Quantenberechnungsverfahren und Vorrichtung mit Spärlicher Codierung

Details

Veröffentlichungs-Nr.

EP4726612

Anmeldetag

9. Oktober 2024

Veröffentlichung

15. April 2026

Rechtsraum

EP

IPC

G06N10/60, // G06N10/20

Abstract

The invention relates to a method and a device for performing a quantum computation, which includes a sparse encoding, wherein, for a number N of particles in a physical system, and a size of a basis set, M∈ℕ∪∞, a set B of M single-particle functions f_j, with each function f_j being a function from a one particle space to , defined as B := {f₁, f₂, ... , f_M}, and two monotonously increasing mappings k and r are defined, which assign a real positive number to each single-particle function f_j, wherein, for a parameter K∈ℝ+ and a parameter R∈ℝ+, a reduced tensor product basis set SK,RT,pq,pr of the single-particle functions f_j is selected as SK,RT,pq,pr:=⊗i=1Nfji:11−T∑i=1Nkfjipq−T/pq∏i=1Nkfji≤Kand∑i=1Nrfjipr1pr≤Rwherej1,…,jN⊆1,…,M, wherein T∈ℝ, with T < 1, defines a reduction in a number of basis functions, p_q and p_r are parameters of selected p_q-norms and p_r-norms and K and R are selected based on an error behavior of the reduced tensor product basis set SK,RT,pq,pr. In particular, this definition includes also special cases, where the mapping k is constant and equal to one and hence can be neglected. Considering as an example the calculation of the FeMoco cofactor of the Nitrogenase enzyme, which involves N = 54 electrons, the number of sufficient qubits can be reduced from 1080 to 256 for plane wave like single-particle basis functions and from 378 to 157 for Gaussian like single-particle basis functions.

Anmelder

Firmen

Fraunhofer-Gesellschaft zur Förderung der angewandten Forschung e.V.
Rheinische Friedrich-Wilhelms-Universität Bonn, Körperschaft des öffentlichen Rechts

Land

🇩🇪 Deutschland

Fachgebiete