Abstract:Quantum machine learning (QML) networks promise to have some computational (or quantum) advantage for classifying supervised datasets (e.g., satellite images) over some conventional deep learning (DL) techniques due to their expressive power via their local effective dimension. There are, however, two main challenges regardless of the promised quantum advantage: 1) Currently available quantum bits (qubits) are very small in number, while real-world datasets are characterized by hundreds of high-dimensional elements (i.e., features). Additionally, there is not a single unified approach for embedding real-world high-dimensional datasets in a limited number of qubits. 2) Some real-world datasets are too small for training intricate QML networks. Hence, to tackle these two challenges for benchmarking and validating QML networks on real-world, small, and high-dimensional datasets in one-go, we employ quantum transfer learning composed of a multi-qubit QML network, and a very deep convolutional network (a with VGG16 architecture) extracting informative features from any small, high-dimensional dataset. We use real-amplitude and strongly-entangling N-layer QML networks with and without data re-uploading layers as a multi-qubit QML network, and evaluate their expressive power quantified by using their local effective dimension; the lower the local effective dimension of a QML network, the better its performance on unseen data. Our numerical results show that the strongly-entangling N-layer QML network has a lower local effective dimension than the real-amplitude QML network and outperforms it on the hard-to-classify three-class labelling problem. In addition, quantum transfer learning helps tackle the two challenges mentioned above for benchmarking and validating QML networks on real-world, small, and high-dimensional datasets.

Quantum transfer component analysis for domain adaptation

Quantum classifiers for domain adaptation

Variational Quantum Domain Adaptation

Quantum kernel subspace alignment for unsupervised domain adaptation

Domain Adaptation via Transfer Component Analysis

Quantum correlation alignment for unsupervised domain adaptation

Quantum Algorithms for Feedforward Neural Networks

Quantum Transfer Learning with Adversarial Robustness for Classification of High-Resolution Image Datasets

An efficient quantum algorithm for independent component analysis

Domain Adaptation with Conditional Transferable Components.

Quantum Transfer Learning for Real-World, Small, and High-Dimensional Datasets

Multi-component transfer metric learning for handling unrelated source domain samples

Transfer learning of optimal QAOA parameters in combinatorial optimization

Classical-to-Quantum Transfer Learning Facilitates Machine Learning with Variational Quantum Circuit

Experimental Quantum Principal Component Analysis via Parametrized Quantum Circuits

Quantum data compression by principal component analysis

Adaptive Quantum State Tomography Via Linear Regression Estimation: Theory and Two-Qubit Experiment

Towards Transfer Learning for Large-Scale Image Classification Using Annealing-based Quantum Boltzmann Machines

Quantum Generative Adversarial Imitation Learning

Topological data analysis on noisy quantum computers

Experimental Demonstration of Real-Time Adaptive One-Qubit Quantum-State Tomography