Single-to-multi-fidelity history-dependent learning with uncertainty quantification and disentanglement: Application to data-driven constitutive modeling
Published in Computer Methods in Applied Mechanics and Engineering, 2026
Data-driven learning is generalized to consider history-dependent multi-fidelity data while quantifying epistemic uncertainty and disentangling it from aleatoric uncertainty. The methodology is demonstrated on data-driven constitutive modeling scenarios for history-dependent plasticity with multiple fidelities and different noise settings.
Recommended citation: Yi, J., Ferreira, B. P., & Bessa, M. A. (2026). Single-to-multi-fidelity history-dependent learning with uncertainty quantification and disentanglement: Application to data-driven constitutive modeling. Computer Methods in Applied Mechanics and Engineering, 448, 118479.
Download Paper | Download Bibtex