IBM scientists say they now have the capability to detect and measure two types of quantum errors (bit-flip and phase-flip) that will occur in any real quantum computer, opening the door for quantum error correction. The scientists also demonstrated a new, square quantum bit circuit design that they say is the only physical architecture that could successfully scale to larger dimensions.
“Just a few weeks ago was the 50th anniversary of Moore’s Law,” says Jerry M. Chow, manager of the Experimental Quantum Computing group at IBM’s T.J. Watson Research Center and the primary investigator on the Intelligence Advanced Research Projects Activity (IARPA) sponsored Multi-Qubit Coherent Operations project. “The whole world knows that Moore’s Law is coming to an end.”
Chow adds, “What’s the next paradigm for computing? What’s beyond Moore’s Law?”
Quantum’s effect on big data
Quantum computing may be the next frontier. The bit is the most basic piece of information typical computers understand. A bit can have one of two values: ‘1’ or ‘0’. But a quantum computer understands quantum bits (qubits) that can hold a value of ‘1’, ‘0’ or both values at the same time (described as a superposition and denoted as ‘0+1’). In theory, the superposition property will allow quantum computers to choose a correct solution among millions of possibilities much faster than a conventional computer.
That property would make quantum computers especially well-suited to big data problems around optimization and simulation. Quantum computers would be able to quickly sort and curate ever larger databases and massive stores of diverse, unstructured data. For instance, Chow says quantum devices could completely transform areas like chemical design, drug design and other biopharma applications by allowing scientists to simulate how a particular molecule interacts with other molecules.
Until now, Chow says, one of the biggest stumbling blocks in quantum computing has been controlling or removing “quantum decoherence,” or the creation of errors in calculations caused by interference from factors such as heat, electromagnetic radiation and material defects.
In conventional computing, you need to worry about bit-flip errors, where a bit that should be ‘0’ presents as a ‘1’ and vice versa. Error correction algorithms are used to detect and correct such errors. That’s why, for instance, a compact disc with some minor scratches can still play.
The problem of bit-flip
In quantum computing, bit-flip errors are still a problem, but so is something called a ‘phase-flip’ error. That’s when the error flips the phase relationship between ‘0’ and ‘1’ in the superposition state — a ‘0+1’ turns into a ‘0-1’ or vice versa.
“Say you have a wave — maybe sine wave or cosine wave,” Chow says. “If you have these two waves and you jostle them together, the sound might add together and constructively interfere. They’d be in phase. If you jostle them opposite each other, the sound might cancel due to destructive interference. In this way, they’re out of phase.”
To correct errors in quantum computing, you need to be able to detect bit-flip errors and phase-flip errors simultaneously. But until now, it has only been possible to address one type of quantum error at a time. That’s especially problematic because quantum information is very fragile — all existing qubit technologies lose their information when interacting with matter and electromagnetic radiation.
To break down that hurdle, the IBM researchers developed a quantum bit circuit, based on a square lattice of four superconducting qubits on a chip roughly 6mm on a side, rather than the linear array of qubits that researchers have used in the past. The researchers say they used a variety of techniques to measure the states of two independent syndrome (measurement) qubits, which each reveal one aspect of the quantum information stored on the other two qubits (called code or data qubits).
Chow says that because these qubits can be designed and manufacturing using standard silicon fabrication techniques — metal on silicon — once a handful of superconducting qubits can be manufactured reliably and repeatedly and controlled with low error rates, there should be no fundamental obstacle to demonstrating error correction in larger lattices of qubits. From there, one of the next milestones may be the creation of quantum algorithms.