On Tuesday, 2 June, student of the University of Tartu Institute of Computer Science Mykhailo Nitsenko defended his thesis Quantum Circuit Fusion in the Presence of Quantum Noise on NISQ Devices, the first master’s thesis defended in the field of quantum computing at the University of Tartu.
In his thesis supervised by Dirk Oliver Theis and Dominique Unruh, Mykhailo Nitsenko studied a concept called “circuit fusion”, which proposes to reduce stochastic noise in estimating the expectation values of measurements at the end of quantum computations. But near-term quantum computing devices are also subject to quantum noise (such as decoherence etc.), and circuit fusion aggravates that problem.
Mykhailo Nitsenko ran thousands of experiments on IBM’s cloud quantum computers and used Fourier analysis techniques to quantify and visualise noise and the resulting information loss.
According to Mykhailo Nitsenko, before he enrolled in the University of Tartu he had a strong opinion that quantum computing is an abstract idea that we will never be able to use or even implement. “I just could not imagine how it is even possible to do computations on things without directly observing them. Quantum computing class showed me how it is done, and it became apparent to me that it is something I want to dedicate my academic efforts to,” said Nitsenko.
“If you don’t want to wait for fault-tolerant quantum computers, you may endeavour to use the noisy quantum computing devices that can be built already now. In that case, researching the effects of quantum noise on computations becomes important: these effects must be mitigated,” said Dirk Oliver Theis, Associate Professor of Theoretical Computer Science at the University of Tartu Institute of Computer Science. Theis added that he had expected that the mathematics which Mykhailo Nitsenko implemented in his thesis would help us understand some aspects of quantum noise which can be devastating to quantum computations, rendering the result pure gibberish.
In near-term quantum computing, one tries to run quantum circuits which are just short enough so that the correct output can be somehow reconstructed from the distorted measurement results. But quantum noise affects the results of computations on near-term quantum computers in complicated ways. “In the mathematical approach based on Fourier analysis that Nitsenko implemented, some effects were predictable, such as a decrease in the amplitudes due to decoherence. What was surprising was that the low frequencies of the quantum noise showed distinct patterns. In future research, this might be exploited to mitigate the effect of quantum noise on the computation,” said Theis.
This year, the Information Technology Foundation for Education (HITSA) granted funding to the University of Tartu Institute of Physics to continue and increase the training and research in the field of quantum computing at the university. With the support of this funding, new interdisciplinary courses focusing on quantum programming will be created.