The Information Extraction Framework of Document Spanners - A Very Informal Survey

This document provides an intuitive and high-level survey of the information extraction framework of document spanners (Fagin, Kimelfeld, Reiss, and Vansummeren (PODS 2013, J. ACM 2015)). Originally, document spanners were presented as a formalisation of the query language AQL, which is used in IBM’s information extraction engine SystemT, and over the last decade this framework is heavily investigated in the database theory community. The research topic of document spanners combines classical results from areas like formal languages, algorithms and database theory, while at the same time posing challenging new research questions. This survey is aimed at a general theoretical computer science audience that is not necessarily familiar with database theory. Its focus are the topics of an invited talk at SOFSEM 2024.