VizNN : Visual Data Augmentation with Convolutional Neural Networks for Cybersecurity Investigation

This week, let’s dive in the use of an unusual method to reach a goal you know much about: automatic attack detection.

The work presented in this article uses a neural network method to graphically represent event logs and learn from them. Once the graphical representation is obtained, the images can be studied using the most relevant newspaper attributes to look for recurring patterns to identify habitual and inhabitual behaviour.

This approach is tested against the DoH attack (DNS over HTTPS) and obtains promising results !

https://arxiv.org/pdf/2112.05657.pdf#page=203