CODE2 Lab

Research Direction 03

Uncertainty-aware reliable design

Quantifying what models know, what they do not know, and how variability affects engineering decisions.

← All research directions
Uncertainty propagation through the process, structure, property, and performance chain
Uncertainty propagates through the process-structure-property-performance chain.

We develop uncertainty-aware methods for materials and mechanical systems whose available data are sparse, noisy, or heterogeneous. Statistical learning and physics-informed models quantify and propagate both aleatoric uncertainty from data and epistemic uncertainty from models, replacing isolated point predictions with calibrated confidence. Out-of-distribution detection identifies unfamiliar conditions, while active learning, adaptive model refinement, and robust optimization use uncertainty to guide new data collection and risk-aware decisions. The goal is to produce transparent, trustworthy predictions and designs that maintain reliable performance under real-world variability.

Selected journal articles

  1. Wang, Z., Bhaduri, A., Xu, H., Wang, L. "An Uncertainty-Aware Deep Learning Framework-Based Robust Design Optimization of Metamaterial Units." Structural and Multidisciplinary Optimization, 2025. Link
  2. Wang, Z., Daeipour, M., Xu, H. "Quantification and Propagation of Aleatoric Uncertainties in Topological Structures." Reliability Engineering & System Safety, 2023. Link
  3. Wang, Z., Xu, H. "Quantitative Representation of Aleatoric Uncertainties in Network-Like Topological Structural Systems." Journal of Mechanical Design, 2021. Link
  4. Cleeman, J., Jackson, A., Patel, A., Wang, Z., et al. "Operational Resilience of Additively Manufactured Parts to Stealthy Cyberphysical Attacks Using Geometric and Process Digital Twins." Journal of Manufacturing Systems, 2025. Link