Quantum Key Distribution Privacy Amplification With Frequency-Space Processing

Resolve Bottlenecks,
Find Innovative Solutions
Generate Solutions

Solution Overview

Problem

Conventional quantum key distribution systems face inefficiencies in privacy amplification processing due to high data transfer requirements between external and local memory, leading to bottlenecks that hinder processing performance.

Innovation Solution

The quantum key distribution apparatus employs transform processing to convert data into a frequency space, allowing element-wise multiplication and inverse transform operations to be performed without accessing external memory, thereby reducing data transfer between memory types.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Loss of information

If conventional privacy amplification processing is used, then encryption key generation is achieved, but data transfer volume between external and local memory becomes excessively large

Engineering Contradiction:
Improvedata transfer volumeVSAvoidprocessing performance
Core Design Contradiction:
Loss of informationVSProductivity

Solution Approach 1:

The patent segments the privacy amplification processing into distinct phases: transform processing phase (where data is converted to frequency domain and processed locally), and inverse transform phase. By segmenting the computation, the system performs element-wise multiplication in the frequency domain using local memory, avoiding repeated external memory accesses and reducing overall data transfer volume.

Inventive Principle:
Principle #1Segmentation

Solution Approach 2:

The patent replaces the conventional time-domain multiplication mechanism with a frequency-domain multiplication mechanism. By applying transform processing (FFT) to convert input vectors and hash function vectors into frequency domain representations, the system substitutes complex time-domain convolutions with simpler element-wise multiplications, reducing computational complexity and memory access requirements.

Inventive Principle:
Principle #28Mechanics substitution (Replace mechanical system)

2Loss of information

If transform processing is applied to perform element-wise multiplication in frequency space, then data transfer volume is reduced, but processing complexity increases

Engineering Contradiction:
Improvedata transfer volumeVSAvoidprocessing complexity
Core Design Contradiction:
Loss of informationVSDevice complexity

Solution Approach 1:

The patent applies preliminary transform processing (FFT) to convert input vectors and hash function vectors into frequency domain representations before performing multiplication. This preliminary action transforms the data structure into a form where element-wise multiplication can be efficiently performed in local memory, avoiding the need for repeated external memory accesses during the computation phase.

Inventive Principle:
Principle #10Preliminary action

Solution Approach 2:

The patent changes the domain parameter of the data from time domain to frequency domain through transform processing. This parameter change enables the system to perform computations using local memory representations of the frequency domain data, fundamentally altering how the multiplication operation is executed and reducing dependency on external memory bandwidth.

Inventive Principle:
Principle #35Parameter changes

Data Source

PatentEP4675970A1Quantum key distribution apparatus, quantum key distribution method, and computer program product
Publication Date: 2026.01.07 KK TOSHIBA
  • EP4675970A1 patent drawingFigure 1~2
  • EP4675970A1 patent drawingFigure 3
  • EP4675970A1 patent drawingFigure 4A

AI summary

A quantum key distribution apparatus includes a processor implemented by at least one processing device. The processor generates input-vector-transformed data and random-number-transformed data by performing transform processing of transforming an input vector based on a photon received through a quantum channel and random number data based on a hash function into data in a frequency space. The processor stores the input-vector-transformed data and the random-number-transformed data in an external memory. The processor reads a portion of the input-vector-transformed data and a portion of the random-number-transformed data from the external memory into a local memory. The processor performs an element-wise multiplication-operation of the portion of the input-vector-transformed data and the portion of the random-number-transformed data and inverse-transform processing of an operation result of the element-wise multiplication-operation without accessing the external memory, and writes an intermediate result of the inverse-transform-processing to the external memory.