CONSTRUCTION OF A BLSTM NEURAL NETWORK TRAINED ON THE UNSW0NB15 DATASET FOR THREAT DETECTION IN IOT NETWORKS

With the rapid growth of the Internet of Things(IoT), connected devices are increasingly exposed to sophisticated threats that compromise network security. This article proposes an approach based on a bidirectional long short-term memory (BLSTM) neural network to effectively detect intrusions in IoT environments.The model is trained and evaluated on the UNSW-NB15 dataset, which offers a wide variety of realistic attack types.BLSTM architecture allows temporal dependencies in network data flows to be captured, thereby improving detection accuracy compared to traditional approaches. Experimental results demonstrate that our model outperforms several conventional detection methods in terms of accuracy, recall and false positive rates. These results highlight the potential of BLSTM networks as a robust solution for enhancing cybersecurity in IoT networks.