While the first generation of quantum computers won’t be ready to compete with modern supercomputers or the best of today’s PCs, they will significantly outperform classical computers when it comes to some tasks. However, one place where they might fall short is in error detection and correction, at least according to new research. Quantum computer scientists are working on ways to develop methods that can detect and correct errors as soon as they happen, but the new research shows that this task might not be possible when dealing with larger quantum computers. Catching Quantum Computer Errors in the Act
What are quantum computers?
Table of Contents
As you probably know, computers operate on binary data. Each 0 or 1 is called a bit, and when your computer sends information over a wire it is encoded as 0s and 1s. However, while a bit can have one of two values, there are more than two possible states for a quantum bit (qubit). In fact, due to something called superpositioning and entanglement, qubits can be in multiple states at once. This makes them incredibly useful for things like database searches and pattern matching.
How will they affect us?
It’s impossible to predict exactly how quantum computers will impact our lives, but there are a few broad impacts we can expect. At their most basic level, quantum computers will be used for calculations that are so complicated and time-consuming that they’re impossible for conventional computers to perform. But when you look at it more closely, you start to see what real promise lies behind these machines.
Where classical computing might be able to model a car travelling across Europe, quantum computing could simulate every atom of that car and every person on board—all at once. That’s massive power.
Where are we now?
Current quantum computers have a very short lifespan and are almost impossible to control. That means they can’t be used to tackle complex problems because any errors or miscalculations made by these early machines simply occur too quickly for us to notice. If a person was looking for an answer to something, it would be like trying to find one needle out of many without even knowing how many there were, let alone where they were. All that could change soon, however; it’s estimated that we’ll see increasingly advanced quantum computers within just 5 years. These won’t be controlled much better than their early predecessors but will still help us catch what’s happening.
What kind of hardware is being used?
The quantum computer uses a process called superposition, wherein qubits can be on or off at any given time, and are capable of processing at insane speeds. If you want to catch errors on a quantum computer, you need to do so before those errors morph into two separate pieces of data. And that means there’s only one way to catch an error—you have to record your quantum system’s state after each calculation and then compare it with what you were looking for (the correct result). To do that, researchers generally use slow but precise measurement devices like electrons or ions. In principle though, any kind of qubit will work because all qubits become classical information (that is, data) once they’re measured.
Where will this lead us?
A common criticism of quantum computers is that they cannot prove their results to be correct. This is usually framed as a problem with trusting quantum computers, but researchers at MIT and UC Santa Barbara have demonstrated a way for a hypothetical future quantum computer to catch its own errors—on any calculation it performs. By using an algorithm called Verify, along with some clever new ways of thinking about computing, a quantum computer would be able to detect errors on any calculation without actually doing it all. It would be impossible for anyone looking at the results of a Verify-powered quantum computer to tell if it were simply returning incorrect answers (as regular computers sometimes do) or reporting uncorrectable errors from before they were even done running.
Will there be a limit to their abilities?
We’ve become accustomed to stories of quantum computers destroying encryption schemes, yet it’s worth remembering that there are many reasons why those attacks might not ever be used. Perhaps a more immediate concern is whether quantum computers will be able to catch their own errors on any calculation. After all, if we can’t even trust them with simple calculations without risking data loss or corruption, then what hope do they have of becoming useful? In two new papers published today, researchers looked at how well quantum computers could detect an error and then fix it. The answer seems to be not very well. With these problems being prevalent (but not insurmountable), perhaps quantum computing won’t be hitting us over our heads anytime soon.
How do you test it?
It’s easier to test an imperfect quantum computer than you might think. By setting up a simple experiment, you can perform multiple calculations and compare them to see how often errors arise. For example, let’s say you want to find prime numbers—numbers that are only divisible by themselves and one. You can run each prime number through a program that will then double check it—if there is an error, it will return a zero instead of another prime number. To test your quantum computer, you could have it run every odd number between one and one million; if it doesn’t find any new primes, your computer passes with flying colors!