Poster: Automated Neural Network Structure Selection for IoT Botnet Detection

IoT botnet attacks are a major concern these days and their detection is an active area of research. Artificial Neural Networks (ANNs) have proven their power and capabilities to detect botnets effectively. However, the process of ANN structure selection and training has been iterative and experimental where one starts with a random number of layers containing an arbitrary number of neurons within them. Experimental results reveal that this work provides massive gains in terms of computational efficiency over the manually selected network structures.