Learning Text Representations for 500K Classification Tasks on Named Entity Disambiguation

By Ander Barrena, Aitor Soroa Etxabe, Eneko Agirre. 2018

7 citations recorded by the source corpus.

Named Entity Disambiguation algorithms typically learn a single model for all target entities. In this paper we present a word expert model and train separate deep learning models for each target entity string, yielding 500K classification tasks. This gives us the opportunity to benchmark popular text representation alternatives on this massive dataset. In order to face scarce training data we propose a simple data-augmentation technique and transfer-learning. We show that bag-of-word-embeddings are better than LSTMs for tasks with scarce training data, while the situation is reversed when having larger amounts. Transferring a LSTM which is learned on all datasets is the most effective context representation option for the word experts in all frequency bands. The experiments show that our system trained on out-of-domain Wikipedia data surpass comparable NED systems which have been trained on in-domain training data.

Subfields: Representation & Self-Supervised Learning, Information Extraction

Citations in the corpus

• Learning Dense Representations for Entity Retrieval (2019, neutral): “To augment the small 1500 example training data in TACKBP, they Barrena et al. (2018) instead train 523k mention-specific deep classifiers—effectively treating entity linking as a special form of word sense disam-biguation.”