Zero-shot Triplet Extraction by Template Infilling

TL;DR

This paper introduces ZETT, a zero-shot triplet extraction framework that leverages template infilling with pre-trained language models, enabling extraction of unseen relations without additional training, and outperforms existing methods.

Contribution

The paper presents a novel zero-shot triplet extraction method using template infilling with pre-trained transformers, eliminating the need for training on specific relations.

Findings

ZETT outperforms previous state-of-the-art methods on FewRel and Wiki-ZSL datasets.

Automatically generated templates are effective for zero-shot relation extraction.

ZETT demonstrates stable performance across different datasets.

Abstract

The task of triplet extraction aims to extract pairs of entities and their corresponding relations from unstructured text. Most existing methods train an extraction model on training data involving specific target relations, and are incapable of extracting new relations that were not observed at training time. Generalizing the model to unseen relations typically requires fine-tuning on synthetic training data which is often noisy and unreliable. We show that by reducing triplet extraction to a template infilling task over a pre-trained language model (LM), we can equip the extraction model with zero-shot learning capabilities and eliminate the need for additional training data. We propose a novel framework, ZETT (ZEro-shot Triplet extraction by Template infilling), that aligns the task objective to the pre-training objective of generative transformers to generalize to unseen relations. Experiments on FewRel and Wiki-ZSL datasets demonstrate that ZETT shows consistent and stable performance, outperforming previous state-of-the-art methods, even when using automatically generated templates. https://github.com/megagonlabs/zett/