A multi-trajectory-prediction loss implementation in PyTorch

This is our customized implementation of the multi-trajectory-prediction (MTP) loss introduced in cui2019multimodal in PyTorch. The source code was developed for the paper alyaev2021direct (see below).

How to cite:

If you want to adopt the code in your research, please cite the original paper:

Alyaev, S., & Elsheikh, A.H. (2021). Direct multi-modal inversion of geophysical logs using a deep neural network. arXiv 2201.01871, submitted to Earth and Space Science. https://arxiv.org/abs/2201.01871

@article{alyaev2021direct,
  doi = {10.48550/ARXIV.2201.01871},
  url = {https://arxiv.org/abs/2201.01871},
  author = {Alyaev, Sergey and Elsheikh, Ahmed H.},
  title = {Direct multi-modal inversion of geophysical logs using deep learning},
  journal = {arXiv 2201.01871, submitted to Earth and Space Science},
  year = {2021},
  copyright = {Creative Commons Attribution 4.0 International}
}

To cite this code

Please use the link/bibtex to its deposited version:

@software{AlyaevMTPloss,
  author       = {Sergey Alyaev},
  title        = {{alin256/multi-mode-prediction-with-mtp-loss: The 
                   release for archiving of the initial release}},
  year         = 2022,
  publisher    = {Zenodo},
  doi          = {10.5281/zenodo.6531510},
  url          = {https://doi.org/10.5281/zenodo.6531510}
}

Acknowledgements

This work is part of the Center for Research-based Innovation DigiWells: Digital Well Center for Value Creation, Competitiveness and Minimum Environmental Footprint (NFR SFI project no. 309589, DigiWells.no). The center is a cooperation of NORCE Norwegian Research Centre, the University of Stavanger, the Norwegian University of Science and Technology (NTNU), and the University of Bergen, and funded by the Research Council of Norway, Aker BP, ConocoPhillips, Equinor, Lundin, Total, and Wintershall Dea.