KI-Lit-Edu – AI Indexing at Literature Information Systems for Educational Research

Project Description

Information is structured, made accessible and found due to systematic literature indexing. Substantial indexing of literature enables researchers to access relevant sources more quickly without having to carry out arduous searches. Indexing is usually an intellectual process and it can be time-consuming. The project therefore aims to develop a system that uses Artificial Intelligence to create keywords for educational research texts using the German Education Index (FIS) vocabulary and suggests them to intellectual indexers once a certain threshold value has been reached. The automatic indexing tool Annif will thus be used and set up in close cooperation with the German National Library (DNB). 
Annif will initially be used as a prototype in the peDOCS open access repository and tested, trained and integrated into the workflow drawing on 26,000 full texts showing specified criteria. Building on this, the aim is to implement it in other internal systems. Following successful implementation, the system might be made available as a web service for users and partners. Application scenarios for further implementation are the two external entry forms of peDOCS and Fachportal Pädagogik ("Publikation melden – report a publication" and "Publikation veröffentlichen – publish a publication") to recommend controlled keywords.

Project Objectives

The overall aim of the project is to set up an AI-based recommender system for keywords in educational research. Further objectives are to train and evaluate the Annif production system and then integrate the tool into systems, including DiDoRe/peDOCS and WinIBW. In addition, conditions for an external web service will be explored.

Project Management

Project Team

• Dr. Alexander Christ
• Henning Hinze
• Robina Mohammad

Project Details