Microstructure-based Variational Neural Networks for Robust Uncertainty Quantification in Materials Digital Twins

2025/12/19 by Robertson, Andreas E., Inman, Samuel B., Lenau, Ashley T. +4
#FOS: Computer and information sciences #FOS: Physical sciences #Machine Learning (cs.LG) #Materials Science (cond-mat.mtrl-sci)

paper · doi:10.48550/arxiv.2512.18104

Abstract

Aleatoric uncertainties - irremovable variability in microstructure morphology, constituent behavior, and processing conditions - pose a major challenge to developing uncertainty-robust digital twins. We introduce the Variational Deep Material Network (VDMN), a physics-informed surrogate model that enables efficient and probabilistic forward and inverse predictions of material behavior. The VDMN captures microstructure-induced variability by embedding variational distributions within its hierarchical, mechanistic architecture. Using an analytic propagation scheme based on Taylor-series expansion and automatic differentiation, the VDMN efficiently propagates uncertainty through the network during training and prediction. We demonstrate its capabilities in two digital-twin-driven applications: (1) as an uncertainty-aware materials digital twin, it predicts and experimentally validates the nonlinear mechanical variability in additively manufactured polymer composites; and (2) as an inverse calibration engine, it disentangles and quantitatively identifies overlapping sources of uncertainty in constituent properties. Together, these results establish the VDMN as a foundation for uncertainty-robust materials digital twins.

Citations

• Orientation-aware interaction-based deep material network in polycrystalline materials modeling
• PSP-GEN: Stochastic inversion of the Process-Structure-Property chain in materials design through deep, generative probabilistic modeling
• Bayesian neural networks for predicting uncertainty in full-field material response
• Probabilistic Neural Networks (PNNs) for Modeling Aleatoric Uncertainty in Scientific Machine Learning
• Estimating Epistemic and Aleatoric Uncertainty with a Single Model
• Clarifying Trust of Materials Property Predictions using Neural Networks with Distribution-Specific Uncertainty Quantification
• Calibrated Reliable Regression using Maximum Mean Discrepancy
• Gaussians on Riemannian Manifolds: Applications for Robot Learning and\n Adaptive Control
• On the Convergence of Adam and Beyond
• SGDR: Stochastic Gradient Descent with Warm Restarts
• Minimax and minimax projection designs using clustering
• Generative Moment Matching Networks
• Adam: A Method for Stochastic Optimization

Related

• Uncertainty Quantification in Multiscale Modeling of Polymer Composite Materials Using Physically Recurrent Neural Networks
• Stochastic Deep-Ritz for Parametric Uncertainty Quantification
• GUST: Quantifying Free-Form Geometric Uncertainty of Metamaterials Using Small Data
• Leveraging Uncertainty from Deep Learning for Trustworthy Materials Discovery Workflows
• Deep learning for size-agnostic inverse design of random-network 3D printed mechanical metamaterials