Repository of the main source code for the assignments of the "Introduction to interpretable AI" course
Presented by: Mara Graziani (PhD student at HES-SO Valais and University of Geneva) Email: mara.graziani@hevs.ch
This repository contains a series of Colab notebooks that you will use as a backbone for the assignmets and practical exercises of the course Introduction to Interpretable AI, as of in the Spring semester of 2021.
The notebooks in this folder present exercises on
- Lecture 2: Feature Visualization with the Lucid toolbox
- Lecture 2: Gradient-weighted Class Activation Mapping (Grad-CAM) on you CNN classifier. This explainability technique generates a heatmap the relevant input features.
- Lecture 4: Concept-based explanations with Regression Concept Vectors. The notebooks are applied on standard computer vision tasks. A specific application is presented for breast histopathology images at high-magnification (40x).
The notebooks can be directly run in the Colab workspace, so no installation / or setup of virtual environments is needed. The notebook data_setup.ipynb will drive you through the data preparation steps. If you encounter any isses please let me know.
You can directly clone the folder to your computer to start developing the project further:
git clone https://github.com/maragraziani/interpretAI_DigiPath- Generate feature activation maximization
- Generate CAM heatmaps for imagenet-pretrained model
- Load your own model and weights to generate CAM
- Apply latest research to interpret your model with Regression Concept Vectors
Some useful links for more information about this hands-on session:
- Funding Project homepage: https://www.ai4media.eu/
- Repository: https://github.com/maragraziani/intro-interpretableAI
- Issue tracker: https://github.com/maragraziani/intro-interpretableAI/issues
- Related projects:
- Someone else's project: https://github.com/maragraziani/rcvtool/
The code in this project is licensed under MIT license.