Abstract:In the objective world, how to deal with the complexity and uncertainty of big data efficiently and accurately has become the premise and key to machine learning. Fuzzy support vector machine (FSVM) not only deals with the classification problems for training samples with fuzzy information, but also assigns a fuzzy membership degree to each training sample, allowing different training samples to contribute differently in predicting an optimal hyperplane to separate two classes with maximum margin, reducing the effect of outliers and noise, Quantum computing has super parallel computing capabilities and holds the promise of faster algorithmic processing of data. However, FSVM and quantum computing are incapable of dealing with the complexity and uncertainty of big data in an efficient and accurate manner. This paper research and propose an efficient and accurate quantum fuzzy support vector machine (QFSVM) algorithm based on the fact that quantum computing can efficiently process large amounts of data and FSVM is easy to deal with the complexity and uncertainty problems. The central idea of the proposed algorithm is to use the quantum algorithm for solving linear systems of equations (HHL algorithm) and the least-squares method to solve the quadratic programming problem in the FSVM. The proposed algorithm can determine whether a sample belongs to the positive or negative class while also achieving a good generalization performance. Furthermore, this paper applies QFSVM to handwritten character recognition and demonstrates that QFSVM can be run on quantum computers, and achieve accurate classification of handwritten characters. When compared to FSVM, QFSVM’s computational complexity decreases exponentially with the number of training samples.

Quantum support vector machine without iteration

Quantum support vector machine based on regularized Newton method

Quantum-Inspired Support Vector Machine

A Quantum Extension of SVM-perf for Training Nonlinear SVMs in Almost Linear Time

Adiabatic Quantum Support Vector Machines

Enhancing Quantum Support Vector Machines through Variational Kernel Training

Quantum kernel estimation-based quantum support vector regression

Towards Efficient Quantum Anomaly Detection: One-Class SVMs using Variable Subsampling and Randomized Measurements

Variational Quantum Linear Solver enhanced Quantum Support Vector Machine

Variational quantum approximate support vector machine with inference transfer

The complexity of quantum support vector machines

A quantum-enhanced support vector machine for galaxy classification

Experimental Realization of a Quantum Support Vector Machine.

Variational Quantum-Neural Hybrid Eigensolver.

Quantum support vector machines for classification and regression on a trapped-ion quantum computer

Quantum Fuzzy Support Vector Machine for Binary Classification

A Single-Step Multiclass SVM based on Quantum Annealing for Remote Sensing Data Classification

Quantum Algorithms for Feedforward Neural Networks

Machine Learning in the Quantum Age: Quantum vs. Classical Support Vector Machines

Quantum machine learning for multiclass classification beyond kernel methods