Information Service Engineering Programming Course Projects

Wintersemester 2018/19, Karlsruhe Institute of Technology (AIFB)

This page presents outcomes of the master student course Information Service Engineering at KIT. Four student groups chose open cultural heritage datasets, developed and evaluated their own web-based applications, thereby utilizing and improving their knowledge in semantic web technologies, natural language processing (NLP), and machine learning (ML). The data used in this course are open cultural heritage datasets made available by the Coding da Vinci initiative, which confronted the students with interesting and mostly uncharted data to explore.

Teaching Team: Russa Biswas, Maria Koutraki, Tabea Tietz, Rima Türker and Prof. Harald Sack

Papers:

A Linked Data Enhanced Museum Chatbot

In this project, a chatbot for the Städel Museum Frankfurt, Germany was developed to make museum visits more interesting for younger generations. The application is based on the Telegram messenger app and lets the user converse with a chatbot to receive information about art pieces and background information about artists, places or paintings integrated using Wikidata and DBpedia.

HistorEx: Exploring Historical Text using Word and Document Embeddings

In this project a framework is presented which combines unsupervised ML techniques and NLP on the example of historical text documents on the 19th century USA. Named entities are identified and extracted from semi-structured text, which is enriched with complementary information from Wikidata. Word embeddings are leveraged to enable the analysis of the text corpus, which is visualized in a web-based application.

Content-based Book Recommender System

A content-based book recommender system was developed using a dataset published by the Bavarian State Library. Recommendations are generated based on semantic similarities of the documents and on relatedness of specific entities in the corpus being connected to Wikidata.

Learning History through Gamification

This project contributes to the area of improving history lessons through gamification. Users interested in studying historical wars in Europe can select from a set of pictures presented on a map. Once chosen, a puzzle game based on historical pictures of European wars opens. Upon completion, the user receives complementary information automatically generated from DBpedia and Wikidata about the event and persons involved. The maps and landscapes used are part of the Hessian State Archives.