Efficient Tensor Network Emulation of Quantum Computers: Large Scale Algorithm Performance Analysis

We develop, implement, and apply numerical tensor network (TN) methods from quantum many-body physics to the simulation of digital quantum computers. Relying on the TN structure, we devise an algorithm to optimally sample from TN states. Employing digital quantum computers, we devise a strategy to generalize the ground-state search algorithm in a hybrid quantum-classical setting. We then focus on the development of HPC-ready, open-source TN libraries, specifically “Quantum Matcha TEA”, tailored to emulate quantum computers efficiently in both noiseless and noisy scenarios. To showcase the library, we first study the entanglement growth in Quantum Neural Networks (QNN) due to random initialization, defining the entanglement speed metric that uniquely characterizes different QNN structures. Finally, we attack the problem of simulating fermionic systems using quantum computers, detailing how the defermionized Hamiltonian is mapped to the qubit’s algebra, focusing on the Fermi-Hubbard model on a 4×2 ladder.