Output-Weighted Optimal Sampling for Bayesian Experimental Design and Uncertainty Quantification
From MaRDI portal
Abstract: We introduce a class of acquisition functions for sample selection that leads to faster convergence in applications related to Bayesian experimental design and uncertainty quantification. The approach follows the paradigm of active learning, whereby existing samples of a black-box function are utilized to optimize the next most informative sample. The proposed method aims to take advantage of the fact that some input directions of the black-box function have a larger impact on the output than others, which is important especially for systems exhibiting rare and extreme events. The acquisition functions introduced in this work leverage the properties of the likelihood ratio, a quantity that acts as a probabilistic sampling weight and guides the active-learning algorithm towards regions of the input space that are deemed most relevant. We demonstrate superiority of the proposed approach in the uncertainty quantification of a hydrological system as well as the probabilistic quantification of rare events in dynamical systems and the identification of their precursors.
Recommendations
- Bayesian optimization with output-weighted optimal sampling
- Output-weighted optimal sampling for Bayesian regression and rare event statistics using few samples
- Maximum Entropy Sampling and Optimal Bayesian Experimental Design
- scientific article; zbMATH DE number 2147402
- Computing optimal experimental designs with respect to a compound Bayes risk criterion
- A Laplace method for under-determined Bayesian optimal experimental designs
- Bayesian \(A\)-optimal design of experiment with quantitative and qualitative responses
Cites work
- A Comparison of Three Methods for Selecting Values of Input Variables in the Analysis of Output from a Computer Code
- ADAPTIVE SAMPLING WITH TOPOLOGICAL SCORES
- Active subspaces. Emerging ideas for dimension reduction in parameter studies
- Bayesian Design and Analysis of Computer Experiments: Use of Derivatives in Surface Prediction
- Bayesian experimental design: A review
- Bayesian optimization in a billion dimensions via random embeddings
- Bayesian optimization with output-weighted optimal sampling
- Can small islands protect nearby coasts from tsunamis? An active experimental design approach
- Design and analysis of computer experiments. With comments and a rejoinder by the authors
- Estimating Percentiles of Uncertain Computer Code Outputs
- Extreme event quantification in dynamical systems with random components
- Extremes and recurrence in dynamical systems
- Gaussian processes for machine learning.
- Independent component analysis: recent advances
- Methods of reducing sample size in Monte Carlo computations
- Near-optimal sensor placements in Gaussian processes: theory, efficient algorithms and empirical studies
- Output-weighted optimal sampling for Bayesian regression and rare event statistics using few samples
- Principal Manifolds and Nonlinear Dimensionality Reduction via Tangent Space Alignment
- Rare Event Estimation for Computer Models
- Reduced-order prediction of rogue waves in two-dimensional deep-water waves
- Sequential Design with Mutual Information for Computer Experiments (MICE): Emulation of a Tsunami Model
- Sequential sampling strategy for extreme event statistics in nonlinear dynamical systems
- Uncertainty propagation of p-boxes using sparse polynomial chaos expansions
- Uncertainty quantification of stochastic simulation for black-box computer experiments
- Visualizing data using t-SNE
Cited in
(7)- Multifidelity Bayesian Experimental Design to Quantify Rare-Event Statistics
- Quality measures for the evaluation of machine learning architectures on the quantification of epistemic and aleatoric uncertainties in complex dynamical systems
- Bayesian optimization for active flow control
- Optimal experimental design: formulations and computations
- Output-weighted optimal sampling for Bayesian regression and rare event statistics using few samples
- APIK: Active Physics-Informed Kriging Model with Partial Differential Equations
- Active search for bifurcations
This page was built for publication: Output-Weighted Optimal Sampling for Bayesian Experimental Design and Uncertainty Quantification
Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q4995116)
