BERT4ETH (PyTorch Version)

This is the PyTorch implementation for the paper BERT4ETH: A Pre-trained Transformer for Ethereum Fraud Detection, accepted by the ACM Web conference (WWW) 2023.

The implementation of fine-tuning on phishing data was finished by Qiustander (https://github.com/Qiustander). Thanks a lot!

If you find this repository useful, please give us a star and cite our paper : ) Thank you!

Getting Start

Requirements

PyTorch > 1.12.0

Preprocess dataset

Step 1: Download dataset from Google Drive.

• Transaction Dataset:

• • Phishing Account
• • De-anonymization(ENS)
• • De-anonymization(Tornado)
• • Normal Account

• ERC-20 Log Dataset (all in one)

Step 2: Unzip dataset under the directory of "BERT4ETH/Data/"

cd BERT4ETH_PyTorch/data; # Labels are already included
unzip ...;

Pre-training

Step 1: Transaction Sequence Generation

cd src;
python gen_seq.py --bizdate=bert4eth_exp

Step 2: Pre-train BERT4ETH

python run_pretrain.py --bizdate="bert4eth_exp" \
                       --ckpt_dir="bert4eth_exp"

Step 3: Output Representation

python run_embed.py --bizdate="bert4eth_exp" \
                       --init_checkpoint="bert4eth_exp/xxx.pth"

Evaluation

Phishing Account Detection

cd eval
python phish_detection_mlp.py --input_dir="../outputs/xxx"

De-anonymization (ENS dataset)

python run_dean_ENS.py --metric=euclidean \
                       --init_checkpoint=bert4eth_exp/model_104000

Fine-tuning for phishing account detection

  Will update later..

---

Citation

@inproceedings{hu2023bert4eth,
  title={BERT4ETH: A Pre-trained Transformer for Ethereum Fraud Detection},
  author={Hu, Sihao and Zhang, Zhen and Luo, Bingqiao and Lu, Shengliang and He, Bingsheng and Liu, Ling},
  booktitle={Proceedings of the ACM Web Conference 2023},
  pages={2189--2197},
  year={2023}
}

---

Q&A

If you have any questions, you can either open an issue or contact me (sihaohu@gatech.edu), and I will reply as soon as I see the issue or email.