Model Uncertainty and Correctability for Directed Graphical Models
From MaRDI portal
Bayesian networkscorrectabilityinformation inequalitiessensitivity analysisstress testsuncertainty quantification
Probabilistic graphical models (62H22) Applications of statistics in engineering and industry; control charts (62P30) Reasoning under uncertainty in the context of artificial intelligence (68T37) Chemical kinetics in thermodynamics and heat transfer (80A30) Sensitivity (robustness) (93B35) Measures of information, entropy (94A17)
Abstract: Probabilistic graphical models are a fundamental tool in probabilistic modeling, machine learning and artificial intelligence. They allow us to integrate in a natural way expert knowledge, physical modeling, heterogeneous and correlated data and quantities of interest. For exactly this reason, multiple sources of model uncertainty are inherent within the modular structure of the graphical model. In this paper we develop information-theoretic, robust uncertainty quantification methods and non-parametric stress tests for directed graphical models to assess the effect and the propagation through the graph of multi-sourced model uncertainties to quantities of interest. These methods allow us to rank the different sources of uncertainty and correct the graphical model by targeting its most impactful components with respect to the quantities of interest. Thus, from a machine learning perspective, we provide a mathematically rigorous approach to correctability that guarantees a systematic selection for improvement of components of a graphical model while controlling potential new errors created in the process in other parts of the model. We demonstrate our methods in two physico-chemical examples, namely quantum scale-informed chemical kinetics and materials screening to improve the efficiency of fuel cells.
Recommendations
- The uncertainty of a selected graphical model
- scientific article; zbMATH DE number 1531986
- Graphical models for imprecise probabilities
- Bayesian Models for Directed Graphs
- GRAPHICAL MODELS IN MACHINE LEARNING, NETWORKS AND UNCERTAINTY QUANTIFICATION
- Variations on undirected graphical models and their relationships
- Structural Markov graph laws for Bayesian model uncertainty
- Inference in a class of directed random graph models with an increasing number of parameters
Cites work
- (f,)-divergences: interpolating between f-divergences and integral probability metrics
- All of Nonparametric Statistics
- Causality and Bayesian network PDEs for multiscale representations of porous media
- Data-driven chance constrained stochastic program
- Data-driven distributionally robust optimization using the Wasserstein metric: performance guarantees and tractable reformulations
- Distinguishing and integrating aleatoric and epistemic variation in uncertainty quantification
- Distributionally Robust Convex Optimization
- Distributionally robust optimization and its tractable approximations
- Distributionally robust optimization under moment uncertainty with application to data-driven problems
- Divergence Estimation for Multidimensional Densities Via k-Nearest-Neighbor Distances
- Elements of Information Theory
- Estimating Divergence Functionals and the Likelihood Ratio by Convex Risk Minimization
- Explanation in artificial intelligence: insights from the social sciences
- Formulation and properties of a divergence used to compare probability measures without absolute continuity
- GINNs: graph-informed neural networks for multiscale physics
- How Biased Is Your Model? Concentration Inequalities, Information and Model Bias
- scientific article; zbMATH DE number 48812 (Why is no real title available?)
- scientific article; zbMATH DE number 1153603 (Why is no real title available?)
- scientific article; zbMATH DE number 2038320 (Why is no real title available?)
- scientific article; zbMATH DE number 4189254 (Why is no real title available?)
- Integral Probability Metrics and Their Generating Classes of Functions
- On deterministic reformulations of distributionally robust joint chance constrained optimization problems
- Optimal uncertainty quantification
- Path-space information bounds for uncertainty quantification and sensitivity analysis of stochastic dynamics
- Pattern recognition and machine learning.
- Probabilistic graphical models.
- Quantifying distributional model risk via optimal transport
- Robust Actuarial Risk Analysis
- Robust Information Divergences for Model-Form Uncertainty Arising from Sparse Data in Random PDE
- Robustness
- ROME
- Scalable information inequalities for uncertainty quantification
- Sensitivity analysis for rare events based on Rényi divergence
- Sensitivity Analysis in Gaussian Bayesian Networks Using a Divergence Measure
- The effect of block parameter perturbations in Gaussian Bayesian networks: sensitivity and robustness
- The vectorization of ITPACK 2C
- Uncertainty quantification for Markov processes via variational principles and functional inequalities
- Uncertainty quantification for Markov random fields
- Variational representations and neural network estimation of Rényi divergences
Cited in
(4)
This page was built for publication: Model Uncertainty and Correctability for Directed Graphical Models
Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q5052911)
