From 221517ff0c9173886894a54c1bca7f7b2ebd809a Mon Sep 17 00:00:00 2001
From: m-julian <52214154+m-julian@users.noreply.github.com>
Date: Thu, 16 Jan 2025 17:08:43 +0200
Subject: [PATCH] add directory

---
 .../water_MLP.ipynb                           | 42 +++++++++++++++++++
 1 file changed, 42 insertions(+)
 create mode 100644 jupyter_notebooks/neural_networks/machine_learning_potentials/water_MLP.ipynb

diff --git a/jupyter_notebooks/neural_networks/machine_learning_potentials/water_MLP.ipynb b/jupyter_notebooks/neural_networks/machine_learning_potentials/water_MLP.ipynb
new file mode 100644
index 0000000..25c9b17
--- /dev/null
+++ b/jupyter_notebooks/neural_networks/machine_learning_potentials/water_MLP.ipynb
@@ -0,0 +1,42 @@
+{
+ "cells": [
+  {
+   "cell_type": "markdown",
+   "metadata": {},
+   "source": [
+    "# Training a simple Neural Network-based Machine Learning Potential for a Water Molecule\n",
+    "\n",
+    "This notebook will cover the construction of a very simple NN-based machine learning potential for water. The goal is to be able to use the resulting model for gas phase simulations of a single water molecule. The model that will be introduced here only uses the energy for training (atomic forces are calculated by backpropagation). In practice, it is a very good idea to also add force information as additional outputs to the NN to make force predictions much more robust. Note that the force predictions will still be calculated by taking the gradient of the energy predictions of the NN with respect to the NN inputs.\n",
+    "\n",
+    "All scripts to generate the dataset and make the models are provided in this repository:\n",
+    "\n",
+    "TODO: insert repository"
+   ]
+  },
+  {
+   "cell_type": "markdown",
+   "metadata": {},
+   "source": [
+    "## Generating Quantum Mechanical Dataset\n",
+    "\n"
+   ]
+  },
+  {
+   "cell_type": "markdown",
+   "metadata": {},
+   "source": []
+  },
+  {
+   "cell_type": "markdown",
+   "metadata": {},
+   "source": []
+  }
+ ],
+ "metadata": {
+  "language_info": {
+   "name": "python"
+  }
+ },
+ "nbformat": 4,
+ "nbformat_minor": 2
+}