The role of entanglement for enhancing the efficiency of quantum kernels towards classification
This work addresses the challenge of achieving quantum advantages in machine learning for text classification, though it appears incremental in optimizing quantum kernel hyperparameters.
The authors tackled the problem of enhancing quantum kernel efficiency for text sentiment classification by introducing a new quantum kernel using linear and fully entangled circuits as hyperparameters. Their results showed that the fully entangled circuit outperformed other quantum circuits and classical algorithms, with efficiency increasing significantly as features grew.
Quantum kernels are considered as potential resources to illustrate benefits of quantum computing in machine learning. Considering the impact of hyperparameters on the performance of a classical machine learning model, it is imperative to identify promising hyperparameters using quantum kernel methods in order to achieve quantum advantages. In this work, we analyse and classify sentiments of textual data using a new quantum kernel based on linear and full entangled circuits as hyperparameters for controlling the correlation among words. We also find that the use of linear and full entanglement further controls the expressivity of the Quantum Support Vector Machine (QSVM). In addition, we also compare the efficiency of the proposed circuit with other quantum circuits and classical machine learning algorithms. Our results show that the proposed fully entangled circuit outperforms all other fully or linearly entangled circuits in addition to classical algorithms for most of the features. In fact, as the feature increases the efficiency of our proposed fully entangled model also increases significantly.