DeepInvisi: A Fileless Malware Detection Tool Using Deep Learning Approach

Abstract

Fileless malware is malicious code operating directly in a computer's memory, bypassing traditional hard disk detection. In the digital age, it's a common threat using legitimate tools to evade security, making it extremely challenging to identify. This is critical as traditional anti-malware solutions like Kaspersky, TotalAV, and Aqua Security often have limited capabilities in comprehensively detecting these sophisticated, memory-resident threats, leaving organizations vulnerable. This project introduces DeepInvisi, a tool addressing this detection gap by using a hybrid deep-learning approach. Transformer models will analyze complex sequential data, while Bidirectional Long Short-Term Memory (BiLSTM) models handle temporal dependencies for comprehensive analysis of PowerShell command histories and memory activities. The proposed solution targets any users or normal people with electronic devices, aiming to protect them from fileless malware. The expected outcome is a real-time detection tool offering automatic updates, instant threat blocking, and real-time notifications for individual users and the general public.