Quantum-inspired Neural Network with Lindblad Master Equation for Sentiment Analysis
Kehuan Yan, Wenjun Wang
Quantum-inspired models have demonstrated superior performance in many downstream language tasks, such as question answering and sentiment analysis. However, recent models have primarily focused on embedding and measurement operations, overlooking the significance of the quantum evolution process. In this work, we present a novel quantum-inspired neural network, LME-QNN, which integrates the Lindblad Master Equation (LME) to model the evolution process. LME-QNN is able to capture rich correlations among words, ensuring a potent representation of the sentence. We conduct comprehensive experiments on six sentiment analysis datasets. Compared to the traditional models, including TextCNN, GRU, ELMo, BERT and RoBERTa, and quantum-inspired models, such as ComplexCNN and ComplexQNN, the proposed method demonstrates superior performance in accuracy and F1-score on six commonly used datasets for sentiment analysis. Additional ablation tests verify the effectiveness of LME.
•Mar 22, 2024•0 citations