evandowning / deepreflect Goto Github PK
View Code? Open in Web Editor NEWDiscovering Malicious Functionality through Binary Reconstruction
License: GNU General Public License v3.0
Discovering Malicious Functionality through Binary Reconstruction
License: GNU General Public License v3.0
Hi,
Sir, I need some assistance if possible for you. As I am trying to use IDAPRO instead of BN tool, so for this first i have understand your "extraction" code of python and found following 4 relevant files.
Please guide me if you find any problem in my understanding of code flow and conversion.
Can you please comment on my understanding of your code flow for BN part and my implementation flow part as well. I will be thankful to you.
Hi, thanks for the interesting research work.
I'm trying to reproduce the ROC results for the ground truth dataset. However, I cannot find the function information for malware-gt from the dataset link (for the malware used in training and testing, there is malicious_unipacker_bndb_function.tar.gz
).
Is there anything I miss? Could you please share the function information of the three ground truth samples?
Many thanks!
Traceback (most recent call last):
File "pca_hdbscan.py", line 16, in
import hdbscan
File "/home/user/anaconda3/envs/dr/lib/python3.7/site-packages/hdbscan/init.py", line 1, in
from .hdbscan_ import HDBSCAN, hdbscan
File "/home/user/anaconda3/envs/dr/lib/python3.7/site-packages/hdbscan/hdbscan_.py", line 21, in
from ._hdbscan_linkage import (single_linkage,
File "hdbscan/_hdbscan_linkage.pyx", line 1, in init hdbscan._hdbscan_linkage
ValueError: numpy.ndarray size changed, may indicate binary incompatibility. Expected 88 from C header, got 80 from PyObject
real 0m0.646s
user 0m2.506s
sys 0m2.723s
what's wrong with me?
Hello! I am very interested in your research work and I am trying to reproduce the research results of your paper. As described in Section 4.2 (Model Setup) of your paper:
We trained the model for a maximum of 10 epochs and we obtained a training MSE of 2.5090e-07 and testing MSE of 2.1575e-07 – recall that a lower the MSE value means a better reconstruction of the benign samples. It took roughly 40 hours to train the model on an NVIDIA GeForce RTX 2080Ti GPU.
I run this AutoEncoder model on an Nvidia Quadro RTX 5000, but after trying a few times I only get an MSE value of about 5.4e-07 on the training set and an MSE value of about 5.1e-07 on the validation, which is twice as large as the result stated in your paper! The resullt of the next step to extract ROIs on the malicious dataset is also far from the results you gave.
I would like to ask, what is the reason for this result? Is there any way to achieve good results as described in your paper?
lose = nan - mse = nan while training the model and not improving the values
I used features download from Dataset.md/feature.tar.gz. Then I tried to run the command "python feature_check.py train.txt", where "train.txt" was copied from "Dataset.md/models.zip". It would raised aboved errors. What's wrong with it?
Hi,
Psycopg2 is installation gives error while installation of required packages for DeepReflect. However, installation of following packages resolve the error of psycogp2 tool.
apt-get install libpq-dev python-dev
Hi,
Sir,
Thanks for sharing wonder work and it is an interesting and tough research area. I am going to start work in this area however right now i am confused in many things. However, i came across this wonderful work.
I have query after reading your paper . As you mentioned that you are going to use malware as a input . However, can i use legitimate binary which contains malicious working, can work in your code or not. I mean malicious working in a legal program like calculator, can we detect its malicious functionality via this method or not. can you give your opinion on it. please
Hi,
Sir. I am getting error of BinaryNinja however i have it in the binja_setup directory. Docker commands were not working. so i am trying to use binja.y file to see the extraction step.
As you answer me in my previous question about IDA Pro usage instead of BinaryNinja tool. Sir, can i perform the same extraction using IDA-PRO. I mean to say, same features can be extracted using any other tool. I know i have to modify the code for this step because of this license issue.
dr$python3 binja.py --exe /data/benign_unpacked/benign/w64.exe --output /data/benign_unpacked_raw//benign/w64.exe.bndb
Traceback (most recent call last):
File "binja.py", line 51, in
_main()
File "binja.py", line 35, in _main
bv = BinaryViewType.get_view_of_file(fn)
File "/home/naveed/Desktop/deepreflect/binja_setup/binaryninja/python/binaryninja/binaryview.py", line 766, in get_view_of_file
view = BinaryView.open(filename)
File "/home/naveed/Desktop/deepreflect/binja_setup/binaryninja/python/binaryninja/binaryview.py", line 1382, in open
binaryninja._init_plugins()
File "/home/naveed/Desktop/deepreflect/binja_setup/binaryninja/python/binaryninja/init.py", line 174, in _init_plugins
raise RuntimeError("License is not valid. Please supply a valid license.")
RuntimeError: License is not valid. Please supply a valid license.
A declarative, efficient, and flexible JavaScript library for building user interfaces.
🖖 Vue.js is a progressive, incrementally-adoptable JavaScript framework for building UI on the web.
TypeScript is a superset of JavaScript that compiles to clean JavaScript output.
An Open Source Machine Learning Framework for Everyone
The Web framework for perfectionists with deadlines.
A PHP framework for web artisans
Bring data to life with SVG, Canvas and HTML. 📊📈🎉
JavaScript (JS) is a lightweight interpreted programming language with first-class functions.
Some thing interesting about web. New door for the world.
A server is a program made to process requests and deliver data to clients.
Machine learning is a way of modeling and interpreting data that allows a piece of software to respond intelligently.
Some thing interesting about visualization, use data art
Some thing interesting about game, make everyone happy.
We are working to build community through open source technology. NB: members must have two-factor auth.
Open source projects and samples from Microsoft.
Google ❤️ Open Source for everyone.
Alibaba Open Source for everyone
Data-Driven Documents codes.
China tencent open source team.