Logo
ScienceStack
Table of Contents
Fetching ...
Sign In

Span-Oriented Information Extraction -- A Unifying Perspective on Information Extraction

Yifan Ding, Michael Yankoski, Tim Weninger

TL;DR

This work offers a unifying perspective centered on what they define to be spans in text and then re-present the wide assortment of information extraction tasks as variants of the same basic Span-Oriented Information Extraction task.

Abstract

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.

Span-Oriented Information Extraction -- A Unifying Perspective on Information Extraction

TL;DR

This work offers a unifying perspective centered on what they define to be spans in text and then re-present the wide assortment of information extraction tasks as variants of the same basic Span-Oriented Information Extraction task.

Abstract

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.
Paper Structure (48 sections, 9 equations, 1 figure, 3 tables)

This paper contains 48 sections, 9 equations, 1 figure, 3 tables.

Table of Contents

  1. Introduction
  2. A Brief History of Spans in Information Extraction
  3. Formal Definition of a Span
  4. Road Map
  5. Information Extraction Tasks
  6. Information Extraction Evaluation
  7. Information Extraction Models
  8. Types of Span-Oriented Information Extraction Tasks
  9. Entity Disambiguation
  10. Entity Linking
  11. Entity Typing
  12. Named Entity Recognition
  13. Attribute Value Extraction
  14. Machine Reading Comprehension
  15. Evaluation of Span-Oriented Information Extraction
  16. ...and 33 more sections

Figures (1)

  • Figure 1: An overview of information extraction tasks. The goal of information extraction is to identify sub-sequences within unstructured or semi-structured text information and link them to certain class-labels, entities in a knowledge base, or other items within some structured database. This structured information plays a central role in many downstream applications such as question answering and recommender systems.