Neural Stance Detection With Hierarchical Linguistic Representations

摘要: Stance detection aims to assign a stance label (i.e., favor or against ) post towards specific target. Recently, there is growing interest in adopting neural models detect of document. However, most these works focus on modeling the sequence words learn document representation, though other linguistic information, such as sentiment and arguments, are correlated with document, may inspire us explore stance. In this article, we propose hierarchical attention model well study various information better represent via representations. addition, network mechanism weight importance kinds mutual between information. Detail evaluation two benchmark datasets demonstrates effectiveness proposed mechanism.