Aspect-base Sentiment Analysis with Dual Contrastive Learning

Abstract

Aspect-based sentiment analysis (ABSA) is a type of sentiment analysis that aims to identify the polarity of sentiment for aspects in a sentence. Also according to the studies, it is an important research area that plays an important role in business intelligence, marketing and psychology. To solve this problem different methods based on dictionary, machine learning and deep learning have been used. Research shows that among the methods based on deep learning, Transformers has been able to achieve good results and help to understand the language better. In this paper we use induced trees from Fine-tuning pre-trained models (FT-PTMs). We also use dual contrastive learning and different pre-trained models such as BERT, RoBERTa and XLNet in our proposed model. The results obtained from the implementation of the model in SemEval2014 benchmarks confirm the performance of our model.