English

Ultra-Fine Entity Typing

Computation and Language 2018-07-16 v1 Artificial Intelligence Machine Learning

Abstract

We introduce a new entity typing task: given a sentence with an entity mention, the goal is to predict a set of free-form phrases (e.g. skyscraper, songwriter, or criminal) that describe appropriate types for the target entity. This formulation allows us to use a new type of distant supervision at large scale: head words, which indicate the type of the noun phrases they appear in. We show that these ultra-fine types can be crowd-sourced, and introduce new evaluation sets that are much more diverse and fine-grained than existing benchmarks. We present a model that can predict open types, and is trained using a multitask objective that pools our new head-word supervision with prior supervision from entity linking. Experimental results demonstrate that our model is effective in predicting entity types at varying granularity; it achieves state of the art performance on an existing fine-grained entity typing benchmark, and sets baselines for our newly-introduced datasets. Our data and model can be downloaded from: http://nlp.cs.washington.edu/entity_type

Keywords

Cite

@article{arxiv.1807.04905,
  title  = {Ultra-Fine Entity Typing},
  author = {Eunsol Choi and Omer Levy and Yejin Choi and Luke Zettlemoyer},
  journal= {arXiv preprint arXiv:1807.04905},
  year   = {2018}
}

Comments

ACL 18

R2 v1 2026-06-23T02:59:52.771Z