Matrix Product State-Based Quantum Classifier.

Area of Science:

• Quantum Computing
• Machine Learning
• Tensor Network Theory

Background:

• Tensor network theory, particularly matrix product states (MPS), is crucial for simulating complex quantum systems.
• MPS are effective in quantum information processing and have emerged as a powerful tool in quantum machine learning.

Purpose of the Study:

• To demonstrate the efficacy of matrix product states (MPS) for classifying both classical and quantum data.
• To evaluate the performance of MPS-based quantum classifiers on real-world datasets using quantum hardware.

Main Methods:

• Encoding classical data (Iris dataset) into quantum states for binary classification.
• Utilizing matrix product states (MPS) as a quantum circuit for data classification.
• Testing MPS classifier performance on a meteorological dataset (evapotranspiration) using historical data.

Main Results:

• MPS demonstrated successful binary classification of the Iris dataset encoded in a quantum state.
• MPS circuits achieved improved accuracy on the ibmqx4 quantum computer.
• The MPS quantum classifier showed learning ability when applied to meteorological data for evapotranspiration prediction.

Conclusions:

• Matrix product states (MPS) are a viable and effective tool for quantum data classification.
• MPS-based quantum classifiers show promise for real-world machine learning tasks, including scientific data analysis.
• The study validates the use of MPS for enhancing classification accuracy in quantum computing applications.