Associate Professor of Linguistics at University of Illinois at Urbana-Champaign

Causal Reasoning in Natural Language: Some Perspectives, Models, and Applications

The acquisition of causal knowledge is an indispensable part of human and machine reasoning, and thus, very useful in many natural language processing applications. Despite its importance, the specification of a satisfactory general analysis of causal relations in language has long proved difficult. In this talk I will present some unsupervised and semi-supervised models for the identification of causal event sequences in various text genres. The experiments give insights into various issues, such as verb semantics, inference rules, causal chains, as well as the perception of causal knowledge. I will conclude with an application to the analysis of interpersonal interactions in social media.