Our lab studies discourse from linguistic and computational points of view. We examine and empirically test the role of discourse structure in different areas of language. In past work, we have worked on computational analysis of opinion (sentiment analysis), and studies of rhetorical/coherence relations.

Current projects involve:

• A study of online news comments, focusing on how we can best describe them in terms of 'good' and 'bad' comments. We describe 'good' comments as those that are constructive, i.e., they contribute to a dialogue about the relevant article, using evidence and providing reasons. 'Bad' comments may be described as toxic, abusive or in general meant to discourage an open exchange of ideas.
• Methods to identify whether a news article is an instance of misinformation.
• The Gender Gap Tracker, which monitors the proportion of women and men quoted in news stories in mainstream Canadian media (in English), with a French version in development.
  

More details in our research section. Recently released data includes the SFU Opinion and Comments Corpus (SOCC), the Constructive Comments Corpus (C3), and MisInfoText, a dataset for detecting fake news and misinformation.  

We gratefully acknowledge the support of the Social Sciences and Humanities Research Council of Canada, the Natural Sciences and Engineering Research Council of Canada, the Alexander von Humboldt Foundation in Germany, and NVIDIA Corporation.