Random Quantum Circuit Noise Classification Across Many Qubits

Resolve Bottlenecks,
Find Innovative SolutionsGenerate Solutions

Solution Overview

Problem

Existing methods for characterizing noise in quantum systems lack scalability and cannot effectively distinguish between coherent and incoherent noise types, which is crucial for quantum computing tasks.

Innovation Solution

A method using random quantum circuits with alternating cycles of single-qubit rotations and entangling gates to generate distributions that are sensitive to coherent and incoherent noise, employing binning techniques based on ideal and experimental probabilities to quantify noise types.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Quantity of substance

If scalable methods for characterizing quantum systems are used, then the system can handle large numbers of qubits, but the ability to discriminate amongst noise types is lost

Engineering Contradiction:

Improvenumber of qubitsVSAvoidnoise type discrimination

Core Design Contradiction:

Quantity of substanceVSMeasurement precision

Solution Approach 1:

The patent segments the noise characterization problem into two distinct measurement approaches: (1) measuring output probabilities to detect incoherent noise, and (2) measuring output distributions to detect coherent noise. This segmentation allows the system to maintain noise type discrimination capability while scaling to many qubits, as each measurement type can be performed independently on the same quantum circuit outputs.

Inventive Principle:

Principle #1Segmentation

Solution Approach 2:

The patent transitions from single-dimension noise characterization to multi-dimensional analysis by examining both the magnitude (probabilities) and the distribution shape (relative probabilities) of quantum circuit outputs. This dimensional expansion enables simultaneous detection of different noise types without requiring separate characterization procedures for each noise type.

Inventive Principle:

Principle #17Another dimension (Dimensionality change)

2Measurement precision

If existing noise classification methods are used, then noise types can be distinguished, but the methods lack scalability beyond a few qubits

Engineering Contradiction:

Improvenoise type discriminationVSAvoidscalability

Core Design Contradiction:

Measurement precisionVSProductivity

Solution Approach 1:

The patent creates a universal noise characterization framework that works across different system sizes by using random quantum circuits as the common testbed. The same circuit execution generates data for both incoherent and coherent noise detection, making the method universally applicable from small to large quantum systems without requiring different characterization techniques for different scales.

Inventive Principle:

Principle #6Universality (Multi-functionality)

Solution Approach 2:

The patent changes the analysis parameters from examining individual quantum state properties to examining statistical properties of output distributions. By focusing on aggregate statistics (probability distributions across many circuit outputs) rather than individual qubit states, the method achieves scalability while maintaining noise discrimination capability.

Inventive Principle:

Principle #35Parameter changes

3Productivity

If random quantum circuits are used for noise classification, then scalability to 10-20+ qubits is achieved, but the complexity of analyzing output distributions increases

Engineering Contradiction:

ImprovescalabilityVSAvoidanalysis complexity

Core Design Contradiction:

ProductivityVSDevice complexity

Solution Approach 1:

The patent uses classical simulation to generate ideal (noiseless) output distributions that serve as reference copies for comparison with experimental data. This copying approach allows the complex quantum system behavior to be replicated classically at the distribution level, enabling scalable analysis without requiring full quantum state simulation which would be computationally intractable for large systems.

Inventive Principle:

Principle #26Copying

Solution Approach 2:

The patent replaces direct quantum state analysis with statistical distribution analysis. Instead of attempting to track and analyze individual quantum amplitudes and phases (the mechanical quantum system), the method substitutes this with classical statistical analysis of measurement outcome distributions, which is computationally tractable for large numbers of qubits.

Inventive Principle:

Principle #28Mechanics substitution (Replace mechanical system)

Data Source

PatentEP4014175B1Noise classification through the output of random quantum circuits

Publication Date: 2025.11.12 INTERNATIONAL BUSINESS MACHINE CORPORATION

Figure 1
Figure 2
Figure 3

AI summary

A method for characterizing noise in a quantum system, the quantum system including a plurality of qubits and a plurality of entangling gates native to the quantum system, includes generating a random quantum circuit on a quantum processor, the random quantum circuit comprising the plurality of entangling gates native to the quantum system. The method includes running a simulation of the random quantum circuit on a classical computer a plurality of times to obtain ideal outcomes, and running the random quantum circuit on the quantum processor a plurality of times to obtain experimental outcomes. The method includes grouping the experimental outcomes based on probabilities of the ideal outcomes to obtain a first distribution, and grouping the experimental outcomes based on probabilities of the experimental outcomes to obtain a second distribution. The method includes characterizing noise in the quantum system based on the first distribution and the second distribution.