Learning Topics and Positions from Debatepedia
journal contributionposted on 2013-10-01, 00:00 authored by Swapna Gottipati, Minghui Qui, Yanchuan Sim, Jing Jiang, Noah A. Smith
We explore Debatepedia, a communityauthored encyclopedia of sociopolitical debates, as evidence for inferring a lowdimensional, human-interpretable representation in the domain of issues and positions. We introduce a generative model positing latent topics and cross-cutting positions that gives special treatment to person mentions and opinion words. We evaluate the resulting representation’s usefulness in attaching opinionated documents to arguments and its consistency with human judgments about positions.