Controversy in Context

TL;DR

This paper demonstrates that the immediate textual context of a concept is highly indicative of its controversiality, and introduces a new large-scale dataset with nuanced controversy ratings for improved prediction.

Contribution

It shows that textual context alone can effectively predict controversiality and provides a new, larger dataset with graded controversy levels for research.

Findings

Context-based features outperform metadata-based methods

Achieved state-of-the-art controversy prediction results

Provided a new dataset with graded controversy labels

Abstract

With the growing interest in social applications of Natural Language Processing and Computational Argumentation, a natural question is how controversial a given concept is. Prior works relied on Wikipedia's metadata and on content analysis of the articles pertaining to a concept in question. Here we show that the immediate textual context of a concept is strongly indicative of this property, and, using simple and language-independent machine-learning tools, we leverage this observation to achieve state-of-the-art results in controversiality prediction. In addition, we analyze and make available a new dataset of concepts labeled for controversiality. It is significantly larger than existing datasets, and grades concepts on a 0-10 scale, rather than treating controversiality as a binary label.