Hack me if you can: Aggregating autoencoders for countering persistent access threats within highly imbalanced data

Sidahmed Benabderrahmane, Ngoc Hoang, Petko Valtchev, James Cheney, Talal Rahwan

Abstract: Highlights•Present the development of a deep learning-based pipeline for the detection of APTs, using various auto-encoder architectures.•Design of a baseline AutoEncoder and five variants thereof (adversarial, recurrent, long short-term memory, gated recurrent units, attention-based).•Design an ensemble learning mechanism on top of AutoEncoder-based classifiers, in a platform-agnostic implementation.•Experimental evaluation of the proposed models using large APT databases.•Comparison of these models to several baseline approaches.