Classical Ansatz Training for Quantum State Purification
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Noisy quantum computers produce mixed states due to noise, which are not suitable for quantum algorithms designed to operate on pure states, leading to inaccuracies in final answers encoded in pure states or observables of such states.
Innovation Solution
A computer-implemented method that uses a classical model with parameters to represent a pure quantum state by selecting measurement pairs from a quantum circuit, updating parameters to minimize distance measures, and producing an updated classical model that approximates the intended pure state.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If quantum state purification is performed using full state tomography, then measurement precision is improved, but device complexity and measurement time increase exponentially
Solution Approach 1:
The patent segments the quantum state representation into a compressed format using classical variables that capture only the essential features of the quantum state. Instead of measuring and representing the entire quantum state vector, the method divides the problem into manageable components that can be processed classically, thereby reducing measurement complexity while maintaining purification accuracy.
Solution Approach 2:
The patent changes the parameters used to represent the quantum state from the full state vector to a compressed set of classical variables. This parameter transformation allows the system to work with a reduced representation that is computationally tractable while still sufficient for calculating quantum advantages, effectively resolving the contradiction between precision and complexity.
2Measurement precision
If more measurement pairs are sampled from the quantum circuit, then the representation accuracy of the pure state is improved, but the loss of time increases
Solution Approach 1:
The patent applies partial action by sampling only a sufficient subset of measurement pairs rather than exhaustively measuring all possible outcomes. The compressed representation format allows the system to capture the essential quantum state information with fewer measurements than would be required for full state tomography, thereby reducing measurement time while maintaining adequate representation accuracy.
Solution Approach 2:
The patent creates a classical copy or representation of the quantum state using compressed variables. This classical copy captures the necessary statistical properties of the quantum state without requiring complete quantum state tomography, enabling accurate computation of quantum advantages with reduced measurement overhead and time investment.
3Device complexity
If a classical model is used to represent the quantum state, then device complexity is reduced, but measurement precision may deteriorate due to noise
Solution Approach 1:
The patent introduces a compressed representation format as an intermediary between the noisy quantum measurements and the classical computation. This intermediary layer processes the measurement data to extract the essential quantum state properties while filtering out noise, allowing the classical model to achieve both low complexity and high fidelity by working with the compressed, denoised representation rather than raw measurement data.
Data Source
AI summary
A computer-implemented method produces a representation of a pure quantum state from a classical model. The classical model has a plurality of parameters. The method includes: (A) selecting a set of outcomes from a library of outcomes of a quantum circuit, wherein the library of outcomes comprises a plurality of measurement pairs sampled from the quantum circuit, each measurement pair comprising a quantum measurement and a corresponding measurement basis; and (B) updating values of the plurality of parameters of the classical model to minimize a value of a distance measure between the classical model and the set of outcomes, thereby producing the updated classical model, wherein the updated classical model has the updated values of the plurality of parameters.


