How to verify that quantum chips are computing correctly:
In a step toward practical quantum computing, researchers from MIT, Google, and elsewhere have designed a system that can verify when quantum chips have accurately performed complex computations that classical computers can't.
Quantum chips perform computations using quantum bits, called "qubits," that can represent the two states corresponding to classic binary bits — a 0 or 1 — or a "quantum superposition" of both states simultaneously. The unique superposition state can enable quantum computers to solve problems that are practically impossible for classical computers, potentially spurring breakthroughs in material design, drug discovery, and machine learning, among other applications.
Full-scale quantum computers will require millions of qubits, which isn't yet feasible. In the past few years, researchers have started developing "Noisy Intermediate Scale Quantum" (NISQ) chips, which contain around 50 to 100 qubits. That's just enough to demonstrate "quantum advantage," meaning the NISQ chip can solve certain algorithms that are intractable for classical computers. Verifying that the chips performed operations as expected, however, can be very inefficient. The chip's outputs can look entirely random, so it takes a long time to simulate steps to determine if everything went according to plan.
In a paper published today in Nature Physics, the researchers describe a novel protocol to efficiently verify that an NISQ chip has performed all the right quantum operations. They validated their protocol on a notoriously difficult quantum problem running on custom quantum photonic chip.
"As rapid advances in industry and academia bring us to the cusp of quantum machines that can outperform classical machines, the task of quantum verification becomes time critical," says first author Jacques Carolan, a postdoc in the Department of Electrical Engineering and Computer Science (EECS) and the Research Laboratory of Electronics (RLE). "Our technique provides an important tool for verifying a broad class of quantum systems. Because if I invest billions of dollars to build a quantum chip, it sure better do something interesting."
[...] The researchers' work essentially traces an output quantum state generated by the quantum circuit back to a known input state. Doing so reveals which circuit operations were performed on the input to produce the output. Those operations should always match what researchers programmed. If not, the researchers can use the information to pinpoint where things went wrong on the chip.
At the core of the new protocol, called "Variational Quantum Unsampling," lies a "divide and conquer" approach, Carolan says, that breaks the output quantum state into chunks. "Instead of doing the whole thing in one shot, which takes a very long time, we do this unscrambling layer by layer. This allows us to break the problem up to tackle it in a more efficient way," Carolan says.
(Score: 0) by Anonymous Coward on Monday January 27 2020, @07:44PM (1 child)
Look at them as they are doing stuff of course... duh!
(Score: 0) by Anonymous Coward on Monday January 27 2020, @08:24PM
"The quantum computer can neither confirm or deny that the answer may or may not be correct." ~Schrodinger's Cat
(Score: 2) by FatPhil on Monday January 27 2020, @08:27PM
Great minds discuss ideas; average minds discuss events; small minds discuss people; the smallest discuss themselves
(Score: 2) by EJ on Monday January 27 2020, @08:44PM (9 children)
Remember the Intel floating point bug? Now imaging that in quantum computers. You have all of this data that tells you one thing, but it's completely wrong. Since you have no other way to test the data's accuracy, you just go about your life trusting it.
Hey, Google. What does the inside of a black hole look like? "Let me simulate that for you. It looks like this..."
Ok. Well, what if it's wrong? What if your protein folding simulation ends up causing millions of babies to grow up to be sterile? Oops. Our bad.
(Score: 2) by ikanreed on Monday January 27 2020, @08:57PM (7 children)
Actually, many difficult computing problems are trivial to verify, once solved.
Determining a private key from arbitrary encrypted information and a public key? NP hard. Validating that a public and private key are coprime to each other? O(log(N)) (And that's still sort of overstating the case)
(Score: 2) by EJ on Monday January 27 2020, @09:39PM
Yes, but we aren't talking about those kinds of problems.
We're talking about the types of problems where you don't know the answer. You don't even know how to verify the answer.
For encryption, you're right. You don't know the key, but you can tell if the key can open the lock.
Lots of the simulations they want to try with quantum computers are ones where we don't even have the lock. We don't even know what the lock looks like.
Lots of the protein folding simulations are being verified by known examples, but that doesn't mean that the unknown ones are going to turn out right.
We use quantum computers to create a new drug. We're confident that it's not improperly-folded like Thalidomide because the babies don't come out with severe deformities, but we have no idea what happens twenty or thirty years down the road when the screwup becomes evident.
(Score: 2) by EJ on Monday January 27 2020, @09:43PM (3 children)
To clarify, I'm talking about quantum simulations of the drug's effects on the human body in place of clinical trials.
I imagine a future where the entire drug development cycle all the way through FDA approval is all done inside computer simulations.
That's what I'm talking about when I say that the screwup could be disastrous once it is finally discovered.
(Score: 2) by ikanreed on Monday January 27 2020, @09:48PM (1 child)
Buddy, how do you think we find out about those things now?
(Score: 1, Touché) by Anonymous Coward on Monday January 27 2020, @10:27PM
Testing in third-world countries so no white people are injured.
(Score: 2) by stormwyrm on Tuesday January 28 2020, @07:43AM
Numquam ponenda est pluralitas sine necessitate.
(Score: 2) by Muad'Dave on Tuesday January 28 2020, @01:35PM
Here's an example that's been dogging me since I was introduced to it in college: Van der Waerden numbers [wikipedia.org]. They're insanely difficult to calculate, but are super-easy to validate (like paper-and-pencil easy) once discovered.
(Score: 0) by Anonymous Coward on Tuesday January 28 2020, @01:37PM
Do you mean O(n log(n))
Typically you can't do better than O(n) as you usually have to read the entire problem to solve it.
(Score: 2) by black6host on Tuesday January 28 2020, @01:11AM
Well, the answer 42 is working for me!