keeeto/ml-training-resources

ml-training-materials

This is a collection of interactive notebooks covering different aspects of machine learning primarily for chemistry/materials science.

Running locally

To run the notebooks on your local machine, you will want to set up the correct conda environment.

conda env create -f ml-mater-conda.yaml

Contents

• Classical ML models for generic classification
  • classification_decision_tree.ipynb
• Classical ML models for property prediction
  • regression_decision_tree.ipynb
  • sulfides_exercise.ipynb
• Deep neural networks for classifying spectra
  • dnn_for_spectra.ipynb
• Interpretable machine learning
  • interpretable_ml.ipynb

Run in Colab

• Navigate to http://colab.research.google.com/
• Sign in with your Google account
• On the menu that appears choose Github
• In the searchbar type: https://github.com/keeeto/ml-training-materials
• Choose the notebook you want to open
• For the practical; open regression_decision_tree_practical.ipynb

Saving notebooks

• If you want to save and reload the notebooks you will need to save a copy of the original notebook to your Google Drive
• Go to File > Save copy in Drive
• When you want to reload navigate to: http://colab.research.google.com/
• Got to File > Open
• Choose the Google Drive tab, you should now see your saved notebooks.

Credits and Attribution

These notebooks were developed for various projects and courses.

classification_decision_tree.ipynb

Is largely based on material from the excellent Python Data Science Handbook by Jake VanDer Plas and is material that I have used teaching the SciML Introduction to Machine Learning course.

regression_decision_tree.ipynb

Is based on material published in the paper Data-Driven Discovery of Photoactive Quaternary Oxides Using First-Principles Machine Learning.

sulfides_evercise.ipynb

Is an excercise that I developed as part of my course - Machine Learning for Chemists, that I teach at the University of Reading

dnn_for_spectra.ipynb

Was developed as part of the SciML Introduction to Machine Learning course.

interpretable_ml.ipynb

Is material to accompany a chapter on Building Trust in Machine Learning in my book Machine Learning in Materials Science.