TacoERE: Cluster-aware Compression for Event Relation Extraction

TL;DR

TacoERE introduces a cluster-aware compression approach that enhances event relation extraction by modeling event dependencies within and across document clusters, reducing redundancy and improving long-range relation detection.

Contribution

It proposes a novel cluster-aware compression paradigm for ERE, combining document clustering and summarization to better handle long-range dependencies and redundancy.

Findings

Improves relation extraction accuracy across multiple datasets.

Effective with both pre-trained and large language models.

Reduces information redundancy and enhances long-range relation modeling.

Abstract

Event relation extraction (ERE) is a critical and fundamental challenge for natural language processing. Existing work mainly focuses on directly modeling the entire document, which cannot effectively handle long-range dependencies and information redundancy. To address these issues, we propose a cluster-aware compression method for improving event relation extraction (TacoERE), which explores a compression-then-extraction paradigm. Specifically, we first introduce document clustering for modeling event dependencies. It splits the document into intra- and inter-clusters, where intra-clusters aim to enhance the relations within the same cluster, while inter-clusters attempt to model the related events at arbitrary distances. Secondly, we utilize cluster summarization to simplify and highlight important text content of clusters for mitigating information redundancy and event distance. We have conducted extensive experiments on both pre-trained language models, such as RoBERTa, and large language models, such as ChatGPT and GPT-4, on three ERE datasets, i.e., MAVEN-ERE, EventStoryLine and HiEve. Experimental results demonstrate that TacoERE is an effective method for ERE.