Quantum-enhanced Support Vector Machine for Large-scale Multi-class Stellar Classification

Published in International Conference on Intelligent Computing, 2024

This paper presents a Quantum-enhanced Support Vector Machine (QSVM) tailored for complex stellar classification tasks using the Harvard system. Key features include:

  • Superior performance compared to classical algorithms like K-Nearest Neighbors and Logistic Regression in both binary and multi-class classification.
  • GPU-accelerated simulation via NVIDIA’s cuQuantum SDK for scalable training on large astronomical datasets.
  • Significant gains in accuracy and computational efficiency, showcasing the strength of quantum machine learning in astrophysics.

The proposed QSVM model sets a new benchmark for stellar classification and demonstrates the growing role of quantum algorithms in scientific data analysis at scale.

Figure: QSVM pipeline and stellar classification results (binary and multi-label). Powered by NVIDIA.

Recommended citation: Chen, Kuan-Cheng; Xu, Xiaotian; Makhanov, Henry; Chung, Hui-Hsuan; & Liu, Chen-Yu. (2024). "Quantum-enhanced Support Vector Machine for Large-scale Multi-class Stellar Classification." In Proceedings of the International Conference on Intelligent Computing, Springer, pp. 155–168.
Download Paper