HIGH PERFORMANCE COMPUTING


Based Quantum Computing Experimental physicists at Lawrence Berkeley have demonstrated error characterisation on their qutrit quantum processor


Raising the Bar: Error Characterisation for Qutrit-


Error rates are currently a considerable


A team of experimental physicists at the Advanced Quantum Testbed (AQT) at Lawrence Berkeley National Laboratory (Berkeley Lab) demonstrated an error characterisation method – randomised benchmarking (RB) – on a superconducting qutrit quantum processor. Scientists have now tested this widely used error characterisation method with qutrits. Their results were published in Physical Review Letters earlier this year, marking a significant milestone towards benchmarking the accuracy of qutrit- based quantum devices and identifying future research barriers to overcome. Quantum processors usually operate on qubits – the quantum version of the classical bit of information with two states, usually labeled ‘0’ and ‘1’. Qutrit makes use of an extra quantum state, ‘2’, increasing the quantity of information stored. Qutrit- based processors promise advantages over qubits at implementing specific algorithms due to increased storage and processing capacity. However, this extra quantum level increases complexity and makes the existing error-characterisation methods unusable. ‘We knew crosstalk was an issue for our qutrit processor, but now with RB, we can quantify its impact on our qutrit device,’ noted Alexis Morvan, a postdoctoral researcher at AQT and in the Quantum Information Science Technology group at Berkeley Lab. ‘This allows us to compare it to a qubit-based architecture and try to find new ideas to minimise these effects and come up with better qutrit processors in the future.’


12 Scientific Computing World Autumn 2021


problem for quantum computers, causing decoherence (loss of information) and, therefore, problems with the execution of quantum logic gates, which corrupt the results. A growing number of qubits or qutrits increases the propensity for errors, so finely describing these errors – error characterisation – allows researchers to overcome them and design better algorithms and processors.


Quantum complexity A key challenge for quantum experimentalists is that, as the number of qubits or qutrit increases, so does the complexity of the error. Techniques like RB have become standard for error characterisation for qubit-based processors. RB averages all possible errors into a single number that enables meaningful comparison across processors and systems. According to Morvan, this experimental


demonstration leveraged key expertise from Berkeley Lab and UC Berkeley and built on previous multidisciplinary research at AQT on superconducting qutrits, enabling them to further explore a qutrit-based architecture and adapt the RB method to characterise the qutrit processor properly. ‘Before we developed qutrit RB,


we didn’t have a way to assess the performance of our qutrit processor in a standardised way,’ Morvan said. ‘Making a comparison between different processors and with other systems is really hard. Furthermore, because of the complexity of the two-qutrit system, we had to change a lot of the qubit method to be able to characterise it properly. Now with these protocols extended to qutrits, we can give a number that represents how well our qutrit quantum processor functions and compare it to other platforms, or compare different generations of processors.’ The team of scientists hopes that


Left to right: a) Illustration of a multi-transmon qubit chip (colored in blue). b) Representation of a qubit and a qutrit. c) Experimental data from randomised benchmarking on a qutrit, where each point represents a random experiment. The exponential decay with circuit depth characterises the qutrit’s performance.


@scwmagazine | www.scientific-computing.com


Page 1  |  Page 2  |  Page 3  |  Page 4  |  Page 5  |  Page 6  |  Page 7  |  Page 8  |  Page 9  |  Page 10  |  Page 11  |  Page 12  |  Page 13  |  Page 14  |  Page 15  |  Page 16  |  Page 17  |  Page 18  |  Page 19  |  Page 20  |  Page 21  |  Page 22  |  Page 23  |  Page 24  |  Page 25  |  Page 26  |  Page 27  |  Page 28  |  Page 29  |  Page 30  |  Page 31  |  Page 32  |  Page 33  |  Page 34