Quantum Arithmetic-based on Quantum Signal Processing

Robin OLLIVE, Stephane LOUISE · 2025

Read Quantum Arithmetic-based on Quantum Signal Processing by Robin OLLIVE, Stephane LOUISE (2025) in the Tower of Babel, a free online library of public-domain works preserved in clean, searchable text. No sign-up. The full text is served as a plain-text file at a predictable URL, where text is present: https://antarctic-labs.com/tower-of-babel/arxiv-2503-18761.txt.

Introduction

Robin Ollive and Stephane Louise reframe quantum arithmetic in 2025. Quantum arithmetic is the set of tools that compute with numbers stored in quantum registers. The classic view treats it as circuits that write results into fresh registers. The authors instead view it through matrices and signal processing.

Their lens treats arithmetic as building the query that encodes the problem. A link between state size and operator phase lets signal methods yield binary values. Those values are then read out by a phase estimation step. The new route may give builders a fresh way to design quantum math blocks.

Catalog record

CREATORRobin OLLIVE, Stephane LOUISE
YEAR2025
COLLECTIONPAPERS
FORMATTXT
FILE SIZE63 KB
VERSION1.0
LICENSEhttp://creativecommons.org/licenses/by/4.0/
RIGHTSCC BY
SOURCEarXiv

Source: https://arxiv.org/abs/2503-18761

Browse the full library · About the Tower of Babel

Full text

The opening of the work, from the library's hosted text:

Quantum Arithmetic-based on Quantum Signal Processing Robin OLLIVE, Stephane LOUISE Abstract As in classical reversible computing, Quantum Arithmetic is typically seen as a set of tools that process binary data encoded into a quantum register to set the value of another quantum register. This article presents another approach to explain the Quantum Arithmetic in quantum computing. Here, Quantum Arithmetic is addressed with a matrix processing point of view. Quantum Arithmetic is a convenient way to construct the Query that implements the mathematical problem of interest. This approach is not only an interpretation with the matrix approach to encode the problem of interest; it also allows us to derive a new original technique to construct Quantum Arithmetic with the framework of embedded Quantum Signal Processing. This work uses the link between the eigenstate amplitude and the operator phase to transform the Quantum Signal Processing processed amplitude into binary value extracted by Quantum Phase Estimation. The explanations allowing the Quantum Signal Processing based Quantum Arithmetic construction let appear natively sub-routines and functions used by well-known algorithms such as Ancilla Quantum Encoding’s Harrow–Hassidim–Lloyd algorithm and Quantum Amplitude Estimation. Methods to implement this circuit are presented in the paper. \standaloneconfig mode=image 1 Introduction Many quantum algorithms need Quantum Arithmetic as a subroutine. Shor’s algorithm [ 1 ] , certain Grover Oracles [ 2 ] , and one of the Harrow–Hassidim–Lloyd algorithm [ 3 ] sub-routines rely on Quantum Arithmetic [ 4 , Table.11] . The ability to compute functions of binary numbers is required to implement or process many classical problems in a quantum computer. Many quantum algorithms process a superposition of binary numbers, which allows the associated problem to be solved efficiently. Quantum Arithmetic is often constructed with two strategies. These strategies are based on either the binary logic bitwise adders [ 5 ] assembly [ 6 , 7 ] or the Quantum Fourier Transphorm phase adders [ 8 ] assembly [ 9 , 10 , 11 , 12 ] 1 1 1 A recent approach proposes to construct a more efficient Quantum Fourier Transphorm-based multiplier directly [ 13 ] ; nevertheless, a deep analysis of the sub-routines seems to cancel this advantage [ 14 ] . . Both these strategies combine these simple blocks to elaborate complex functions. It can lead to polynomial approximations of arbitrary functions [ 15 ] . Many developments to improve, categorize and evaluate them exist in the literature [ 16 , 4 ] . This part of the algorithm has an extensive cost of quantum resources due to the high number of required ancilla qubits. These ancilla qubits are needed to store the intermediate results [ 15 ] . The trade-off between the number of ancilla qubits and the number of operations is one of the key aspects in designing a quantum circuit for Quantum Arithmetic. Recent developments in quantum algorithms allow to interpret the implemented problem as qubitized matrices and bring a routine, the Quantum Signal Processing [ 17 , 18 , 19 , 20 , 21 ] , that competes with Harrow–Hassidim–Lloyd algorithm to process the matrices. The Quantum Signal Processing and its generalization, the Quantum Singular Value Transformation [ 22 ] routine allow the processing of real part of matrix amplitudes. Quantum Signal Processing does polynomial functions of the amplitudes with the polynomial order equal to the number of matrix repetitions [ 21 ] . Composing layers of Quantum Signal Processing and therefore function composition is possible using the more restrictive embedded Quantum Signal Processing or Quantum Singular Value Transformation [ 23 ] . It gives access to more complex functions. Using these tools in order to construct Quantum Arithmetic with a smooth function process thanks to Quantum Signal Processing is the core idea of this paper. The idea of avoiding Quantum Arithmetic to construct a function was already developed in [ 24 ] to prepare a quantum state. This article proposes two contributions. The first contribution is to show the links between Quantum Arithmetic and the other techniques of problem implementation. It explains the Quantum Fourier Transphorm-based adder structure directly. This part shows why it implies that projector constructs from Quantum Arithmetic are a particular Harrow–Hassidim–Lloyd algorithm instance. The second contribution is an original strategy to construct the Quantum Arithmetic routines. This strategy is not based on a combination of simple operations. Instead, thanks to well-tuned Quantum Signal Processing techniques, the function of the binary register is computed in the amplitudes of the ancilla qubit. Depending on the reason why the Quantum Arithmetic is constructed, this amplitude can be read into a third register thanks to a Quantum Phase Estimation variant, or directly transformed into a Grover oracle, or state initialization. We give an example with the complete set of parameters that allow the implementation. Depending on the function of the binary register, this method can be more efficient, especially regarding the number of ancilla qubits. 2 Quantum Arithmetic based query To solve problems with quantum algorithms, the first thing that must be checked is whether or not the classical problem can be formulated with quantum information. If so, then this quantum information must be implemented as a quantum unitary gate, the query, which can be processed by the Quantum Processing Unit. Two very popular queries are Hamiltonian Simulation eq. Equation 1 and Block-Encoding 2 2 2 Different types of Block-Encoding exist. As soon as the matrix of interest is probabilistically applied to the vector of interest with a control register, the gate is a Block-Encoding. The example of this paper is a specific case called signal operator in the Y -convention due to his behavior when called in a Quantum Signal Processing algorithm. eq. Equation 2 . A third one, that encodes the same quantum information is the results from a Quantum Phase Estimation algorithm eq. Equation 3 . It is often more complicated to construct it. The Quantum Phase Estimation result is often used to read the information because it contains the binary representation of the matrix eigen-phases \lambda_{i} with the associated eigenstate \ket{\lambda_{i}} , which can easily be read. \widehat{V}_{H}=e^{i2\pi\frac{1}{\lambda_{max}}\widehat{H}} (1) \widehat{W_{y}}_{H}=\begin{bmatrix}\widehat{H}&-\sqrt{\widehat{I}-\widehat{H}^{2}}\\ \sqrt{\widehat{I}-\widehat{H}^{2}}&\widehat{H}\end{bmatrix}\begin{matrix}\ket{\psi}\ket{\psi_{\parallel}}\\ \ket{\psi}\ket{\psi_{\perp}}\end{matrix} (2) with \ket{\psi_{\parallel}} and \ket{\psi_{\perp}} some control states. \widehat{QPE}_{H}=\sum_{i}\ket{{\lambda_{i}}}\bra{{\lambda_{i}}}\otimes\ket{\mathrm{bin}[(b+\lambda_{i})\mathrm{mod}[2^{F}]]}\bra{\mathrm{bin}[b]} (3) Since all these queries contain the same information, switching from query to query with another query assembly Fig. 1 is possible. Figure 1 : Conversion between the different Query used to process information into Quantum Processing Unit. The dotted and dashed lines represent the classical data implementation, while the filled lines represent data processing on the Quantum Processing Unit. Another way to load the problem’s data into the Quantum Processing Unit, the Quantum Arithmetic Equation 4 , can be used for binary information: \displaystyle\widehat{U}_{\mathrm{f}}=\sum_{x=0}^{2^{F}}\ket{\mathrm{bin}[x]}\bra{\mathrm{bin}[x]}\otimes\ket{\mathrm{bin}[(b+\mathrm{f}(x))\mathrm{mod}[2^{F}]]}\bra{\mathrm{bin}[b]} (4) \displaystyle\widehat{U}_{\mathrm{f}}\ket{\mathrm{bin}[x]}\ket{0}=\ket{\mathrm{bin}[x]}\ket{\mathrm{bin}[\mathrm{f}(x)\mathrm{mod}[2^{F}]]} While Hamiltonian Simulation and Block-Encoding are based on a decomposition of the matrix formulation of the problem on a family of matrices accessible by the quantum algorithm ( Fig. 1 , dashed lines), Quantum Arithmetic is based on the assembly of basic logical operations. Using the Toffolly gate and X gate, the universal gate of reversible classical logic, all the logical operations can be implemented similarly to the classical logic circuit Table 1 [ 6 , Section 3] . Table 1 : Classical logic gates with a quantum computer. The

Read the complete text