Discourse Marker

Discourse markers are words or phrases that signal relationships between sentences or clauses, shaping the overall discourse structure of a text. Current research focuses on computationally modeling these markers, often using large language models and novel architectures to improve tasks like argument structure extraction and sentiment analysis, particularly in low-resource settings. This work is significant because accurately identifying and interpreting discourse markers enhances the ability of natural language processing systems to understand nuanced meaning and context, leading to improved performance in various applications, including cross-genre argument mining and cross-modality discourse analysis. The development of robust methods for handling ambiguity and domain adaptation in discourse marker processing remains a key challenge.