Table 2 Intrusion detection systems based on federated learning.
From: A federated transformer-enhanced double Q-network for collaborative intrusion detection
Refs. | ML Algorithm | Dataset | Aggregation Function | non-IID | Generalization |
|---|---|---|---|---|---|
Stacked-Unsupervised43 | Stacking Deep Autoencoder | TON_IoT, Bot-IoT, CICIDS2018, UNSW-NB15 | FedAvg | Â | \(\surd\) |
VAN-IDS47 | LightGBM | Vein | FedAvg | \(\surd\) | Â |
FELIDS44 | DNN, CNN, RNN | MQTTset | Fed+ | Â | Â |
FL for Industrial IoT45 | DNN | CIC-ToN_IoT | FedAvg, Fed+ | \(\surd\) | Â |
FDRL-IDS46 | Q-net | NSL-KDD, ISOT-CID | Majority Vote | Â | Â |
Ours | Q-Network, Transformer | TON_IoT-v2, Bot-IoT-v2, CICIDS2018-v2 UNSW-NB15-v2 UNSW-MG24 CIC IIoT 2025 | Dynamic Contextual weights | \(\surd\) | \(\surd\) |
