A Capsule Network-based Embedding Model for Knowledge Graph Completion and Search Personalization

NAACL 2019 Dai Quoc NguyenThanh VuTu Dinh NguyenDat Quoc NguyenDinh Phung

In this paper, we introduce an embedding model, named CapsE, exploring a capsule network to model relationship triples (subject, relation, object). Our CapsE represents each triple as a 3-column matrix where each column vector represents the embedding of an element in the triple... (read more)

PDF Abstract

Evaluation Results from the Paper


TASK DATASET MODEL METRIC NAME METRIC VALUE GLOBAL RANK COMPARE
Link Prediction FB15k-237 CapsE MRR 0.523 # 2
Link Prediction FB15k-237 CapsE [email protected] 0.593 # 2
Link Prediction FB15k-237 CapsE MR 303 # 11
Link Prediction WN18RR CapsE MRR 0.415 # 21
Link Prediction WN18RR CapsE [email protected] 0.56 # 1
Link Prediction WN18RR CapsE MR 719 # 2