Generating Entity Embeddings for Populating Wikipedia Knowledge Graph by Notability Detection

Knowledge graphs (KGs) have been playing a crucial role in leveraging information on web for several downstream tasks. Despite previous efforts in populating KGs, these methods typically do not focus on analyzing entity-specific content exclusively but rely on a fixed collection of documents. We define an approach to populate such KGs by utilizing entity-specific content on the web, for generating entity embeddings. We empirically prove our approach’s effectiveness, by utilizing it for a downstream task of Notability detection, associated with the Wikipedia Knowledge graph. To moderate content uploaded to Wikipedia, “Notability” guidelines are defined by its editors to identify entities warranting article on Wikipedia. So far notability is enforced by humans, which makes scalability an issue. In this paper, we define a multipronged approach based on web-based entity features, to construct entity embeddings for determining an entity’s notability. We distinguish entities based on their categories and utilize neural networks for classification. Our system outperforms machine learning-based classifiers and handcrafted entity salience detection algorithms, by achieving performance accuracy of around 88%. Our system provides a scalable alternative to manual decision-making about the importance of a topic, which could be extended to other such KG-based tasks.