zkSummit6: Ongoing challenges in systematizing universal SNARKs Mikerah Quintyne-Collins - Hashcloak

Transcript

[Music] hello i'm akira and together with carl we formed the hash clock research team the goal of this lightning talk is to give a very very broad overview of our ongoing work on systematizing universal snark constructions um so as many of you know there has been an onslaught of new universal strike constructions over the past year and a half almost two years now and it can be like very daunting for even experts to try and figure out like what do all of these schemes have in common

instead of etch out uh the landscape of universal snarks um you may notice that on this slide we include constructions that most people wouldn't typically consider universal snacks such as darks and bulletproofs so we do include constructions that require no trusted setup within this framework if this this is something the audience think is sort of not advisable please reach out um but the goal really is to find out what do all of these schemes have in common um what makes them different so currently what we have is

sort of this incomplete table um in the interest of brevity we couldn't include everything but at a very very high level we separated these these constructions into two columns those relying on transparent arguments and those relying on universal structured reference strings and then for each of these we've separated them into their proof types security models and mathematical building blocks in our paper we've also included details on their how they handle pre-processing and what kind of compilers they may acquire so as you can see for proof types

we include buddha proofs which we got from protrusion and zk study club so voter proofs are constructions based on the inner pairing uh sorry inner product um argument that was uh mainly like worked on by jonathan boodle um then we have iop base constructions and then we have pc plus ahp pc stands for polynomial commitments and ahp stands for algebraic holographic proofs and then we have other and this is sort of like a catch-all term for those that we couldn't really edge out a

proof type uh for example mirage requires a new construction that they call separated zk snarks um if you have any suggestions for different kinds of proof types we're open to that for cryptographic model it's all the usual ones discrete log random oracle models algebraic group model engineer group model we have a few examples of constructions that fit each each of those models and then we have the mathematical building blocks that these different constructions rely on so you have bilinear groups polynomial commitments

groups of unknown order hash functions and others in our paper and other there are other sort of properties that we consider as well that's not on the slide um we consider the compiler the prover and verifier efficiencies in terms of space time and time and we take interest in the pre-processes and steps in addition to any implementation details so our current blockers for this paper are finding any intrinsic differences between

the zero knowledge versions of these snug constructions it's easy to like write them down and sort of not have a high level overview but we really want to uh summarize and etch out these differences um second we want to go deeper into the types of setups and how they construct their arguments and then go into implementation details such as their algorithms any like they use whether they construct new languages etc so that's it if you want to reach out to us you can reach out to us on twitter or you can send us an email thank you very much hey okay i think that actually it looks like

dara had some comments about some of the things that you've just spoken about um that might be something worth talking about in the uh party afterwards when we can all chat together but i guess like my my uh oh yeah jordan's recognizing you here i think you're so famous man you're famous um so what do you like what do you think's next with this project what are you going to be focusing on to make it happen uh i guess we we need to like better i guess

but internally for this we've been just like reading papers as they come out in addition to like the previous ones uh so one like a big issue is when do we stop um like just make us reading papers that are good or new then the second issue is what are some things that people haven't really sympathized in other surveys that we should focus more on in this paper what do you think's the trick what do you think's the trickiest part about um diving into all of this uh well for me it was just like revisiting computational complexity theory

i mean i took i took a course in university and then i read a book but in like industry don't use it that much but then with the ck stuff if you're reading papers it mentions that stuff a lot for me i just had to like reread all that other background um i guess for carl it's probably similar but he graduated relatively recently so he probably has that knowledge fresh a little fresher yeah cool um so i guess people should get in touch to work with you on this is there any kind of call to action i know you you had one but maybe you want to say it again a call to action uh like no request for comments i guess yeah i guess send me a

twitter dm or telegram vm or an email or something cool thanks a lot then