Span-Oriented Information Extraction -- A Unifying Perspective on Information Extraction
This work addresses a foundational problem for NLP researchers by offering a novel perspective to unify disparate information extraction tasks, potentially streamlining progress in the field.
The paper tackles the heterogeneity among information extraction tasks in NLP by proposing a unifying perspective centered on spans in text, reorienting various tasks into a single Span-Oriented Information Extraction framework.
Information Extraction refers to a collection of tasks within Natural Language Processing (NLP) that identifies sub-sequences within text and their labels. These tasks have been used for many years to link extract relevant information and to link free text to structured data. However, the heterogeneity among information extraction tasks impedes progress in this area. We therefore offer a unifying perspective centered on what we define to be spans in text. We then re-orient these seemingly incongruous tasks into this unified perspective and then re-present the wide assortment of information extraction tasks as variants of the same basic Span-Oriented Information Extraction task.