Which One to Choose: Random Walks or Spreading Activation?

  • Serwah Sabetghadam
  • Mihai Lupu
  • Andreas Rauber
Conference paper
Part of the Lecture Notes in Computer Science book series (LNCS, volume 8849)

Abstract

Modeling data as a graph of objects is increasingly popular, as we move away from the relational DB model and try to introduce explicit semantics in IR. Conceptually, one of the main challenges in this context is how to “intelligently” traverse the graph and exploit the associations between the data objects. Two highly used methods in retrieving information on structured data are: Markov chain random walks, as is the basic method for page rank, and spreading activation, which originates from the artificial intelligence area. In this paper, we compare these two methods from a mathematical point of view. Random walks have been preferred in information retrieval, while spreading activation has been proposed before, but not really adopted. In this study we find that they are very similar fundamentally under certain conditions. However, spreading activation has much more flexibility and customization options, while random walks holds concise mathematics foundation.

Keywords

Information retrieval Graph Spreading activation Random Walks 

Preview

Unable to display preview. Download preview PDF.

Unable to display preview. Download preview PDF.

References

  1. 1.
    Berkhin, P.: A survey on pagerank computing. Internet Mathematics 2 (2005)Google Scholar
  2. 2.
    Berthold, M.R., Brandes, U., Kötter, T., Mader, M., Nagel, U., Thiel, K.: Pure spreading activation is pointless. In: Proc. of CIKM (2009)Google Scholar
  3. 3.
    Chib, S., Greenberg, E.: Understanding the metropolis-hastings algorithm. The American Statistician 49(4) (1995)Google Scholar
  4. 4.
    Clements, M., De Vries, A.P., Reinders, M.J.T.: The task-dependent effect of tags and ratings on social media access. ACM Trans. Inf. Syst. 28(4) (2010)Google Scholar
  5. 5.
    Collins-Thompson, K., Callan, J.: Query expansion using random walk models. In: Proc. of CIKM (2005)Google Scholar
  6. 6.
    Craswell, N., Szummer, M.: Random walks on the click graph. In: Proc. of SIGIR (2007)Google Scholar
  7. 7.
    Crestani, F.: Application of spreading activation techniques in information retrieval. Artificial Intelligence Review 11 (1997)Google Scholar
  8. 8.
    Golub, G.H., Van Loan, C.F.: Matrix computations, 3rd edn. Johns Hopkins University Press (1996)Google Scholar
  9. 9.
    Hussein, T., Ziegler, J.: Adapting web sites by spreading activation in ontologies. In: Int. Workshop on Recommendation and Collaboration (in Conjunction with IUI 2008) (2008)Google Scholar
  10. 10.
    Richardson, M., Domingos, P.: The Intelligent Surfer: Probabilistic Combination of Link and Content Information in PageRank. In: Proc. of NIPS (2002)Google Scholar
  11. 11.
    Rocha, C., Schwabe, D., Aragao, M.P.: A hybrid approach for searching in the semantic web. In: Proc. of WWW (2004)Google Scholar
  12. 12.
    Sabetghadam, S., Lupu, M., Rauber, A.: Astera - a generic model for multimodal information retrieval. In: Proc. of Integrating IR Technologies for Professional Search Workshop (2013)Google Scholar
  13. 13.
    Sabetghadam, S., Bierig, R., Rauber, A.: A hybrid approach for multi-faceted IR in multimodal domain. In: Kanoulas, E., Lupu, M., Clough, P., Sanderson, M., Hall, M., Hanbury, A., Toms, E. (eds.) CLEF 2014. LNCS, vol. 8685, pp. 86–97. Springer, Heidelberg (2014)CrossRefGoogle Scholar
  14. 14.
    Salton, G., Buckley, C.: On the use of spreading activation methods in automatic information. In: Proc. of SIGIR (1988)Google Scholar
  15. 15.
    Szummer, M., Jaakkola, T.: Partially labeled classification with markov random walks. In: NIPS (2001)Google Scholar

Copyright information

© Springer International Publishing Switzerland 2014

Authors and Affiliations

  • Serwah Sabetghadam
    • 1
  • Mihai Lupu
    • 1
  • Andreas Rauber
    • 1
  1. 1.Institute of Software Technology and Interactive SystemsVienna University of TechnologyViennaAustria

Personalised recommendations