Improving Automated Controversy Detection on the Web

Abstract

Automatically detecting controversy on the Web is a useful capability for a search engine to help users review web content with a more balanced and critical view. The current state-of-the art approach is to find K-Nearest-Neighbors in Wikipedia to the document query, and to aggregate their controversy scores that are automatically computed from the Wikipedia edit-history features. In this paper, we discover two major weakness in the prior work and propose modifications. First, the generated single query from document to find KNN Wikipages easily becomes ambiguous. Thus, we propose to generate multiple queries from smaller but more topically coherent paragraph of the document. Second, the automatically computed controversy scores of Wikipedia articles that depend on "edit war" features have a drawback that without an edit history, there can be no edit wars. To infer more reliable controversy scores for articles with little edit history, we smooth the original score from the scores of the neighbors with more established edit history. We show that the modified framework is improved by up to 5% for binary controversy classification in a publicly available dataset.

Explaining Controversy on Social Media via Stance Summarization

“…Controversy Analysis on the Web: To identify controversial topics in Web documents, some work has demonstrated that identifying relevant Wikipedia pages as well as building a controversy language model is effective [3,9,11]. Several studies then have formally defined a model for controversy detection [10,21].…”

Predicting Contradiction Intensity

“…Other works have also exploited Wikipedia to detect and to identify controversial topics on the web [4,9,10]. Dori-Hacohen and Allan in [4] and Jang and Allan in [9] proposed to align web pages to Wikipedia pages on the assumption that a page deals with a controversial topic if the Wikipedia page describing this topic is itself controversial. The controversial or non-controversial nature of a Wikipedia page is automatically detected based on the metadata and discussions associated with the page.…”

Finding and Quantifying Temporal-Aware Contradiction in Reviews

“…Other studies try to detect controversies on specific domains, for example in news or in debate analysis [18]. Other studies try to be more generic and detect the controversy on web [11]. Ennals et al [5] addressed the problem as a search of conflicting topics on the web through text patterns like "It is not correct that...".…”