Systems and methods for enhanced eigenvalue inversion using quantum conditional logic

Patent metadata

Field	Value
Publication number	US-12242562-B2
Application number	US-202117331445-A
Country	US
Kind code	B2
Filing date	May 26, 2021
Priority date	May 26, 2021
Publication date	Mar 4, 2025
Grant date	Mar 4, 2025

Abstract

Embodiments use quantum conditional logic in the Quantum Phase Estimation Algorithm (QPEA) to compute eigenvalues prior to inversion. Embodiments estimate the eigenvalues of a unitary, U=e iÂt , generated by a N×N Hermitian matrix Â. The binary representations of the n-bit estimations of eigenvalues of  may be encoded in these states: |λ i =|b 1 b 2 . . . b n ; λ i is an estimation of the i-th eigenvalue, excluding degeneracy, and .b 1 b 2 . . . b n is its binary representation. To perform the eigenvalue inversion, an n-qubit controlled Ry rotation with angle λ i /2 (n−1) conditioned on seeing |b 1 b 2 . . . b n is applied for each possible n-bit binary string b 1 b 2 . . . b n (2 n values). The overall unitary is called a “uniformly controlled Ry rotation” in literature.

First claim

What is claimed is:

1. A classical computer, comprising:
a computer processor; and
a memory storing a classical computer program;
wherein the classical computer program is configured to generate a quantum computer program comprising instructions to:
initialize a register to an initial state;
initialize an ancilla qubit and apply a Hadamard gate to the ancilla qubit;
determine a number of iterations n to estimate n-bit eigenvalues;
perform a first unitary gate raised to 2 n−1 and controlled by the ancilla qubit;
perform a Hadamard gate on the ancilla qubit;
measure the ancilla qubit and save in a classical register C 0 ;
clear the ancilla qubit;
initialize a counter i to n−2;
perform a second unitary gate raised to 2 i and controlled by the ancilla qubit;
perform a phase gate on the ancilla qubit conditioned on a value in a classical register C (n−1)−i−1 and apply a Hadamard gate on the ancilla qubit;
measure the ancilla qubit and save in a classical register C (n−1)−i ;
clear the ancilla qubit; and
decrease i by 1 and repeating the performing of the second unitary gate, the performing of the phase gate, the applying of the Hadamard gate,

the measuring of the ancilla qubit, the saving of the ancilla qubit, and the decreasing of i by 1 until i is equal to zero;
wherein the classical computer program sends the quantum computer program to a quantum computer, the quantum computer executes the quantum computer program, receives results from the quantum computer in the classical registers C 0 , C 1 , . . . C n−1 , and outputs the results.

2. The classical computer of claim 1 , wherein the results from the quantum computer in the classical registers C 0 , C 1 , . . . C n−1 are binaries of the n-bit eigenvalues.

3. A method for n-bit eigenvalue estimation, comprising:
receiving, by a quantum computer, a quantum computer program from a classical computer;
initializing, by the quantum computer program, a register on the quantum computer to an initial state;
initializing, by a quantum computer program, an ancilla qubit and apply a Hadamard gate to the ancilla qubit;
determining, by a quantum computer program, a number of iterations n to estimate n-bit eigenvalues;
performing, by a quantum computer program, a first unitary gate raised to 2 n−1 and controlled by the ancilla qubit;
performing, by a quantum computer program, a Hadamard gate on the ancilla qubit;
measuring, by a quantum computer program, the ancilla qubit and save in a classical register C 0 ;
clearing, by a quantum computer program, the ancilla qubit;
initializing, by a quantum computer program, a counter i to n−2;
performing, by a quantum computer program, a second unitary gate raised to 2 i and controlled by the ancilla qubit;
performing, by a quantum computer program, a phase gate on the ancilla qubit conditioned on a value in a classical register C (n−1)−i−1 and apply a Hadamard gate on the ancilla qubit;
measuring the ancilla qubit and save in a classical register C (n−1)−i ;
clearing the ancilla qubit;
decreasing i by 1 and repeating the performing of the second unitary gate, the performing of the phase gate, the applying of the Hadamard gate, the measuring of the ancilla qubit, the saving of the ancilla qubit, and the decreasing of i by 1 until i is equal to zero; and
outputting the classical registers C 0 , C 1 , . . . C n−1 to the classical computer, wherein the classical registers C 0 , C 1 , . . . C n−1 comprise the n-bit eigenvalues.

4. A system, comprising:
a classical computer comprising a memory storing a classical computer program, wherein the classical computer program is configured to generate a quantum computer program comprising instructions to:
initialize a register in the quantum computer to an initial state;
initialize an ancilla qubit and apply a Hadamard gate to the ancilla qubit;
determine a number of iterations n to estimate n-bit eigenvalues;
perform a first unitary gate raised to 2 n−1 and controlled by the ancilla qubit;
perform a Hadamard gate on the ancilla qubit;
measure the ancilla qubit and save in a classical register C 0 ;
clear the ancilla qubit;
initialize a counter i to n−2;
perform a second unitary gate raised to 2 i and controlled by the ancilla qubit;
perform a phase gate on the ancilla qubit conditioned on a value in a classical register C (n−1)−i−1 and apply a Hadamard gate on the ancilla qubit;
measure the ancilla qubit and save in a classical register C (n−1)−i ;
clear the ancilla qubit; and
decrease i by 1 and repeating the performing of the second unitary gate, the performing of the phase gate, the applying of the Hadamard gate, the measuring of the ancilla qubit, the saving of the ancilla qubit, and the decreasing of i by 1 until i is equal to zero;

a quantum computer that is configured to receive the quantum computer program from the classical computer, to execute the quantum computer program, and to output the classical registers C 0 , C 1 , . . . C n−1 to the classical computer program, wherein the classical registers C 0 , C 1 , . . . C n−1 comprise the n-bit eigenvalues.

Assignees

• Jpmorgan Chase Bank Na

Inventors

• Yalovetzky Romina
• Herman Dylan
• Minssen Pierre
• Pistoia Marco
• Buts Alexander
• Hu Shaohan

Classifications

• G06N10/00

  Quantum computing, i.e. information processing based on quantum-mechanical phenomena · CPC title

• G06F17/16

  Matrix or vector computation {, e.g. matrix-matrix or matrix-vector multiplication, matrix factorization (matrix transposition G06F7/78)} · CPC title

• G06N10/60

  Quantum algorithms, e.g. based on quantum optimisation, quantum Fourier or Hadamard transforms · CPC title

• G06F17/12 Primary

  Simultaneous equations {, e.g. systems of linear equations} · CPC title

• G06N10/20 Primary

  Models of quantum computing, e.g. quantum circuits or universal quantum computers · CPC title