🗝️

Fully Homomorphic Encryption (FHE)

The original way I found about this was when I was searching for a way to use an individual's Crypto Wallet (without the server actually knowing the key) and trade, stake, etc.
So, the question I had was is it possible to do computations and calculations on data without actually not knowing what you (aka server) actually is computing. Is it even possible?
Well, the answer turns out that it is possible and there are teams working on trying to do that.
And it is called Homomorphic Encryption. Although, I am not qualified enough to even understand the math behind it - it seems realllly cool.
Fully Homomorphic Encryption (FHE) is an emerging cryptographic technique that allows developers to perform computations on encrypted data. This represents a paradigm shift in how data processing and data privacy relate to each other.
https://6min.zama.ai/
 
Imagine Algorithms recomending you your favourite websites, without knowing about you, nor about the websites they recomended. Genuinenly amazing and bigger implications for healthcare, research, privacy in general usecases, and more.
 
The only problem as far as I have understood is that it takes really long to process a single gate (AND, OR, XOR, etc.) operation. As for addition itself - based on one algorithm: you need Matrices, being multiplied and added, filtering noise from it. All which takes lots of cycles even on modern computer.
 

Resources to Learn more about it

 
 
Document page of the TFHE github page. Has some interesting links, including reference to research papers and existing implementations.
 
 
 
Library that converts normal C++ code into FHE compatible code.
 
Using CUDA to accelerate FHE computation.