Machine-learning Tutorials in Molecular Simulations -made simple-

Welcome to the GitHub repository for Machine-learning Tutorials in Molecular Simulations (ML4molecularSims) This is a collection of Tutorials and Solution to Exercises for ML enthusiasts, focused on ML algorithms in Molecular Simulation. The Tutorials are aimed to reach multiple backgrounds in physics, chemistry biology, computer and material sciences.

Contents

• Tutorial 1: Fundamentals of deep learning
• Tutorial 2: Free energy calculation: SchNet regression arquitecture
• Tutorial 3: SchNet regression with PyTorch Geometric
• Tutorial 4: Comparison between PyTorch Geometric and TensorFlow implementations
• Tutorial 5: DimeNet architecture
• Tutorial 6: Molecular kinetics using Neural Networks - VAMPnets
• Tutorial 7: How to train a neural network potential
• Tutorial 8: Comparison between ML algorithms of generations 3rd and 4th
• Tutorial 9: Methods for dimensionality reduction
• Tutorial 10: Generative artificial intelligence for Molecular Simulations

We aim to describe the basic and practical aspects of the ML algorithms by showing all possible math details and the most common pitfalls. Note also that current *References to research articles have been included, which are currently dynamically evolving and will be updated periodically. The root project PyF4all has been presented at multiple conferences (PyCon, SciPy, and ScipyLA).

Getting Started

Head over to the official website!

Building community

If you find the material useful, then support this initiative and their contributors by:

1. Sharing it around on Twitter!
2. Reporting any Bugs or Issues
3. Sharing it with your colleagues
4. Citing the SciPy 2020 presentation!