Scholarlink

Abstract <p>Tightly coupled multi-compartment epidemic models tend to expose a weakness of standard Physics-Informed Neural Networks (PINNs). When collocation points are placed uniformly, the network spends much of its capacity on smooth regions while leaving the sharp transients poorly resolved. We work around this by pairing Failure-Informed PINNs (FI-PINNs) with a Self-Adaptive Importance Sampling (SAIS) refinement strategy, and we apply the combination to a nine-equation Susceptible/Vaccinated/Exposed/Infected/Recovered (SVEIR) model that follows three viral strains together with a vaccination compartment. The idea behind the construction is simple. We build a residual-based limit-state function, estimate the failure probability associated with it, and let the network steer its own sampling toward the time intervals where the governing equations are not yet satisfied to within a prescribed tolerance. On the same temporal domain, with identical initial conditions and the same network architecture, SAIS-enhanced FI-PINNs reach a relative <InlineEquation ID="IEq1"><EquationSource Format="TEX">\(L_2\)</EquationSource></InlineEquation> error of <InlineEquation ID="IEq2"><EquationSource Format="TEX">\(1.04\times 10^{-6}\)</EquationSource></InlineEquation>, against <InlineEquation ID="IEq3"><EquationSource Format="TEX">\(7.36\times 10^{-3}\)</EquationSource></InlineEquation> for uniform sampling and <InlineEquation ID="IEq4"><EquationSource Format="TEX">\(3.42\times 10^{-4}\)</EquationSource></InlineEquation> for Residual-Based Adaptive Refinement (RAR)&#xa0;[Lu et. al,. 63:208-228, 2021], an improvement of two to three orders of magnitude. The benefits go beyond raw accuracy. The failure-probability estimate gives an interpretable stopping criterion, and the truncated Gaussian proposal of SAIS keeps the sampling-bias risk under control, a risk that affects purely greedy residual-based schemes. The framework is general enough to be transported to other coupled compartmental systems and to data-assimilation settings in which partial observations are available.</p> Mathematics subject classification <p>92D30 , 65L05 , 68T07 , 00A71</p>

错误：搜索内容不能为空，请输入英文关键词
错误：关键词超出字数限制，请精简
高级检索

Failure-informed PINNs for a multi-strain SVEIR epidemic model

  • 期刊论文

  • 出版模式：

    开放获取

  • 发表日期：

    2026年8月3日

查看全文
  • Kawtar Idhammou Ouyoussef,
  • Jaafar El Karkri,
  • Léon Matar Tine,
  • Rajae Aboulaich

摘要

Abstract

Tightly coupled multi-compartment epidemic models tend to expose a weakness of standard Physics-Informed Neural Networks (PINNs). When collocation points are placed uniformly, the network spends much of its capacity on smooth regions while leaving the sharp transients poorly resolved. We work around this by pairing Failure-Informed PINNs (FI-PINNs) with a Self-Adaptive Importance Sampling (SAIS) refinement strategy, and we apply the combination to a nine-equation Susceptible/Vaccinated/Exposed/Infected/Recovered (SVEIR) model that follows three viral strains together with a vaccination compartment. The idea behind the construction is simple. We build a residual-based limit-state function, estimate the failure probability associated with it, and let the network steer its own sampling toward the time intervals where the governing equations are not yet satisfied to within a prescribed tolerance. On the same temporal domain, with identical initial conditions and the same network architecture, SAIS-enhanced FI-PINNs reach a relative \(L_2\) error of \(1.04\times 10^{-6}\), against \(7.36\times 10^{-3}\) for uniform sampling and \(3.42\times 10^{-4}\) for Residual-Based Adaptive Refinement (RAR) [Lu et. al,. 63:208-228, 2021], an improvement of two to three orders of magnitude. The benefits go beyond raw accuracy. The failure-probability estimate gives an interpretable stopping criterion, and the truncated Gaussian proposal of SAIS keeps the sampling-bias risk under control, a risk that affects purely greedy residual-based schemes. The framework is general enough to be transported to other coupled compartmental systems and to data-assimilation settings in which partial observations are available.

Mathematics subject classification

92D30 , 65L05 , 68T07 , 00A71