Google’s mind-boggling claim—posted on the NASA Ames Research Center web site and then quickly taken down–to have achieved “quantum supremacy” continues to stir controversy and skepticism in the nerdier corners of the web.
As we explained here a few days ago, quantum supremacy means a calculation made by a computer designed and built to employ the principles of quantum mechanics to solve problems that even the most powerful conventional supercomputers cannot.
The notion of quantum computing was born in the early 1980s when the renowned physicist Richard Feynman proposed that a quantum computer (which didn’t yet exist) would be an effective tool to solve large-scale problems in physics and chemistry because it is enormously costly to simulate large quantum systems with traditional computers.
But, building a high-fidelity quantum computer capable of quantum algorithms in an exponentially large computational space poses monumental experimental, theoretical, and engineering problems. Quantum computing chips are very unstable and subject to interference from heat and electricity. An illustration of just how big a challenge may be that the random numbers test used by Google had its mathematical theoretical origins at IBM labs in 2004—15 years ago.
Big companies and governments have spent billions of dollars over the past two decades without proving that a quantum computer could solve any problem faster than a traditional computer could solve it. They had failed to do so. Until now.
In a research paper titled “Quantum Supremacy Using a Programmable Superconducting Processor” that was posted and removed (if it hasn’t been taken down, there’s a copy here), a researcher named Eleanor G. Rieffel, who is Senior Research Scientist Lead, Quantum Artificial Intelligence Laboratory (QuAIL) at the NASA Ames Research Center, made exactly that claim about a test done with Google:
Here, we report using a processor with programmable superconducting qubits to create quantum states on 53 qubits, occupying a state space 253 ˘1016. Measurements from repeated experiments sample the corresponding probability distribution, which we verify using classical simulations. While our processor takes about 200 seconds to sample one instance of the quantum circuit 1 million times, a state-of-the-art supercomputer would require approximately 10,000 years to perform the equivalent task.
To put it another way, Google and NASA created a 53-qubit quantum processor (actually 54 but one didn’t work), called Sycamore, that was able to solve a calculation–proving the randomness of numbers produced by a random number generator–in 3 minutes and 20 seconds that would take the world’s fastest traditional supercomputer, IBM’s Summit, around 10,000 years to compute. As Rieffel explained in the now withdrawn paper:
This dramatic speedup relative to all known classical algorithms provides an experimental realization of quantum supremacy on a computational task and heralds the advent of a much-anticipated computing paradigm…Our experiment marks a milestone towards full scale quantum computing: quantum supremacy.
Forget all that, did they really do it?
Both Google and NASA have refused to answer questions about the paper since it disappeared but no one doubts its authenticity. An updated version of the paper is expected to be published in the next couple of months in one of the most respected scientific publications at which time we can expect Google to light the candles and launch the balloons.
My take is that it is a milestone achievement and an important moment in the evolution of quantum computing. At the same time, it is a reminder that we are still years away from stable quantum computers that can perform all the miraculous tasks its proponents foresee.
The best assessment I’ve seen is in a blog post by Scott Aaronson, a well-respected theoretical computer scientist and Professor of Computer Science at the University of Texas at Austin. Aaronson wrote:
It’s not an everythingburger, but it’s certainly at least a somethingburger!
.(tagsToTranslate)Machine intelligence and AI
This is a syndicated post. Read the original post at Source link .