Antigranular: How to Access Sensitive Datasets Without Looking At Them

Author(s): Bex T.

Originally published on Towards AI.

and developing privacy-first data projects
Image by me with Leonardo AI

Today, open-source data represents the tip of the world’s data iceberg. Most of the world’s data is below the ocean surface because of privacy concerns. Imagine if we all could access that great mound of underwater data without any privacy concerns. How would it be if we could create software and tools that allowed us to use that data without violating anyone’s privacy?

This was a question asked long ago (early 2000s) and many solutions have been developed since then. A widely adopted among them is differential privacy (DP).

DP is a powerful framework to mask the contributions of individuals in datasets. It ensures that computations performed on sensitive data do not expose any specific information about its participants. Leveraging DP correctly would allow us to work with even the most sensitive datasets (like census data) without actually looking at individual rows.

In this article, we will learn how to use differential privacy in practice using a platform called Antigranular.

Image used with permission

Antigranular is a Kaggle-like data science competition platform with a twist: instead of using open-source or publicly available data, participants are given sensitive datasets to solve a machine learning challenge.

The data itself is protected by the… Read the full blog for free on Medium.

Join thousands of data leaders on the AI newsletter. Join over 80,000 subscribers and keep up to date with the latest developments in AI. From research to projects and ideas. If you are building an AI startup, an AI-related product, or a service, we invite you to consider becoming a sponsor.

Published via Towards AI