Vecchia approximated Bayesian heteroskedastic Gaussian processes

Many computer simulations are stochastic and exhibit input dependent noise. In such situations, heteroskedastic Gaussian processes (hetGPs) make ideal surrogates as they estimate a latent, non-constant variance. However, existing hetGP implementations are unable to deal with large simulation campaigns and use point-estimates for all unknown quantities, including latent variances. This limits applicability to small experiments and undercuts uncertainty. We propose a Bayesian hetGP using elliptical slice sampling (ESS) for posterior variance integration, and the Vecchia approximation to circumvent computational bottlenecks. We show good performance for our upgraded hetGP capability, compared to alternatives, on a benchmark example and a motivating corpus of more than 9-million lake temperature simulations. An open source implementation is provided as bhetGP on CRAN.

• Publication year

  2025

• Venue

  Unknown venue

• Publication date

  2025-07-10

• Fields of study

  Mathematics, Computer Science

• Identifiers
  arXiv 2507.07815
• External record

  Open on Semantic Scholar

• Source metadata

  Semantic Scholar

• No claims are published for this paper.

• No concepts are published for this paper.

• hetGPy: Heteroskedastic Gaussian Process Modeling in Python

  2025cited by this paper
• Modernizing full posterior inference for surrogate modeling of categorical-output simulation experiments

  2025cited by this paper
• Data assimilation experiments inform monitoring needs for near‐term ecological forecasts in a eutrophic reservoir

  2024cited by this paper
• Synthesizing data products, mathematical models, and observational measurements for lake temperature forecasting

  2024influential reference
• Progress and opportunities in advancing near‐term forecasting of freshwater quality

  2023cited by this paper
• Contour location for reliability in airfoil simulation experiments using deep Gaussian processes

  2023cited by this paper
• Data-driven models for predicting community changes in freshwater ecosystems: A review

  2023cited by this paper
• Near‐term forecasts of NEON lakes reveal gradients of environmental predictability across the US

  2023cited by this paper
• Vecchia-Approximated Deep Gaussian Processes for Computer Experiments

  2022influential reference
• The Reanalysis for the Global Ensemble Forecast System, version 12.

  2021cited by this paper
• Lake heatwaves under climate change

  2021cited by this paper
• Nearest‐neighbor sparse Cholesky matrices in spatial statistics

  2021cited by this paper
• Advancing lake and reservoir water quality management with near-term, iterative ecological forecasting

  2021cited by this paper
• Locally induced Gaussian processes for large-scale simulation experiments

  2020cited by this paper
• Surrogates: Gaussian Process Modeling, Design, and Optimization for the Applied Sciences

  2020cited by this paper
• Knot selection in sparse Gaussian processes with a variational objective function

  2020cited by this paper
• Active Learning for Deep Gaussian Process Surrogates

  2020influential reference
• Scaled Vecchia Approximation for Fast Computer-Model Emulation

  2020influential reference
• Analyzing Stochastic Computer Models: A Review with Opportunities

  2020cited by this paper
• A General Lake Model (GLM 3.0) for linking with high-frequency sensor data from the Global Lake Ecological Observatory Network (GLEON)

  2019cited by this paper
• hetGP: Heteroskedastic Gaussian Process Modeling and Sequential Design in R

  2019cited by this paper
• Vecchia Approximations of Gaussian-Process Predictions

  2018cited by this paper
• Large-Scale Heteroscedastic Regression via Gaussian Process

  2018influential reference
• A multi-lake comparative analysis of the General Lake Model (GLM): Stress-testing across a global observatory network

  2018cited by this paper
• A General Framework for Vecchia Approximations of Gaussian Processes

  2017cited by this paper
• Calibrating a Stochastic, Agent-Based Model Using Quantile-Based Emulation

  2017cited by this paper
• Replication or Exploration? Sequential Design for Stochastic Simulation Experiments

  2017cited by this paper
• Chained Gaussian Processes

  2016influential reference
• laGP: Large-Scale Spatial Modeling via Local Approximate Gaussian Processes in R

  2016cited by this paper
• Practical Heteroscedastic Gaussian Process Modeling for Large Simulation Experiments

  2016influential reference
• Read and Write MAT Files and Call MATLAB from Within R

  2016cited by this paper
• Sequential Design for Ranking Response Surfaces

  2015cited by this paper
• The rise of novelty in ecosystems.

  2015cited by this paper
• Modeling satellite drag coefficients with response surfaces

  2014cited by this paper
• Simulating 2368 temperate lakes reveals weak coherence in stratification phenology

  2014cited by this paper
• RcppArmadillo: Accelerating R with high-performance C++ linear algebra

  2014cited by this paper
• Hierarchical Nearest-Neighbor Gaussian Process Models for Large Geostatistical Datasets

  2014cited by this paper
• Estimating Ocean Circulation: An MCMC Approach With Approximated Likelihoods via the Bernoulli Factory

  2014cited by this paper
• Bayesian and Maximum Likelihood Estimation for Gaussian Processes on an Incomplete Lattice

  2014cited by this paper
• Statistical Properties of Covariance Tapers

  2013cited by this paper
• Local Gaussian Process Approximation for Large Computer Experiments

  2013cited by this paper
• Eco-physiological adaptations that favour freshwater cyanobacteria in a changing climate.

  2012cited by this paper
• Bayesian Kriging Analysis and Design for Stochastic Simulations

  2012cited by this paper
• Blueprint for a unifying framework for synthesis of aquatic ecodynamics

  2012cited by this paper
• Matrix Computations

  2011cited by this paper
• Variational Heteroscedastic Gaussian Process Regression

  2011cited by this paper
• Spatio‐temporal smoothing and EM estimation for massive remote‐sensing data sets

  2011cited by this paper
• Rcpp: Seamless R and C++ Integration

  2011cited by this paper
• Bayesian Gaussian Process Latent Variable Model

  2010cited by this paper
• Cases for the nugget in modeling computer experiments

  2010cited by this paper
• Elliptical slice sampling

  2009cited by this paper
• The kriging update equations and their application to the selection of neighboring data

  2009cited by this paper
• Climate change: a catalyst for global expansion of harmful cyanobacterial blooms.

  2009cited by this paper
• Kernel Conditional Quantile Estimation via Reduction Revisited

  2009cited by this paper
• Gaussian predictive process models for large spatial data sets

  2008cited by this paper
• Fixed rank kriging for very large spatial data sets

  2008cited by this paper
• Covariance Tapering for Likelihood-Based Estimation in Large Spatial Data Sets

  2008cited by this paper
• Stochastic kriging for simulation metamodeling

  2008influential reference
• Strictly Proper Scoring Rules, Prediction, and Estimation

  2007cited by this paper
• Most likely heteroscedastic Gaussian process regression

  2007cited by this paper
• Discrete Optimization via Simulation Using COMPASS

  2006cited by this paper
• Microcolony and biofilm formation as a survival strategy for bacteria.

  2006cited by this paper
• Covariance Tapering for Interpolation of Large Spatial Datasets

  2006cited by this paper
• Approximating likelihoods for large spatial data sets

  2004cited by this paper
• Modern applied statistics with S, 4th Edition

  2002cited by this paper
• Objective Bayesian Analysis of Spatially Correlated Data

  2001cited by this paper
• Ecological forecasts: an emerging imperative.

  2001cited by this paper
• Bayesian calibration of computer models

  2001cited by this paper
• Interpolation of Spatial Data: Some Theory for Kriging

  2000cited by this paper
• Slice Sampling

  2000cited by this paper
• Design and analysis of computer experiments

  1998cited by this paper
• Efficient Global Optimization of Expensive Black-Box Functions

  1998cited by this paper
• Regression with Input-dependent Noise: A Gaussian Process Treatment

  1997influential reference
• Estimation and model identification for continuous spatial processes

  1988cited by this paper
• Maximum likelihood from incomplete data via the EM - algorithm plus discussions on the paper

  1977cited by this paper
• Principles of geostatistics

  1963cited by this paper

• No citing papers are available for this paper.