Adaptive weighting of Bayesian physics informed neural networks for multitask and multiscale forward and inverse problems
DOI10.1016/j.jcp.2023.112342arXiv2302.12697OpenAlexW4383558850MaRDI QIDQ6095075
Suryanarayana Maddu, Sarah Perez, Philippe Poncet, Ivo F. Sbalzarini
Publication date: 27 November 2023
Published in: Journal of Computational Physics (Search for Journal in Brave)
Full work available at URL: https://arxiv.org/abs/2302.12697
artificial intelligenceuncertainty quantificationHamiltonian Monte CarloBayesian physics-informed neural networksadaptive weight learningmulti-objective training
Parametric inference (62Fxx) Artificial intelligence (68Txx) Probabilistic methods, stochastic differential equations (65Cxx)
Cites Work
- Unnamed Item
- The No-U-Turn Sampler: Adaptively Setting Path Lengths in Hamiltonian Monte Carlo
- Discovering governing equations from data by sparse identification of nonlinear dynamical systems
- Using DC PSE operator discretization in Eulerian meshless collocation methods improves their robustness in complex geometries
- Sobolev training of thermodynamic-informed neural networks for interpretable elasto-plasticity models with level set hardening
- Optimally weighted loss functions for solving PDEs with neural networks
- A velocity-vorticity method for highly viscous 3D flows with application to digital rock physics
- B-PINNs: Bayesian physics-informed neural networks for forward and inverse PDE problems with noisy data
- Multi-fidelity Bayesian neural networks: algorithms and applications
- Physics-informed semantic inpainting: application to geostatistical modeling
- When and why PINNs fail to train: a neural tangent kernel perspective
- Multi-variance replica exchange SGMCMC for inverse and forward problems via Bayesian PINN
- Surrogate modeling for fluid flows based on physics-constrained deep learning without simulation data
- Adaptive multi-fidelity polynomial chaos approach to Bayesian inference in inverse problems
- Adversarial uncertainty quantification in physics-informed neural networks
- Physics-informed neural networks: a deep learning framework for solving forward and inverse problems involving nonlinear partial differential equations
- On the geometric ergodicity of Hamiltonian Monte Carlo
- The geometric foundations of Hamiltonian Monte Carlo
- Bayesian physics informed neural networks for real-world nonlinear dynamical systems
- Uncertainty quantification in scientific machine learning: methods, metrics, and comparisons
- Spatially distributed control for optimal drag reduction of the flow past a circular cylinder
- Riemann Manifold Langevin and Hamiltonian Monte Carlo Methods
- Learning partial differential equations via data discovery and sparse optimization
- Understanding and Mitigating Gradient Flow Pathologies in Physics-Informed Neural Networks
- Hidden fluid mechanics: Learning velocity and pressure fields from flow visualizations
- An Adaptive Surrogate Modeling Based on Deep Neural Networks for Large-Scale Bayesian Inverse Problems
- Contribution to the Discussion of the Paper ‘Geodesic Monte Carlo on Embedded Manifolds’
This page was built for publication: Adaptive weighting of Bayesian physics informed neural networks for multitask and multiscale forward and inverse problems