Automatic Reverse Compilation & Semantic Comprehension
The machine code to source code problem can be simplified by relaxing the problem constraints and allowing translation from machine code to any reasonably sound higher-level code interpretation instead of expecting a close approximation of the original source code. As a corollary, any solution to the problem also demonstrates semantic comprehension of the machine code.
Automatic reverse compilation of machine code to source code and the semantic comprehension of machine code without code execution or emulation can be achieved by combining techniques used in data compression, automatic differentiation, linear algebra, mathematical logic, syntactic analysis and statistical classification. A simplified application of the same tools can be applied to predict pseudo-random numbers and sequences of other mathematical objects.
Speakers
 |
Christopher Abad (aempirei) |