The task of mining the Web for analytical insights is still too complex for the average user. Our research aims at simplifying this task. We focus on three research areas:

  • Interactive Entity Linking and Understanding of Large Texts
    What search intentions does a user have while reading a text? How can we support an author with complementary information during the process of writing? How can we adapt semantic models of language using interaction?
    (See project TASTY)
  • In-Database Analytics
    How do we explore information beyond a simple Google lookup search? Can we leverage transactions from thousands of humans for on-the-fly domain adaption of information extraction algorithms?  Can we execute these algorithms in a RDBMS that already holds domain and context data? How can we measure Open Information Extraction and where should we execute OIE? How can we link text data to relational data from Data Warehouses to incorporate more fresh information from text into decision making?
    (See projects INDREX , RelVIS and IDEL)
  • Explorative Analytics and Data Veracity
    Who is always the first? And who removes valuable information? Can we trace back the origin of information on the Web? Which data manipulation operations are executed by whom and for which purpose?
    (See project GoOLAP -finished-)