Authors: Mert Nakıp, Erol Gelenbe

Published on: June 22, 2023

Impact Score: 7.8

Arxiv code: Arxiv:2306.13030

Summary

• What is new: Introduces a novel Self-Supervised Intrusion Detection (SSID) framework that operates fully online without human intervention or prior off-line learning.
• Why this is important: Traditional intrusion detection systems require offline data collection and manual data labeling, which can introduce human errors and are not adaptable to rapidly changing network environments.
• What the research proposes: The proposed SSID framework uses an Auto-Associative Deep Random Neural Network for automatically analyzing and labeling incoming traffic packets, using an online estimation of its statistical trustworthiness without prior training on offline datasets.
• Results: Experimental evaluation on public datasets shows the SSID framework outperforms existing machine learning and deep learning models in terms of accuracy and efficiency in an IoT environment.

Technical Details

Technological frameworks used: Auto-Associative Deep Random Neural Network, Self-Supervised Learning

Models used: Deep Learning models

Data used: Public datasets

Potential Impact

Cybersecurity companies focusing on intrusion detection, IoT security service providers, network administrators, and businesses requiring robust security solutions may find this technology particularly disruptive.

Want to implement this idea in a business?

We have generated a startup concept here: SecureNet AI.