Herbie Lee

Associate Professor Department of Applied Math and Statistics, University of California, Santa Cruz

office: 151 Baskin Engineering, phone: 831-459-1655
mailing address: UC Santa Cruz, School of Engineering (MS: SOE2), 1156 High Street, Santa Cruz, CA 95064

My books on multiscale modeling and neural networks:


Multiscale Modeling: A Bayesian Perspective. Published by Springer, this book reviews a variety of methods for Bayesian modeling and statistical computation in multiscale settings.		Bayesian Nonparametrics via Neural Networks. Published by SIAM, in the SIAM-ASA series on statistics and applied probability, the goal of this book is to treat neural networks as a statistical model, not just a black box. errata

Teaching

Winter 2008: AMS 206, Bayesian Statistics

2007: Winter: AMS 206, Bayesian Statistics; Spring: AMS 162, Design and Analysis of Computer Simulation Experiments, AMS 245, Spatial Statistics; Fall: AMS 7, Statistics for the Biological, Environmental, and Health Sciences

2006: Winter: AMS 206, Bayesian Statistics; Spring: AMS 162, Design and Analysis of Computer Simulation Experiments

2005: Winter: AMS 113, Managerial Statistics; Spring: AMS 162, Design and Analysis of Computer Simulation Experiments; Fall: AMS 5, Statistics, AMS 285, Seminar in Career Skills

2004: Winter: ENGR 113, Managerial Statistics; Spring: Jaynes reading group; Fall: AMS 205, Mathematical Statistics

2003: Winter: ENGR 113, Managerial Statistics; Fall: ENGR 5, Statistics, ENGR 205, Mathematical Statistics

2002: Fall: ENGR 205, Mathematical Statistics

Statistical Society Links

CSNA Classification Society of North America
ISBA International Society for Bayesian Analysis
ASA American Statistical Association
IMS Institute of Mathematical Statistics

My Research

I work in the field of Bayesian statistics, with current primary emphases on computer models (e.g., spatial inverse problems) and connections between statistics and machine learning.

While I was a post-doc at ISDS at Duke, I was part of the NSF/KDI funded project Multi-Scale Modeling and Simulation in Scientific Inference: Hierarchical Methods for Parameter Estimation in Porous Flow, which also involved the Center for Multi-Scale Modeling and Distributed Computing.

Publications

Chocolate Chip Cookies as a Teaching Aid (The American Statistician 2007, Volume 61, Issue 4, pp. 351-355)
Bayesian Treed Gaussian Process Models with an Application to Computer Modeling with Robert Gramacy (to appear in the Journal of the American Statistical Association 2007)
Inference for a Proton Accelerator Using Convolution Models with Bruno Sansó, Weining Zhou, and Dave Higdon (to appear in the Journal of the American Statistical Association 2007; this version is UCSC TR ams2005-31)
Multiscale Modeling: A Bayesian Perspective. With Marco Ferreira (2007) (see above)
Default Priors for Neural Network Classification (Journal of Classification, 2007, pp. 53-70; this version is UCSC TR ams2005-15)
Multi-resolution Genetic Algorithms and Markov Chain Monte Carlo with Chris Holloman and Dave Higdon (Journal of Computational and Graphical Statistics 2006, pp. 861-879; this version is Duke ISDS TR #02-06)
Multi-Scale and Hidden Resolution Time Series Models with Marco A. R. Ferreira, Mike West, and Dave Higdon (Bayesian Analysis, 2006, pp. 947-968)
Gaussian Processes and Limiting Linear Models with Robert Gramacy (Proceedings of the American Statistical Association, Section on Bayesian Statistical Science, 2006)
Inferring Particle Distribution in a Proton Accelerator Experiment with Bruno Sansó, Weining Zhou, and Dave Higdon (Bayesian Analysis, 2006, pp. 249-264)
Neural Networks and Default Priors (Proceedings of the American Statistical Association, Section on Bayesian Statistical Science, 2005)
Efficient Models for Correlated Data via Convolutions of Intrinsic Processes with Dave Higdon, Kate Calder, and Chris Holloman (Statistical Modelling, 2005; this version is UCSC TR ams2004-03)
Bayesian Nonparametrics via Neural Networks. (2004) (see above)
Priors for Neural Networks (2004, in Classification, Clustering, and Data Mining Applications, pp. 141-150; this version is UCSC TR ams2003-09)
Parameter Space Exploration With Gaussian Process Trees with Robert Gramacy and William Macready (2004, in Proceedings of the International Conference on Machine Learning, pp. 353-360)
Lossless Online Bayesian Bagging with Merlise Clyde (Journal of Machine Learning Research, February 2004)
Markov chain Monte Carlo-based approaches for inference in computationally intensive inverse problems with Dave Higdon and Chris Holloman (2003, in Bayesian Statistics 7, pp. 181-197; this version is Duke ISDS #02-10)
Multi-scale Modeling of 1-D Permeability Fields with Marco A. R. Ferreira, Zhuoxin Bi, Mike West, and Dave Higdon (2003, in Bayesian Statistics 7, pp. 519-527; this version is Duke ISDS TR #02-08)
A Noninformative Prior for Neural Networks (Machine Learning, 2003; this version is Duke ISDS TR #00-04)
Markov Random Field Models for High-Dimensional Parameters in Simulations of Fluid Flow in Porous Media with David Higdon, Zhuoxin Bi, Marco Ferreira, and Mike West (Technometrics, August 2002; this version is Duke ISDS #00-35), a version of which also won Best Contributed Paper in the Statistical Computing Section sessions at the 2000 Joint Statistical Meetings
A Bayesian Approach to Characterizing Uncertainty in Inverse Problems Using Coarse and Fine Scale Information with Dave Higdon and Zhuoxin Bi (IEEE Transactions on Signal Processing, February 2002; this version is Duke ISDS TR #01-02)
Did Lennox Lewis Beat Evander Holyfield? Methods for Analyzing Small-sample Inter-rater Agreement Problems with Daniel Cork and David Algranati (The Statistician, July 2002; this version is CMU Stats TR #732)
Difficulties in Estimating the Normalizing Constant of the Posterior for a Neural Network (Journal of Computational and Graphical Statistics, March 2002)
Model Selection for Neural Network Classification (Journal of Classification, 2001; this version is Duke ISDS TR #00-18)
Bagging and the Bayesian Bootstrap with Merlise Clyde (In Artificial Intelligence and Statistics 2001, T. Richardson and T. Jaakkola eds.; this version is Duke ISDS TR #00-34)
Loglinear Models and Goodness-of-Fit Statistics for Train Waybill Data, with Kert Viele (Journal of Transportation and Statistics, April 2001)
Consistency of Posterior Distributions for Neural Networks (Neural Networks, July 2000; this version is CMU Stats TR #676, 1998)
Model Selection and Model Averaging for Neural Network Regression (Proceedings of the American Statistical Association, Section on Bayesian Statistical Science, 1999; this version is Duke ISDS TR #00-32)

Technical Reports

Bayesian Guided Pattern Search for Robust Local Optimization with Matthew Taddy, Genetha Gray, and Joshua Griffin (2008 - UCSC TR ams2008-02)
Fast Bayesian Inference for Computer Simulation Inverse Problems with Matthew Taddy and Bruno Sansó (2008 - UCSC TR ams2008-03)
Adaptive Design of Supercomputer Experiments with Robert Gramacy (2006 - UCSC TR ams2006-13)
Bayesian Treed Gaussian Process Models with Robert Gramacy (2006 - UCSC TR ams2006-08)
Adaptive Exploration of Computer Experiment Parameter Spaces with Robert Gramacy and William Macready (2005 - UCSC TR ams2005-16)
Multi-scale Random Field Models with Marco Ferreira, Dave Higdon, and Mike West (2005)
Upscaling Tensorial Permeability Fields Based on Gaussian Markov Random Field Models and the Hybrid Mixed Finite Element Method with Zhuoxin Bi, John Trangenstein, and Dave Higdon (2005 -- UCSC TR ams2005-14)
Flexible Gaussian Processes via Convolution with Chris Holloman, Kate Calder, and Dave Higdon (2002 -- Duke ISDS #02-09)
Model Selection for Consumer Loan Application Data (1996 -- CMU Stats TR #650)

My Education

I finished my Ph.D. in statistics at Carnegie Mellon University in December, 1998. The title of my thesis is Model Selection and Model Averaging for Neural Networks and my advisor was Larry Wasserman.

My B.S. in mathematics is from Yale University. It was an intensive major, with a specialty in statistics.

I attended Punahou School for primary and secondary school in Honolulu, Hawaii.

Check out this web page on Squirrel fishing.

You can email me at herbie at ams.ucsc.edu.

Last modified on January 11, 2008.