Python package to model and to perform topology optimization for graphene kirigami using deep learning
Python package to model and to perform topology optimization for graphene kirigami using deep learning. We use convolutional neural networks (similar to VGGNet architecure) for regression.
See our published paper:
generate_LAMMPS_input/generate_LAMMPS_configuration_input.ipynb
. New methods to generate parallel cuts are now avalaible.models/regression_CNN/convert_coarse_to_fine.ipynb
models/regression_CNN/tf_fgrid_dnn_validtrain.py
models/regression_CNN/tf_cnn_search_large_v2.py
models/simple/simple_machine_learning.ipynb
mddata
. This dataset generated using AIREBO potential with 1.7 mincutoff which is the default of CH.airebo.models_supervisedAutoencoder_forwardInverseDesign/supervisedAE_for_kirigamiDesign.ipynb
. See notebook for details of the code.This package is still under developement. More features will be added soon.
git clone https://github.com/phanakata/ML_for_kirigami_design.git
Paul Hanakata
If you use this package/code/dataset, build on or find our research is useful for your work please cite
@article{hanakata-PhysRevLett.121.255304,
title = {Accelerated Search and Design of Stretchable Graphene Kirigami Using Machine Learning},
author = {Hanakata, Paul Z. and Cubuk, Ekin D. and Campbell, David K. and Park, Harold S.},
journal = {Phys. Rev. Lett.},
volume = {121},
issue = {25},
pages = {255304},
numpages = {6},
year = {2018},
month = {Dec},
publisher = {American Physical Society},
doi = {10.1103/PhysRevLett.121.255304},
url = {https://link.aps.org/doi/10.1103/PhysRevLett.121.255304}
}
@article{PhysRevResearch.2.042006,
title = {Forward and inverse design of kirigami via supervised autoencoder},
author = {Hanakata, Paul Z. and Cubuk, Ekin D. and Campbell, David K. and Park, Harold S.},
journal = {Phys. Rev. Research},
volume = {2},
issue = {4},
pages = {042006},
numpages = {6},
year = {2020},
month = {Oct},
publisher = {American Physical Society},
doi = {10.1103/PhysRevResearch.2.042006},
url = {https://link.aps.org/doi/10.1103/PhysRevResearch.2.042006}
}