PSINN is a neural-network-based dynamics estimation project. It uses a neural network (NN) to learn low-dimensional shape functions (denoted with the Greek letter psi) that are composited together into an estimate of the dynamics operator.
The neural network infrastructure for PSINN is TensorFlow (https://www.tensorflow.org/)
PSINN was developed by Majerle Reeves (LLNL & UC Merced; https://appliedmath.ucmerced.edu/content/majerle-reeves).
TBD
TBD
Please contact Majerle Reeves or Harish Bhat if interested in PSINN and related work.
PSINN is distributed under the terms of the MIT license. All new contributions must be made under this license.
No version of the code shall appear without the LICENSE and NOTICE files. No version of the code shall appear without the LLNL Release Number as given in the section "Release" below.
PSINN was developed under the auspices of the U. S. Department of Energy by Lawrence Livermore National Laboratory under Contract DE-AC52-07NA27344 (see NOTICES).
LLNL Release Number: LLNL-CODE-790441
SPDX-License-Identifier: MIT