Few-Shot Learning-Based Network Intrusion Detection through an Enhanced Parallelized Triplet Network

Figure 1

The architecture of building few-shot intrusion detection using the triplet network. The model with the same background colour in the diagram is functionally identical. For example, the refinement training module inherits the model trained by the initial training module and transfers that model to the detection module. The ellipsis between models in the detection module indicates where the model can be deployed in parallel.