Abstract: Variational quantum algorithms (VQA) are widely speculated to deliver quantum advantages for practical problems under the quantum-classical hybrid computational paradigm in the near term. In this talk, I review some of our recent works which attempt to further incorporate neural modules into such quantum-classical hybrid paradigm. The combination of parameterized quantum circuits (PQC) and neural networks, also known as variational quantum-neural hybrid schemes, can greatly enhance the power of VQAs.

　　We show how quantum-neural hybrid schemes can be applied in the scenarios of quantum architecture search (QAS) and variational quantum eigensolvers (VQE). In the QAS case, we demonstrate how neural architecture search ideas can be adopted by two examples: differentiable QAS and neural predictor based QAS. Such automatic architecture engineerings of quantum circuits have various applications in VQA design and quantum error mitigation. In the VQE case, we demonstrate how neural network post-processing module can be efficiently attached with the PQC, giving an exponential acceleration than all previous methods. Such variational quantum-neural hybrid eigensolvers consistently and significantly outperform VQE in simulating ground-state energies of quantum spins and molecules given the same amount of quantum hardware resources.

　　Bio: Shixin Zhang is a Senior Researcher at Tencent Quantum Laboratory. He obtained Ph.D. at Institute for Advanced Study, Tsinghua University, under the supervision of Prof. Hong Yao. Before that, he received his bachelor's degree in physics from Tsinghua University. His research interests include quantum computation, quantum simulation, variational quantum algorithms and non-equilibrium quantum physics.