Latest joint work of Ferenc Béres, István András Seres, András A. Benczúr and Mikerah Quintyne-Collins on Ethereum user profiling and deanonymization.
In this work we assess the privacy shortcomings of Ethereum's account-based model. We collect and analyze a wide source of Etherum related data, including Ethereum name service, Etherscan blockchain explorer, Tornado Cash mixer contracts, and Twitter. To the best of our knowledge, we are the first to propose and implement Ethereum user profiling techniques based on user quasi-identifiers. By learning Ethereum address representations we deanonymize accounts that belong to the same user.
In this repository we publish our data and code for further research, in the from of a Python package (ethprivacy).
You can find our pre-print paper on arXiv. Please cite our work if you use our code or the related data set.
@misc{beres2020blockchain,
title={Blockchain is Watching You: Profiling and Deanonymizing Ethereum Users},
author={Ferenc Béres and István András Seres and András A. Benczúr and Mikerah Quintyne-Collins},
year={2020},
eprint={2005.14051},
archivePrefix={arXiv},
primaryClass={cs.CR}
}
- UNIX environment
- This package was developed in Python 3.6 (conda environment)
After cloning the repository you can install the ethprivacy package with pip.
git clone https://github.com/ferencberes/ethereum-privacy.git
cd ethereum-privacy
python setup.py install
pip install karateclubYou must download our Ethereum data in order to use our code!
You can choose to use our download script below or just simply use this link.
bash download_data.sh
ls data- By running the following script you can check your setup.
bash -e run_tests.sh- We also provide a script to run every experiment from our paper. We recommend you to parallelize the tasks as it could take days to execute them on a single thread.
We thank Daniel A. Nagy, David Hai Gootvilig, Domokos M. Kelen and Kobi Gurkan for conversations and useful suggestions. Support from Project 2018-1.2.1-NKP-00008: Exploring the Mathematical Foundations of Artificial Intelligence and the “Big Data–Momentum” grant of the Hungarian Academy of Sciences.