We address an important gap in detection of political bias in news articles.
Previous works that perform supervised document classification can be biased
towards the writing style of each news outlet, leading to overfitting and
limited generalizability. Our approach overcomes this limitation by considering
both the sentence-level semantics and the document-level rhetorical structure,
resulting in a more robust and style-agnostic approach to detecting political
bias in news articles. We introduce a novel multi-head hierarchical attention
model that effectively encodes the structure of long documents through a
diverse ensemble of attention heads. While journalism follows a formalized
rhetorical structure, the writing style may vary by news outlet. We demonstrate
that our method overcomes this domain dependency and outperforms previous
approaches for robustness and accuracy. Further analysis demonstrates the
ability of our model to capture the discourse structures commonly used in the
journalism domain.

本文提出了一种新的多头分层注意力模型，通过考虑句子层面的语义和文档层级的修辞结构，有效地编码长文档的结构，从而实现了对新闻文章政治倾向的更为健壮和风格无关的检测，并证明了该方法的优越性和准确性。