current Issue / Paper 1706244
AI-Driven Neuromorphic Computing for Energy-Efficient Anomaly Detection in IoT Networks
Subject area: Science,Engineering and Technology · Area of research: Artificial Intelligence
Abstract
The propagation of Internet of Things (IoT) networks has led to an increasing need for real-time anomaly detection to ensure system reliability and security. However, traditional deep learning models employed for this task often come with significant energy consumption and latency challenges, particularly when deployed on resource-constrained edge devices. This research explores the use of neuromorphic computing, specifically spiking neural networks (SNNs), to develop an energy-efficient anomaly detection system for IoT networks. A novel architecture is proposed where SNNs operate at the edge, leveraging their event-driven nature to provide ultra-low-power, real-time anomaly detection. The designed system reduces energy consumption and minimizes detection latency, making it suitable for deployment in energy-sensitive IoT environments. A comprehensive analysis is conducted, comparing the performance of the neuromorphic model against traditional deep learning approaches, focusing on metrics such as energy efficiency, detection accuracy, and latency. The findings demonstrate that SNN-based anomaly detection can significantly enhance the energy efficiency of IoT systems while maintaining or even improving detection performance, paving the way for more sustainable and responsive IoT deployments.
Keywords
Neuromorphic Computing, Spiking Neural Networks (SNNs), Real-Time Anomaly Detection, Low-Power AI
How to cite this paper
@article{1706244,
author = {Srujana Maddula, Himanshu Gupta, Shaik Mohammad Jani Basha, Gayathri S},
title = {AI-Driven Neuromorphic Computing for Energy-Efficient Anomaly Detection in IoT Networks},
journal = {Iconic Research And Engineering Journals},
year = {2024},
volume = {8},
number = {3},
pages = {122-129},
issn = {2456-8880},
url = {https://www.irejournals.com/formatedpaper/17062441.pdf},
abstract = {The propagation of Internet of Things (IoT) networks has led to an increasing need for real-time anomaly detection to ensure system reliability and security. However, traditional deep learning models employed for this task often come with significant energy consumption and latency challenges, particularly when deployed on resource-constrained edge devices. This research explores the use of neuromorphic computing, specifically spiking neural networks (SNNs), to develop an energy-efficient anomaly detection system for IoT networks. A novel architecture is proposed where SNNs operate at the edge, leveraging their event-driven nature to provide ultra-low-power, real-time anomaly detection. The designed system reduces energy consumption and minimizes detection latency, making it suitable for deployment in energy-sensitive IoT environments. A comprehensive analysis is conducted, comparing the performance of the neuromorphic model against traditional deep learning approaches, focusing on metrics such as energy efficiency, detection accuracy, and latency. The findings demonstrate that SNN-based anomaly detection can significantly enhance the energy efficiency of IoT systems while maintaining or even improving detection performance, paving the way for more sustainable and responsive IoT deployments.},
keywords = {Neuromorphic Computing, Spiking Neural Networks (SNNs), Real-Time Anomaly Detection, Low-Power AI},
month = {September},
}