Universal effectiveness of high-depth circuits in variational eigenproblems

2020/10/31 by Joonho Kim, Jaedeok Kim, Dario Rosa
Computer Science · Mathematics · Physics and Astronomy · #Combinatorics #Electronic circuit #Euclidean geometry #Geometry #Hamiltonian (control theory) #Ising model #Mathematical analysis #Mathematical optimization #Mathematics #Maxima and minima #Physics #Quantum #Quantum Computing Algorithms and Architecture #Quantum Information and Cryptography #Quantum circuit #Quantum computer #Quantum many-body systems #Quantum mechanics #Quantum network #Statistical physics #Topology (electrical circuits) #Transverse field #cond-mat.str-el #cs.LG #hep-th #quant-ph

paper · pdf · doi:10.1103/physrevresearch.3.023203

published as Phys. Rev. Research 3, 023203 (2021) · v1: revtex4-1, 14 pages, 11 figures; v2: 13 pages, revised for journal submission

openalex created_date 2020/10/08 · arxiv created 2021/03/08 · openalex publication_date 2021/06/11 · arxiv updated 2021/06/16 · openalex updated_date 2026/08/06

Abstract

This paper shows that generic high-depth variational circuits can approximate the ground state of any given one-dimensional quantum many-body Hamiltonian by carrying out the local gradient search of the circuit parameters.

Citations

• Adam: A Method for Stochastic Optimization
• Gapless spin-fluid ground state in a random quantum Heisenberg magnet
• A variational eigenvalue solver on a photonic quantum processor
• Comments on the Sachdev-Ye-Kitaev model
• No bad local minima: Data independent training error guarantees for multilayer neural networks
• Towards Understanding Generalization of Deep Learning: Perspective of Loss Landscapes
• Eigenvalues of the Hessian in Deep Learning: Singularity and Beyond
• Quantum Computing in the NISQ era and beyond
• Evaluating analytic gradients on quantum hardware
• Quantum Natural Gradient
• Symmetry-adapted variational quantum eigensolver
• Learning Unitaries by Gradient Descent
• Loss landscapes and optimization in over-parameterized non-linear systems and neural networks
• Learning to learn with quantum neural networks via classical neural networks