An Attentive Neural Architecture for Fine-grained Entity Type Classification

Sonse Shimaoka, Pontus Stenetorp, Kentaro Inui, Sebastian Riedel

2016 · 88 citations · 14 references

DOIFull text

Open access

Concepts

Abstract

In this work we propose a novel attentionbased
\nneural network model for the task of
\nfine-grained entity type classification that unlike
\npreviously proposed models recursively
\ncomposes representations of entity mention
\ncontexts. Our model achieves state-of-theart
\nperformance with 74.94% loose micro F1-
\nscore on the well-established FIGER dataset,
\na relative improvement of 2.59% . We also investigate
\nthe behavior of the attention mechanism
\nof our model and observe that it can learn
\ncontextual linguistic expressions that indicate
\nthe fine-grained category memberships of an
\nentity.

References

14