Research Interests and Selected Papers

• Computational materials science

  collaborators:   Ed Fuller (NIST),   Bob Derr (SAS Institute)

  statistical models for microstructures, estimation of materials properties, Bayesian hierarchical models for fracture mechanics and cracking evolution

  • `` Statistical Modelling of Microstructure with Applications to Effective Property Computation in Materials Science '',
    Ph.D. thesis of Robert Derr (1999)   ( zipped ps file ).
  • ``Fitting microstructural models in materials science'' (with Robert Derr)   ( pdf file ).
  • ``Hard-core and soft-shell models for microstructures'' (with Robert Derr) (a preliminary version available upon request).
  • ``From microstructures to properties: statistical aspects of computational materials science''  
    (pdf file ).
  
   

• Computational finance

  collaborators:   Xin Ge (eBay),   Beom-Seok Lee (Univ. of Alabama),   Ai-ru Cheng (UC-Santa Cruz),  
  Ron Gallant (Duke),   Eric Renault (UNC-CH),   Jungyeon Yoon (UNC-CH)

  stochastic volatility models, Bayesian calibration, term structures of interest rate

  • ``Markov chain Monte Carlo calibration of stochastic volatility models'' (with Xin Ge) (2000)   ( zipped ps file ) ( pdf file )
  • ``A Gaussian approximation scheme for computation of option prices in stochastic volatility models''
    (with Ai-ru Cheng, Ron Gallant and Beom Lee) (2007)   ( pdf file )
  • ``Inference and computation for stochastic volatility models related to option pricing'' (2007)   ( pdf file )
  
   

• Inference of stochastic geometry models (Markov random fields, Markov point processes of geometric objects)

  collaborators:   Lynne Seymour (Univ. of Georgia),   Bob Derr (SAS Institute)

  parameter estimation and model selection of Gibbs measures, applications in image analysis, e.g. texture synthesis, spatial birth-and-death processes, Monte Carlo likelihood

  • ``Approximate Bayes model selection criteria for Markov random fields'' (with Lynne Seymour), J. Stat. Planning and Inference 51 (1996), 75-97.
  • ``A consistent model selection procedure for Markov random fields based on penalized pseudo-likelihood'' (with Lynne Seymour), Ann. Appl. Prob. 6 (1996), 423-443.
  
   

• Computational complexity of the Gibbs sampler on lattice

  collaborators:   Beti Kira (IME, Univ. of Sao Paulo, Brazil)
  • ``Rates of convergence for the Gibbs sampler'' (with Elisabeti Kira), Markov Processes and Related Fields 3 (1997), 89-102   ( zipped ps file ).
  
   

• Statistical shape models and applications to object detection

  • ``Object detection by partial deformable templates'' (with Ulf Grenander).
  
   

• Chaotic dynamical systems and smooth ergodic theory

  collaborators:   Donna Molinek (Davidson College)

  1D Gibbs measures as invariant measures, non-hyperbolic dynamical systems

  • ``Estimating functionals of one-dimensional Gibbs states'', Prob. Th. Related Fields 82 (1989), 155-175.
  • ``Asymptotic physical measures for nearly Brownian skew products'' (with Donna Molinek), Ergodic Th. and Dynamical Sys. 20 (2000), 473-481.
  
   

• Computer networks and sequential scheduling

  • ``Sequential scheduling of priority queues and arm-acquiring bandits'', J. Appl. Math. and Simulation 1 (1987), 115-135.
  
   

• Psychology

  collaborators:   Hongbin Gu (UNC-CH, Dept. of Psychiatry)

  Bayesian sequential decision approach for speeded categorization

  • ``A unified view on speeded categorization'' (with Hongbin Gu)   ( pdf file )

( This page is still under construction. )