Protecting national security software against unauthorized distribution and exploitation is necessary for the protection of intellectual property (IP) and secrets in our nation. Modern software attack and analysis tools have become, along with their users, more skilled at analyzing and modifying programs. Currently, technologies that protect, detect, and react (PDR) to the exploitation of binary executables are inserted in a manual or semi-automated fashion. This method of insertion, especially manual insertion, is labor intensive, time-consuming, and subject to human error. This proposal describes a method for utilizing current PDR insertion and measurement technologies which can be used to build an automated tool that quickly and securely generates binary executable protections that fit within given performance and size constraints. The approach described is unique in that it uses artificial intelligence to evolve optimal protections. Such an approach could pave the way for future software security standards.
Keywords: SENSORS, ELECTRONICS, BATTLESPACE, SOFTWARE PROTECTION, ARTIFICIAL INTELLIGENCE
