Duke University
Department of Statistical Science
Box 90251
Durham, NC 27708
I'm currently a postdoc in statistics at Duke University, working with David Dunson. My research interests are:
- robustness to model misspecification,
- nonparametric Bayesian models,
- frequentist analysis of Bayesian methods,
- efficient algorithms for inference in complex models, and
- applications in genomics, ecology, and neuroscience.
Robust Bayesian inference via coarsening, J. W. Miller and D. B. Dunson, (Submitted), 2015. (pdf) (arXiv) (code)
Mixture models with a prior on the number of components, J. W. Miller and M. T. Harrison, (Revision submitted), 2015. (pdf) (arXiv)
Importance sampling for weighted binary random matrices with specified margins, M. T. Harrison and J. W. Miller, (In preparation). (pdf) (arXiv)
Microclustering: When the cluster sizes grow sublinearly with the size of the data set, J. W. Miller, B. Betancourt, A. Zaidi, H. Wallach, and R. C. Steorts, Advances in Neural Information Processing Systems (NIPS), Bayesian Nonparametrics: The Next Generation workshop, 2015. (pdf) (arXiv)
Inconsistency of Pitman-Yor process mixtures for the number of components, J. W. Miller and M. T. Harrison, Journal of Machine Learning Research, Vol. 15, 2014, pp. 3333−3370. (pub) (pdf) (arXiv)
A simple example of Dirichlet process mixture inconsistency for the number of components, J. W. Miller and M. T. Harrison, Advances in Neural Information Processing Systems (NIPS), Vol. 26, 2013. (pub) (pdf) (arXiv)
Exact sampling and counting for fixed-margin matrices, J. W. Miller and M. T. Harrison, The Annals of Statistics, Vol. 41, No. 3, 2013, pp. 1569-1592. (pub) (pdf) (arXiv)
Reduced criteria for degree sequences, J. W. Miller, Discrete Mathematics, Vol. 313, Issue 4, 2013, pp. 550–562. (pub) (pdf) (arXiv)
Nonparametric and Variable-Dimension Bayesian Mixture Models: Analysis, Comparison, and New Methods,
J. W. Miller, Brown University, Division of Applied Mathematics, 2014.
(pdf)
(Received the Brown University Outstanding Dissertation Award in the Physical Sciences, generously sponsored by the Joukowsky Family Foundation.)
Exact enumeration and sampling of matrices with specified margins, J. W. Miller and M. T. Harrison, Unpublished report (2011). (pdf) (arXiv)
A practical algorithm for exact inference on tables, J. W. Miller and M. T. Harrison, Proceedings of the Joint Statistical Meetings 2010, Statistical Computing Section.
8th International Conference of the ERCIM WG on Computational and Methodological Statistics, Dec 14, 2015, London, UK. Robust Bayesian inference via coarsening.
Bayesian Nonparametrics: The Next Generation, NIPS 2015 Workshop, Dec 12, 2015, Montreal, CA. Non-standard approaches to nonparametric Bayes (talk), Microclustering: When the cluster sizes grow sublinearly with the size of the data set (poster). (One of 5 winners of travel award)
Harvard Statistics Departmental Colloquium, Sept 21, 2015, Harvard University. Robust Bayesian inference via coarsening.
Joint Statistical Meetings (JSM), August 11, 2015, Seattle, WA. Robust Bayesian inference via coarsening.
Bayesian Nonparametrics: Synergies between Statistics, Probability and Mathematics, June 30, 2015, SAMSI. Robust Bayesian inference via coarsening.
10th Conference on Bayesian Nonparametrics (BNP10), June 23, 2015, Raleigh, NC. An approach to inference under misspecification.
G70: A Celebration of Alan Gelfand's 70th Birthday, April 20, 2015, Duke University. The small clustering problem: What if the clusters don't grow with N?
Texas A&M Statistics Departmental Colloquium, October 31, 2014, Texas A&M University. Combinatorial stochastic processes for variable-dimension models.
International Society for Bayesian Analysis (ISBA) World Meeting, July 14 - 18, 2014, Cancun, Mexico. Combinatorial stochastic processes for variable-dimension models.
New England Statistics Symposium (NESS), April 26, 2014, Harvard School of Public Health. Combinatorial stochastic processes for variable-dimension models.
Dissertation defense, April 15, 2014, Brown University.
Duke Statistical Science Seminar, February 7, 2014, Duke University. Combinatorial stochastic processes for variable-dimension models.
Neural Information Processing Systems (NIPS) 2013, Lake Tahoe, NV. A simple example of Dirichlet process mixture inconsistency for the number of components. (Full oral presentation) (slides) (poster)
Pattern Theory Seminar, November 6, 2013, Brown University. Dirichlet process mixture inconsistency for the number of components, and dimension mixture models.
9th Conference on Bayesian Nonparametrics (BNP9), Amsterdam, 2013. Dimension mixtures of finite-dimensional models. (Winner of 1st place in poster competition) (poster)
New England Machine Learning day (NEML) 2013, Cambridge, MA. Posterior consistency for the number of components in a finite mixture.
New England Statistics Symposium (NESS) 2013, Storrs, CT. Posterior consistency for the number of components in a finite mixture.
Brown University Symposium for Undergraduates in the Mathematical Sciences (SUMS), 2013. High-dimensional parameter spaces and Fisher information.
Neural Information Processing Systems (NIPS) 2012, Lake Tahoe, NV (Workshop on Modern Nonparametric Methods in Machine Learning). Posterior consistency for the number of components in a finite mixture.
ICERM Bayesian Nonparametrics Workshop, 2012, Providence, RI. Dirichlet process mixtures are inconsistent for the number of components in a finite mixture.
New England Statistics Symposium (NESS) 2011, Storrs, CT. A practical algorithm for exact inference on tables. (One of four winners of the IBM Thomas J. Watson Research Center Student Research Award)
Joint Statistical Meetings (JSM) 2010, Vancouver, BC. A practical algorithm for exact inference on tables.