Description
Binary type inference is a crucial technique for understanding the behavior of compiled code. While type
information is essential for understanding the behavior of code, it is erased during the compilation process,
making it challenging to recover. However, by analyzing the patterns of data access and usage within
binary code, type inference methods can reconstruct this lost type information.
This thesis aims to provide a comprehensive overview of binary type inference techniques. We have
collected, summarized, and evaluated common approaches, exploring their theoretical capabilities and
practical usefulness. By understanding the strengths and limitations of different methods, we can identify
areas for improvement and inform future research in this field.
In addition to theoretical analysis, we will also conducted practical experiments to evaluate the performance
of different binary type inference methods. We used real-world binaries as test cases and assessed
the accuracy, efficiency, and scalability of these methods. Our goal is to provide a clear and objective comparison
of the available techniques.
|
