JEL: Applying End-to-End Neural Entity Linking in JPMorgan Chase

Knowledge Graphs have emerged as a compelling abstraction for capturing key relationship among the entities of interest to enterprises and integrating data from heterogeneous sources. JPMorgan Chase (JPMC) is leading this trend by leveraging knowledge graphs across organization multiple mission critical applications such risk assessment, fraud detection, investment advice, etc. A core problem in graph link mentions (e.g., company names) that are encountered textual sources graph. Although several techniques exist entity linking, they tuned Wikipedia, fail generalize an enterprise. In paper, we propose novel end-to-end neural linking model (JEL) uses minimal context information margin loss generate embeddings, Wide & Deep Learning match character semantic respectively. We show JEL achieves state-of-the-art performance names financial news with our report on efforts deploy company-wide system alerts response news. The methodology used directly applicable usable other who need solutions unique their respective situations.