A Kriging or Gaussian Process emulator has: an unadjusted mean (frequently a least squares fit: ), a correction / adjustment to the mean based on data,

Slides:



Advertisements
Similar presentations
Pattern Recognition and Machine Learning
Advertisements

Neural Networks and Kernel Methods
Modeling of Data. Basic Bayes theorem Bayes theorem relates the conditional probabilities of two events A, and B: A might be a hypothesis and B might.
Generation of Pareto Optimal Ensembles of Calibrated Parameter Sets for Climate Models Keith Dalbey, Ph.D. Sandia National Labs, Dept 1441, Optimization.
Insert Date HereSlide 1 Using Derivative and Integral Information in the Statistical Analysis of Computer Models Gemma Stephenson March 2007.
Running a model's adjoint to obtain derivatives, while more efficient and accurate than other methods, such as the finite difference method, is a computationally.
Designing Ensembles for Climate Prediction
Pattern Recognition and Machine Learning
Kriging.
Cost of surrogates In linear regression, the process of fitting involves solving a set of linear equations once. For moving least squares, we need to form.
Ensemble Emulation Feb. 28 – Mar. 4, 2011 Keith Dalbey, PhD Sandia National Labs, Dept 1441 Optimization & Uncertainty Quantification Abani K. Patra, PhD.
Cost of surrogates In linear regression, the process of fitting involves solving a set of linear equations once. For moving least squares, we need to.
Optimal Design Laboratory | University of Michigan, Ann Arbor 2011 Design Preference Elicitation Using Efficient Global Optimization Yi Ren Panos Y. Papalambros.
Basis Expansion and Regularization Presenter: Hongliang Fei Brian Quanz Brian Quanz Date: July 03, 2008.
Analysis of Simulation Input.. Simulation Machine n Simulation can be considered as an Engine with input and output as follows: Simulation Engine Input.
Linear and generalised linear models
GAUSSIAN PROCESS REGRESSION FORECASTING OF COMPUTER NETWORK PERFORMANCE CHARACTERISTICS 1 Departments of Computer Science and Mathematics, 2 Department.
Linear and generalised linear models
Improving Contaminant Mixing Models For Water Distribution Pipe Networks Siri Sahib S. Khalsa University of Virginia Charlottesville, VA
Gaussian process modelling
PATTERN RECOGNITION AND MACHINE LEARNING
Machine Learning CUNY Graduate Center Lecture 3: Linear Regression.
Sandia National Laboratories is a multi-program laboratory managed and operated by Sandia Corporation, a wholly owned subsidiary of Lockheed Martin Corporation,
Biointelligence Laboratory, Seoul National University
Applications of Bayesian sensitivity and uncertainty analysis to the statistical analysis of computer simulators for carbon dynamics Marc Kennedy Clive.
1/26 LOÏC LE GRATIET Bayesian analysis of hierarchical codes with different levels of accuracy Masters thesis conducted from 07/04/2010 to 15/09/2010 in.
Progress Towards Nested Space and Sub-Space Filling Latin Hypercube Sample Designs July 25 – July 28, 2011 Keith Dalbey, PhD Sandia National Labs, Dept.
WB1440 Engineering Optimization – Concepts and Applications Engineering Optimization Concepts and Applications Fred van Keulen Matthijs Langelaar CLA H21.1.
PATTERN RECOGNITION AND MACHINE LEARNING CHAPTER 3: LINEAR MODELS FOR REGRESSION.
CSC321: 2011 Introduction to Neural Networks and Machine Learning Lecture 11: Bayesian learning continued Geoffrey Hinton.
Strategies for Solving Large-Scale Optimization Problems Judith Hill Sandia National Laboratories October 23, 2007 Modeling and High-Performance Computing.
Center for Radiative Shock Hydrodynamics Fall 2011 Review Assessment of predictive capability Derek Bingham 1.
Investigating Adaptive Compilation using the MIPSpro Compiler Keith D. Cooper Todd Waterman Department of Computer Science Rice University Houston, TX.
EXTENSION OF LATIN HYPERCUBE SAMPLES WITH CORRELATED VARIABLES C. J. SALLABERRY, a J. C. HELTON b – S. C. HORA c aSandia National Laboratories, New Mexico.
LAMMPS Users’ Workshop
Selecting Input Probability Distribution. Simulation Machine Simulation can be considered as an Engine with input and output as follows: Simulation Engine.
Fast Generation of Nested Space-filling Latin Hypercube Sample Designs
ECE-7000: Nonlinear Dynamical Systems Overfitting and model costs Overfitting  The more free parameters a model has, the better it can be adapted.
Linear Models for Classification
July 11, 2006Bayesian Inference and Maximum Entropy Probing the covariance matrix Kenneth M. Hanson T-16, Nuclear Physics; Theoretical Division Los.
Overview of Optimization in Ag Economics Lecture 2.
Sandia is a multi-program laboratory operated by Sandia Corporation, a Lockheed Martin Company, for the United States Department of Energy’s National Nuclear.
Data Modeling Patrice Koehl Department of Biological Sciences National University of Singapore
1  The Problem: Consider a two class task with ω 1, ω 2   LINEAR CLASSIFIERS.
Speech Lab, ECE, State University of New York at Binghamton  Classification accuracies of neural network (left) and MXL (right) classifiers with various.
Over-fitting and Regularization Chapter 4 textbook Lectures 11 and 12 on amlbook.com.
Derivative-Enhanced Variable Fidelity Kriging Approach Dept. of Mechanical Engineering, University of Wyoming, USA Wataru YAMAZAKI 23 rd, September, 2010.
Multifidelity Optimization Using Asynchronous Parallel Pattern Search and Space Mapping Techniques Genetha Gray*, Joe Castro i, Patty Hough*, and Tony.
INTRO TO OPTIMIZATION MATH-415 Numerical Analysis 1.
Learning Theory Reza Shadmehr Distribution of the ML estimates of model parameters Signal dependent noise models.
Jack Flicker, Robert Kaplar, Matt Marinella, and Jennifer Granata Sandia National Laboratories Acknowledgements Contact Sandia National Laboratories is.
Multidisciplinary Design and Optimization (MDO) Natural Evolution of that Other Engineering Activity. Dr. Rob McDonald Lockheed Martin Endowed Professor.
8 Sept 2006, DEMA2006Slide 1 An Introduction to Computer Experiments and their Design Problems Tony O’Hagan University of Sheffield.
Sandia National Laboratories is a multi-program laboratory managed and operated by Sandia Corporation, a wholly owned subsidiary of Lockheed Martin Corporation,
Kriging - Introduction Method invented in the 1950s by South African geologist Daniel Krige (1919-) for predicting distribution of minerals. Became very.
3. Linear Models for Regression 後半 東京大学大学院 学際情報学府 中川研究室 星野 綾子.
1 C.A.L. Bailer-Jones. Machine Learning. Data exploration and dimensionality reduction Machine learning, pattern recognition and statistical data modelling.
LECTURE 11: Advanced Discriminant Analysis
CH 5: Multivariate Methods
Maximum Likelihood Estimation
Roberto Battiti, Mauro Brunato
CSCI 5822 Probabilistic Models of Human and Machine Learning
10701 / Machine Learning Today: - Cross validation,
OVERVIEW OF LINEAR MODELS
Biointelligence Laboratory, Seoul National University
OVERVIEW OF LINEAR MODELS
LECTURE 21: CLUSTERING Objectives: Mixture Densities Maximum Likelihood Estimates Application to Gaussian Mixture Models k-Means Clustering Fuzzy k-Means.
Multivariate Methods Berlin Chen
Multivariate Methods Berlin Chen, 2005 References:
Probabilistic Surrogate Models
Presentation transcript:

A Kriging or Gaussian Process emulator has: an unadjusted mean (frequently a least squares fit: ), a correction / adjustment to the mean based on data, and an estimated distribution about the adjusted mean of possible true surfaces Challenge: Should use Gaussian corr. function to exploit smoothness, but using gradients GREATLY AMPLIFIES ill-conditioning due to poor spacing. Insight: A matrix is ill-conditioned because it contains duplicate info. Solution: Use pivoted Cholesky on (not ) to efficiently sort points so that those with the most new information come first. Apply same ordering to whole points (a function evaluation plus its gradient as a block) in, and use LAPACK’s (level 3 BLAS) Cholesky. Use bisection with LAPACK rcond estimate to drop equations off the end of until it’s not ill-conditioned (rcond > ). Risk: depends on inputs not outputs; if there is a sharp jump in the output and very nearby points on either side of it, 1 could be discarded. Protection???: Should modify likelihood during maximization, but how? Suggestion: Use the Mahalanobis distance. Okay, but how??? Reward: A better surrogate for the same simulation budget. Simulators are frequently used to predict the performance of physical systems. Often, only estimated distributions are available for the simulator’s inputs. These must be propagated through the simulator to produce distributions for the system’s performance. Typically, engineering systems have a large number of input dimensions, M, and are computationally expensive to simulate. The “ Curse of Dimensionality ” states that the number of samples or simulations needed to quantify the uncertainty in a simulator’s output is exponential in M. A strategy for circumventing the curse is to construct a low cost surrogate, such as a Kriging emulator (a.k.a. a Gaussian Process), for the simulator from a “small” number of simulations, N. However, this “small” number is still subject to the Curse of Dimensionality. One can reduce the required number of simulations by generating additional information, for example gradient and Hessian equations, with each simulation, but that comes with a new set of challenges. This work addresses the challenge of an extremely ill-conditioned correlation matrix for gradient enhanced Kriging. Gradient Enhanced Kriging / Gaussian - Process Models Keith R. Dalbey, PhD Sandia National Laboratories, Albuquerque, NM SAND Number : P Introduction: Kriging ACKNOWLEDGEMENTS REFERENCES Abstract B.A. Lockwood and M. Anitescu. “Gradient-Enhanced Universal Kriging for Uncertainty Propagation.” Nuclear Science and Engineering, To Appear. A. O’Hagan, P. Challenor, D. Cornford, H.P. Wynn, M. Goldstein, and J. Oakley. Managing uncertainty in complex models: a step change in understanding how models perform. accessed July 9, C.E. Rasmussen and C.K.I. Williams. Gaussian processes for machine learning. Adaptive computation and machine learning. MIT Press, The smoothness, or differentiability, of a Kriging model is determined by the choice of correlation function. Two popular choices are the Matern =3/2 function, which makes Kriging once differentiable; Gaussian correlation function, which makes it infinitely differentiable The Gaussian correlation function’s infinite smoothness makes it a good choice when there are few and well-spaced data points (the anticipated case in high dimensions), but can also make numerically singular when the number of sample points, N, is large or the points are poorly spaced. The correlation parameters,, can be found by maximizing the likelihood of the model given the data under the constraint that is non-singular. Correlation Function Challenge, Risk, & Reward To break the Curse of Dimensionality via a GEK emulator, one must exploit: 1.low cost gradients (from automatic differentiation or adjoints), 2.smoothness (if present) in the simulator’s outputs, and 3.high quality sample designs, for example space-filling Latin Hypercube Sampling (LHS) or adaptive sampling. And overcome the challenge of an ill-conditioned matrix. Where is the sample design matrix; contains evaluations of the simulator,, at ; contains the least squares basis functions evaluated at. The correlation matrix,, is real, symmetric, and positive-definite and it contains the correlation function,, evaluated at all pair-wise combinations of points in,. The vector contains the correlations between an arbitrary point,, and all points in,. The maximum likelihood estimates for and are The emulator’s adjusted mean is and the adjusted variance of the estimated distribution is Gradient Enhanced Kriging (GEK) To get gradient enhanced Kriging, start from Kriging and make the following substitutions The superscript on X indicates whether it’s the 1st or 2nd input to r(X 1,X 2 ) A. O’Hagan suggested using the Mahalanobis distance. Sandia National Laboratories is a multi-program laboratory managed and operated by Sandia Corporation, a wholly owned subsidiary of Lockheed Martin Corporation, for the U.S. Department of Energy’s National Nuclear Security Administration under contract DE-AC04-94AL85000.