ไม่สามารถเล่นวิดีโอนี้
ขออภัยในความไม่สะดวก
Zero Knowledge Proof - ZKP
ฝัง
- เผยแพร่เมื่อ 18 ส.ค. 2024
- Protecting your online privacy could become a lot easier with Zero Knowledge Proof. A technique that allows you to prove things without disclosing what it is you're proving.
Need to have a certain age before you can signup for a service? You can deliver a proof that you meet the criteria without sharing your actual age. Or you could prove to your bank that your salary is sufficient to get a loan, without revealing your real salary.
📚 Sources
savjee.be/vide...
🌍 Social
Twitter: / savjee
Facebook: / savjee
Blog: savjee.be
💌 Newsletter: newsletter.savj... (no more than once a month)
❤️ Become a Simply Explained member: / @simplyexplained
Hey ! We need a math example to understand the ZKP.
Yes!
You should make a part 2 because I am still confused. How does the non-interactive zk proof work?
I totally agree with you that there was a lack of explaination regarding the non-interactive zk-proof. Without knowing if it is correct, I see the non-interactive zk-proof as something that is easily tested for truth rather than using a process to convince. For instance in the case of the color blind guy with the green and red ball, in that case it was interactive as you had to prove that you could see a difference. Imagine if the green ball was made of some magnetic material and the red was non-magnetic, you could easily prove the balls were different without the testing back and forth. This is a fact you easily can share with many other in the same situation.
I come to this conclusion as he says that blockchain have to use non-interactive proof to work and in the case with blockchains you don't have to show every nounce you where trying to show everyone you have a solution, only the last value that was a solution as it is a very easy computation for everyone to to test if it is true. I do see this is 100% sure though, and he keeps saying that with zk-proof you can never be 100% sure. I can't fit that part into my reasoning for non-interactive zk-proof...
Would this be a plausible explaination?
Same
零知識證明 → 禪定 悟道 修證非果位 筆記叔叔日誌 . ..
Same, I'm still confused, probably because it didn't dive into the math and talked too much at the high level.
The answer is digital signatures. When you sign a message or chunk of data, the resulting signature can be used to determine who signed the message, but only if you know what the original message is, that got signed
imagine Fermat used zero knowledge proof to prove that he knew the proof to his theorem (Fermat's last theorem)
I still don't get it :D How does the ball example transfer to atomic weapons? How does the non-interactive proof work?
They use "a non-electronic fast neutron differential radiography technique using superheated emulsion detectors" to "confirm that two objects are identical without revealing their geometry or composition".
And here I thought sci-fi's were spouting BS. It's real!
Great Explanation. Waiting for part 2, maybe more in-depth on non-interactive zk Proof
I finally get the two balls example. It is the skeptical color blind friend who is holding the balls, and he is secretly switching or not switching. Only the skeptical color blind friend knows whether he switched behind his back. So in order for you to prove to him they are different colors you would have to point out to him exactly each time they switch color, corresponding exactly to each of his secret switches.
That color blind analogy is amazing!
In a zero-knowledge proof of the first case, you should use more iterations, demonstrating that the limit of the correct answer approaches infinity. This will help convince the learner.
Awesome, but can you please explain about non interactive zero knowledge proof in another video?
Thanks~
I find that I can understand this video more easily than the one provided by that computer scientist who explained this concept to 5 levels of people.
I need to review this again. I'm lost!
Great explantation of interactive proof. I really like your color scheme. An explanation on non- interactive proof would be nice.
Can you provide clear example of non interactive knowledge proof ? Specifically the ING bank application
aren't the buzzwords enough? just buy the crypto currency and make me rich.
I can think of one zero-knowledge proof that is non-interactive that involves proving that one knows the prime factors of a number without revealing said prime factors, basically it's RSA signing with a twist. It goes like this:
Say you have the number N = 128423, and you want to prove that you know the prime factors without revealing any information about the primes themselves. Here's what you do
1. Since you know the factors are 769 * 167 = 128423, you can calculate the Euler totient to get (769-1) * (167-1) = 127488, we'll call this value _phi._
2. Now pick a public number _e_ which is coprime to phi, the easiest is a number like 65537 which is prime but greater that the square-root of phi.
3. Now calculate a secret number _d_ by taking the modlular inverse of _e_ modulo phi, this gives us 1/65537 = 54785 (mod 127488).
4. If we take any number _m_ and raise it to the power _d_ then do modulo N to get a new numer _s,_ doing the same with _s_ but with _e_ instead of _d_ we arrive back at _m._
m ^ d (mod N) = s
s ^ e (mod N) = m
For any m < N
We can calculate _s_ only by knowing _d,_ and knowing _d_ (combined with _e,_ which is public) is equivalent to knowing the factors of N.
Finally, let's try a practical example to demonstrate:
Here is a proof I know the prime factors of the number 5609654452760709630006622756892418841890735941208888880805793565319940075849799980043767420213243893:
N = 5609654452760709630006622756892418841890735941208888880805793565319940075849799980043767420213243893
e = 65537
s = 4691398041935726800648528142580678766790089967215103950511154552099580646783686977287224608076011797
The signature _s_ is of the number you get by coverting "Hello, World!" into ASCII-encoded binary taking the result as a number in base 2. That is, _m_ = encode("Hello, World!"), _s_ = m ^ d (mod N). You can check this proof by taking _m_ = s ^ e (mod N), then decoding m.
Unless I was *extremely* good at guessing, calculating _s_ is impossible without first knowing the prime factors of N.
@@deidara_8598 that's basically how you have the private keys to a wallet which uses a different algorithm that uses the public key to decrypt while the private key encrypts but idk how you could make a block chain where the account balances are private
Much better than the computerphile vid
it would be good if you mention real life examples of solving problems by mentioning them in another videos.
one thing is clear you really explain in simple way, simplicity is mother of complexity.
So your channel is like a boss of other channels, who do explain things in complex manner.
I really liked the example using color blindness, it was a neat solution. An alternative to repeating the shuffle and question process to get higher probability of correctness would be to cut a piece out of one ball, then cut up the other ball into many pieces, then mix the pieces in such a way that the person doing the mixing knew where the one piece was, while the person asserting that the balls are a different color does not. This is more destructive, and more complex, but possibly faster if you needed high probabilities of correctness quickly.
As many other comments have said, I’d be interested in further videos on this topic.
amazing staff right there.
The probability is now low enough to convince the friend? This is confusing
Dominik Sosnowski ,
If someone kept knowing exactly what you did to the balls, how long would it take you to believe that they are different?
You see them both as the same but he always knew what you did behind your back...
What are the chances he guessed it right 5 times? Still not sure? Go for 10... even 20... do you still think he guessed it 20 times? What are the chances he guessed correctly 20 times?
There is almost no way someone can guess something 20 times or even 10 times in a row.. the probability of him doing so is extremely small.
That's the meaning there
2:50
The probability of guessing right *10 times in a row* is low. But the probability of each guess is still 50% (unless you swapped to having 2 identical balls, using sleight of hand).
The second method is not mathematically convincing. If you prove that n is in the interval [a, b], you can use dichotomy as a tool to approximate the number to the desired precision. However, as the video points out, you can't be certain that it is the exact number, only that it falls within the range.
Happy new year to you and all your family Xavier :)
All the best for 2019! Keep it up the great job
I Like your video explanation
Zero Knowledge Proof = logan paul?...
great video
The problem with ZKP is that it over states it’s utility. There’s only limited use cases in the real world, without incurring more complexity in system design. Intellectually interesting. More hype than practical.
3:01 small correction. the probably is not going down each time, it is 1/2 every time.
The probability of guessing correct every time is goin down. I mean that is clearly what you mean but the way it is expressed is not ideal
This video could be better explained.
> You have normal eyesight and claim that the two balls are differently colored red and green.
> Your color-blind friend sees both balls as green (a common color-blind problem is to see a red rose as a green one) and suspects that you are trying to trick him.
> Your color-blind friend knows when he switches the balls around and whether you are right in saying that he switched or didn't.
> Your many multiple correct guesses convince your friend that the balls do indeed have different colors.
> There is no need to specify which color is which to your color blind friend to prove (based on a high statistical probability) that the balls have different colors.
> This interactive proof requires many efforts but blockchain requires proof with many nodes and therefore needs a non-interactive proof (not explained here).
Personal data, privacy, alternative way to share data so that we do not have to share more than we wanted
Proving sth is true without revealing it
This the name zero knowledge
Zkp is not error-proof - correct in the probabilistic sense
Categories: interactive and non-interactive- pros and cons - and why blockchain uses zksnarcks
Zkp allows the adoption of blockchain technologies in banks because of the otherwise concerns over privacy . Ref JS - getting more attention due to the ways of application in certain blockchains.
Wow, given what i heard at 1 min in i realize why its called zero knowledge proof. Let's see what else..
I understand zkp has 2 forms, but what algorithm or encryption technique does it use to send and verify the private data??
love having quite some Ethereum therefore.
Can you make an episode about ZK roll-ups vs Polkadot's solution?
Great video. Thanks a lot!
Great
what if two balls acutally have same color but different smell, just by get it right 1000 time doesn't prove to your friend that they have different color, maybe they just have different smell.
In that case, the person would have to be smell-blind in addition to color blind. The point of this zk example is to start with an actual difference between the 2 balls (e.g. color, or smell) but which the person holding them cannot detect. And then you convince them -- to any degree of certainty they wish -- that the 2 balls are different in that aspect, without revealing to them what that 'knowledge' is; thus a 'zk' proof.
Could you explain Monero
ING? Wasn't that developed in Bitcoin, as Bullet Proofs range proves first?
I don't understand.
J ai fait la meme avec un dji ronin et un recepeur de drone et une telecommande de drone
Thanks to Zcash !
I do not understand how bank can verify my salary. I can lie about it, right? It somehow resembles me of the Oracle limitation in blockchain: what's outside blockchain can not be trusted (not at first, at least).
It isnt just getting more attention cos of block chain. It is widely used in bank/credit card transactions because it is optimised to run on low power devices such as the chip on a bank card. So this video isnt entirely accurate.
Nice video, what video tool do you use?
This isn't a proof; you just reassured him. In the example, there are other ways to observe his hands, such as using a camera or having another person watch. For instance, one could judge his hand movements by observing the balls of the same color.
I liked the zero knowledge proof explanation. Ask the currency Zcash and BitcoinZ already do this?
The start of this video sounded like a sponsored segment for a vpn.
Wonder if this is reminiscent of the panther protocol thats coming put on xrpl?
nice, but how does the non-interactive really works?
can you please give us the link of "The Knowledge Complexity of Interactive Proof-System"?
Well explained, Savjee 🙏🏼
Can you do a video on Mimble Wimble, please? I have a hard time with it
Awesome
u were awesome .
But then why is ZCash not performing well? 😆
This is panther protocol?
Damn what software did you use for this video? It seems easy but conveys the message clearly
So cool to see that you are a metalhead also, Hail bro!, what genres and bands? you seem as polite as most of my thrash and death metal friends, as heavier the genre the nicer the people haha, I could be wrong...
I like all sorts of metal as long as there is not too much screaming. Hah, thanks for the compliment! I go to a metal festival every year. Indeed, the more bad ass people look, the nicer they are!
Nice content
Hi your mind blesy and good habit
I understand the concept now after watching the video twice but the part where you introduce the analogy about the colored balls always gets me confused.
In colorblind balls example, what if I always lie to friend and always say that ball is green for example? He may switch hands however he wants but he will reach conclusion that both balls are same.
You don't say the color, you say if he switched or not. Your friend is colorblind but he's not dumb. He knows if he switched the balls or not.
You did not understand the point.
If the both balls are the same you cannot gues if the blind frend swich hands or not.
Great video, Xavier! I can't wait to see your IOT devices.
can you give as the steps of the algorithm
🔥
is there any application that zkp can do but digital signature cannot do?
Can u make a video on FlyClient and Blockquick sir!
A possible use on patents?
k, tnx, bye! -slayermusiq1
Not everything is correct on this video... Ergo content quality is quite bad... It's a shame because graphics and exposition is quite well designed.
fucking simplified! :) thx dudes! loved it!
Confused.. Not a good explanation!!
start a frontend framework series please man. Angular would be nice :)
Still confused, the example of blinding colors doesn't help much though :D
What if the colorblind mathematician digitally signed a statement verifying the distinguisibility of the balls ?
A great video. but couldn't understand the concept
Wow this is a pretty bad example
SUB EARNED
At 2:45 if the balls would be the same color... This i do not understand. How does probability of 1/2 changes to 1/4 o.0
Dominik Sosnowski in the example you’re trying to prove to your friend that the balls are a different color. So asks you a question that you wouldn’t be able to reliably answer if the balls were the same color - you would just have to guess. But because they are a different color, you don’t have to rely on guessing. Think of it like this - if someone asked you to guess a number between 1 and 10, you have a 10% chance of being right. But if you knew for a fact that the number was 7, then it wouldn’t even require guessing.
The probability goes from 1/2 to 1/4 because when you’re determining the probability of getting the right answer for repeated events, you multiply the probability of the individual events together. In this case, the probability for the first and second event are both 1/2 (50% chance of getting the right answer by guessing), so the chance of getting it right both times is 1/2 * 1/2 = 1/4
This probability gets successively lower by repeating the event more times.
@@noahmccann4438 aha, thank you for lengthy explanation. So we talking about a lower probability of me lying about the color cause for the color blind friend it doesn't matter what's the color but he cares about the biggest probability of me telling the correct answer?
Dominik Sosnowski precisely. In the example, your friend is already skeptical of your answers. But the probability of you correctly guessing the right answer goes down with each repetition of shuffling the balls around and asking for your guess. The only way that you could be right every time after a large number of repetitions is that you aren’t guessing - you’re using some additional information to inform your answer. In this case, that additional information is knowledge about color.
Now, something the video doesn’t touch on is the fact that you could be using knowledge other than color - for instance maybe you’re able to see behind his back by using a reflection. Or maybe you know that your friend will do a certain shuffle order - thus not being completely random and giving you a slight advantage above and beyond the 50/50 odds one would expect. The example assumes randomness and what kind of knowledge you have access to.
I hate the colorblind example. It is utterly worthless.
But you haven't actually explained anything
Kurwa, nie kumam
Waste of time, you didnt explain non-interactive zk proof