Daily Arxiv

This is a page that curates AI-related papers published worldwide.
All content here is summarized using Google Gemini and operated on a non-profit basis.
Copyright for each paper belongs to the authors and their institutions; please make sure to credit the source when sharing.

Exploring an implementation of quantum learning pipeline for support vector machines

Created by
  • Haebom

Author

Mario Bifulco, Luca Roversi

Outline

This paper presents a fully quantum approach to support vector machine (SVM) training by integrating gate-based quantum kernel methods and quantum annealing-based optimization. Quantum kernels are constructed using various feature maps and qubit configurations, and their fitness is evaluated via kernel-target alignment (KTA). The SVM dual problem is reformulated as a quadratic unconstrained binary optimization (QUBO) problem and solved using a quantum annealer. Experimental results demonstrate that the high alignment of the kernels and appropriate regularization parameters contribute to competitive performance, with the best-performing model achieving an F1 score of 90%. These results highlight the feasibility of an end-to-end quantum training pipeline and the potential of hybrid quantum architectures in quantum high-performance computing (QHPC) environments.

Takeaways, Limitations

•
Takeaways:
◦
We demonstrate the feasibility of SVM training via a hybrid quantum approach that combines gate-based quantum computing and quantum annealing.
◦
We present an efficient method for constructing and evaluating quantum kernels using various feature maps and qubit configurations.
◦
Solving the QUBO problem using a quantum annealer suggests possibilities for practical quantum machine learning applications.
◦
It demonstrates competitive performance by achieving an F1-score of 90%.
◦
We demonstrate the potential of hybrid quantum architectures in the field of quantum high-performance computing (QHPC).
•
Limitations:
◦
Only experimental results for specific problems and datasets are presented, so further research is needed to determine generalizability.
◦
There is a lack of detailed description of the performance and constraints of the quantum annealer used.
◦
There is a lack of comparative analysis with other quantum machine learning algorithms.
◦
Further research is needed on applicability and scalability to large datasets.
👍
Made with Slashpage