This is the implementation of GEMAL.
For reverse a binary and extract function call graphs(FCGs), please refer frtools.
To train a word2vec model by yourself, check train_word2vec.py, which you'll have to reverse the binaries according to your needs. And we have already provided our pre-trained one in model_saved/
To train your own detector, the whole training progress is in implement_Wu.ipynb, follow the steps inside, training a new one would not be a big deal.
As the original work is only implemented on x86 family, we follow its setting. And since our reverse tool is different from the original work, it might vary a little compare to that work.