Quantum Neural Network Classification via Parameterized Gates
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing methods for classification on quantum computing systems require specialized quantum versions of classical artificial neural networks and are limited to classical input states, lacking the ability to effectively classify quantum states.
Innovation Solution
A classification method using a quantum computing system that employs parameterized quantum gates to learn and classify input states, including both classical and quantum states, without requiring specialized quantum neural networks, and can be implemented on near-term quantum computing systems.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Adaptability or versatility
If specialized quantum versions of classical artificial neural networks are used for classification, then classification capability is improved, but device complexity and implementation difficulty increase
Solution Approach 1:
The patent replaces specialized quantum neural network architectures with a universal quantum computing approach using parameterized quantum circuits. Instead of designing complex quantum-specific neural networks, the invention uses standard quantum gates with learnable parameters that can be optimized through classical gradient descent, substituting complex quantum architecture design with a more manageable parameter optimization problem.
Solution Approach 2:
The patent creates a universal classification framework that can handle both classical and quantum input states using the same quantum circuit architecture. The parameterized quantum circuits serve multiple functions: they can process classical data encoded in quantum states, process native quantum states, and be trained using classical optimization methods, making the system universally applicable without requiring specialized quantum neural network designs.
2Ease of operation
If quantum computing systems are designed for near-term implementation, then ease of operation and accessibility are improved, but functional limitations and precision constraints worsen
Solution Approach 1:
The patent employs variational quantum circuits with a limited but sufficient number of parameterized gates that provide enough expressive power for classification tasks while remaining implementable on near-term quantum hardware with constrained qubit counts and gate fidelities. The approach uses partial quantum processing combined with classical optimization to achieve useful classification precision without requiring full fault-tolerant quantum computing capabilities.
Solution Approach 2:
The invention optimizes classification precision by adjusting and learning the parameters of quantum gates through classical gradient descent. This parameter optimization approach allows the system to achieve high classification accuracy by fine-tuning gate parameters without requiring additional quantum hardware resources or higher physical precision, making it suitable for near-term quantum processors.
3Ease of manufacture
If classical machine learning methods are used, then ease of implementation is improved, but the ability to classify quantum states and exploit quantum advantages is lost
Solution Approach 1:
The patent introduces quantum encoding as an intermediary step that transforms classical data into quantum states, allowing classical machine learning optimization methods to work with quantum representations. This intermediary quantum encoding layer enables the system to process both classical and quantum inputs while maintaining compatibility with classical optimization algorithms, bridging the gap between classical ease of implementation and quantum state classification capability.
Applied Scientific Principles
This section explains which scientific principles are used to turn an abstract innovation direction into a practical engineering solution.
Function Achieved in This Case
Enables efficient classification of both classical and quantum states, facilitating applications in quantum metrology and other fields, while being suitable for near-term quantum processors.
Implementation Method 1
Quantum computers are computing devices that exploit quantum superposition and entanglement to solve certain types of problem faster than a classical computer.
Implementation Method 2
Quantum computers are computing devices that exploit quantum superposition and entanglement to solve certain types of problem faster than a classical computer.
Data Source
Figure 1
Figure 2
Figure 3
AI summary
This disclosure relates to classification methods that can be implemented on quantum computing systems. According to a first aspect, this specification describes a method for training a classifier implemented on a quantum computer, the method comprising: preparing a plurality of qubits in an input state with a known classification, said plurality of qubits comprising one or more readout qubits; applying one or more parameterised quantum gates to the plurality of qubits to transform the input state to an output state; determining, using a readout state of the one or more readout qubits in the output state, a predicted classification of the input state; comparing the predicted classification with the known classification; and updating one or more parameters of the parameterised quantum gates in dependence on the comparison of the predicted classification with the known classification.