Representation Learning via Quantum Neural Tangent Kernels
Open Access
- 17 August 2022
- journal article
- research article
- Published by American Physical Society (APS) in PRX Quantum
- Vol. 3 (3) , 030323
Abstract
Variational quantum circuits are used in quantum machine learning and variational quantum simulation tasks. Designing good variational circuits or predicting how well they perform for given learning or optimization tasks is still unclear. Here we discuss these problems, analyzing variational quantum circuits using the theory of neural tangent kernels. We define quantum neural tangent kernels, and derive dynamical equations for their associated loss function in optimization and learning tasks. We analytically solve the dynamics in the frozen limit, or lazy training regime, where variational angles change slowly and a linear perturbation is good enough. We extend the analysis to a dynamical setting, including quadratic corrections in the variational angles. We then consider a hybrid quantum classical architecture and define a large-width limit for hybrid kernels, showing that a hybrid quantum classical neural network can be approximately Gaussian. The results presented here show limits for which analytical understandings of the training dynamics for variational quantum circuits, used for quantum machine learning and optimization problems, are possible. These analytical results are supported by numerical simulations of quantum machine-learning experiments.Keywords
All Related Versions
Funding Information
- International Business Machines
- University of Chicago
- AFOSR (FA9550-21-1-0209, W911NF-18-1-0020, W911NF-18-1-0212, W911NF-16-1-0349, FA9550-19-1-0399, FA9550-21-1-0209, EFMA-1640959, OMA-1936118, EEC-1941583)
This publication has 21 references indexed in Scilit:
- Barren plateaus in quantum neural network training landscapesNature Communications, 2018
- Hardware-efficient variational quantum eigensolver for small molecules and quantum magnetsNature, 2017
- Quantum machine learningNature, 2017
- Chaos and complexity by designJournal of High Energy Physics, 2017
- The theory of variational hybrid quantum-classical algorithmsNew Journal of Physics, 2016
- Quantum Support Vector Machine for Big Data ClassificationPhysical Review Letters, 2014
- Quantum principal component analysisNature Physics, 2014
- A variational eigenvalue solver on a photonic quantum processorNature Communications, 2014
- Quantum Algorithm for Data FittingPhysical Review Letters, 2012
- Quantum Algorithm for Linear Systems of EquationsPhysical Review Letters, 2009