Neuromorphic Cybersecurity with Lifelong Learning

A bio-inspired SNN architecture for Network Intrusion Detection Systems (NIDS) capable of lifelong learning.

Overview

This project presents a Spiking Neural Network (SNN) architecture for lifelong Network Intrusion Detection Systems (NIDS), inspired by the brain’s hierarchical processing and energy efficiency.

Key Innovations

  1. Hierarchical Architecture:

    • Static SNN: Efficiently identifies potential intrusions.
    • Dynamic SNN: Adaptive classifier that learns specific attack types.

  2. Bio-Plausible Learning:

    • Utilizes Grow When Required (GWR)-inspired structural plasticity.
    • Implements a novel Adaptive Spike-Timing-Dependent Plasticity (Ad-STDP) learning rule.

  3. Lifelong Learning Capabilities:

    • Enables the network to learn new threats incrementally.
    • Preserves existing knowledge (mitigating catastrophic forgetting).
    • Achieved 85.3% overall accuracy on the UNSW-NB15 benchmark in a continual learning setting.

  4. Hardware Efficiency:

    • Demonstrates high operational sparsity.
    • Optimized for low-power deployment on neuromorphic hardware (validated using Intel Lava framework).

Publication

This work was presented at the International Conference on Neuromorphic Systems (ICONS) 2025. Read the paper

References