ILF: A Quantum Semi-Supervised Learning Approach for Binary Classification

Abstract: The lack of enough labeled data is a great issue when designing a real-life scheme. Data labeling is time-consuming as well as costly. Semi-supervised learning (SSL) is a way to solve the issues of data labeling. SSL uses a tiny quantity of labeled data to find labels of massive quantities of unlabeled data. This paper presents a quantum-classical SSL mechanism named "Iterative Labels Finding (ILF)" by combining the Quantum Support Vector Machine algorithm (QSVM) and Ising Models Based Binary Clustering algorithm. The proposed method performs a matching and iteration process to discover the labels of unlabeled data. ILF is designed for binary classification purposes. We have illustrated the experimental result of ILF with a real-time dataset and with a practical example. From experimental results, we have found ILF as a highly efficient approach for quantum SSL.

Keywords: Quantum Computing, QSVM, Semi-Supervised Learning, QML Works Cited: Reshma Ahmed Swarna, Mohammed Ibrahim Hussain, Md. Sadiq Iqbal, Mohammad Mamun, Safiul Haque Chowdhury " ILF: A Quantum Semi-Supervised Learning Approach for Binary Classification ", IJARCCE International Journal of Advanced Research in Computer and Communication Engineering, vol. 12, no. 11, pp. 88-96, 2023. Crossref https://doi.org/10.17148/IJARCCE.2023.121112