โก Quick Summary
This study introduces a stacking-based Tiny Machine Learning (TinyML) model designed for attack detection in IoT networks. The model achieved an impressive accuracy rate of 99.98% with minimal latency and power consumption, showcasing its potential for real-time security applications. ๐
๐ Key Details
- ๐ Dataset: ToN-IoT dataset with 461,008 labeled instances
- ๐งฉ Features used: 10 types of attack categories
- โ๏ธ Technology: Stacking ensemble learning with Decision Trees and Neural Networks
- ๐ Performance: Accuracy 99.98%, Inference latency 0.12 ms, Power consumption 0.01 mW
๐ Key Takeaways
- ๐ Enhanced security for IoT networks through advanced machine learning techniques.
- ๐ก Stacking ensemble learning combines multiple models for improved detection performance.
- ๐ Superior performance compared to traditional machine learning methods.
- โก Real-time processing capabilities with low computational overhead.
- ๐ Potential applications in various IoT environments, enhancing overall security.
- ๐ Comprehensive dataset allows for robust model training and evaluation.
- ๐งช Data preprocessing techniques such as label encoding and feature selection were utilized.
๐ Background
As the Internet of Things (IoT) continues to expand, the security of these interconnected devices has become increasingly critical. Traditional methods of attack detection often struggle with the complexities and real-time demands of IoT systems. This study addresses these challenges by leveraging Tiny Machine Learning (TinyML) to create a more efficient and effective detection model.
๐๏ธ Study
The research conducted by Sharma et al. utilized the publicly available ToN-IoT dataset, which includes a diverse range of attack types. The study focused on developing a stacking-based TinyML model that integrates lightweight Decision Trees and small Neural Networks to enhance detection capabilities while minimizing computational demands.
๐ Results
The results of the experiments were remarkable, with the stacking-based TinyML model achieving an accuracy rate of 99.98%. Additionally, the model demonstrated an average inference latency of just 0.12 ms and an estimated power consumption of 0.01 mW. These metrics highlight the model’s efficiency and effectiveness in real-time attack detection.
๐ Impact and Implications
The implications of this study are significant for the future of IoT security. By employing advanced machine learning techniques, we can enhance the security posture of IoT networks and provide timely responses to potential threats. This research paves the way for broader applications of TinyML in various sectors, ensuring that IoT devices can operate securely and efficiently.
๐ฎ Conclusion
This study exemplifies the transformative potential of machine learning in cybersecurity, particularly within the realm of IoT. The stacking-based TinyML model not only achieves high accuracy but also operates with minimal latency and power consumption, making it a promising solution for real-time attack detection. Continued research in this area is essential to further enhance the security of IoT networks and protect against evolving threats.
๐ฌ Your comments
What are your thoughts on the advancements in IoT security through machine learning? We would love to hear your insights! ๐ฌ Share your comments below or connect with us on social media:
An optimized stacking-based TinyML model for attack detection in IoT networks.
Abstract
With the expansion of Internet of Things (IoT) devices, security is an important issue as attacks are constantly gaining more complex. Traditional attack detection methods in IoT systems have difficulty being able to process real-time and access limitations. To address these challenges, a stacking-based Tiny Machine Learning (TinyML) models has been proposed for attack detection in IoT networks. This ensures detection efficiently and without additional computational overhead. The experiments have been conducted using the publicly available ToN-IoT dataset, comprising a total of 461,008 labeled instances with 10 types of attacks categories. Some amount of data preprocessing has been done applying methods such as label encoding, feature selection, and data standardization. A stacking ensemble learning technique uses multiple models combining lightweight Decision Tree (DT) and small Neural Network (NN) to aggregate power of the system and generalize. The performance of the model is evaluated by accuracy, precision, recall, F1-score, specificity, and false positive rate (FPR). Experimental results demonstrate that the stacked TinyML model is superior to traditional ML methods in terms of efficiency and detection performance, and its accuracy rate is 99.98%. It has an average inference latency of 0.12โms and an estimated power consumption of 0.01 mW.
Author: [‘Sharma A’, ‘Rani S’, ‘Shabaz M’]
Journal: PLoS One
Citation: Sharma A, et al. An optimized stacking-based TinyML model for attack detection in IoT networks. An optimized stacking-based TinyML model for attack detection in IoT networks. 2025; 20:e0329227. doi: 10.1371/journal.pone.0329227