A survey of the extraction and applications of causal relations

• DOI: 10.1017/s135132492100036x
• Publication Date: 2022-01-20T10:17:32Z
• Authors: Brett Drury; Hugo Gonçalo Oliv...; Alneu de Andrade ...
• Keywords: Causation; Causality; Relationship extraction

Abstract Causationin written natural language can express a strong relationship between events and facts. Causation in the form be referred to as causal relation where cause event entails occurrence of an effect event. A is stronger than correlation events, therefore aggregated relations extracted from large corpora used numerous applications such question-answering summarisation produce superior results traditional approaches. Techniques like logical consequence allow niche practical prediction which useful for diverse domains security finance. Until recently, use was relatively unpopular technique because extraction techniques were problematic, returned incomplete, error prone or simplistic. The recent adoption models improved extractors Transformer-XL (Dai et al . (2019). Transformer-xl: Attentive beyond fixed-length context arXiv preprint arXiv:1901.02860 ) has seen surge research interest possibilities using applications. now, there not been extensive survey relations; therefore, this intended precisely demonstrate potential relations. It comprehensive work on their applications, while also discussing nature causation its representation text.