Recent years, prototypes of quantum computers have been built to the scale of around 100 qubits. Due to noise and length of coherence time, none of them have shown the capability of fault tolerant computing yet. This stage of quantum computing is dubbed as Noisy Intermediate-Scale Quantum (NISQ). One of the most interesting topics with NISQ devices is the possibility of achieving useful results with these noisy devices. A prominent method is Quantum Error Mitigation (QEM) which aims to reduce or cancel the error in NISQ quantum circuits. It has been shown that QEM has good performance in serval conditions. However, most of them rely on full knowledge of the noise model or have strong assumptions of the Hilbert space structure of the device. It is no harm to have these assumptions when doing experiments for oneself unless someone is accessing quantum computing service through cloud service, which could be prevailing in the future. In this talk, we present a formalism of QEC in a black box manner, where nothing is assumed except for the experimental statistics. We demonstrate the efficacy by applying it to a real Variational Quantum Algorithm (VQA) which solves the eigenstate energy of a chemical molecule. Our formalism will also survive the verification of Blind Quantum Computing (BQC).

Speaker Intro：

Dr. Wu Xingyao is currently a Quantum Algorithm Scientist in JD Explore Academy. He received his Ph.D. from Centre for Quantum Technologies (CQT) at National University of Singapore in 2017 and is also an alumnus of the USTC Special Class for the Gifted Young. After his Ph.D., he went to QuICS at University of Maryland for postdoc research. In 2019, he joined Huawei as a senior research staff on quantum algorithms and later in 2021 moved to JD. His main research interest lies in NISQ algorithms and its intersection with machine learning, quantum entanglement theory and verification of quantum computing.