International Workshop on Maximum Entropy and Bayesian Methods Santa Fe, July 31 - August 4, 1995 ################### MONDAY 8:30 TUTORIAL ON BAYESIAN METHODS AND THE MAXENT PRINCIPLE Peter Cheeseman and Wray Buntine NASA Ames 10:00 Break - Refreshments 10:30 Tutorial continued 12:30 Lunch 1:30 Laboissiere, R. INPG (France) A Bayesian approach to acoustics-to-articulatory inversion in speech: Detecting the position of the velum from speech sounds 2:00 Hawkins, R. Consultant, San Francisco Reconstruction of the probability density function implicit in option prices from incomplete and noisy data 2:30 Ramasmami, A. Washington Univ. Model selection and parameter estimation for exponential signals 3:00 Break - Refreshments ------------------------------------------------------ TIME-SERIES ANALYSIS organized by Mark Berliner, Ohio State University 3:30 Berliner, M. Ohio State Univ. Bayesian time-series analysis 4:10 Tsay, R. Univ. Chicago Bayesian specification of multivariate time series 4:50 West, M. Duke Univ. Bayesian time series: Models and computations for the analysis of time series in the physical sciences ------------------------------------------------------ 5:30 End 6:00 Kickoff Dinner - Richard Silver's house #################### TUESDAY 8:30 INVITED REVIEW Imry Csiszar Institute of Mathematical Sciences (Budapest, Hungary) The maximum-entropy principle and information theory 9:30 Dowe, D. Monash Univ. (Australia) Bayesian estimation of the von Mises concentration parameter 10:00 Break - Refreshments ------------------------------------------------------ MACHINE LEARNING organized by David Wolpert, Santa Fe Institute and TXN Inc. 10:30 Heckerman, D. Microsoft Priors for learning Bayesian networks 11:00 Madigan, D. Univ. of Washington Bayesian model averaging for acyclic directed graphical models. 11:30 Rosen, D. New York Medical College Issues in selecting empirical performance measures for probabilistic classifiers 12:00 Keren, D. Univ. of Haifa (Israel) A Bayesian framework for regularization ------------------------------------------------------ 12:30 Lunch 1:30 Tchoumatchenko, I. Laforia-IBP-CNRS (France) Mixture modeling to incorporate meaningful constraints into learning 2:00 Kreinovich, V. Univ. of Texas, El Paso Maximum-entropy methods in expert systems and intelligent control: new possibilities and limitations 2:30 Press, J. Univ. of California, Riverside The de Finetti transform 3:00 Short Break - Refreshments 3:15 Poster Prologue 4:15 Poster Session I 5:45 End 6:15 Reception - Santa Fe Institute #################### WEDNESDAY 8:30 INVITED REVIEW Michael Miller Washington Univ. Deformable models ------------------------------------------------------ DEFORMABLE MODELS organized by Ken Hanson, Los Alamos National Laboratory 9:30 Gee, J. Univ. of Pennsylvania Probabilistic matching of deformed images 10:00 Break - Refreshments 10:30 Rangarajan, A. State Univ. New York - Stonybrook Mechanical models as priors in Bayesian reconstruction of nuclear medical images 11:00 Hanson, K. Los Alamos National Laboratory The Bayes inference engine ------------------------------------------------------ 11:30 Mohammad-Djafari, A. CNRS-ESE-UPS (France) A full Bayesian approach for inverse problems 12:00 Puetter, R.C. Univ. of California, San Diego Pixon-based multiresolution image reconstruction and the quantification of image information content 12:30 Lunch 1:30 Trip to Los Alamos Options - Learn about the science done at LANL at the Bradbury Science Museum Learn about the early years at the Los Alamos Historical Museum and Fuller Lodge Walk through Indian Ruins at Bandelier National Monument Hike to the top of Pajarito Mountain (Los Alamos Ski Area), alt. 10000 ft. 6:30 Dinner at Fuller Lodge #################### THURSDAY 8:30 INVITED REVIEW Julian Besag Univ. of Washington The Bayesian inference machine: an introduction to Markov-chain Monte Carlo 9:30 Wu, M-D. Univ. of Cambridge (England) Bayesian evidence computation for multimodal distributions by adaptive tempering and thermodynamic integration MCMC methods 10:00 Break - Refreshments ------------------------------------------------------ DATA ANALYSIS OF PHYSICS SIMULATIONS organized by Richard Silver, Los Alamos National Laboratory 10:30 Gubernatis, J. Los Alamos National Laboratory Bayesian inference and the analytic continuation of imaginary-time quantum Monte carlo data 11:00 Preuss, R. Universitaet Wuerzburg (Germany) Consistent application of maximum entropy to quantum-Monte-carlo data 11:30 Pang, H. Univ. of Cincinnati Ground state dynamical properties: an approach from density matrix renormalization group and maximum entropy 12:00 Silver, R. Los Alamos National Laboratory Maximum-entropy and kernel-polynomial methods for moment problems ------------------------------------------------------ 12:30 Lunch 1:30 Lafferty, J. Carnegie Mellon Univ. Efficient iterative scaling of class of maximum-entropy language models 2:00 Vignaux, A. Victoria Univ. (New Zealand) A MaxEnt tomography method for estimating fish densities in a commercial fishery 2:30 Wagner, R, Food and Drug Administration, Rockville Toward optimal human and algorithmic observer performance of detection and discrimination tasks on reconstruction from sparse data 3:00 Short Break 3:15 Poster Prologue 4:15 Poster Session II - Refreshments 5:45 End #################### FRIDAY 8:30 Montgomery, D. Dartmouth College Entropies for dissipative fluids and magnetofluids without discretization 9:00 Fischer, R. Max-Planck-Institute (Germany) On the importance of correct alpha marginalization in inversion problems 9:30 Youssef, S. Florida State Univ. Quantum mechanics as an exotic probability theory 10:00 Break - Refreshments 10:30 Dose, V. Max-Planck Institut (Germany) Parameter estimation of nuclear-fusion confinement-time scaling laws 11:00 Boulanger, P. Nat. Research Council (Canada) Hierarchical segmentation of registered range and coloured images based on Bayesian decision theory 11:30 Skilling, J. Univ. of Cambridge (England) Prior distributions on measure space 12:00 Wolpert, D. Santa Fe Institute and TXN Inc. Bayesian tests of whether two data sets were generated by the same process 12:30 Lunch 1:30 Balasubramanian, V. Princeton Univ. A geometric formulation of Occam's razor for inference of parametric distributions 2:00 Borland, L. Univ. of California, Berkeley Inferring nonlinear dynamics of stochastic systems in continuous and discrete time 2:30 Dubertret, B. Massachusetts Inst. Tech. Prior probabilities, maximum entropy, and universality of random cellular structures 3:00 Break - Refreshments 3:30 Snapp, R. Univ. of Vermont Predicting the accuracy of Bayes classifiers 4:00 Shapiro, J. Univ. of Manchester (England) Maximum entropy analysis of genetic algorithms 4:30 Gautier, S. CNRS-ESE-UPS (France) Data fusion in the field of nondestructive testing 5:00 Sastry 5:30 End