Neural Metric Learning for Fast End-to-End Relation Extraction

TL;DR

This paper introduces a neural architecture for end-to-end relation extraction that leverages 2D convolutions and metric-based features, achieving state-of-the-art accuracy with significantly faster training and testing times.

Contribution

The proposed neural model eliminates the need for global optimization, improving accuracy and efficiency in end-to-end relation extraction tasks.

Findings

Achieves approximately 1% F-score improvement over previous methods.

Training and testing times are seven to ten times faster.

Validated on ADE and CoNLL04 datasets with superior performance.

Abstract

Relation extraction (RE) is an indispensable information extraction task in several disciplines. RE models typically assume that named entity recognition (NER) is already performed in a previous step by another independent model. Several recent efforts, under the theme of end-to-end RE, seek to exploit inter-task correlations by modeling both NER and RE tasks jointly. Earlier work in this area commonly reduces the task to a table-filling problem wherein an additional expensive decoding step involving beam search is applied to obtain globally consistent cell labels. In efforts that do not employ table-filling, global optimization in the form of CRFs with Viterbi decoding for the NER component is still necessary for competitive performance. We introduce a novel neural architecture utilizing the table structure, based on repeated applications of 2D convolutions for pooling local dependency and metric-based features, that improves on the state-of-the-art without the need for global optimization. We validate our model on the ADE and CoNLL04 datasets for end-to-end RE and demonstrate $\approx 1\%$ gain (in F-score) over prior best results with training and testing times that are seven to ten times faster --- the latter highly advantageous for time-sensitive end user applications.