Experimental progress of quantum machine learning based on spin systems
At a Glance
Section titled âAt a Glanceâ| Metadata | Details |
|---|---|
| Publication Date | 2021-01-01 |
| Journal | Acta Physica Sinica |
| Authors | Yu Tian, Zidong Lin, Xiangyu Wang, Liangyu Che, Dawei Lu |
| Institutions | Southern University of Science and Technology |
| Citations | 1 |
| Analysis | Full AI Review Included |
Executive Summary
Section titled âExecutive SummaryâThis paper reviews the experimental progress of Quantum Machine Learning (QML) algorithms implemented on two primary spin-based quantum platforms: Nuclear Magnetic Resonance (NMR) systems and Nitrogen-Vacancy (NV) centers in diamond.
- Core Value Proposition: QML offers the potential for exponential speedup (quantum speedup) over classical algorithms for specific computational bottlenecks, particularly in linear algebra and data analysis (e.g., HHL algorithm, qPCA).
- Platform Validation: Both NMR (liquid-state multi-qubit systems) and NV centers (solid-state, room-temperature systems) have been successfully used as experimental platforms to validate foundational QML algorithms.
- Key Algorithm Demonstrations: Experiments successfully demonstrated the Harrow-Hassidim-Lloyd (HHL) algorithm for solving linear equations, Quantum Support Vector Machines (QSVM) for classification, and Quantum Principal Component Analysis (qPCA) for data dimensionality reduction.
- Coherence Enhancement Achievement: A Quantum Autoencoder (QAE) implemented on an NV center system achieved a greater than 1000x increase in entanglement lifetime, boosting the Bell state coherence from 2 ”s to 3 ms by encoding information into the long-coherence nuclear spin.
- Hybrid Optimization: The research validates the use of Hybrid Classical-Quantum Algorithms (HQCA) for optimizing Parameterized Quantum Circuits (PQC), which is crucial for training QML models on current noisy intermediate-scale quantum (NISQ) devices.
- Topological Classification: QML, specifically a 3D Convolutional Neural Network (CNN), was demonstrated on NV centers to classify topological phases with high accuracy (~98%), showcasing QMLâs utility in fundamental physics research.
Technical Specifications
Section titled âTechnical Specificationsâ| Parameter | Value | Unit | Context |
|---|---|---|---|
| NV Center Ground State Splitting (Dgs) | 2.87 | GHz | Zero-field splitting of the 3A2 state |
| 14N Quadrupole Splitting (PN) | -4.95 | MHz | 14N nuclear spin in NV center |
| NV Electron Gyromagnetic Ratio (Îłe) | 2.082 | MHz/G | Electron Zeeman splitting |
| 14N Nuclear Gyromagnetic Ratio (ÎłN) | -0.308 | kHz/G | 14N nuclear Zeeman splitting |
| NV Center Initialization/Excitation | 532 | nm | Standard green laser wavelength |
| NV Center Zero Phonon Line (ZPL) | 637 | nm | Optical transition wavelength |
| NV Center Fluorescence Readout Range | 637 - 750 | nm | Wavelength range for spin-dependent readout |
| QAE Entanglement Lifetime (Protected) | 3 | ms | Achieved using Quantum Autoencoder |
| QAE Entanglement Lifetime (Unprotected) | 2 | ”s | Baseline Bell state lifetime |
| QSVM Recognition Accuracy (NMR) | >96 | % | 4-qubit system, handwritten digit recognition |
| qPCA Feature Extraction Fidelity (NMR) | 99 | % | 4-qubit system, human face recognition |
| CNN Topological Classification Accuracy | ~98 | % | NV center system, using 10x10x10 density matrix input |
| HHL Algorithm Qubit Count (NMR) | 4 | bits | Demonstrated on 2x2 linear system |
Key Methodologies
Section titled âKey MethodologiesâThe experimental work relies on precise quantum control and hybrid classical-quantum optimization techniques across two distinct spin platforms:
-
Nuclear Magnetic Resonance (NMR) Systems (Liquid State):
- Qubit Definition: Nuclear spins (e.g., 13C, 19F) within complex molecules (e.g., Deuterated Chloroform, Crotonic Acid) are used as qubits.
- System Hamiltonian: The system is governed by Zeeman terms (single-qubit energy) and indirect scalar coupling (J-coupling, two-qubit interaction), as direct dipole-dipole coupling is averaged out in liquid samples.
- Initialization: Achieved by preparing the system into a pseudo-pure state (PPS) using techniques like spatial or temporal averaging, necessary because NMR systems typically operate near room temperature.
- Control and Gates: Universal quantum logic gates (e.g., CNOT, single-qubit rotations) are implemented using precisely shaped Radio Frequency (RF) pulses, often optimized via pulse sequence compilation or optimal control methods.
- Readout: Quantum state information is extracted by reconstructing the state from the Free Induction Decay (FID) signal.
-
Nitrogen-Vacancy (NV) Centers in Diamond (Solid State):
- Qubit Definition: The electron spin (S=1) and nearby nuclear spins (e.g., 14N, 13C) are used as qubits. The electron spin is typically used for fast operations and readout, while nuclear spins provide long coherence times.
- Initialization and Readout: Electron spin is initialized and read out optically using a 532 nm green laser and measuring spin-dependent fluorescence (637-750 nm).
- Control: Electron spin transitions (e.g., between ms = 0 and ms = ±1) are controlled using resonant microwave (MW) fields. Nuclear spins are controlled using RF pulses.
- QML Optimization (HQCA): Parameterized Quantum Circuits (PQC) are optimized using a Hybrid Classical-Quantum Approach (HQCA). The quantum processor measures the cost function and its gradient, while a classical computer handles parameter storage, gradient descent, and update steps.
- Entanglement Protection (QAE): The Quantum Autoencoder encodes fragile entanglement from the electron spin to the robust, long-coherence nuclear spin subspace, effectively protecting the quantum information from environmental decoherence.
Commercial Applications
Section titled âCommercial ApplicationsâThe demonstrated QML capabilities on spin systems are foundational for several high-value engineering and commercial sectors:
- Quantum Computing Hardware: NV diamond and NMR systems serve as viable platforms for building specialized quantum processors. NV centers, in particular, are promising for scalable, solid-state quantum architectures operating at room temperature.
- Quantum Sensing and Metrology: NV centers are leading candidates for high-resolution magnetometry and sensing. QML techniques (like qPCA) can be applied to efficiently process and analyze complex sensor data, improving the speed and accuracy of solid-state quantum sensors.
- Advanced Data Analytics: The demonstrated speedup for linear algebra (HHL) and dimensionality reduction (qPCA) is critical for processing massive datasets in finance, logistics, and scientific simulation, potentially accelerating tasks currently limited by classical GPU/CPU power.
- Quantum Memory and Communication: The Quantum Autoencoder technique, which achieved a 1000x increase in entanglement lifetime, is directly applicable to developing robust quantum memory solutions and enhancing the fidelity of quantum communication channels.
- Material Science (Diamond): The successful implementation of NV-based QML requires ultra-high purity, low-strain diamond substrates with controlled NV concentration. This drives demand for advanced diamond growth techniques (like those provided by 6ccvd.com) to meet the stringent material specifications for quantum technology.
- Drug Discovery and Chemistry: NMR-based QML can accelerate complex molecular simulations and quantum chemistry calculations, aiding in the design and optimization of new materials and pharmaceuticals.
View Original Abstract
Machine learning is widely applied in various areas due to its advantages in pattern recognition, but it is severely restricted by the computing power of classic computers. In recent years, with the rapid development of quantum technology, quantum machine learning has been verified experimentally verified in many quantum systems, and exhibited great advantages over classical algorithms for certain specific problems. In the present review, we mainly introduce two typical spin systems, nuclear magnetic resonance and nitrogen-vacancy centers in diamond, and review some representative experiments in the field of quantum machine learning, which were carried out in recent years.