Sustaining Future Digital Forensics Through Intelligent Automation

This thesis examines the application of machine learning to policing. First, it identifies and `maps' online criminality, forming a framework in order to allow software to `learn' criminal behaviour. Then, it demonstrates automated detection and classification of child exploitation imagery, via open source software capable of running on low-end hardware. This capability represents a major leap in the investigation and combating of online radicalisation and trade in child exploitation materials.