We construct a new annotated dataset of fallacies, using user comments containing fallacy mentions as noisy labels, and cleaning the data via crowdsourcing.
確定! 回上一頁