Google Attempts to Build Quantum Computer

 Pages 1 2 NEXT
 

Google Attempts to Build Quantum Computer

D-Wave One

If successful, Google's goal of creating a quantum computer could change the field of computing drastically.

On top of Project Wing and attempting to create an encrypted email service with Yahoo!, Google has placed another endeavor directly on their plate. The search engine giant has begun efforts to design and build hardware for a quantum computer.

Previously, Google began work back in 2009 with "quantum computing company" D-Wave Systems, whose claims of having made the first quantum computer have been proven illegitimate. Now, John Martinis, professor of physics at the University of California, Santa Barbara, has joined the company to set up a new quantum hardware lab near the university. This lab will function as a place to attempt to create his own versions of the chips found in D-Wave's computers.

Quantum computers operate using quantum physics to solve problems- hence their appropriate name. Their rate of problem solving is at a level that a normal computer would accomplish the same amount of work if given millions of years to do so. Martinis has spent over ten years working on a more substantial approach to quantum computing and, based on his findings, built some of the largest, most accurate systems of qubits (the basic building blocks that encode information in a quantum computer).

Qubits working together can use advantages of quantum mechanics to quickly discard incorrect paths to a solution and identify the correct one. However, qubits are tricky to operate due to the delicacy of quantum states. "We would like to rethink [D-Wave's] design and make the qubits in a different way," said Martinis. "We think there's an opportunity in the way we build our qubits to improve the machine."

Google still plans on collaborating with D-Wave. The leader of Google's quantum research stated that the two companies will continue to work together, and that Google's D-Wave computer will be upgraded with a new 1,000 qubit processor when it becomes available.

Source: MIT Technology Review

Like what you read? Check out Dell's new 5K (yes, 5K) 27 inch monitor.

image

Permalink

I'm really amazed at what Google is spending their income on, namely things that can greatly benefit humanity in the long run. Hats off.

And then Google will proceed to have it build the most accurate identities of everyone on the internet so they can be marketed to.

Unless I'm mistaken, no one has yet to solve the no cloning rule which basically makes quantum computing impossible. To elaborate:

In a computer, you have a binary or base 2 counting system represented in bits. These bit are 1 or 0, on or off. In a quantum computer, bits are replaced with qubits; photons with measured polarities. With a qubit, the photon can be 1 or 0, in this case represented by the photon being in either a vertical or horizontal spin and can also be in a superposition of both 1 and 0 so long as you haven't yet measured the spin it is currently in. Now what the no cloning rule states is that while you can have a fundamental particle in a horizontal spin, vertical spin, or superposition of both, once you've measured it and locked it into a state, you can never put it back into a superposition of both. In other words, once you've used a qubit, it is effectively destroyed. When a computer makes a calculation, it uses a system of ones and zeroes, concepts that by their nature are identical. With qubits, the superposition is unique and once lost, is lost to all future calculations.

Here is an example: In a game of baseball, when the batter comes up to the plate, he is in a superposition of hitting and not hitting the ball being pitched to him. Once the pitch occurs and he swings, he is locked into that position; he has either hit it or he has missed it. Once that calculation is made, it can never be unmade. You can recreate the situation to near perfection, but it will never be THAT pitch, THAT swing, THAT one moment in time and space that is now gone forever.

Shdwrnr:
Unless I'm mistaken, no one has yet to solve the no cloning rule which basically makes quantum computing impossible. To elaborate:

In a computer, you have a binary or base 2 counting system represented in bits. These bit are 1 or 0, on or off. In a quantum computer, bits are replaced with qubits; photons with measured polarities. With a qubit, the photon can be 1 or 0, in this case represented by the photon being in either a vertical or horizontal spin and can also be in a superposition of both 1 and 0 so long as you haven't yet measured the spin it is currently in. Now what the no cloning rule states is that while you can have a fundamental particle in a horizontal spin, vertical spin, or superposition of both, once you've measured it and locked it into a state, you can never put it back into a superposition of both. In other words, once you've used a qubit, it is effectively destroyed. When a computer makes a calculation, it uses a system of ones and zeroes, concepts that by their nature are identical. With qubits, the superposition is unique and once lost, is lost to all future calculations.

Here is an example: In a game of baseball, when the batter comes up to the plate, he is in a superposition of hitting and not hitting the ball being pitched to him. Once the pitch occurs and he swings, he is locked into that position; he has either hit it or he has missed it. Once that calculation is made, it can never be unmade. You can recreate the situation to near perfection, but it will never be THAT pitch, THAT swing, THAT one moment in time and space that is now gone forever.

So like, why even use "qubits" then? What advantage do they have over regular bits? If they can only be in the 3 states of on/off/undetermined. Sorry if its a dumb question, your post was readable enough for me to get interested. Thanks

Rosiv:

Shdwrnr:
Unless I'm mistaken, no one has yet to solve the no cloning rule which basically makes quantum computing impossible. To elaborate:

In a computer, you have a binary or base 2 counting system represented in bits. These bit are 1 or 0, on or off. In a quantum computer, bits are replaced with qubits; photons with measured polarities. With a qubit, the photon can be 1 or 0, in this case represented by the photon being in either a vertical or horizontal spin and can also be in a superposition of both 1 and 0 so long as you haven't yet measured the spin it is currently in. Now what the no cloning rule states is that while you can have a fundamental particle in a horizontal spin, vertical spin, or superposition of both, once you've measured it and locked it into a state, you can never put it back into a superposition of both. In other words, once you've used a qubit, it is effectively destroyed. When a computer makes a calculation, it uses a system of ones and zeroes, concepts that by their nature are identical. With qubits, the superposition is unique and once lost, is lost to all future calculations.

Here is an example: In a game of baseball, when the batter comes up to the plate, he is in a superposition of hitting and not hitting the ball being pitched to him. Once the pitch occurs and he swings, he is locked into that position; he has either hit it or he has missed it. Once that calculation is made, it can never be unmade. You can recreate the situation to near perfection, but it will never be THAT pitch, THAT swing, THAT one moment in time and space that is now gone forever.

So like, why even use "qubits" then? What advantage do they have over regular bits? If they can only be in the 3 states of on/off/undetermined. Sorry if its a dumb question, your post was readable enough for me to get interested. Thanks

If they find a way to make it work, we would have computers that are millions of times faster than what we have today.

Google, as far as I know, have always tried to improve upon or change technology for the better. They have all this money which they COULD be sitting on to make the next consumer electronic product. Though they do that too, I always see them fiddling with something cool to bring some of the crazy stuff that we've all imagined to to reality. Whether they succeed in changing things, at least they're trying.

Meanwhile, Apple releases a slimmer phone and are lauded as innovative geniuses.

Rosiv:

So like, why even use "qubits" then? What advantage do they have over regular bits? If they can only be in the 3 states of on/off/undetermined. Sorry if its a dumb question, your post was readable enough for me to get interested. Thanks

The reason is because the data is in multiple superimposed states, you can actually use other qbits to work on each of those states simultaneously.
Suppose you're searching for something, you need to check the calculation of A B or C
A normal computer just goes A B and then C in sequence, a quantum computer does all three simultaneously.
When you actually read the qbits, the it will show you the results of A or B or C
By using the correct algorithm, you can fix it so that the correct answer appears with a high probability.
You then use a normal computer to check that its the right one.

This might sound needlessly complicated, but as far as we know there are alot of problems that are difficult to solve, but easy to check if you have a right answer (This is called NP-Hard).
One of these is Integer factorization.
Why would you want to factor an integer?
Well, almost all the encryption you used online today uses RSA encryption, it take advantage of the difficulty of integer factorization to secure your logins,etc.
And we already have an algorithm called Shor's algorithm that makes integer factorization easy, we just need a quantum computer to run it on.

So if Google build a quantum computer, a huge chunk of online security goes out the window.

Yan007:
]If they find a way to make it work, we would have computers that are millions of times faster than what we have today.

and then you've got the singularity

as much as I hate to say it I find myself becoming more of a luddite in the face of the uncertainty of our future

Vault101:

Yan007:
]If they find a way to make it work, we would have computers that are millions of times faster than what we have today.

and then you've got the singularity

as much as I hate to say it I find myself becoming more of a luddite in the face of the uncertainty of our future

Simply realizing that some lines shouldn't be crossed doesn't make you a Luddite.

Yan007:
If they find a way to make it work, we would have computers that are millions of times faster than what we have today.

This isn't exactly accurate; A qubit is more efficient in larger calculations, but less efficient with smaller calculations. Quantum computers will never replace standard computers for the consumer. Modeling weather or running the algorithms that run a countries economy would be much easier on a quantum computer though.

Vault101:

and then you've got the singularity

as much as I hate to say it I find myself becoming more of a luddite in the face of the uncertainty of our future

The singularity is reached when we create a system that can iterate on itself; build a better version of itself without our input. The fastest computer in the universe won't bring the singularity to us with the proper software. That software, while getting closer all the time, is still a ways out.

TheSYLOH:
So if Google build a quantum computer, a huge chunk of online security goes out the window.

But also introduces some really novel new security possibilities, correct? I remember reading that a quantum encryption system (or maybe certain ones) could be "unhackable," since looking at the data would change the key used to encrypt it. It was based on the uncertainty principle in some manner, I think, but physics isn't my bag, so I could be wrong.

https://www.youtube.com/watch?v=IcrBqCFLHIY&list=PLkahZjV5wKe_dajngssVLffaCh2gbq55_

Veritasium did a wonderful series that went through how quantum computers work and they they are effective at doing.

seeing that the article mentions 1000 qubits. According to one of those videos, when determining what information those qubits can hold, the math works out to N qubits is equivelent to 2^N conventional bits. It only has the capability to have that amount of information within it while all of the qubits are in their super position. Once the calculation is done, a standard 1/0 set of numbers, which then need to be calculated through a regular computer.

however, within that superpositional state, you end up with a binary number with 1000 digits.

In a quantum computer working properly, during the calculation, 1000 binary digits becomes 1.07 x 10^301 digits worth of information.

at the end of the day, after the calculation is complete it comes back down to 1000 digits. But during the calculation... there really is nothing that can't be simulated with that much computing power.

the relevant video is number 3 "How a quantum computer works", but the first 2 are good as well, for context

Lets be honest here, Google is really just run by a bunch of stoners who like to work with words that are fun to say like Google or quibits.
They probably have a whole think tank somewhere devoted to making the most Bulbous Bouffant as well.

I'm not really sure I want the company behind such privacy destroying abominations as 'Buzz' to obtain computers powerful enough to render most / all modern encryption obsolete.

That being said, I'm going to project that quantum computing will take a similar path of progression to classical computing. The first quantum computers will be room-sized behemoths with (relatively) little computing power due to the size of the cooling systems required to make them run properly (as current quantum bits require near absolute-zero temperatures to function without collapsing). Eventually these systems will be miniaturized and the stability of the quantum computers will be increased to the point where smaller machines are viable.

But even then, there remains a crap-ton of software work to be done as well. Ignoring the possible challenges of designing software for an environment with four states instead of two; you don't want some asshole company or government coming along to every classical computer system currently in the world and going "lol security broken in .2 seconds, we have all your nude pics (and everything else) now".

OH MY GOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOD!

Does millions of times faster = sentience?!

Skynet!

It's happening!

Duh-da-duh-dada

duh-da-duh-dada

dew dew dew.... dew, dew, dew...

michael87cn:
OH MY GOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOD!

Does millions of times faster = sentience?!

Skynet!

It's happening!

Duh-da-duh-dada

duh-da-duh-dada

dew dew dew.... dew, dew, dew...

you might possibly get sentience inside the calculation.

but once the calculation's done, the result it farts out would not be complex enough to have sentience.

Quantum Computers are only useful for dealing with problems which it's said regular computers would take a ridiculously long time to do. Like when we try large scale simulations, and the math works out that our fastest conventional computers thousands or millions of years of constant calculating to computate, a quantum computer could run that calculation in minutes. Once the calculation's complete though, the answer that it returns will basically not be any more complex then what our conventional computers could achieve in that larger timeframe, it just exponantially shortens the number of calculations that need to be completed in order to reach the answer.

Yeah maybe now we can have worthwhile games and gaming computers!.... inb4 they cost your soul and three sacrifices to the Google god. (worth it)

They call it a quantum computer, but that implies some rather untrue stuff. It's more like a quantum calculator, with some oddly specific buttons and an enter button that you can press once.

That this news has less interest in it than a thread about 4-fucking-chan, makes me weep for this website.

Svarr:
Yeah maybe now we can have worthwhile games and gaming computers!.... inb4 they cost your soul and three sacrifices to the Google god. (worth it)

I'll take two.

Alcom1:
They call it a quantum computer, but that implies some rather untrue stuff. It's more like a quantum calculator, with some oddly specific buttons and an enter button that you can press once.

The ENIAC was a calculator too. I don't. Think anyone would argue that it wasn't a computer too.

TheSYLOH:
So if Google build a quantum computer, a huge chunk of online security goes out the window.

There are cryptosystems that do not rely on integer factoring or the discrete logarithm problem, which are the two most common trapdoor one-way functions in use these days and which are vulnerable to Shor's algorithm; check out lattice-based cryptography for one of them if you are interested.

Alcom1:
They call it a quantum computer, but that implies some rather untrue stuff. It's more like a quantum calculator, with some oddly specific buttons and an enter button that you can press once.

A quantum computer is a computer, because it can emulate a universal Turing machine, because any conventional circuit can be emulated by Hadamard gates. (With an arbitrarily small probability of getting the wrong answer if there is an answer, natch.)

From what I've read, the D-wave style chip isn't a "true" quantum computer. Not being a physicist, the details of the technology go over my head, but I definitely remember computer scientists and physicists insisting that, while what D-wave achieved was valuable and interesting, it wasn't a proper quantum computer.

If Google just build on the system D-wave invented, then they aren't creating a real quantum computer either.

Because if there's one thing Google needs, it's more power amirite?

Shdwrnr:
Snip

The no-cloning rule is a hindrance in that it means we can't verify computation the way that people traditionally used to (i.e. run it a few dozen times and see what happens). This was worked around by quantum error correction.

I kind of hope that quantum computing is possible, given that my University claims to have one that you can run code on.

The jury's still up on whether D-Wave's better at what it does (that is, annealing) than a traditional computer. It hasn't been "proven" either way so check your facts.

Quantum computation allows for solving SOME problems with much better time complexity. Take into account quantum decoherence also.

Piece author, pelase check your facts.

sorsa:
I'm really amazed at what Google is spending their income on, namely things that can greatly benefit humanity in the long run. Hats off.

And eventually they will create true artificial intelligence and use it for marketing :D

I just want to sound silly and say that "Martinis has spent over ten years..." could well be substituted by "Martinis were spent over ten years..." just to better show us how much work it has been. :)

"Previously, Google began work back in 2009 with "quantum computing company" D-Wave Systems, whose claims of having made the first quantum computer have been proven illegitimate."

This is false, the statement from the linked article reads

"But independent tests published earlier this year found no evidence that D-Wave's computer uses quantum physics to solve problems more efficiently than a conventional machine."

This does not say the D-wave's computer isn't a quantum computer, it says that the quantum computer is no faster than a normal computer which is true of everything if you compare a first generation of a new tech with late gen of an old tech.

While D-wave is now calling it's machine a quantum optimizer rather than a quantum computer you have to remember that by todays standards the first computers are less capable than a scientific calculator of today as well.

So Google is building Skynet? They also own a drone company and AI company. Future doesnt look good. ;-)

Shdwrnr:
The singularity is reached when we create a system that can iterate on itself; build a better version of itself without our input. The fastest computer in the universe won't bring the singularity to us with the proper software. That software, while getting closer all the time, is still a ways out.

I feel that's a rather narrow conception of the singularity.

If you were looking at a black hole from the outside, you would see literally a black spot in space but that's not the singularity. What you're seeing is the event horizon, the point beyond which no light can escape so no light makes it back to your eyes to tell you what the hole actually looks like. In the technological singularity metaphor, the event horizon is the point at which human beings begin augmenting intelligence to superhuman levels. That might not be a computer at all, it might be implants, or genetic engineering, or even drugs. That's not the singularity itself, but it's the point at which future technological development becomes completely unpredictable because it will be governed by intelligence greater than that of modern human beings. It's not the end of development, it's merely the point at which we can no longer guess what happens next.

But if the singularity is not the event horizon, then you have to expand the definition of the singularity to include all the gravitational effects around it. We're already enhancing the way we use human intelligence, and even if we're still reliant on human intelligence to build and operate the devices we create, it still enables things which people in history might have found impossible to imagine or predict.

Zombie_Fish:

Shdwrnr:
Snip

The no-cloning rule is a hindrance in that it means we can't verify computation the way that people traditionally used to (i.e. run it a few dozen times and see what happens). This was worked around by quantum error correction.

I kind of hope that quantum computing is possible, given that my University claims to have one that you can run code on.

Thanks for that; I suppose entangling the qubits would act as a kind of correction so long as the noise doesn't propagate through the entangled pairs. I've also read that a group has been able to use the principles of quantum tunneling to determine the spin of a particle without actually measuring it thus leaving it in its superposition but that it's still not 100% accurate. I'm guessing that our early quantum computers will not necessarily be 100% accurate, just accurate enough to do what we need them to do.

michael87cn:
OH MY GOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOD!

Does millions of times faster = sentience?!

Skynet!

It's happening!

Duh-da-duh-dada

duh-da-duh-dada

dew dew dew.... dew, dew, dew...

No, it doesn't sorry. Sentience is more about how we process information.

So first, Google is developing aircrafts, and now they aim to develop quantum computers. Whelp. I watched Jörmungand, Google. I know what you're planning... >_> Except they're not launching satalites, but that will come. Soon, Google will control the flow of information!

Serious reply: Pretty amazing, fairly scary implications for the future not withstanding.

Simply realizing that some lines shouldn't be crossed doesn't make you a Luddite

Indeed, it just makes you wrong.
Just because something challenges us existentially doesn't mean we shouldn't do it. It certainly doesn't mean we wont.

 Pages 1 2 NEXT

Reply to Thread

Posting on this forum is disabled.