LightRel SemEval-2018 Task 7: Lightweight and Fast Relation Classification

TL;DR

LightRel is a fast, lightweight relation classification method that uses minimal features and external knowledge, serving as a high baseline for relation extraction tasks.

Contribution

The paper introduces a simple, efficient linear classifier with minimal features and external knowledge for relation classification, emphasizing speed and simplicity.

Findings

Achieves high baseline performance with minimal features

Operates efficiently with fast training and inference

Utilizes word clusters and embeddings for external knowledge

Abstract

We present LightRel, a lightweight and fast relation classifier. Our goal is to develop a high baseline for different relation extraction tasks. By defining only very few data-internal, word-level features and external knowledge sources in the form of word clusters and word embeddings, we train a fast and simple linear classifier.