Generalization in Quantum Machine Learning: a Quantum Information Perspective

Abstract

Quantum classification and hypothesis testing are two tightly related subjects, the main difference being that the former is data driven: how to assign to quantum states $\rho (x)$ the corresponding class $c$ (or hypothesis) is learnt from examples during training, where $x$ can be either tunable experimental parameters or classical data "embedded" into quantum states. Does the model generalize? This is the main question in any data-driven strategy, namely the ability to predict the correct class even of previously unseen states. Here we establish a link between quantum machine learning classification and quantum hypothesis testing (state and channel discrimination) and then show that the accuracy and generalization capability of quantum classifiers depend on the (Rényi) mutual informations $I(C:Q)$ and $I_2(X:Q)$ between the quantum state space $Q$ and the classical parameter space $X$ or class space $C$. Based on the above characterization, we then show how different properties of $Q$ affect classification accuracy and generalization, such as the dimension of the Hilbert space, the amount of noise, and the amount of neglected information from $X$ via, e.g., pooling layers. Moreover, we introduce a quantum version of the Information Bottleneck principle that allows us to explore the various tradeoffs between accuracy and generalization. Finally, in order to check our theoretical predictions, we study the classification of the quantum phases of an Ising spin chain, and we propose the Variational Quantum Information Bottleneck (VQIB) method to optimize quantum embeddings of classical data to favor generalization.

19 pages, 7 figures. Updated version with better connection to quantum hypothesis testing, new numerical applications for quantum phase recognition and for training quantum embeddings

References in corpus (12)

• The power of quantum neural networks
• Quantum Illumination with Gaussian States
• The Quantum Chernoff Bound
• Quantum Illumination at the Microwave Wavelengths
• Microwave Quantum Illumination
• Absence of Barren Plateaus in Quantum Convolutional Neural Networks
• Quantum state discrimination and its applications
• Information-theoretic bounds on quantum advantage in machine learning
• Renyi generalizations of the conditional quantum mutual information
• Entanglement-Assisted Absorption Spectroscopy
• Entanglement-enhanced testing of multiple quantum hypotheses
• Quantum-enhanced barcode decoding and pattern recognition

Cited by in corpus (72)

• Challenges and Opportunities in Quantum Machine Learning
• Generalization in quantum machine learning from few training data
• Quantum Computing for High-Energy Physics: State of the Art and Challenges. Summary of the QC4HEP Working Group
• Quantum Machine Learning for Chemistry and Physics
• Efficient measure for the expressivity of variational quantum algorithms
• Theory for Equivariant Quantum Neural Networks
• Theoretical Guarantees for Permutation-Equivariant Quantum Neural Networks
• Biology and medicine in the landscape of quantum advantages
• Towards quantum enhanced adversarial robustness in machine learning
• Understanding quantum machine learning also requires rethinking generalization
• Out-of-distribution generalization for learning quantum dynamics
• NISQ Computers: A Path to Quantum Supremacy
• Subtleties in the trainability of quantum machine learning models
• Importance of Kernel Bandwidth in Quantum Machine Learning
• Classical-to-quantum convolutional neural network transfer learning
• Quantum machine learning of large datasets using randomized measurements
• Analytic theory for the dynamics of wide quantum neural networks
• Generalization despite overfitting in quantum machine learning models
• Generative quantum machine learning via denoising diffusion probabilistic models
• Quantum Neural Network Classifiers: A Tutorial
• Quantum phase detection generalisation from marginal quantum neural network models
• Overfitting in quantum machine learning and entangling dropout
• Structural risk minimization for quantum linear classifiers
• Exponential data encoding for quantum supervised learning
• Fock State-enhanced Expressivity of Quantum Machine Learning Models
• Quantum kernels to learn the phases of quantum matter
• Hierarchical quantum circuit representations for neural architecture search
• Noisy Quantum Kernel Machines
• Complexity of quantum circuits via sensitivity, magic, and coherence
• Generalization of Quantum Machine Learning Models Using Quantum Fisher Information Metric
• Tensor networks for interpretable and efficient quantum-inspired machine learning
• Numerical evidence against advantage with quantum fidelity kernels on classical data
• A General Approach to Dropout in Quantum Neural Networks
• Training robust and generalizable quantum models
• Efficient algorithms for quantum information bottleneck
• Randomness-enhanced expressivity of quantum neural networks
• Statistical Complexity of Quantum Learning
• Incremental Data-Uploading for Full-Quantum Classification
• When Federated Learning Meets Quantum Computing: Survey and Research Opportunities
• Expressivity of Variational Quantum Machine Learning on the Boolean Cube
• Optimizing Quantum Convolutional Neural Network Architectures for Arbitrary Data Dimension
• Quantum support vector data description for anomaly detection
• Generalization Error Bound for Quantum Machine Learning in NISQ Era -- A Survey
• Satellite image classification with neural quantum kernels
• Variational Quantum Circuit Design for Quantum Reinforcement Learning on Continuous Environments
• Accuracy vs Memory Advantage in the Quantum Simulation of Stochastic Processes
• An invitation to the sample complexity of quantum hypothesis testing
• Quantum neural compressive sensing for ghost imaging
• Generalized quantum Arimoto-Blahut algorithm and its application to quantum information bottleneck
• Quantum Kerr Learning
• Single-shot quantum machine learning
• Optimal lower bounds for Quantum Learning via Information Theory
• Supervised Learning Guarantee for Quantum AdaBoost
• Encoding-dependent generalization bounds for parametrized quantum circuits
• QKAN: quantum Kolmogorov-Arnold networks with applications in machine learning and multivariate state preparation
• Quantum network tomography of Rydberg arrays by machine learning
• Data-Dependent Generalization Bounds for Parameterized Quantum Models Under Noise
• Quantum-enhanced cluster detection in physical images
• Attainability of quantum state discrimination bounds with collective measurements on finite copies
• Out-of-distribution generalisation for learning quantum channels with low-energy coherent states
• Optimal Qubit Mapping Search for Encoding Classical Data into Matrix Product State Representation with Minimal Loss
• Double descent in quantum kernel methods
• Quantum-enhanced learning with a controllable bosonic variational sensor network
• Analysis of Quantum Image Representations for Supervised Classification
• A method for quantifying the generalization capabilities of generative models for solving Ising models
• Noise-Resilient Quantum Reinforcement Learning
• Superior resilience to poisoning and amenability to unlearning in quantum machine learning
• Quantum Machine Learning for Climate Modelling
• Quantum phase classification via partial tomography-based quantum hypothesis testing
• Quantum Machine Learning with SQUID
• Chernoff Information Bottleneck for Covert Quantum Target Sensing
• Disentangling quantum autoencoder