Riemannian optimization and automatic differentiation for complex quantum architectures
I. A. Luchnikov, Mikhail Krechetov, Sergey N. Filippov · arXiv (Cornell University) · 2020
Optimization methods on Riemannian manifolds are a powerful class of optimization methods that allow performing constrained optimization. We apply first-order Riemannian optimization methods to various problems of quantum physics including entanglement renormalization of local many-body Hamiltonians, quantum control, and quantum tomography. We show that the Riemannian optimization forms a new powerful numerical tool for solving different problems of quantum technologies. Besides, we provide a package written on top of TensorFlow for the Riemannian optimization in quantum physics