# Project Title: Sybil Hunter
Empowering Decentralization: Machine Learning's Sybil Actor Solution
## Project Overview
In the pursuit of enhancing fairness and integrity in airdrop campaigns within blockchain ecosystems,
our project aims to tackle the rising issue of sybil attacks. Sybil actors exploit airdrop distribution
mechanisms, posing a threat to decentralization and community engagement. Our solution leverages machine
learning techniques to detect clusters of similarly behaving wallets and identify potential airdrop hunters,
preserving fairness and privacy.
## Table of Contents
- [Background](#background)
- [Features](#features)
- [Technologies Used](#technologies-used)
- [Usage](#usage)
- [Challenges and Solutions](#challenges-and-solutions)
- [Future Enhancements](#future-enhancements)
## Background
Airdrops have become a popular way for projects to distribute tokens and engage the community. However, this
has led to the emergence of sybil attacks, where malicious actors exploit the system to gain undue rewards.
Our project addresses this challenge by applying machine learning to identify suspicious wallet clusters,
mitigating the risk of airdrop hunters.
## Features
- Machine Learning Detection: Employing machine learning models to identify clusters of wallets that can be attributed to one airdrop hunter.
- Data Collection and Analysis:Gathering on-chain transaction data, preprocessing, and applying clustering algorithms.
- Privacy-Preserving: Unlike traditional KYC methods, our approach maintains users' privacy while identifying malicious actors.
## Technologies Used
- K-Means Clustering
- DeepWalk Algorithm
- Louvain Method
- Graph2Gauss
## Usage
1. Collect on-chain transaction data.
2. Preprocess the data using provided scripts.
3. Apply the clustering algorithm to identify wallet clusters.
Due to the time constraints of the hackathon and Github's file size limit, we weren't able to publish the datasets here. If you want to skip the data collection step and run our scripts on the prerpared data, please reach out to any of the authors and we will be happy to provide you with the files.
## Challenges and Solutions
- Data Gathering: Slow RPC speeds and unreliable connections were tackled by developing a framework for efficient dataset downloads.
- Computational Complexity: The DeepWalk algorithm's resource consumption was mitigated by focusing on smaller datasets.
- Large Sparse Matrices: Applying Louvain and Connected Components methods was enhanced with increased computational resources.
## Future Enhancements
- Incorporating more advanced machine learning models.
- Enhancing the identification of specific (similar within the clusters) behaviors associated with airdrop hunters.
- Integrating the detection mechanism into a user-friendly interface.
## Authors
- Stefanie Mayer ([email protected])
- Denys Kosianchuk ([email protected])
- Andrii Yakovyna ([email protected])
- Jakub Warmuz ([email protected])
-
Notifications
You must be signed in to change notification settings - Fork 1
YaKalmar0/sybil_hunting
Folders and files
Name | Name | Last commit message | Last commit date | |
---|---|---|---|---|
Repository files navigation
About
No description, website, or topics provided.
Resources
Stars
Watchers
Forks
Releases
No releases published
Packages 0
No packages published