no code implementations • COLING 2022 • Zhihua Su, Qiang Zhou
We propose a speaker clustering model for textual dialogues, which groups the utterances of a multi-party dialogue without speaker annotations, so that the actual speakers are identical inside each cluster.