Intrusion Detection System with Federated Learning using the Flower framework

Example of training of a simple Neural Network model as an Intrusion Detection System for Cybersecurity defense using Federated Learning with the Flower framework and the IEC 60870-5-104 Intrusion Detection Dataset.

Instructions

1. Run the Flower_IDS_IEC104_preprocessing.ipynb notebook for preprocessing the dataset and extracting individual straggler training data and test data.
2. Train and test the model with one of the following 2 options

• Option 1: Use the docker-compose.yaml to launch an aggretation server and three training stragglers.
• Option 2: Manually launch the aggregation server and then manually launch the three training stragglers.