Structural risk minimization for quantum linear classifiers

Abstract

Quantum machine learning (QML) models based on parameterized quantum circuits are often highlighted as candidates for quantum computing's near-term ``killer application''. However, the understanding of the empirical and generalization performance of these models is still in its infancy. In this paper we study how to balance between training accuracy and generalization performance (also called structural risk minimization) for two prominent QML models introduced by Havlíček et al. (Nature, 2019), and Schuld and Killoran (PRL, 2019). Firstly, using relationships to well understood classical models, we prove that two model parameters -- i.e., the dimension of the sum of the images and the Frobenius norm of the observables used by the model -- closely control the models' complexity and therefore its generalization performance. Secondly, using ideas inspired by process tomography, we prove that these model parameters also closely control the models' ability to capture correlations in sets of training examples. In summary, our results give rise to new options for structural risk minimization for QML models.

27 pages, 3 figures

References in corpus (15)

• Supplementary information for "Quantum supremacy using a programmable superconducting processor"
• Variational Quantum Algorithms
• A Quantum Approximate Optimization Algorithm
• The power of quantum neural networks
• Power of data in quantum machine learning
• A rigorous and robust quantum speed-up in supervised machine learning
• Classical simulation of commuting quantum computations implies collapse of the polynomial hierarchy
• Information-theoretic bounds on quantum advantage in machine learning
• Generalization in Quantum Machine Learning: a Quantum Information Perspective
• Efficient measure for the expressivity of variational quantum algorithms
• On the statistical complexity of quantum circuits
• Prediction, Learning, Uniform Convergence, and Scale-sensitive Dimensions
• Practical application improvement to Quantum SVM: theory to practice
• The Learnability of Quantum States
• Rademacher complexity of noisy quantum circuits

Cited by in corpus (14)

• Quantum machine learning beyond kernel methods
• Understanding quantum machine learning also requires rethinking generalization
• Out-of-distribution generalization for learning quantum dynamics
• Shadows of quantum machine learning
• Generalization despite overfitting in quantum machine learning models
• Dynamical simulation via quantum machine learning with provable generalization
• Learning Quantum Processes and Hamiltonians via the Pauli Transfer Matrix
• On the expressivity of embedding quantum kernels
• Statistical Complexity of Quantum Learning
• Expressivity of Variational Quantum Machine Learning on the Boolean Cube
• Kernel-based quantum regressor models learn non-Markovianity
• Quantum Kernel Machine Learning With Continuous Variables
• Double descent in quantum kernel methods
• Quantum Algorithms for Causal Estimands