Home > News content

200 seconds = 10,000 years overrun, Google achieved "quantum hegemony" papers on the shelf were withdrawn immediately

via:博客园     time:2019/9/22 15:33:55     readed:189

data-ratio=0.5625

Bian Ce Chestnut from Aofei Temple

Quantum number QbitAI

The quantum computer took 10,000 years to complete a calculation in 3 minutes and 20 seconds.

The results, from Google's latest quantum computing research, are published on NASA's website. The paper announced that,

Quantum Technical Adviser to the British GovernmentSteve BrierleySaid:

This is the first time that quantum computers can really outperform classical computers.

It is a great achievement.

In addition, Yang Anze, the presidential candidate of the Democratic Party of the United States, tweeted this message, arousing strong concern, and 5,100 people praised it one day:

Big things, at least that means, there is no password that can't be cracked.

data-ratio=0.8251273344651953

Although, NASA did not have much time to do so.undercarriageThis paper, but because of this, humans are more curious about Google's new results.

A glimpse of the thesis

NASA dropped the paper in a hurry, and we can still get a glimpse of the content of the paper from the web cache.

In its abstract, Google said:

We use a programmable processor with 53 superconducting qubits, which takes up state space of 2Sixteen。 The results of repeated experiments will sample the corresponding probability distribution.

We use classical simulations to verify. Although our processor takes about 200 seconds to sample.

Millions of quantum circuit examples, but an advanced supercomputer will take about 10,000 years to perform equivalent tasks.

Compared with all known classical algorithms, this huge speed increase realizes the quantum hegemony of computing tasks in experiments, and indicates the emergence of long-anticipated computing paradigm.

In the abstract, Google reveals why the quantum computer is powerful. Because the state of an object in quantum mechanics evolves in Hilbert space, only 53 quantum bits are needed to simulate 10.state, and this number is already beyond the computational power of today ’ s supercomputers (typically equivalent to 50 quantum bits).

The main point is that although Google has implemented 72 qubit chips, this is different from 72-bit quantum computers. Google Bristlecone chip uses nine qubits of the same mode to couple, and then extends them in turn, rather than realizing the coupling between pairwise qubits.

The practical application of quantum computer also faces many problems. Because the energy difference between 0 and 1 is too small, it needs to be reduced to near absolute zero in order to prevent being destroyed by heat.

In addition, there is a time limit for the coupling of states between particles, and for a long time, the two particles will no longe

In March this year, Google presented the following conceptual map of quantum computer evolution in a paper:

data-ratio=0.4342948717948718

This graph shows the relationship between the error rate of quantum computation and the number of quantum bits. Google's Quantum Artificial Intelligence Laboratory is expected to study the red curve in the figure. They hope to reduce the error rate by building error-correcting quantum computers, thus pushing the technology into the green available area in the lower right corner.

What is quantum hegemony?

Quantum hegemonyAlso calledQuantum advantageThat is, at some point in the future, powerful quantum computers will be able to accomplish tasks that are almost impossible for classical computers.

For example, in a day to crack the original tens of thousands of years to crack the password, the implementation of general artificial intelligence, fast simulation of molecular models.

The reason for this hypothesis is that the development of quantum computers seems to follow.And classic computers follow

data-ratio=0.66484375

Hartmut Neven, who proposed the internal law

Moore's Law is well known that the transistor density of computer chips doubles every 18 months and the computational power doubles, which is a law of exponential growth. However, in recent years, as the size of transistors approaches the physical limit, this law has slowed down or even failed.

Hartmut Neven, head of Google's Quantum Artificial Intelligence Laboratory, argues that the speed of quantum computers is growing exponentially. A double index is an index plus a layer of index in the following form:

Neven believes that quantum computers have two exponential advantages over classical computers:

Firstly, quantum bits are more efficient than ordinary bits. If a quantum circuit has four quantum bits, a classical circuit with 16 ordinary bits is needed to achieve the equivalent computing power.

Second, quantum chips are also improving rapidly. Google quantum chips are developing at an exponential rate, a rapid improvement due to lower error rates in quantum circuits. Reducing the error rate can help us build larger quantum chips.

Biexponential growth is much faster than exponential function, so Google believes that although quantum computers are far less fast than classical computers, they will one day surpass the latter.

data-ratio=1

This is not just the boasting of Google researchers, but Google Labs is actually pushing at a double-exponential pace: in December last year, a laptop could simulate Google's best quantum computer; by January this year, a powerful desktop would be comparable; and by February, the speed of a classic computer would no longer be comparable to that of a quantum computer.

Why NASA released

You might wonder why Google's paper was posted on NASA. In fact, Google already worked with NASA last year, and set up flag: to achieve so-called quantum hegemony this year, that is, to make quantum computers far more computing power than classical computers.

In July 2018, Google announced a partnership with NASA to compare the results run on quantum computers with classic simulations to achieve the so-called

The name of the quantum chip used by the two sides is calledA total of 72 qubits. Because Bristlecone needs to keep its superconducting circuit near absolute zero, it cannot be removed from Google's lab. NASA researchers can only remotely connect to Bristlecone via Google's cloud API.

data-ratio=0.525

The two sides will work together to study

At the beginning of this year, they coded the software needed to run these simulations on NASA's strongest supercomputer Pleiades, and compared the results of quantum circuit simulation with those of Google's quantum computer hardware 12 months after the contract was signed, that is, in July this year.

Although Google and NASA are very optimistic, the industry also has a flag to reverse.

Researchers from Alibaba's data infrastructure and search technology department have published a paper arguing that quantum chips with lower error rates may be needed to achieve quantum hegemony.

Daniel Lidar, director of the Center for Quantum Information Science and Technology at the University of Southern California, is also sceptical. In response to the MIT Technology Review, he said:

If this paper passes peer review, it means that Google and NASA flags are not falling, and quantum computing will enter a new era.

Make history

For decades, the field of quantum computing has been shrouded in a powerful hypothesis:

Any other type of computer that can do calculations efficiently,Classical ComputerThey can also be accomplished efficiently.

This assumption, from

Today, Google's Quantum Computer3 minutes and 20 secondsCompleted calculations, handed over to Summit, the world's number one supercomputer, probably needTen thousand years。 This breaks the human conjecture.

Google said:

This marks the first operation that can only be performed with a quantum processor.

This is a milestone on the way to full quantum computing.

The computational power of quantum machines will grow at a double exponential rate.

Of course, this unprecedented achievement is not just Google's own carnival.

As an adviser to the British government on quantum technology.Steve BrierleyHe has worked in the field for 20 years and is the founder of Riverlane, a quantum software startup. He stressed that:

For the first time, it has been proved that quantum computers can really outperform classical computers.

This is a great achievement.

Beyond the realm, Yang Anze, the presidential candidate for the Democratic Party of the United States, said this:

Google's quantum hegemony is a big deal. Not to mention anything else, at least that means there's no code that can't be cracked.

He's got 5100 likes on this tweet.

Shortly afterwards, Yang added one more item:

Our encryption technology has to keep up.

This Twitter has won 3,200 more praises.

What's next?

Quantum hegemony has been achieved, but the story is not over.

In the eyes of many people, quantum hegemony is an artificial milestone: on any task, it is possible to prove that quantum computers outperform classical computers.

Any task can be done, and it is not necessarily of practical significance. For example, Google's task for quantum computers is to identify whether a random number generator is really random.

So, after one milestone, there is another milestone to rush to.

data-ratio=0.66640625

Practicability

Robert Suor, head of the IBM's quantum computing strategy, mentioned aQuantum superiority(Quantum Advantage)

Quantum advantage is that in a real application scenario (such as financial services, AI, chemistry), quantum computers have done significantly better work than any classical computer.

As a rival to Google, IBM has been exploring applications for quantum computing, working with JPMorgan Chase and Mercedes-Benz. Recently, they also opened a 53-bit universal quantum computer online.

IBM says that the term quantum hegemony does not apply to them.No needAlso pay no attention to

data-ratio=1

But Fernando Brand, a theoretical physicist at the California Institute of Technology, Google Research Institute

Quantum computers can't do anything interesting until they achieve quantum hegemony.

Now, even if the tasks done by quantum computers are of no practical significance, researchers can still learn from them and develop more useful quantum computers in the future.

In 2018, Boston Consulting Group (BCG) reported that quantum computers can change the rules of the game in many areas:

for examplecryptographyAnd chemistry, the impact on chemistry can widely spread to materials, as well as agriculture and pharmaceuticals and other fields.

Artificial intelligence, not to mention machine learning.

In addition, logistics, manufacturing, finance, energy

The whole world is looking forward to this future.

error-tolerant rate

The next milestone is the creation of a fault-tolerant quantum computer.

data-ratio=0.6654545454545454

Such a computer can correct errors in a calculation in real time, and in principle can achieve error-free quantum computing.

At present, the mainstream method is called

This is far beyond the current quantum computing power.

Therefore, whether quantum computing needs fault-tolerant ability or not is also a question worthy of discussion.

Fernando Brand from Google.

There are many ideas to go, but none of them is definite.

One More Thing

However, before moving to the future, there is also a question of hitting the soul directly about this study:

If it takes 10,000 years to calculate the excess, how can we know that the results of quantum computers are right?

data-ratio=0.6997885835095138

Paper cache address:

Https://filebin.net/k5rr4l0p6ldde7hr/Quantum_Supremacy_Using_a_Programmable_Superconducting_Processor.html

China IT News APP

Download China IT News APP

Please rate this news

The average score will be displayed after you score.

Post comment

Do not see clearly? Click for a new code.

User comments