Korean researchers introduce quantum classifiers in machine learning

By INDIAai

Highlights

The non-linear quantum kernels provide new insights for improving the accuracy of quantum machine learning.

Share this Article:

Korean researchers have proposed a quantum classifier based on quantum state fidelity by using a different initial state and replacing the Hadamard classification with a swap test. This new method is deemed to boost classification tasks in small datasets, by utilising the quantum advantage in finding non-linear features in a large feature space.

While machine learning recognises objectives based on its training dataset, which feeds labelled data which is classified, quantum machine learning, via the kernel method, can identify non-linear relationships in complex data. 

The kernel method has only been recently introduced in quantum machine learning. Quantum computers’ ability to proficiently access and manipulate data in the quantum feature space provides ample opportunities for quantum techniques to improve various existing machine learning methods. 

Classification algorithm with a non-linear kernel in a quantum test state will have the protocol calculate the weighted power sum of the fidelities of quantum data in quantum parallel via a swap-test circuit followed by two single-qubit measurements. This method needs a handful of quantum data operations, no matter how big or small the size of the data is. This approach allows large training data which is labelled to be packed densely into a quantum state and then compared to the test data.

The Korea Advanced Institute of Science and Technology (KAIST) have pioneered this approach in collaboration with the University of KwaZulu-Natal (UKZN) in South Africa and Data Cybernetics in Germany. The team has made advancements in the field of quantum machine learning by introducing quantum classifiers with tailored quantum kernels.

The input data is either constituted by classical data via a quantum feature map or intrinsic quantum data, and the classification is based on the kernel function that measures the closeness of the test data to training data.

Dr Daniel Park at KAIST, one of the lead authors of this research, said that the quantum kernel can be tailored systematically to an arbitrary power sum, which makes it an excellent candidate for real-world applications.

Professor Francesco Petruccione from UKZN explained, “The state fidelity of two quantum states includes the imaginary parts of the probability amplitudes, which enables the use of the full quantum feature space.” Korean researchers have proposed a quantum classifier based on quantum state fidelity by using a different initial state and replacing the Hadamard classification with a swap test. This new method is deemed to boost classification tasks in small datasets, by utilising the quantum advantage in finding non-linear features in a large feature space.

While machine learning recognises objectives based on its training dataset, which feeds labelled data which is classified, quantum machine learning, via the kernel method, can identify non-linear relationships in complex data. 

The kernel method has only been recently introduced in quantum machine learning. Quantum computers’ ability to proficiently access and manipulate data in the quantum feature space provides ample opportunities for quantum techniques to improve various existing machine learning methods. 

Classification algorithm with a non-linear kernel in a quantum test state will have the protocol calculate the weighted power sum of the fidelities of quantum data in quantum parallel via a swap-test circuit followed by two single-qubit measurements. This method needs a handful of quantum data operations no matter how big or small the size of the data is. This approach allows large training data which is labelled to be packed densely into a quantum state and then compared to the test data.

The Korea Advanced Institute of Science and Technology (KAIST) have pioneered this approach in collaboration with the University of KwaZulu-Natal (UKZN) in South Africa and Data Cybernetics in Germany. The team has made advancements in the field of quantum machine learning by introducing quantum classifiers with tailored quantum kernels.

The input data is either constituted by classical data via a quantum feature map or intrinsic quantum data, and the classification is based on the kernel function that measures the closeness of the test data to training data.

Dr Daniel Park at KAIST, one of the lead authors of this research, said that the quantum kernel can be tailored systematically to an arbitrary power sum, which makes it an excellent candidate for real-world applications.

Professor Francesco Petruccione from UKZN explained, “The state fidelity of two quantum states includes the imaginary parts of the probability amplitudes, which enables the use of the full quantum feature space.” 

By INDIAai


Previous News

NVIDIA’s AI supporting Ampere GPUs are now available for Google Cloud

Next News

A new global coalition to help policymakers leverage AI against COVID-19

Suggested Articles