ReSel: N-ary Relation Extraction from Scientific Text and Tables by Learning to Retrieve and Select

TL;DR

ReSel is a two-stage method for extracting N-ary relation tuples from scientific texts and tables, combining retrieval of relevant components with cross-modal entity selection, significantly outperforming existing methods.

Contribution

ReSel introduces a novel two-stage approach with a specialized feature set and cross-modal graph architecture for N-ary relation extraction from complex scientific documents.

Findings

ReSel outperforms state-of-the-art baselines on three datasets.

The method effectively retrieves relevant document parts for relation extraction.

Cross-modal entity correlation improves selection accuracy.

Abstract

We study the problem of extracting N-ary relation tuples from scientific articles. This task is challenging because the target knowledge tuples can reside in multiple parts and modalities of the document. Our proposed method ReSel decomposes this task into a two-stage procedure that first retrieves the most relevant paragraph/table and then selects the target entity from the retrieved component. For the high-level retrieval stage, ReSel designs a simple and effective feature set, which captures multi-level lexical and semantic similarities between the query and components. For the low-level selection stage, ReSel designs a cross-modal entity correlation graph along with a multi-view architecture, which models both semantic and document-structural relations between entities. Our experiments on three scientific information extraction datasets show that ReSel outperforms state-of-the-art baselines significantly.