Distributed prediction of relations for entities: the Easy, the Difficult, and the impossible

Boleda, Gemma; Gupta, Abhijeet; Padó, Sebastian

Distributed prediction of relations for entities: the Easy, the Difficult, and the impossible

Citation

Gupta A, Boleda G, Pado S. Distributed prediction of relations for entities: the easy, the difficult, and the impossible. In: Proceedings of the 6th Joint Conference on Lexical and Computational Semantics (*SEM 2017); 2017 Aug 3-4; Vancouver, Canada. Stroudsburg (PA): ACL; 2017. p. 104–9.

Abstract

Word embeddings are supposed to provide easy access to semantic relations such as “male of” (man–woman). While this claim has been investigated for concepts, little is known about the distributional behavior of relations of (Named) Entities. We describe two word embedding-based models that predict values for relational attributes of entities, and analyse them. The task is challenging, with major performance differences between relations. Contrary to many NLP tasks, high difficulty for a relation does not result from low frequency, but from (a) one-to-many mappings; and (b) lack of context patterns expressing the relation that are easy to pick up by word embeddings.

Description

Comunicació presentada a la 6th Joint Conference on Lexical and Computational Semantics (*SEM 2017), celebrat els dies 3 i 4 d'agost de 2017 a Vancouver, Canada.