Logo
ScienceStack
Table of Contents
Fetching ...
Sign In

Know Your Account: Double Graph Inference-based Account De-anonymization on Ethereum

Shuyi Miao, Wangjie Qiu, Hongwei Zheng, Qinnan Zhang, Xiaofan Tu, Xunan Liu, Yang Liu, Jin Dong, Zhiming Zheng

TL;DR

This paper tackles Ethereum account de-anonymization under label scarcity and evolving transaction patterns. It introduces DBG4ETH, a double-graph framework combining a Global Static Graph and a Local Dynamic Graph, complemented by an adaptive confidence calibration mechanism. The approach, including contrastive learning, temporal graph encoding, and a LightGBM classifier, achieves state-of-the-art results across six account types and demonstrates robustness to novel categories like bridge and DeFi with limited labeled data. The work advances regulatory and security-oriented analyses in Web3 by providing reliable, explainable predictions in dynamic blockchain environments.

Abstract

The scaled Web 3.0 digital economy, represented by decentralized finance (DeFi), has sparked increasing interest in the past few years, which usually relies on blockchain for token transfer and diverse transaction logic. However, illegal behaviors, such as financial fraud, hacker attacks, and money laundering, are rampant in the blockchain ecosystem and seriously threaten its integrity and security. In this paper, we propose a novel double graph-based Ethereum account de-anonymization inference method, dubbed DBG4ETH, which aims to capture the behavioral patterns of accounts comprehensively and has more robust analytical and judgment capabilities for current complex and continuously generated transaction behaviors. Specifically, we first construct a global static graph to build complex interactions between the various account nodes for all transaction data. Then, we also construct a local dynamic graph to learn about the gradual evolution of transactions over different periods. Different graphs focus on information from different perspectives, and features of global and local, static and dynamic transaction graphs are available through DBG4ETH. In addition, we propose an adaptive confidence calibration method to predict the results by feeding the calibrated weighted prediction values into the classifier. Experimental results show that DBG4ETH achieves state-of-the-art results in the account identification task, improving the F1-score by at least 3.75% and up to 40.52% compared to processing each graph type individually and outperforming similar account identity inference methods by 5.23% to 12.91%.

Know Your Account: Double Graph Inference-based Account De-anonymization on Ethereum

TL;DR

This paper tackles Ethereum account de-anonymization under label scarcity and evolving transaction patterns. It introduces DBG4ETH, a double-graph framework combining a Global Static Graph and a Local Dynamic Graph, complemented by an adaptive confidence calibration mechanism. The approach, including contrastive learning, temporal graph encoding, and a LightGBM classifier, achieves state-of-the-art results across six account types and demonstrates robustness to novel categories like bridge and DeFi with limited labeled data. The work advances regulatory and security-oriented analyses in Web3 by providing reliable, explainable predictions in dynamic blockchain environments.

Abstract

The scaled Web 3.0 digital economy, represented by decentralized finance (DeFi), has sparked increasing interest in the past few years, which usually relies on blockchain for token transfer and diverse transaction logic. However, illegal behaviors, such as financial fraud, hacker attacks, and money laundering, are rampant in the blockchain ecosystem and seriously threaten its integrity and security. In this paper, we propose a novel double graph-based Ethereum account de-anonymization inference method, dubbed DBG4ETH, which aims to capture the behavioral patterns of accounts comprehensively and has more robust analytical and judgment capabilities for current complex and continuously generated transaction behaviors. Specifically, we first construct a global static graph to build complex interactions between the various account nodes for all transaction data. Then, we also construct a local dynamic graph to learn about the gradual evolution of transactions over different periods. Different graphs focus on information from different perspectives, and features of global and local, static and dynamic transaction graphs are available through DBG4ETH. In addition, we propose an adaptive confidence calibration method to predict the results by feeding the calibrated weighted prediction values into the classifier. Experimental results show that DBG4ETH achieves state-of-the-art results in the account identification task, improving the F1-score by at least 3.75% and up to 40.52% compared to processing each graph type individually and outperforming similar account identity inference methods by 5.23% to 12.91%.

Paper Structure

This paper contains 46 sections, 25 equations, 9 figures, 6 tables.

Table of Contents

  1. INTRODUCTION
  2. BACKGROUND AND RELATED WORK
  3. Ethereum Account Model
  4. De-anonymization Inference
  5. Graph Representation Learning
  6. Graph Embedding Methods
  7. GNN-based Methods
  8. GNN-based De-anonymization Methods
  9. DOUBLE GRAPH CONSTRUCTION
  10. Problem Definition
  11. Ethereum Data Processing
  12. Transaction Data Filtering
  13. Node Feature Construction
  14. Interaction Merging and Edge Feature Construction
  15. METHODOLOGY
  16. ...and 31 more sections

Figures (9)

  • Figure 1: Details about the challenge (i): (a) and (b) have similar static graphs, but completely different dynamic graphs. (b) and (c) have similar dynamic graphs, but different static graphs.
  • Figure 2: The overview of our DBG4ETH.
  • Figure 3: Adaptive calibration method for two types of prediction values: global static graph and local dynamic graph.
  • Figure 4: Heat map of 15-dimensional attribute correlation of node features.
  • Figure 5: Scatter plot of account category features.
  • ...and 4 more figures