Shards of Knowledge - Modeling Attributions for Event-Centric Knowledge Graphs

Florian Plötzky, Katarina Britz, Wolf-Tilo Balke

Abstract: Recently the usage of narratives as a means of fusing information from large knowledge graphs (KGs) into a coherent line of argumentation has been proposed. Narratives are especially useful in event-centric knowledge graphs in that they provide a means to categorize real-world events by well-known narrations. However, specifically for controversial events a problem in information fusion arises. Namely, the existence of multiple viewpoints regarding the validity of certain event aspects, e.g., regarding the role a participant takes in an event. Expressing those viewpoints into large KGs is challenging, because disputed information provided by different viewpoints may introduce inconsistencies. Hence, most KGs only feature a single view on the contained information, hampering the effectiveness of narrative information access. In this paper, we introduce attributions, i.e., parameterized predicates that allow for the representation of facts that are only valid in a certain viewpoint. For this, we develop a conceptual model that allows for the representation of viewpoint-dependent information and further describes how such information can be fused for querying and reasoning consistently.