# Constructions of PRF (Pseudo Random Function)

I was taught only GGM based PRF construction in class. It's very inefficient. I am just curious about various PRF constructions from standard assumptions. Please provide a few PRF constructions from various assumptions.

• Well, one standard assumption is that the SHA-256 compression function is a PRF, from which we can conclude that HMAC-SHA256 is a PRF (and a reasonably efficient one at that), but maybe you meant to restrict the domain of ‘standard assumptions’? – Squeamish Ossifrage Apr 26 at 18:44

The most common efficient PRFs from specific assumptions are:

The Naor-Reingold PRF, which is based on the decision Diffie-Hellman assumption (DDH), and

The BPR PRF, which is based on the learning with error assumption (LWE).

Perhaps slightly less well-known is the NRR PRF, which is based on the hardness of factoring.