October 25th! Happy St. Crispin’s Day! (Not to be confused with CRISPR)
Earlier this week a paper was published in the journal Nature wherein computer scientists from Google report achieving “quantum supremacy” with a quantum computer called “Sycamore.” This means that a computer using quantum physics technology was able to outperform a regular (or “electronic”) computer on a calculation. To understand how quantum computers work and why quantum supremacy is a big deal we first need to understand a bit about how electronic computers work and about quantum mechanics.
How Electronic Computers Work
Electronic computers use electronic switches to store and process information using zeros and ones – so-called binary code. A single representation of a zero or a one is a “bit.” Every word, number, sound and image can be stored as a combination of zeros and ones. Electronic computers use a vast amount of electronic circuits to process algorithms using zeros and ones as the language. A key concept is that every calculation a computer makes is incredibly simple but a vast amount of simple calculations put together and performed really fast can perform very complex processes.
Electronic computers are speed limited by how fast electrons can move through wires and the heat generated by that movement. A key way computers have increased in speed is that they’ve made transistors extremely small and squeezed them very close together. As transistors get smaller they get faster due to less travel time for the electrons and use of fewer electrons which means faster switching. Smaller size also leads to use of less electricity and generation of less heat. However, we’ve pretty much tapped out how small and close together we can get transistors to each other. According to Quartz:
At the present, companies like Intel are mass-producing transistors 14 nanometers across—just 14 times wider than DNA molecules. They’re made of silicon, the second-most abundant material on our planet. Silicon’s atomic size is about 0.2 nanometers. Today’s transistors are about 70 silicon atoms wide, so the possibility of making them even smaller is itself shrinking. We’re getting very close to the limit of how small we can make a transistor.
So, big increases in computer speed likely will need to come from computing efficiency (i.e. to make more efficient algorithms), using photons instead of electrons as the medium for computer circuits or to move to quantum computing.
How Do Quantum Computers Work?
Quantum computers rely on some strange properties of sub-atomic particles. According to quantum mechanics, subatomic particles such as electrons exist in a cloud of probability where they can be thought to exist in multiple locations simultaneously until observed. This ability of tiny particles to exist in multiple states simultaneously is called “superposition.” Related IFODs: The Fuzzy Existence of Electrons and Heisenberg’s Uncertainty Principle.
The concept of superposition is key to how quantum computers work because superposition frees quantum computers from binary constraints. Quantum computers don’t use binary bits. Instead, they use qubits which are quantum particles in superposition. This means that instead of being either a zero or a one, a qubit can be a zero, a one, simultaneously a zero and one or neither a zero nor a one. So, a qubit has four possible states. A thing that most (maybe all) of us cannot comprehend is that these four options occur in a single instance of a qubit due to superposition.
By having four options, which all occur at the same time, rather than two binary possibilities allows such a computer to be exponentially faster. “Because a quantum computer can contain these multiple states simultaneously, it has the potential to be millions of times more powerful than today’s most powerful supercomputers.” Source. What takes a regular computer years might take a quantum computer seconds. What Sycamore did was complete a calculation in 200 seconds that Google estimates would take a powerful regular computer 10,000 years. (Note that IBM disputes this and estimates that the calculation would take a classical computer 2.5 days.)
Similar to how electronic computers link circuits to calculate, quantum computers link qubits. However, as more qubits are linked the harder it is to maintain their fragile states. Sycamore linked a mere 54 qubits to perform its calculation. It is thought that a useful quantum computer will require about one million linked qubit circuits. So, the technology used by Sycamore is just a tiny fraction of what will need to be achieved. But its a start!
What Having Quantum Computers Might Mean
The speed of quantum computers just doesn’t mean that existing software will run millions of times faster. Rather, what’s exciting about quantum computers is that they will be able to perform functions at which electronic computers struggle. Here’s some things that quantum computers are expected to do quite well:
Artificial Intelligence. Electronic computers have given us basic AI, but for AI to jump to the next level of true self-learning the power of quantum computing is likely necessary.
Molecular Modeling. Quantum computers could give rise to “quantum chemistry.” Calculating the electronic structure of a molecule is “a fiendish problem that requires modelling multiple quantum interactions.” Source. Having quantum molecular models will allow scientists to calculate the optimum configurations for chemical interactions which previously has been done partially using trial and error.
Cryptography. A key function a quantum computer should be able to easily complete that electronic computers struggle at is factoring very large numbers. Calculating the factors for numbers with hundreds of digits is considered impossible for even the most powerful electronic computers. This difficulty in factoring very large numbers is the basis for encryption. Here’s a great explanation from the University of Waterloo:
RSA encryption, the method used to encrypt your credit card number when you’re shopping online, relies completely on the factoring problem. The website you want to purchase from gives you a large “public” key (which anyone can access) to encode your credit card information.This key actually is the product of two very large prime numbers, known only to the seller.
The only way anyone could intercept your information is to know those two prime numbers that multiply to create the key. Since factoring is very hard, no eavesdropper will be able to access your credit card number and your bank account is safe.
A fully functional quantum computer could relatively quickly break the encryption. At that point encryption systems will need to change as existing methods will be obsolete.
Modeling Complex Systems. Quantum computers should be able to model complex or chaotic systems such as the weather and financial markets. Interesting article on use of quantum computers in finance here.
Thanks. At least I now have a slightly basic idea of what quantum computing is.
It has been said that “if you think you understand quantum computing, you don’t understand quantum computing.” I can see why.