Dynamic Multi-granularity Attribution Network for Aspect-based Sentiment Analysis
Abstract: Aspect-based sentiment analysis (ABSA) aims to predict the sentiment polarity of a specific aspect within a given sentence. Most existing methods predominantly leverage semantic or syntactic information based on attention scores, which are susceptible to interference caused by irrelevant contexts and often lack sentiment knowledge at a data-specific level. In this paper, we propose a novel Dynamic Multi-granularity Attribution Network (DMAN) from the perspective of attribution. Initially, we leverage Integrated Gradients to dynamically extract importance scores for each token, which contain underlying reasoning knowledge for sentiment analysis. Subsequently, we aggregate attribution representations from multiple semantic granularities in natural language, enhancing profound understanding of the semantics. Finally, we integrate attribution scores with syntactic information to more accurately capture the relationships between aspects and their relevant contexts during the sentence understanding process. Extensive experiments on five benchmark datasets demonstrate the effectiveness of our proposed method.
Paper Type: Long
Research Area: Sentiment Analysis, Stylistic Analysis, and Argument Mining
Research Area Keywords: Sentiment Analysis; Pre-trained Language Models; Argument Mining
Contribution Types: NLP engineering experiment
Languages Studied: English
Submission Number: 4657
Loading