The aim of the fellowship is to devise computer-based tools that will accelerate the discovery of security flaws. These tools will automate the labour-intensive tasks that a white-hat hacker has to undertake when searching for vulnerabilities. Vulnerability discovery is notoriously time-consuming because a white-hat often does not have access to the human-readable version of a program - only the binary executable (assembler) version. Thus a major impediment to finding vulnerabilities is the time required to understand a program given its executable. The project will address this problem by developing (program analysis) techniques that will automatically recover information about the behaviour of a program, and then present it in a digestible form to the white-hat.