Hi-Graph / README.md
hzcheney's picture
Update README.md
00a0971 verified
metadata
configs:
  - config_name: default
    data_files:
      - split: train
        path: data/*.parquet

HiGraph: A Large-Scale Hierarchical Graph Dataset

Hierarchical Graph Dataset for Malware Analysis with Function Call Graphs and Control Flow Graphs

Hugging Face Dataset Size Function Call Graphs Control Flow Graphs Time Period License

A comprehensive hierarchical graph-based dataset for malware analysis and detection.

OverviewDataset StatisticsInteractive ExplorerDownload

Overview

HiGraph is a novel, large-scale dataset that models each application as a hierarchical graph: a local Control Flow Graph (CFG) capturing intra-function logic and a global Function Call Graph (FCG) capturing inter-function interactions.

Graph-based methods have shown great promise in malware analysis, yet the lack of large-scale, hierarchical graph datasets limits further advances in this field. This hierarchical design facilitates the development of robust detection models that are more resilient to obfuscation, model aging, and malware evolution.

Dataset Overview

Key Features

  • 🔍 Hierarchical Graph Structure: Two-level representation with FCGs and CFGs
  • 📈 Large Scale: 200M+ Control Flow Graphs and 595K+ Function Call Graphs
  • 🏷️ Rich Semantic Information: Preserves crucial structural details for malware analysis
  • 📊 Comprehensive Coverage: 11-year temporal span (2012-2022)
  • 🎯 Benchmark Ready: Designed for advancing hierarchical graph learning in cybersecurity

Dataset Statistics

Class Time Period # Apps Function Call Graph Control Flow Graph
Avg # Nodes Avg # Edges # Graphs Avg # Nodes Avg # Edges
Malicious 2012.01-2022.12 57,184 266.48 491.67 6,925,406 12.29 14.94
Benign 2012.01-2022.12 538,027 791.54 1,414.51 194,866,679 12.17 13.94

Interactive Visualization

Explore the hierarchical structure of malware samples through our interactive visualization tool:

🔗 Launch Interactive Explorer

Click to explore the complete dataset structure and sample graphs

Sample Function Call Graph Sample Function Call Graph (FCG) visualization

Download Dataset

Access the complete HiGraph dataset through multiple platforms:

Platform Description Link
🤗 Hugging Face Primary dataset repository View on Hugging Face
🌐 Project Page Interactive explorer HiGraph Explorer

Requirements

  • Python >= 3.9
  • torch==2.6.0
  • torch-geometric==2.6.1

Install dependencies:

pip install -r requirements.txt

📄 License

This dataset is licensed under the Creative Commons Attribution-NonCommercial-ShareAlike (CC-BY-NC-SA) license. See the LICENSE file for details.