We propose a novel method for hierarchical entity classification that embraces ontological structure at both training and during prediction. At training, our novel multi-level learning-to-rank loss compares positive types against negative siblings according to the type tree. During prediction, we define a coarse-to-fine decoder that restricts viable candidates at each level of the ontology based on already predicted parent type(s). We achieve state-of-the-art across multiple datasets, particularly with respect to strict accuracy.
@article{arxiv.2004.02286,
title = {Hierarchical Entity Typing via Multi-level Learning to Rank},
author = {Tongfei Chen and Yunmo Chen and Benjamin Van Durme},
journal= {arXiv preprint arXiv:2004.02286},
year = {2020}
}