Citation
D. D. Bourgin, J. T. Abbott, T. L. Griffiths, K. A. Smith, E. Vul. "Empirical Evidence for Markov Chain Monte Carlo in Memory Search". Annual Conference of the Cognitive Science Society, 2014.

Abstract
Previous theoretical work has proposed the use of Markov chain Monte Carlo as a model of exploratory search in memory. In the current study we introduce such a model and evaluate it on a semantic network against human performance on the Remote Associates Test (RAT), a commonly used creativity metric. We find that a family of search models closely resembling the Metropolis-Hastings algorithm is capable of reproducing many of the response patterns evident when human participants are asked to report their intermediate guesses on a RAT problem. In particular we find that when run our model produces the same response clustering patterns, local dependencies, undirected search trajectories, and low associative hierarchies witnessed in human responses.

Electronic downloads

• http://cocosci.berkeley.edu/josh/papers/MCMC_MemorySearch.pdf

• HTML

  D. D. Bourgin, J. T. Abbott, T. L. Griffiths, K. A. Smith,
  E. Vul. <a
  href="http://robotics.eecs.berkeley.edu/pubs/11.html"
  >Empirical Evidence for Markov Chain Monte Carlo in
  Memory Search</a>, Annual Conference of the Cognitive
  Science Society, 2014.

• Plain text

  D. D. Bourgin, J. T. Abbott, T. L. Griffiths, K. A. Smith,
  E. Vul. "Empirical Evidence for Markov Chain Monte
  Carlo in Memory Search". Annual Conference of the
  Cognitive Science Society, 2014.

• BibTeX

  @inproceedings{BourginAbbottGriffithsSmithVul14_EmpiricalEvidenceForMarkovChainMonteCarloInMemorySearch,
      author = {D. D. Bourgin and J. T. Abbott and T. L. Griffiths
                and K. A. Smith and E. Vul},
      title = {Empirical Evidence for Markov Chain Monte Carlo in
                Memory Search},
      booktitle = {Annual Conference of the Cognitive Science Society},
      year = {2014},
      abstract = {Previous theoretical work has proposed the use of
                Markov chain Monte Carlo as a model of exploratory
                search in memory. In the current study we
                introduce such a model and evaluate it on a
                semantic network against human performance on the
                Remote Associates Test (RAT), a commonly used
                creativity metric. We find that a family of search
                models closely resembling the Metropolis-Hastings
                algorithm is capable of reproducing many of the
                response patterns evident when human participants
                are asked to report their intermediate guesses on
                a RAT problem. In particular we find that when run
                our model produces the same response clustering
                patterns, local dependencies, undirected search
                trajectories, and low associative hierarchies
                witnessed in human responses.},
      URL = {http://robotics.eecs.berkeley.edu/pubs/11.html}
  }
  

Posted by Ehsan Elhamifar on 30 May 2014.
Groups: ehumans
For additional information, see the Publications FAQ or contact webmaster at robotics eecs berkeley edu.