A Glimpse into Quantum Computing
The term “Quantum Computing” is surfacing popular science and media outlets more and more. So what are Quantum Computers? What does quantum mean? How are companies tackling quantum computing research? This post will introduce quantum computing, quantum mechanical phenomena and the status of the quantum computing industry today.
Topics covered include:
- Moore’s law
- Quantum computing today
To start, let us briefly review hardware fundamentals of computers. Computer chips are composed of basic modules, which are made up of logic gates. Using Boolean algebra, logic gates allow computers to make simple decisions by manipulating transistors.
A transistor is the smallest component of a computer that can processes data. It can be in two different states of 0 or 1, meaning it either blocks or allows information from passing through.
In 1964, Gordon Moore, Intel co-founder, famously made an observation and prediction about transistors. He said, “The number of transistors incorporated in a chip will approximately double every 24 months.” Moore’s law means that every 2 years the size of a transistor is cut in half. By 2020-2025, estimates suggest that transistors will generate so much heat from their microscopic size that the current silicon technology may collapse. This brings us into creating transistors at the quantum level.
What does Quantum mean?
Quantum refers to a general quantity or size. In physics, quantum is the measure of the smallest amount of energy that an object can possess. Quantum computers process data at the smallest measurable levels, such as with electrons.
The most famous example of Quantum Superposition is Schrödinger’s cat. In this hypothetical experiment, a cat is placed in a box along with a bottle of radioactive poison that will randomly decay and instantly kill the cat. Let us assume the kitty was already at the end of a full happy life.
In a classical macroscopic environment, when we look at the closed box we may not know if the cat is dead or alive, but it still is either dead or alive. But Schrödinger’s thought experiment assumes a state of quantum superposition, so what is both wacky and amazing is that with the box closed the cat is both dead and alive at the same time.
Interestingly, once we open the box, the cat loses it’s quantum state and we are in a classical state again, where again know if it is either dead or alive.
Schrödinger’s cat illustrates that quantum superposition allows for processing many classical states in parallel (i.e. being dead/alive at once, or being a 0 or 1 bit). Once a measurement of a particle is made, the quantum superposition is lost and the particle collapses into one known state.
So if there are N cats in a box, there exists 2^N possible states of being dead/alive. This is different from probability, which would also have 2^N possible states, but the cats would actually only be in N states of dead/alive. With quantum mechanics, there can be 2^N states of cat life at once.
Qubits (Quantum Bits)
Quantum superposition illuminates the significance of a qubit.
A qubit is a unit of quantum information.
In a classical system, a bit is in a state of 0 or 1, but in a quantum-mechanical system, a qubit is in a superposition of 0 and 1 at the same time.
Qubits allow for exponentially larger computation space. In mathematical terms, a qubit can be described as a point on the surface of a Bloch Sphere. Transformations to a qubit are called quantum gates. Transformations of qubits can be visualized through rotations of the Bloch Sphere. There is a lot of amazing math that is involved in merely visualizing a qubit that I would encourage you to explore, but for now we will focus on gaining a high level sense of qubits and their quantum mechanical properties.
Like quantum superposition, quantum entanglement is another key quantum mechanical phenomenon that allows for extra computing power in quantum computers. Entanglement happens when particles’ quantum states must be described for all of the particles as a whole, instead of individually. In other words, the particles’ quantum states are entangled!
Entanglement can occur even when particles are very far apart. However a very important distinction and common misconception is that entanglement involves only correlation between individually random behaviors of entangled particles, not action. For example, you cannot use quantum entanglement to send a message from one entangled particle to another. You can only detect a correlation between particles after observations of both particles have happened.
Quantum Computing Today
All of these properties of quantum computers are being vehemently reserached in a myriad of approaches by academic and corporate institutions alike. Quantum Computing development opens up a whole new realm of solvable problems for computers.
Mathematicians have already developed 50+ quantum algorithms that if run on a quantum computer, could disrupt an incredible amount of security infrastructure. Today, much of internet security lays on the foundation of the RSA Cryptosystem, which secures your information based on the mathematical difficulty of finding prime factors of very large numbers (used to encrypt/decrypt public keys). A very notable quantum algorithm, Shor’s Algorithm, finds prime factors incredibly quickly in polynomial time. This would wreak havoc on RSA.
Quantum computers take the square root of computational time of a normal computer. For instance, an arduous computation that would take a normal computer 1 million seconds (~11.5 days) would only take a quantum computer 16.6 minutes. The bigger the computation time, the more drastic power a quantum computer would have compared to a classical computer.
Today, there are two types of Quantum Computing research and development that are vastly different and often confounded as the same: Quantum Annealing and Universal Quantum Computing.
Quantum annealing is a form of quantum computing, but even that statement is a heated debate within the quantum community. A computer that uses quantum annealing is called an adiabatic quantum computer.
Adiabatic computing company D-Wave released it’s 2000 Qubit commercial computer in January 2017. While their qubit count is the highest to date, their machine has been described as a “…a black box, literally and figuratively.” The only thing users can do is see what the machine outputs. Critics complain their qubit quality is low, and their $15 million chips that have the processing power of a laptop make it a far reach from a cost effective investment.
Greg Kuperberg, a mathematician, Quantum Algebra researcher, and professor at the University of California, Davis points out,
“Just because [their chips] are quantum, that doesn’t make them a quantum computer… That’s like saying that any invention that is influenced by air must be an airplane. Of course, it’s not true; it might instead be bagpipes.”
Adiabatic computers cannot run Shor’s Algorithm and other quantum algorithms. Quantum annealing uses quantum tunnelling, where there is no direct manipulation of how a quantum state develops. Configurations of experiments are set up at the beginning, a quantum state evolves naturally, and the result at the end is the solution. This also means that computers using quantum annealing suffer from severe noise and errors because error correction cannot take place during a computation.
Universal Quantum Computing
Unlike quantum annealing, Universal Quantum Computing can control and manipulate quantum states. Therefore error correction can take place while a calculation is being processed.
UQC has the capability to run all quantum algorithms including Shor’s Algorithm and Grover’s Algorithm (for faster search).
Quantum state control and manipulation is a steep feat, so research and development are presently at a significantly smaller qubit scale compared to quantum annealers. Intensive research of this unchartered field comes with intensive costs, so unsurprisingly some of the biggest pioneers in the quantum computing field are those who have the most to gain and lose with commercial quantum computers: IBM, Google and Microsoft.
Quantum Computers Today
IBM’s quantum research approach utilizes loops of superconducting wire to make quantum bits (as does Google). They currently have a 17 qubit quantum computer. IBM believes in exposing quantum computers to developers as early as possible to progress the field, and they have an free SDK (Software Development Kit) called QISKIT (Quantum Information Software Kit) available on GitHub and on their cloud platform Bluemix. If you are interested in learning more about Quantum Computing in general, both are excellent resources to understand the fundamentals of quantum computing and create experiments of your own.
Google is joining forces with academia, and the head of it’s Quantum Computing Lab is John Martinis – a physics professor at UC Santa Barbara. In January 2017 Martinis made the bold claim that Google would achieve quantum supremacy – performing a calculation on a quantum computer beyond the reach of any classical computer – by the end of this year.
Google’s latest chip only has 6 qubits, but they are arranged in a 2×3 configuration. This is a necessary structure and important breakthrough for larger scales when qubits must be nestled side by side. Martinis says he needs a grid of 49 qubits for his quantum supremacy experiment. Currently the most qubits they have are 9 qubits arranged in a line, but Martinis is confident in their goal.
Microsoft is taking a unique and long term approach compared to IBM and Google – they are researching topological quantum computers. For qubits they are harnessing the topological properties of non-material quasiparticles called non-abelian anyons.
To give you some insight into how cutting edge this topic is, 3 physicists won the Nobel Prize on topological states of matter in October 2016.
One fascinating and somewhat easy to visualize property of topological computers is a technique called braiding:
Braiding is when information is encoded in the order that anyons swap positions. The paths that the sequence of swaps in space and time literally look braided.
While Microsoft’s approach is further from physical qubit development, it has a higher promise of achieving significantly more accurate calculations. In topological computing, information changes with macroscopic movements compared to small fluctuations in other research approaches. Thus calculations on a topological quantum computer are predicted to be computed with 99.999999(9)% accuracy. They call this six or seven nines because there are six to seven decimal places of precision. Other approaches have the capacity of only three to five nines. That much more precision will allow for significantly more formidable processing and error correction.
We are still quite a few years away from seeing commercial universal gate quantum computers. There is concurrent research utilizing all sorts of materials for qubits ranging from electrons, diamonds, and optics, to quasiparticles.
Alex Bocharov, a researcher in Microsoft’s Research Quantum Architectures and Computation lab, illustrates how capricious the field is:
“In the early 2000s… people thought it would take about 24 billion years to calculate on a quantum computer the energy levels of ferredoxin, which plants use in photosynthesis. Now, through a combination of theory, practice, engineering and simulation, the most optimistic estimates suggest that it may take around an hour.”
Even research that may not ultimately be used for quantum computing is exploring completely new parameters in unchartered territory. There are many unforeseeable and potentially powerful discoveries to be made, whether it be unrelated side effects or quantum computers themselves. The quantum computing industry has an enormous capacity to disrupt cryptography, chemistry, biology and technological innovation as a whole.