Abstract:This work considers the computation of risk measures for quantities of interest governed by PDEs with Gaussian random field parameters using Taylor approximations. While efficient, Taylor approximations are local to the point of expansion, and hence may degrade in accuracy when the variances of the input parameters are large. To address this challenge, we approximate the underlying Gaussian measure by a mixture of Gaussians with reduced variance in a dominant direction of parameter space. Taylor approximations are constructed at the means of each Gaussian mixture component, which are then combined to approximate the risk measures. The formulation is presented in the setting of infinite-dimensional Gaussian random parameters for risk measures including the mean, variance, and conditional value-at-risk. We also provide detailed analysis of the approximations errors arising from two sources: the Gaussian mixture approximation and the Taylor approximations. Numerical experiments are conducted for a semilinear advection-diffusion-reaction equation with a random diffusion coefficient field and for the Helmholtz equation with a random wave speed field. For these examples, the proposed approximation strategy can achieve less than $1\%$ relative error in estimating CVaR with only $\mathcal{O}(10)$ state PDE solves, which is comparable to a standard Monte Carlo estimate with $\mathcal{O}(10^4)$ samples, thus achieving significant reduction in computational cost. The proposed method can therefore serve as a way to rapidly and accurately estimate risk measures under limited computational budgets.

Approximating the distributions of estimators of financial risk under an asymmetric Laplace law

Conditional Tail-Related Risk Estimation Using Composite Asymmetric Least Squares and Empirical Likelihood

Data-driven and distribution-free estimation of tailed-related risks for GARCH models using composite asymmetric least squares regression

Risk measures with non-Gaussian fluctuations

Nonasymptotic Estimation of Risk Measures Using Stochastic Gradient Langevin Dynamics

A Non-Gaussian Approach to Risk Measures

Asymptotic Analysis for Spectral Risk Measures Parameterized by Confidence Level

On approximations of value at risk and expected shortfall involving kurtosis

New Bayesian method for estimation of Value at Risk and Conditional Value at Risk

Asymptotic representations for importance-sampling estimators of value-at-risk and conditional value-at-risk

Statistical Learning of Value-at-Risk and Expected Shortfall

Estimating financial risk using piecewise Gaussian processes

Distributional uncertainty of the financial time series measured by G-expectation

Accounting for risk of non linear portfolios: a novel Fourier approach

A hybrid nonparametric multivariate density estimator with applications to risk management

Semi-parametric Dynamic Asymmetric Laplace Models for Tail Risk Forecasting, Incorporating Realized Measures

Smoothing the Conditional Value-at-Risk based Pickands Estimators

Estimating Value at Risk and Expected Shortfall: A Brief Review and Some New Developments

Adaptive Robust Large Volatility Matrix Estimation Based on High-Frequency Financial Data

Nonparametric Expectile Shortfall Regression for Complex Functional Structure

Gaussian mixture Taylor approximations of risk measures constrained by PDEs with Gaussian random field inputs