Tutorials on explainable machine learning, artifical intelligence in general and in bioinformatics in biomedical research
In this repository, you will find tutorials about artifical intelligence applied in medical research.
The repository is divided into differents sections where each section contains tutorials about an erea in data science field like NLP, Image processing and analysis, graph etc. Fill free to propose update or to collaborate. I tried to do all the code in colab to make it easy for reproducibility. Some code are not mine!.
- Documentation: https://laminetourelab.github.io
- Contributing: https://github.com/LamineTourelab/Tutorial/graphs/contributors
The Mkdocs was used for the creation of the website.
First install mkdocs tools with pip:
conda create -n mkdoc python=3.11
conda activate mkdoc
pip install -U mkdocs
pip install mkdocs-jupyter
pip install python-markdown-math
pip install markdown-callouts
pip install mdx-gh-links
pip install mkdocs-click
To create your docs source project, you will need the repository source cloned:
mkdocs new GITHUB_repo
Now if you go at this repo you will fine these files:
./
-- docs/
-- mkdocs.yml
Run mkdocs serve to test your browser.
The deployment can be done inside your github.io repository as follow:
mkdocs gh-deploy --config-file ../GITHUB_repo/mkdocs.yml --remote-branch main
For more details visit the mkdocs.org website.
- RNA Velocity analysis scVelo
- SingleCell Data Preprocessing with scanpy
- Trajectory inference With CellRank
- Introduction to Autoencoder
- Decision Trees using iris dataset
- Transfer learning using VGG16
- Introduction to tensorflow API tutorial
- Introduction to tensor basics operation
-
Explainable tensorflow multiclass model using Shap and model weights
-
Chest X Ray Images classification with DenseNet121 and explain with GradCam on tensorflow 2
There is an ongoing project in explainable AI applied to healthcare in collaboration with Salvatore RAIELI (his github). The source code can be found at https://github.com/SalvatoreRa/explanaibleAI. Don't hesitate to reach out for collaboration or any question.