betaFeedback

WorksPhysics and Astronomy

Nonlinear function regression using quantum ensemble learning

Abstract

Recent advances in artificial intelligence and quantum technologies have driven the emergence of quantum machine learning, yet imperfections in near-term quantum devices remain a major obstacle for achieving practical quantum advantage. In the absence of fault-tolerant machines, ensemble learning, which combines multiple imperfect quantum devices, may unlock new quantum computing applications. Here, we introduce two ensemble-learning algorithms, namely Bagging and AdaBoost, for enhancing the accuracy of nonlinear regression problems. While Bagging trains variational quantum regressors independently, AdaBoost employs a sequential training strategy and achieves significantly better prediction accuracy. By incorporating systematic feature selection into our ensemble learning algorithms, we enhance their predictive power for highly nonlinear functions. We further improve prediction accuracy through a feature-to-qubit mapping strategy based on repeated feature encoding. The performance of our ensemble-based algorithms improves with both the number of regressors in the ensemble and the size of each regressor. We demonstrate the effectiveness of our approach using four benchmarking functions with varying degrees of nonlinearity by relying on a feature pool limited to first- and second-order polynomial terms.

Paper details

vlnc:W_e4gvprzc
Version
v1 · latest
Published
Fresh — added to the Hub within the last 30 days
Added
License
arXiv nonexclusive license
Usage

views · 0

downloads · 0

A view counts once per browser per day, a download each time the PDF is requested.

  • Physics and Astronomy

Not recorded for this paper

View statement
Looking for related papers…