Torsten Kilias is a Research Associate with the DATEXIS team and at the Beuth University of Applied Sciences Berlin. He graduated from Ilmenau University in 2012. During his studies he worked as intern for IBM in Böblingen, later he joined the team at TU-Berlin. 

His research is focused on Text Mining in Databases.  He worked on the problem of representing and querying text data with the span data model, in particular for the shared nothing paradigm (see also the INDREX).

Currently, he works in the BMWi funded Smart Data Web project on managing knowledge graphs and text data in a relational database.


Relevant publications

  1. Torsten Kilias, , Felix A. Gers, Richard Koopmanschap, Ying Zhang, Martin Kersten:   In-Database Entity Linking with Neural Embeddings       IDEL 2018 (see also the TASTY feat. IDEL)
  2. : INDREX: In-database relation extraction. Inf. Syst. 53124-144 ()
  3. : A Marketplace for Web Scale Analytics and Text Annotation Services. COLING (Demos) 
  4. :
    INDREX: in-database distributional relation extraction. DOLAP 93-100
  5. GiST scan acceleration using coprocessors. DaMoN 63-69



E-Mail: <firstname>.<lastname> (at)