Multi-Fidelity Physics-Informed Neural Networks with Bayesian Uncertainty Quantification and Adaptive Residual Learning for Efficient Solution of Parametric Partial Differential Equations

Olaf Yunus Laitinen Imanov · 2026

Read Multi-Fidelity Physics-Informed Neural Networks with Bayesian Uncertainty Quantification and Adaptive Residual Learning for Efficient Solution of Parametric Partial Differential Equations by Olaf Yunus Laitinen Imanov (2026) in the Tower of Babel, a free online library of public-domain works preserved in clean, searchable text. No sign-up. The full text is served as a plain-text file at a predictable URL, where text is present: https://antarctic-labs.com/tower-of-babel/arxiv-2602-01176.txt.

Description

Yunus Laitinen Imanov and co-authors ask when cheap physics truly saves costly high-grade runs for partial differential equations. High-grade runs cost much, while cheap coarse or plain-physics runs skew results, and past mixed plans fail in silence when cheap runs mislead. Their sharp sample bound shows learning the gap from cheap runs cuts high-grade runs by a set factor tied to gap size. Their gated GaMF-PINN learns where to trust cheap runs, falls back to pure physics elsewhere, and adds honest doubt bands.

Catalog record

CREATOROlaf Yunus Laitinen Imanov
YEAR2026
COLLECTIONPAPERS
FORMATTXT
FILE SIZE31860
LICENSEhttp://creativecommons.org/licenses/by/4.0/
RIGHTSCC BY
SOURCEarXiv

Source: https://arxiv.org/abs/2602.01176

Browse the full library · About the Tower of Babel