Deep Learning for Anomaly Detection in IoT Devices

EasyChair Preprint 14346

Abstract

The rapid proliferation of Internet of Things (IoT) devices has led to an exponential increase in data generation, making it crucial to develop effective methods for anomaly detection to ensure system reliability and security. Traditional anomaly detection techniques often struggle with the high-dimensional, dynamic, and heterogeneous nature of IoT data. This paper explores the application of deep learning methods for anomaly detection in IoT devices, emphasizing their ability to automatically learn and extract complex patterns from large datasets. We review various deep learning architectures, including autoencoders, convolutional neural networks (CNNs), and recurrent neural networks (RNNs), and their effectiveness in identifying anomalies across different types of IoT data, such as sensor readings and network traffic. The paper also addresses the challenges and limitations of applying deep learning in this context, including the need for large labeled datasets and the potential for overfitting. We propose a novel hybrid approach that combines deep learning with domain-specific knowledge to improve detection accuracy and robustness. Experimental results demonstrate the effectiveness of these methods in real-world IoT environments, highlighting their potential for enhancing the reliability and security of IoT systems.

Keyphrases: Cybersecurity, learning, machine

@booklet{EasyChair:14346,
  author    = {Obaloluwa Ogundairo and Peter Broklyn},
  title     = {Deep Learning for Anomaly Detection in IoT Devices},
  howpublished = {EasyChair Preprint 14346},
  year      = {EasyChair, 2024}}