References & Citations
Quantum Physics
Title: Quantum Annealing for Semi-Supervised Learning
(Submitted on 27 Mar 2020 (v1), last revised 12 Apr 2020 (this version, v2))
Abstract: Recent advances in quantum technology have led to the development and the manufacturing of programmable quantum annealers that promise to solve certain combinatorial optimization problems faster than their classical counterparts. Semi-supervised learning is a machine learning technique that makes use of both labeled and unlabeled data for training, which enables a good classifier with only a small amount of labeled data. In this paper, we propose and theoretically analyze a graph-based semi-supervised learning method with the aid of the quantum annealing technique, which efficiently utilize the quantum resources while maintaining a good accuracy. We illustrate two classification examples, suggesting the feasibility of this method even with a small portion (20%) of labeled data is involved.
Submission history
From: Yulin Zheng [view email][v1] Fri, 27 Mar 2020 15:09:44 GMT (789kb,D)
[v2] Sun, 12 Apr 2020 15:57:24 GMT (789kb,D)
Link back to: arXiv, form interface, contact.