Aspect-based sentiment analysis (ABSA), which aims to identify fine-grained opinion polarity towards a specific aspect, is a challenging subtask of sentiment analysis (SA). In this paper, we construct an auxiliary sentence from the aspect and convert ABSA to a sentence-pair classification task, such as question answering (QA) and natural language inference (NLI). We fine-tune the pre-trained model from BERT and achieve new state-of-the-art results on SentiHood and SemEval-2014 Task 4 datasets.
@article{arxiv.1903.09588,
title = {Utilizing BERT for Aspect-Based Sentiment Analysis via Constructing Auxiliary Sentence},
author = {Chi Sun and Luyao Huang and Xipeng Qiu},
journal= {arXiv preprint arXiv:1903.09588},
year = {2019}
}