- tags
- Cryptography
- resources
- Vitalik Buterin’s blog
Principle
The idea of homomorphic encryption is to encrypt data in such a way that, given a function \(f\) and a message to encrypt \(x\), \(\text{enc}(f(x)) = f(\text{enc}(x))\).
This idea is similar in spirit to Privacy-preserving machine learning, or federated learning, where one wants to obfuscate data while still being able to use it in a learning model. Here, one considers arbitrary functions.