Efficient Retrieval of Knowledge Graph Fact Evidences

  • Dhruv GuptaEmail author
  • Klaus Berberich
Conference paper
Part of the Lecture Notes in Computer Science book series (LNCS, volume 11762)


We report preliminary results on the problem of efficient retrieval of evidences for knowledge graph (KG) facts from large document collections. KGs are rich repositories of human knowledge and real-world events. To verify and validate facts about entities, it is often required to spot their evidences in large news archives or on the Web. To do so, KG facts can be translated to their natural language equivalent by using surface forms. Naïvely, attempting to search for all combinations of the aliases in large document collections is a time-consuming solution. We show that by using a combination of inverted indexes over n-grams and skip-grams we can return evidences in the form of sentences for KG facts within seconds.


  1. 1.
    Bhatia, S., Dwivedi, P., Kaur, A.: That’s interesting, tell me more! finding descriptive support passages for knowledge graph relationships. In: Vrandečić, D., et al. (eds.) ISWC 2018. LNCS, vol. 11136, pp. 250–267. Springer, Cham (2018). Scholar
  2. 2.
    Cohen, S., et al.: Computational journalism: a call to arms to database researchers. In: CIDR (2011)Google Scholar
  3. 3.
    Ercan, G., Elbassuoni, S., Hose, K.: Retrieving textual evidence for knowledge graph facts. In: Hitzler, P., et al. (eds.) ESWC 2019. LNCS, vol. 11503, pp. 52–67. Springer, Cham (2019). Scholar
  4. 4.
    Gupta, D., Berberich, K.: GYANI: an indexing infrastructure for knowledge-centric tasks. In: CIKM (2018)Google Scholar
  5. 5.
    Gerber, D., et al.: DeFacto - temporal and multilingual deep fact validation. J. Web Semant. 35, 85–101 (2015)CrossRefGoogle Scholar

Copyright information

© Springer Nature Switzerland AG 2019

Authors and Affiliations

  1. 1.Max Planck Institute for InformaticsSaarbrückenGermany
  2. 2.Graduate School of Computer ScienceSaarbrückenGermany
  3. 3.htw saarSaarbrückenGermany

Personalised recommendations