This talk will explore strategies for simulating large quantum circuits on a classical computer that has access to a small quantum device. We show that a quantum circuit, represented by a tensor network, can be cut into smaller pieces and each piece executed independently on a smaller device by simulating contraction of the corresponding smaller tensor network. Assuming a partition with not too many edges between different parts can be found, we provide efficient algorithms for simulating such circuits. While in general the simulation cost scales exponentially in the total number of edges between different parts, the size of the quantum memory required scales only linearly in the degree of each part.

This talk is based on joint work with:

Aram Harrow (MIT)

Tianyi Peng (Tsinghua University)

Xiaodi Wu (University of Oregon)