Hierarchical Entity Typing via Multi-level Learning to Rank
This addresses the problem of accurately classifying entities in hierarchical ontologies for natural language processing applications, representing an incremental improvement over existing methods.
The paper tackles hierarchical entity classification by incorporating ontological structure in training with a multi-level learning-to-rank loss and during prediction with a coarse-to-fine decoder, achieving state-of-the-art results, especially in strict accuracy, across multiple datasets.
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.