DEBIN uses machine learning to recover debug information (e.g., names and types) of stripped binaries (x86, x64, ARM).
This is helpful for various binary analysis tasks such as decompilation, malware inspection and similarity.
DEBIN is a novel system for predicting debug information in stripped binaries. It is able to distinguish register-allocated and memory-allocated variables with decision-tree-based classification. Moreover, it is capable of predicting meaningful names and types for variables and functions through structured prediction with probabilistic graphical models. These models are learned from thousands of non-stripped binary in open source packages. The system can be further used for malware inspection.
If you have any feedback, suggestions or want to use DEBIN for larger binaries, please email Martin Vechev by firstname.lastname@example.org.
To use our application with all its potential, please use a tablet device display or larger. Our prediction results cannot be properly displayed on a small screen.
Thank you, DEBIN Team
or try samples:
Please address the following errors:
The file you uploaded is not in the correct file format. Please upload a binary of correct format.
Error transforming binary code into intermediate representation.
Please select a binary.
Error getting file.
Error while processing the binary file.
Error uploading binary file. Please retry uploading a file.
Error getting prediction status.
Error generating output binary.
The architecture of the binary is not supported. It should be any of x86, x64 and ARM (without thumb instructions).
The format of the binary is not supported. It should be ELF format. Please retry another one.
There is no .text section in the input binary or section headers are removed. Please retry another one.
The file you uploaded is larger than 2MB. Please upload a smaller file.