ML4Science 2025 Workshop

This repository contains all the notebooks for the ML4Science 2025 Workshop.

Getting Started with Google Colab

If you're new to Google Colab, watch this helpful tutorial: Introduction to Google Colab

How to Run the Notebooks

1. Download the notebook from this GitHub repository to your local system
2. Go to Google Colab
3. Upload the notebook: File → Upload notebook and select the downloaded file
4. Choose your runtime: Runtime → Change runtime type and select CPU or GPU as needed
5. Run the cells sequentially using Shift + Enter or run them all together using the commands from the menu.

Datasets

Some notebooks required you to download and upload datasets to Google Colab manually. You can find all such datasets in this Google Drive folder.

Notebooks

Introductory Notebook

The introductory notebook covers Python fundamentals, NumPy, cheminformatics with RDKit, and basic PyTorch concepts. It's designed as pre-work to help you get familiar with the tools and concepts we'll use throughout the workshop.

Please work through this notebook at your own convenience before the workshop. Don't worry if certain sections (especially PyTorch and neural networks) feel unclear—we'll cover all these topics in much greater detail during the workshop with additional notebooks and hands-on exercises. You can always revisit this material later as concepts become clearer!

Day 2

Decision Trees Without Fingerprints

• The Dataset we use for these notebooks is ESol.
• In these notebooks, we will explore decision trees using raw molecular representations instead of traditional fingerprints.
• We will learn to predict properties like LogS and LogP using decision trees.