Publications Repository - Helmholtz-Zentrum Dresden-Rossendorf

1 Publication

Data publication: Inverse-Dirichlet Weighting Enables Reliable Training of Physics Informed Neural Networks

Maddu, S.; Sturm, D.; Müller, C. L.; Sbalzarini, I. F.

Scripts and notebooks to reproduce the results presented in the paper "Inverse-Dirichlet Weighting Enables Reliable Training of Physics Informed Neural
Networks", Maddu et al., 2021

Keywords: physics informed neural networks; sobolev training; computational physics; deep learning; data-driven modeling; multi-objective optimization; catastrophic forgetting; active turbulence

Related publications

  • Software in external data repository
    Publication year 2021
    Programming language: Python
    System requirements: python, numpy, scipy, sklearn, torch, matplotlib, seaborn, pandas
    License: MIT
    Hosted on GitHub: Link to location

Permalink: https://www.hzdr.de/publications/Publ-33337
Publ.-Id: 33337