Double descent in quantum kernel methods

Abstract

The double descent phenomenon challenges traditional statistical learning theory by revealing scenarios where larger models do not necessarily lead to reduced performance on unseen data. While this counterintuitive behavior has been observed in a variety of classical machine learning models, particularly modern neural network architectures, it remains elusive within the context of quantum machine learning. In this work, we analytically demonstrate that linear regression models in quantum feature spaces can exhibit double descent behavior by drawing on insights from classical linear regression and random matrix theory. Additionally, our numerical experiments on quantum kernel methods across different real-world datasets and system sizes further confirm the existence of a test error peak, a characteristic feature of double descent. Our findings provide evidence that quantum models can operate in the modern, overparameterized regime without experiencing overfitting, potentially opening pathways to improved learning performance beyond traditional statistical learning theory.

References in corpus (43)

• Quantum Machine Learning
• Variational Quantum Algorithms
• Supervised learning with quantum enhanced feature spaces
• Machine learning and the physical sciences
• Noisy intermediate-scale quantum (NISQ) algorithms
• Quantum machine learning in feature Hilbert spaces
• Reconciling modern machine learning practice and the bias-variance trade-off
• Circuit-centric quantum classifiers
• The effect of data encoding on the expressive power of variational quantum machine learning models
• Power of data in quantum machine learning
• Data re-uploading for a universal quantum classifier
• A rigorous and robust quantum speed-up in supervised machine learning
• Prevalence of Neural Collapse during the terminal phase of deep learning training
• Information-theoretic bounds on quantum advantage in machine learning
• Exploring entanglement and optimization within the Hamiltonian Variational Ansatz
• Barren Plateaus in Variational Quantum Computing
• Quantum machine learning beyond kernel methods
• Theory of overparametrization in quantum neural networks
• Scaling description of generalization with number of parameters in deep learning
• Generalization in Quantum Machine Learning: a Quantum Information Perspective
• Training Quantum Embedding Kernels on Near-Term Quantum Computers
• ADAPT-VQE is insensitive to rough parameter landscapes and barren plateaus
• Evaluating the noise resilience of variational quantum algorithms
• On the Quantum versus Classical Learnability of Discrete Distributions
• A jamming transition from under- to over-parametrization affects loss landscape and generalization
• Representation Learning via Quantum Neural Tangent Kernels
• Understanding quantum machine learning also requires rethinking generalization
• Memorizing without overfitting: Bias, variance, and interpolation in over-parameterized models
• Potential and limitations of quantum extreme learning machines
• Analytic theory for the dynamics of wide quantum neural networks
• Adaptive pruning-based optimization of parameterized quantum circuits
• Generalization despite overfitting in quantum machine learning models
• Universal Effectiveness of High-Depth Circuits in Variational Eigenproblems
• Overfitting in quantum machine learning and entangling dropout
• Structural risk minimization for quantum linear classifiers
• On fundamental aspects of quantum extreme learning machines
• Generalization of Quantum Machine Learning Models Using Quantum Fisher Information Metric
• A General Approach to Dropout in Quantum Neural Networks
• A super-polynomial quantum-classical separation for density modelling
• Potential and limitations of random Fourier features for dequantizing quantum machine learning
• On the expressivity of embedding quantum kernels
• Characterization of overparametrization in the simulation of realistic quantum systems
• Single-shot quantum machine learning