Poster – Paper 522

Frame Embeddings for Event-Based Knowledge Reconciliation Mehwish Alam, Diego Reforgiato, Misael Mongiovi, Aldo Gangemi and Petar Ristoski Poster October 23, 2017, Poster and Demo Reception, 18:30-21:20 Festsaal 1 Download paper (preprint)

Abstract

This paper focuses on reconciling knowledge graphs generated from two text documents about similar events described differently. We present an approach that employs and extends MERGILO, a tool for reconciling knowledge graphs extracted from text, using word similarity and graph alignment. Our approach is able to effectively handle events using Frame Embeddings and Frame Based Similarities. Our approach is evaluated over a coreference resolution task.