The First Reprogrammable Quantum Computer Has Been Created
Adam Rifkin stashed this in Quantum Computers
The beginning of reprogrammable quantum computers:
While several other teams and companies, including computer technology giant IBM, are in on the race towards quantum computing, all the quantum computers presented thus far can only run one type of operation—which is ironic, seeing as quantum computers can theoretically run more operations than there are atoms in the universe.
Finally, a team of researchers from the University of Maryland say they have developed the first ever fully programmable and reprogrammable quantum computer. It is made up of just five ytterbium atoms standing as quantum bits or qubits of information, which are electrically charged in a magnetic field.
“Until now, there hasn’t been any quantum-computing platform that had the capability to program new algorithms into their system. They’re usually each tailored to attack a particular algorithm,” said study lead author Shantanu Debnath, a quantum physicist and optical engineer at the University of Maryland, College Park.
The ions are manipulated using lasers through a process called “optical pumping” in order to set them into the desired quantum energy state. The researchers program and reprogram the ions with a variety of algorithms using lasers to infuse them with precise amounts of energy, which dictates how they interact with each other.
The five-qubit quantum computer was tested on three algorithms previous quantum computers were able to execute quickly: the Deutsch-Jozsa algorithm, the Bernstein-Vazirani algorithm, and the quantum Fourier transform algorithm. The system scored 95%, 90%, and 70% in each of the algorithms respectively.
BABY STEPS TOWARDS SOMETHING COLOSSAL:
While “five qubits” sounds cute—like we’re being taken to a journey back to primitive computer technology, the implications of this development are huge.
Quantum computing follows the Schrödinger’s Cat thought experiment which dictates that a particle is simultaneously in all possible states. This means that in quantum computing, every qubit can be in superposition (both one and zero at the same time), unlike the binary system wherein each bit can only be either 1 or 0. This leads to an exponentially faster computing power.
Top Reddit comment:
This is a pretty interesting development, but the article is a little too optimistic with what it means, exactly. This is not the fault of the author, quantum computing is a super mind-bendy subject and I don't know anybody who's comfortable saying: "I know how it works." I'll try to explain what this development means, by the best of my ability, and I'll give a short explanation about the details of quantum computing that the author got slightly wrong.
Quantum computing isn't a quantitative improvement over normal computers, it is a qualitative change. Without going into details (details which I don't claim to understand), you can say that quantum computers will not help us solve any possible computer problem, but it will be vastly superior to current computers in a small set of important problems. So they won't help you play an mp3, but they might allow Google to train its AI's a billion times faster in some esoteric fashion, something like that (hand waving occurs).
The reason quantum computers only help us with a small set of problems is complicated to explain as well, but the result is that we only have a few processess to run on a quantum computer at all. Discovering more of these processes is the difficult part. These processes are so few, and so specific, that up to now, quantum computers were specifically designed with one of these processes in mind. This paper tells us of the first computer that doesn't work like that, and that can run essentially any such process.
This is helpful for researchers, because when a new quantum process is "discovered", the researchers won't need to build a new computer to be able to run this new process. This can really speed up the quantum computing field. Imagine back in the day people would have to build a new processor because someone invents a computer algorithm to forecast the weather, and then again build a new processor because someone learns to predict ocean currents. Then imagine somebody makes a computer like we have now, which can do all those things and more. This discovery is similar to that.
This is super cool of course, but the problem still lies with discovering the processes. That's the hard part, because we can already discover these processes via a mathematical way, we only need a computer to test correctness, evaluate performance and actually run the damn thing. (again AFAIK).
Source and 45 Reddit comments:
Quantum computers ready to leap out of the lab in 2017:
Everytime I have investigated claims of actual quantum computers I've been disappointed.
Strangely enough, though, I spent 6 very tough weeks in college writing programs for non-existent quantum computers.
Actual quantum computers seem quite primitive right now.
The UMaryland one seems legit.
Sounds like a fun college class.