Noise tailoring for robust amplitude estimation

A universal fault-tolerant quantum computer holds the promise to speed up computational problems that are otherwise intractable on classical computers; however, for the next decade or so, our access is restricted to noisy intermediate-scale quantum (NISQ) computers and, perhaps, early fault tolerant (EFT) quantum computers. This motivates the development of many near-term quantum algorithms including robust amplitude estimation (RAE), which is a quantum-enhanced algorithm for estimating expectation values. One obstacle to using RAE has been a paucity of ways of getting realistic error models incorporated into this algorithm. So far the impact of device noise on RAE is incorporated into one of its subroutines as an exponential decay model, which is unrealistic for NISQ devices and, maybe, for EFT devices; this hinders the performance of RAE. Rather than trying to explicitly model realistic noise effects, which may be infeasible, we circumvent this obstacle by tailoring device noise using randomized compiling to generate an effective noise model, whose impact on RAE closely resembles that of the exponential decay model. Using noisy simulations, we show that our noise-tailored RAE algorithm is able to regain improvements in both bias and precision that are expected for RAE. Additionally, on IBM's quantum computer ibmq_belem our algorithm demonstrates advantage over the standard estimation technique in reducing bias. Thus, our work extends the feasibility of RAE on NISQ computers, consequently bringing us one step closer towards achieving quantum advantage using these devices.

Automated summary

A universal fault-tolerant quantum computer is currently unavailable, but the development of near-term quantum algorithms like robust amplitude estimation (RAE) can optimize the performance of noisy intermediate-scale quantum (NISQ) computers and early fault tolerant (EFT) quantum computers. The lack of realistic error models has been a challenge for RAE, but we solve this by tailoring device noise using randomized compiling to generate an effective noise model that closely simulates the impact of device noise on RAE. By conducting noisy simulations, we demonstrate that our noise-tailored RAE algorithm can achieve improvements in bias and precision, even showing an advantage over standard estimation techniques on IBM's quantum computer ibmq_belem.