Abstract:We consider the quantum implementations of the two classical iterative solvers for a system of linear equations, including the Kaczmarz method which uses a row of coefficient matrix in each iteration step, and the coordinate descent method which utilizes a column instead. These two methods are widely applied in big data science due to their very simple iteration schemes. In this paper we use the block-encoding technique and propose fast quantum implementations for these two approaches, under the assumption that the quantum states of each row or each column can be efficiently prepared. The quantum algorithms achieve exponential speed up at the problem size over the classical versions, meanwhile their complexity is nearly linear at the number of steps.

What problem does this paper attempt to address?

The problem that this paper attempts to solve is to implement the Kaczmarz method and the coordinate descent method in classical iterative methods on a quantum computer to solve the linear equation system \(Ax = b\). Specifically, the goals of the paper are: 1. **Propose fast quantum algorithms**: Using block - encoding technology, the paper proposes fast quantum implementations of the Kaczmarz method and the coordinate descent method. These quantum algorithms achieve exponential acceleration in the problem size \(n\), while maintaining an approximately linear complexity in the number of iterations \(T\). 2. **Overcome the limitations of classical algorithms**: Although the classical Kaczmarz method and the coordinate descent method are simple and efficient, there are still problems of relatively high computational complexity in large - scale data processing. Through the method of quantum computing, the paper aims to significantly reduce the computational complexity of these algorithms, making them more practical in big data science. 3. **Ensure the feasibility of the algorithms**: The paper assumes that the quantum states of each row or column can be efficiently prepared, and on this basis, constructs specific quantum algorithms. These algorithms are not only theoretically superior but also feasible in practical applications. ### Main contributions - **Application of block - encoding technology**: The paper uses block - encoding technology to implement the quantum versions of the Kaczmarz method and the coordinate descent method. Block - encoding technology allows matrix operations to be converted into unit operations in quantum circuits, thereby achieving efficient quantum computing. - **Exponential acceleration**: The quantum algorithms achieve exponential acceleration in the problem size \(n\), which means that for large - scale linear equation systems, the computing time of the quantum algorithms is much less than that of the classical algorithms. - **Approximately linear iteration complexity**: Although exponential acceleration is achieved in the problem size \(n\), the complexity of the quantum algorithms in the number of iterations \(T\) is still approximately linear, which makes the algorithms more efficient in practical applications. ### Specific implementation - **Quantum implementation of the Kaczmarz method**: By defining the unit operator \(U_t\) and the swap operator \(SWAP\), the paper describes in detail how to implement the iterative steps of the Kaczmarz method on a quantum computer. These steps include preparing the initial state, selecting a random row, updating the solution vector, etc. - **Quantum implementation of the coordinate descent method**: Similarly, the paper also describes how to implement the iterative steps of the coordinate descent method on a quantum computer. These steps include preparing the initial state, selecting a random column, updating the solution vector and the residual vector, etc. ### Conclusion The paper shows the implementation methods of the quantum versions of the Kaczmarz method and the coordinate descent method on a quantum computer, and proves that these quantum algorithms achieve exponential acceleration in the problem size \(n\), while maintaining an approximately linear complexity in the number of iterations \(T\). These results provide new ideas and methods for the application of quantum computing in large - scale data processing.

Randomized Row and Column Iterative Methods with a Quantum Computer

Quantum multi-row iteration algorithm for linear systems with non-square coefficient matrices

Fast and Practical Quantum-Inspired Classical Algorithms for Solving Linear Systems.

Experimental Quantum Computing to Solve Systems of Linear Equations

Qubit-Efficient Randomized Quantum Algorithms for Linear Algebra

Experimental Realization Of Quantum Algorithm For Solving Linear Systems Of Equations

Quantum Gram-Schmidt Processes and Their Application to Efficient State Readout for Quantum Algorithms

Quantum algorithms for matrix operations and linear systems of equations

Quantum speedup of leverage score sampling and its application

Iterative quantum algorithm for combinatorial optimization based on quantum gradient descent

Quantum Iterative Methods for Solving Differential Equations with Application to Computational Fluid Dynamics

An improved quantum-inspired algorithm for linear regression

Quantum Krylov-Subspace Method Based Linear Solver

Scalable Program Implementation and Simulation of the Large-Scale Quantum Algorithm: $1024\times 1024$ Quantum Linear Solver and Beyond

Hybrid quantum-classical and quantum-inspired classical algorithms for solving banded circulant linear systems

Real-Time Krylov Theory for Quantum Computing Algorithms

A near-term quantum algorithm for solving linear systems of equations based on the Woodbury identity

Quantum machine learning of large datasets using randomized measurements

Quantum Algorithm For Solving Nonlinear Algebraic Equations

Quantum Linear System Solver Based on Time-optimal Adiabatic Quantum Computing and Quantum Approximate Optimization Algorithm

Compiling basic linear algebra subroutines for quantum computers