GLaRA: Graph-based Labeling Rule Augmentation for Weakly Supervised Named Entity Recognition

TL;DR

GLaRA introduces a graph neural network framework that automatically augments labeling rules for weakly supervised NER, significantly improving performance with minimal manual rule creation.

Contribution

The paper presents a novel graph-based approach to automatically generate and augment labeling rules for NER, reducing manual effort and enhancing weak supervision effectiveness.

Findings

Achieves +20% F1 score improvement over baselines.

Effectively learns new labeling rules from unlabeled data.

Demonstrates robustness across three NER datasets.

Abstract

Instead of using expensive manual annotations, researchers have proposed to train named entity recognition (NER) systems using heuristic labeling rules. However, devising labeling rules is challenging because it often requires a considerable amount of manual effort and domain expertise. To alleviate this problem, we propose \textsc{GLaRA}, a graph-based labeling rule augmentation framework, to learn new labeling rules from unlabeled data. We first create a graph with nodes representing candidate rules extracted from unlabeled data. Then, we design a new graph neural network to augment labeling rules by exploring the semantic relations between rules. We finally apply the augmented rules on unlabeled data to generate weak labels and train a NER model using the weakly labeled data. We evaluate our method on three NER datasets and find that we can achieve an average improvement of +20\% F1 score over the best baseline when given a small set of seed rules.