Machine Learning Based Approach to Recommend MITRE ATT&CK Framework for Software Requirements and Design Specifications : AIMS (Artificial Intelligence, Machine Learning, and Security) Research Group

Abstract

As cyber-attacks increase in volume and sophistication, the current state of cybersecurity solutions is inadequate. According to Red Canary, Advanced Persistent Threat (APT) attacks have increased from approximately 500 attacks per year in 2009 to almost 2,500 APT attacks per year in 2019. The lack of timely detection and response is mainly caused by the insufficient support of attack action correlations and prediction to allow for proactive intrusion, investigation, and mitigation. There are lots of different ways to try and get past security, and it is difficult to create a system that protects against every single one. With the vast amount of data available online, it is necessary to integrate security-related activities and deliverables into each phase of the software development life cycle.