Novel Approach on Error Suppression for Quantum Computers
A new study published in PRX reports on a novel technique based on a practically feasible quantum circuit, to control and suppress errors in current and near-term quantum devices.
Quantum computers, made essentially of qubits, use superposition and entanglement to improve their processing capability beyond what is possible with standard classical computers. This improvement is still in the midst of achieving its milestones simply because qubit instability is the main challenge that this field is facing nowadays.
Qubits are very sensitive because any interaction with their environment tends to add decoherence to the system and hampers the quality of measurements and operations. This noise translates into what scientists refer to as “errors” and, thus, being able to control and correct these quantum errors is essential to achieving a fault-tolerant quantum computational system, including quantum gates, preparations of states, measurements, that can process, if needed, faulty information in a very efficient way by using quantum error-correcting codes (QECs).
Correcting errors and noise
In a recent study published in Physical Review X, Oxford researcher Bálint Koczor, partner of the Quantum Flagship AQTION project, has reported on a novel theoretical approach for exponential error suppression in near-term quantum devices. The main challenge the study has addressed is that the bias that comes from imperfections of the experimental quantum gates in estimating an expectation value[1] hinders the practical applicability of quantum computers. While estimating these expectation values is the basis of almost all near-term quantum computations, only true QECs were thought to be able to remove such biases. On the other hand, QECs are incompatible with near-term technology as they will require mature, more advanced quantum computers.
Now, with this approach, one can obtain a much better estimate of the expectation values in quantum computers by suppressing errors via a so-called derangement circuit. That is, this technique prepares n identical copies of the computational states and uses derangement operators to protect the collective permutation symmetry of the states. Most of the noise events that occur during the preparation of the quantum states break the permutation symmetry and, thus, can be easily filtered out by the algorithm. These n copies of the computational state can be prepared completely independently and only need to be “bridged” by a shallow derangement circuit (by shallow we mean that it can be decomposed into a linear number of primitive gates), picking up much less noise than in the state-preparation phase.
The main idea is that we use multiple noisy quantum processors to perform multiple copies of the same computation and we use the copies to validate each other.
Bálint Koczor / Oxford researcher at AQTION
As Koczor comments “This approach has given us an entirely new way of thinking about errors and noise in quantum computers. The main idea is that we use multiple noisy quantum processors to perform multiple copies of the same computation and we use the copies to validate each other. The reason why comparing quantum states is much more delicate than comparing everyday objects is that quantum states get destroyed whenever we try to `read' them. What we see here is that this approach actually circumvents this, giving us a much better result”.
In the long term, there is much expectation awaiting for quantum computers, aimed to improve, for example, the field of biomedicine, in the modelling of molecules with much more precision than with current computers, an essential tool for discovering new drugs. Similarly, we expect to see a major improvement in the field of material sciences, where, again, the modelling of materials and understanding and controlling their properties at the smallest scale possible will enable discovering better materials for batteries or even superconductors.
[1] Expectation value: the average result of a number of repetitive measurements of an operator or quantum circuit.
More information
Cited article: B. Koczor, Exponential Error Suppression for Near-Term Quantum Devices, 2021, Phys. Rev. X 11, 031057
