Towards a Federated Intrusion Detection System based on Neuromorphic Computing

Intrusion Detection Systems (IDSs) have emerged as essential tools for detecting cyber attacks and safeguarding sensitive data. Over time, there has been a shift towards designing IDSs that leverage Federated Learning (FL) methods, enabling them to detect attacks across distributed environments while upholding privacy-preserving manner. Concurrently, selecting the appropriate algorithm for Host execution, ensuring data privacy, low power consumption, and swift execution, has become a promising challenge. Recently, there has been a growing interest in Spike Neural Networks (SNNs) due to their ability to directly generate spikes and closely emulate human brain functions. SNN-based models are optimized to achieve energy efficiency by representing computations through asynchronously generated spikes. To tackle these challenges, we propose a theoretical approach for implementing a Federated Intrusion Detection System (IDS) that gathers data from different geographical locations, based on Neuromorphic Computing principles.