Linking Entities to Unseen Knowledge Bases with Arbitrary Schemas

TL;DR

This paper introduces a zero-shot entity linking method that effectively links mentions to unseen knowledge bases with arbitrary schemas by converting entities into flat strings and employing regularization techniques.

Contribution

The work presents a novel approach for linking to unseen KBs with unknown schemas, enabling zero-shot generalization and combining multiple datasets seamlessly.

Findings

Outperforms baseline models by over 12 points in accuracy.

Effective in linking to unseen KBs with arbitrary schemas.

Able to combine multiple training datasets successfully.

Abstract

In entity linking, mentions of named entities in raw text are disambiguated against a knowledge base (KB). This work focuses on linking to unseen KBs that do not have training data and whose schema is unknown during training. Our approach relies on methods to flexibly convert entities from arbitrary KBs with several attribute-value pairs into flat strings, which we use in conjunction with state-of-the-art models for zero-shot linking. To improve the generalization of our model, we use two regularization schemes based on shuffling of entity attributes and handling of unseen attributes. Experiments on English datasets where models are trained on the CoNLL dataset, and tested on the TAC-KBP 2010 dataset show that our models outperform baseline models by over 12 points of accuracy. Unlike prior work, our approach also allows for seamlessly combining multiple training datasets. We test this ability by adding both a completely different dataset (Wikia), as well as increasing amount of training data from the TAC-KBP 2010 training set. Our models perform favorably across the board.