IBM Paves The Way Towards Scalable Quantum Computing
Alex Knapp, Forbes Staff
Three superconducting qubits. (Credit: IBM Research)
And now that I’ve thrown a ton of information at you in one tiny sentence, let’s break it all down. I had a chance to talk with IBM scientist Matthias Steffen about this new technology, and he broke it down for me. Let’s start with the qubit. Classical computing, as you probably know, is based on the bit. A bit can exist in one of two possible states, which are typically referred to as “0″ or “1″. A qubit is the equivalent of a bit for quantum computing. It can be in three possible states – “0″ or “1″ or both. The “both” state is known as the superposition. Now, the difference may seem subtle, but mathematically, it’s huge. A few hundred qubits can contain more classical bits of information than the the universe has atoms.
IBM Shrinks Computer Memory Into Only Twelve Atoms
Alex Knapp Forbes StaffWhat makes quantum computing challenging is the problem of decoherence. When a qubit is moved from the 0 state to either 1 or the superposition, it will decohere to state 0 due to interference from other parts of the computer. In order for quantum computing to be scalable and practical, the qubits have to be coherent for a long enough time that error-correction techniques can be employed to make sure that the decoherence doesn’t prevent accurate computation.
“In 1999, coherence times were about 1 nanosecond,” Steffen told me. “Last year, coherence times were achieved for as long as 1 to 4 microseconds. With these new techniques, we’ve achieved coherence times of 10 to 100 microseconds. We need to improve that by a factor of 10 to 100 before we’re at the threshold we want to be. But considering that in the past ten years we’ve increased coherence times by a factor of 10,000, I’m not scared.”
Alex Knapp Forbes Staff
MIT's Scott Aaronson Explains Quantum Computing
The IBM team has taken two approaches to quantum computing, both of which factor into the breakthroughs announced here. The first approach is building a 3-D qubit made from superconducting, microfabricated silicon. Steffen notes that the benefit of using silicon for these qubits is that the manufacturing equipment and know-how already exists – new techniques don’t have to be developed. 3-D qubits were pioneered by the Schoelkopf Lab at Yale, and Steffen expressed his admiration for that work. Building on the Yale techniques, the IBM team was able to maintain coherence for 95 microseconds. (“But you could round that to 100 for the piece if you want,” Steffen joked.)
MIT's Scott Aaronson Explains Quantum Computing
The IBM team has taken two approaches to quantum computing, both of which factor into the breakthroughs announced here. The first approach is building a 3-D qubit made from superconducting, microfabricated silicon. Steffen notes that the benefit of using silicon for these qubits is that the manufacturing equipment and know-how already exists – new techniques don’t have to be developed. 3-D qubits were pioneered by the Schoelkopf Lab at Yale, and Steffen expressed his admiration for that work. Building on the Yale techniques, the IBM team was able to maintain coherence for 95 microseconds. (“But you could round that to 100 for the piece if you want,” Steffen joked.)
Alex Knapp Forbes StaffThe second approach involved a traditional 2-D qubit, which IBM’s scientists used to build a “Controlled NOT gate” or CNOT gate, which is a building block of quantum computing. A CNOT gate connects two qubits such that the second qubit will change state if the first qubit changes its state to 1. For example, if qubit A’s state is changed from 0 to 1, and qubit B’s state is 1, it will flip to state 0. But if qubit A’s state is changed from 1 to 0, qubit B is unaffected. That seems simple enough, but when you scale multiple logic gates like this together, you have a very real basis for computation. The CNOT gates were able to maintain coherence times of 10 microseconds, which is long enough to show a 95% accuracy rate. The previous accuracy record for CNOT gates was 81% accuracy, so this is a huge step. Of course, Steffen was quick to note that there’s still a ways to go before this can be implemented as a computing solution. That makes common sense, since 95% is accurate, but in the long run you need the accuracy to be as close to 100% as possible.
Given the rapid progress that IBM has made, scalable quantum computing is starting to look like a real possibility. As error-correction protocols improve and coherence times lengthen, accurate quantum computing becomes a real possibility. But don’t expect to have a quantum smartphone anytime soon using this technique. In order to get the results the IBM team has seen in either the 2-D or 3-D configuration, the qubits have to be cooled down to less than a degree above absolute zero.
“There’s a growing sense that a quantum computer can’t be a laptop or desktop,” said Steffen. “Quantum computers may well just being housed in a large building somewhere. It’s not going to be something that’s very portable. In terms of application, I don’t think that’s a huge detriment because they’ll be able to solve problems so much faster than traditional computers.”
The next steps for the team is to improve coherence and error-correction protocols to the point where the accuracy is over 99.9%. That means they’ll have achieved a “logical qubit” – one that, for practical purposes, doesn’t experience decoherence. From that point, the next step is to develop a quantum computing architecture. IBM is considering some possibilities here, including developing some quantum memory architechture. But what encourages Steffen in these endeavors is that these are questions of engineering, not of theory.
“We are very excited about how the quantum computing field has progressed over the past ten years,” he told me. “Our team has grown significantly over past 3 years, and I look forward to seeing that team continue to grow and take quantum computing to the next level.”
Newscribe : get free news in real time
