Neural networks for parameter estimation in intractable models
From MaRDI portal
Abstract: We propose to use deep learning to estimate parameters in statistical models when standard likelihood estimation methods are computationally infeasible. We show how to estimate parameters from max-stable processes, where inference is exceptionally challenging even with small datasets but simulation is straightforward. We use data from model simulations as input and train deep neural networks to learn statistical parameters. Our neural-network-based method provides a competitive alternative to current approaches, as demonstrated by considerable accuracy and computational time improvements. It serves as a proof of concept for deep learning in statistical parameter estimation and can be extended to other estimation problems.
Recommendations
- Solving inverse problems in stochastic models using deep neural networks and adversarial training
- Bayesian estimation of large-scale simulation models with Gaussian process regression surrogates
- Parameter estimation for computationally intensive nonlinear regression with an application to climate modeling
- Learning Summary Statistic for Approximate Bayesian Computation via Deep Neural Network
- Fast approximate Bayesian computation for estimating parameters in differential equations
Cites work
- A spectral representation for max-stable processes
- A two-step approach to model precipitation extremes in California based on max-stable and marginal point processes
- Adaptive approximate Bayesian computation
- An Introduction to Random Sets
- An overview of composite likelihood methods
- Approximate Bayesian computing for spatial extremes
- Bayesian inference for the Brown-Resnick process, with an application to extreme low temperatures
- Bayesian inference from composite likelihoods, with an application to spatial extremes
- Bivariate extreme value theory: Models and estimation
- Constructing summary statistics for approximate Bayesian computation: semi-automatic approximate Bayesian computation. With discussion and authors' reply
- Efficient inference and simulation for elliptical Pareto processes
- Exploiting occurrence times in likelihood inference for componentwise maxima
- scientific article; zbMATH DE number 2042687 (Why is no real title available?)
- Learning Summary Statistic for Approximate Bayesian Computation via Deep Neural Network
- Likelihood estimators for multivariate extremes
- Likelihood-based inference for max-stable processes
- Models for stationary max-stable random fields
- On the likelihood function of Gaussian max-stable processes
- On the robustness of maximum composite likelihood estimate
- Scalable Bayesian inference for the inverse temperature of a hidden Potts model
- Stationary max-stable fields associated to negative definite functions
- Statistical inference for max-stable processes in space and time
- Statistical modeling of spatial extremes
- Strictly Proper Scoring Rules, Prediction, and Estimation
- The elements of statistical learning. Data mining, inference, and prediction
- The frontier of simulation-based inference
- Variograms for spatial max-stable random fields
Cited in
(26)- Global estimation of feedforward networks with a priori constraints
- Likelihood ratio of unidentifiable models and multilayer neural networks
- scientific article; zbMATH DE number 49542 (Why is no real title available?)
- Robust estimation of parametric membership regions
- Comparison of four gradient-learning algorithms for neural network Wiener models
- Parameter inference with deep jointly informed neural networks
- scientific article; zbMATH DE number 7745105 (Why is no real title available?)
- Change point detection via feedforward neural networks with theoretical guarantees
- Likelihood-Free Parameter Estimation with Neural Bayes Estimators
- Fast parameter estimation of generalized extreme value distribution using neural networks
- A physics-inspired neural network for short-wave radiation parameterization
- A supervised learning approach to estimating IRT models in small samples
- A generalized unified skew-normal process with neural Bayes inference
- Properties and estimation of a novel skewed generalized normal distribution
- Neural Bayes Estimators for Irregular Spatial Data Using Graph Neural Networks
- Towards black-box parameter estimation
- Flexible Modeling of Nonstationary Extremal Dependence using Spatially Fused LASSO and Ridge Penalties
- Efficient Large-Scale Nonstationary Spatial Covariance Function Estimation Using Convolutional Neural Networks
- Vecchia Likelihood Approximation for Accurate and Fast Inference with Intractable Spatial Max-Stable Models
- Modern extreme value statistics for Utopian extremes. EVA (2023) Conference Data Challenge: Team Yalla
- Modeling anisotropy and non-stationarity through physics-informed spatial regression
- Mathematics, statistics, and geometry of extreme events in high dimensions. Abstracts from the workshop held August 19--23, 2024
- Neural networks for Bayesian inverse problems governed by a nonlinear ODE
- Neural Bayes estimation and selection of complex bivariate extremal dependence models
- A new mixture model for spatiotemporal exceedances with flexible tail dependence
- Demonstrating the power and flexibility of variational assumptions for amortized neural posterior estimation in environmental applications
This page was built for publication: Neural networks for parameter estimation in intractable models
Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q6115548)
