Beyond Pairwise: Global Zero-shot Temporal Graph Generation

TL;DR

This paper introduces a zero-shot method for temporal relation extraction that generates complete temporal graphs in a single step, improving efficiency and consistency over pairwise methods, and introduces a new dataset for evaluation.

Contribution

It presents a novel zero-shot approach for TRE that produces global temporal graphs and enforces consistency, along with a new comprehensive dataset for evaluation.

Findings

Outperforms existing zero-shot TRE methods

Offers a competitive alternative to supervised models

Enhances global consistency in temporal relation extraction

Abstract

Temporal relation extraction (TRE) is a fundamental task in natural language processing (NLP) that involves identifying the temporal relationships between events in a document. Despite the advances in large language models (LLMs), their application to TRE remains limited. Most existing approaches rely on pairwise classification, where event pairs are classified in isolation, leading to computational inefficiency and a lack of global consistency in the resulting temporal graph. In this work, we propose a novel zero-shot method for TRE that generates a document's complete temporal graph in a single step, followed by temporal constraint optimization to refine predictions and enforce temporal consistency across relations. Additionally, we introduce OmniTemp, a new dataset with complete annotations for all pairs of targeted events within a document. Through experiments and analyses, we demonstrate that our method outperforms existing zero-shot approaches and offers a competitive alternative to supervised TRE models.