One sentence description:

• Two parties collaborate on correlation finding without sharing private data.

Example use case:

• Imagine two government organizations: Health Ministry and Census Ministry
• They want to collaborate in identifying risk of diseases by making correlations on factors like place of living, age, occupation, etc. (for this use case we´ll use a toy example of an area with 2 regions)
• They want to collaborate without sharing any of their data with the other organization.
• The Health Ministry has information about people´s diseases.
• The Census Ministry has information about people´s place of living, age, occupation, etc.
• We want to combine the data sets

Deeper Explanation:

This is a toy example meant to be used as proof of concept (POC).

• The Health Ministry has the list of people with diseases (0 or 1 to simplify)
• The Census Ministry has the list of people and their location (0 or 1 to simplify)
• An MPC protocol relates Alice records on Health Ministry and Census Ministry.
• MPC does A0B0+A1B1+….+An*Bn
• At the end we have a number of people who got a disease per region.
• This has to be normalized per region to give us a percentage.
• If we find a high correlation, we got a target to look at!
• i.e. In a population of 200 people where 100 people live in each region. 80 people from region 0 got the disease vs 2 people from region 1. This means region 0 has 80% chance of getting a disease vs 2% chance from region 1.