Skip to main content

Food Security Sandbox

A collaborative machine learning platform for agricultural data analysis and model training with privacy-preserving features.

The Food Security Sandbox is a comprehensive web application that enables farmers and researchers to collaborate on machine learning models while preserving data privacy. The platform provides tools for dataset management, collaborative model training, and privacy risk analysis in agricultural contexts.

GitHub Repo License

Key Features:

  • Dataset upload and management
  • Collaborative machine learning model training
  • Privacy-preserving data sharing
  • Model repository with risk analysis
  • Chat for collaboration
  • Similar farmer identification

References​

Key Technologies and Libraries​

  • Frontend: React.js, Material-UI, Axios
  • Backend: Flask, Python, MongoDB
  • Machine Learning: TensorFlow, Scikit-learn, Adversarial Robustness Toolbox
  • Privacy: Differential Privacy, Membership Inference Attack Detection
  • Authentication: TACC Tapis Authentication

External Documentation​

Key Concepts​

  • Differential Privacy: A mathematical framework for providing privacy guarantees when analyzing data
  • Membership Inference Attack: An attack that determines whether a particular data point was used to train a machine learning model
  • Collaborative Learning: A machine learning approach where multiple parties contribute to model training without sharing raw data
  • Federated Learning: A machine learning technique that trains an algorithm across multiple decentralized edge devices or servers

Acknowledgements​

This research was supported in part by the National Science Foundation (NSF) under awards OAC-2112606 and 2112533. Also, this research was partly supported by the United States Department of Agriculture (USDA) under grant number NR233A750004G019.