ZKV presents... Privacy on Polkadot - Plug and Play Privacy for Polkadot w Shumo Chu, Manta Network

Transcript

[Music] yeah first uh thanks anna swore like organizing such a went for event i'm going to talk about manta which is a plug and play privacy for polka dot um so first tell you guys a little bit about me so i'm a co-founder of mental network also assistant professor at ucsb so a little bit history for me into the blockchain space i actually ran the first blockchain seminar at the university of washington where i still a graduate student there and then i become a research scientist at algorand

and for my own research i'm very interested in applied crypto private privacy preserving ml and the broadcast systems so since in the last few years have been seattle and then boston and santa barbara uh please don't ask me which one is my favorite city um so i don't think we i need to motivate you uh you guys about the cryptocurrency it's really eating a word and i think for the good i think a common problems of the cryptocurrency is that uh we probably all know that pseudo anonymity doesn't mean privacy right it it's actually pretty against privacy in the sense that like we have a public ledger and uh like everyone's transaction record is on chen

and probably forever right if you think about that that's pretty scary and um don't take my word i actually take the mr or mrs centas santoshi nakamoto's word from himself right um he probably argues that we need a new address for each transactions um so actually what santoshi nakamoto didn't envision that that's probably not enough so for example even if you have a new address each time right so the machine learning techniques are becoming extremely powerful these days so they can do the link analysis basically tracking the lineage of the transaction record

and even if you get a new address so so like all the privacy privacy mining companies still can probably get you right so um so to to motivate you a little bit more so we actually did a privacy survey right so a lot of people argue that hey like privacy is that thing that everyone is talking about nobody really cares but at least from our survey results right we survey a wide range of users from with a different kind of like experience and different years of in into crypto right i think

a majority of users that actually concern the linkage between their identity and their addresses and the next result is even more interesting which is like 90 percent of the users have inspect others wallet addresses and there are like more than like a 70 percent of users have avoid or hesitate to make a transaction because of privacy so i think the problem here is not really people don't want privacy it's like we don't have a easy to use privacy for the users and i don't think the tech is there right so people usually talk about the privacy kind of like a utility trade-offs i think there are some in inherent trade-offs

but i do think we can push the perito kind of like a pareto frontier in a sense that we can make the privacy techniques more friendly to the users so that's brought you to mantas mission so we are trying to be the plug-and-player plug-and-play privacy layer for polka dots so manta is going to um going through many uh three phases basically so the first uh mental 1.0 we call the decentralized anonymous payment system right so this is actually the fundamental of uh of mantas tech so the first thing we are going to build and have actually have already been partially built is a decentralized anonymous payment system

which is essentially a privacy layer for the founder foundable token transfer right so the process that you can use menta is that uh you can take your own like a public crypto token and then through the mint process and the mint in mentor and will give you a private token and you can use this private token to transfer to any like a shielded address you want and also if you want to get back to your puppy crypto tokens you can go through the reclaim as well so the meta protocol itself is pretty similar to zcash with some improvements uh for example we're using like shorty the merkle trees

there are like a few technical technicians i won't get into super deep here but the key differentiator here is that we want to build the plug and play layer for all the polka dots and the perishing asset in the sense that we're going to support all the parity and relation assets including dots like ausd mccullough like usd from tesla engine btc and all the uh and all the coins that uh probably a lot of projects participating here you are going to issue right so um so that's so that's really the uh mental 1.0 right so um just dive dive into a little bit uh deeper into attack so i in a very highly

healthy point of view i will tell you how does the private transfer work all right so think of the public blockchain is a public ledger right the the in a public transfer you basically send a transaction for example alex and bought 50 coins and then like every single validator is actually re-executing the transactions and the mentor right it's very similar except the both ledger state and the transactions is encrypted so in the sense that the transaction are actually not just the plain text but encoded by using the zero number proof here is pi transfer basically this pi transfer is proving that the largest trans ledger

state transformation from the l1 to l2 is valid and then all the validators is running on zk snark and all the validators validating this like uh the validity of this uh zero launch proof right this is a very high level point of view so um so i will give you a little bit like a very very slight overview of zk snark right zk snark is such a wonderful thing that basically can prove you fact about computation without leaking any information so for example um so there is alice and bob and bob want to prove to alex for example i love you and then essentially um and without leaking any information

right basically you you go through this k generation and this is like a proof and validation right so so this is so this is uh again this is very high level we probably don't have much time so for the for the mental 1.0 we have already built a prototype uh you can see there uh the studio and the balance whistles but basically we build a prototype on substrate and uh it's already working and we're still working on the user interface and the front end right so um so for the mentos zkp concrete performance right we require trusted setup but we don't think that's a big issue here and our approval efficiency is about 10 seconds and our verifier efficiency about like a

10 millisecond and i mean thanks for the one for things that uh the pog dot ecosystem build a very efficient watson runtime so that's what mental 1.0 basically plug and the play layer for the foundable token transfer for all proctor ecosystems and then i'm going to talk about a little bit about mental 2.0 so mental 2.0 is going to be a decentralized anonymous exchange right so um let me remind you the one mental 1.0 is basically you take your token and the mint to a private token and do the transfer and reclaim and this meta 2.0 is actually very similar right you can take your favorite token and mint it to a private verb of the token and then you can do exchange exchange to the private token

another kind of private token elementa and then perhaps if you want to claim back or you can use a new private token to do the transfer right so again the the the working cost here is use zk snark and for the exchange mechanisms we are actually using the uh automation market maker right if you are not super familiar with automated market maker and essentially it's a you can think of instead of uh trading against some part trading against some users you're actually trading against the pool right and there is a curve this is called x times y equals to k and essentially every every the pool has two assets

and every point of time this pool maintains this invariant uh known as curve right so let's say if you want to trade some trade something right so basically uh the the the position will move moving on the curve right so i will again i won't dive super into details um but i think why the amm is great for for several reasons right first is like a it actually brought you a little bit of front-running resistance and second and actually most importantly right is there a meat will translate to gas efficiency which means the the amm is extremely simple to construct and in poke dodge and also in like a kind of like a zero node proof term it will actually

transferring to like a zero node proof efficiency in the sense that the circuit will be way more simpler compare you for example you build a like a other book based uh yeah and also it will have like a few like a market efficiency and please uh so there is just kind of like a stanford literature about a detailed analysis right again in a very high level point of view uh the the difference of the dax scheme here is that instead of uh so we're basically adding the am invariant to the zero number proof right so then um i then um i'll also give you a very very preview of the manta 3.0 which is a privacy preserving smart contract platform

so it takes a nap take advantage of amazon right so basically i'm going to first build the products and then become a platform and we want to go in the same way so from the architecture point of view so probably so once we build the mentor like a private payment and deck it will become kind of like a library layer and it will support like a smart contract for example zk row up banking non synthetic acid and ft and so on right so a little bit of i'm probably running out of time yeah actually sumo i was going to sort of jump in um just one one one thing i want to mention is that at the end when we do the breakout sessions actually if people want to have a bit more

conversation about this we can actually do that at that point and just so you know there's actually a question in the chat that maybe after your talk you can answer in that chat as well but do you shall i just say thank you then thank you so much for the talk okay great