Noise Mitigation for Neural Entity Typing and Relation Extraction

TL;DR

This paper tackles noise in information extraction by introducing neural multi-instance multi-label learning for entity typing and improving relation extraction through probabilistic predictions and joint training, achieving competitive results.

Contribution

It presents novel neural algorithms for multi-instance multi-label learning in entity typing and enhances relation extraction by integrating probabilistic predictions and joint training.

Findings

Probabilistic predictions outperform discrete ones.

Joint training yields the best performance.

Models achieve results comparable to state-of-the-art supervised methods.

Abstract

In this paper, we address two different types of noise in information extraction models: noise from distant supervision and noise from pipeline input features. Our target tasks are entity typing and relation extraction. For the first noise type, we introduce multi-instance multi-label learning algorithms using neural network models, and apply them to fine-grained entity typing for the first time. This gives our models comparable performance with the state-of-the-art supervised approach which uses global embeddings of entities. For the second noise type, we propose ways to improve the integration of noisy entity type predictions into relation extraction. Our experiments show that probabilistic predictions are more robust than discrete predictions and that joint training of the two tasks performs best.