csinva/transformation-importance
Using / reproducing TRIM from the paper "Transformation Importance with Applications to Cosmology" 🌌 (ICLR Workshop 2020)
This tool helps scientists and researchers understand which parts of their complex data inputs are most influential to a machine learning model's predictions. You input your scientific data (like cosmological images, audio recordings, or text), your pre-trained model, and a mathematical transformation (like a Fourier Transform or NMF), and it tells you the 'importance' of different components in the transformed data. This is useful for scientists, data analysts, and researchers working with complex data and deep learning models.
No commits in the last 6 months.
Use this if you need to explain why your machine learning model made a specific prediction by understanding the importance of features in a transformed version of your original data.
Not ideal if you want to interpret model predictions without transforming your data, or if you are looking for a simple, out-of-the-box model interpretation for basic datasets.
Stars
9
Forks
1
Language
Jupyter Notebook
License
MIT
Last pushed
Dec 16, 2020
Commits (30d)
0
Get this data via API
curl "https://pt-edge.onrender.com/api/v1/quality/ml-frameworks/csinva/transformation-importance"
Open to everyone — 100 requests/day, no key needed. Get a free key for 1,000/day.
Higher-rated alternatives
obss/sahi
Framework agnostic sliced/tiled inference + interactive ui + error analysis plots
tensorflow/tcav
Code for the TCAV ML interpretability project
MAIF/shapash
🔅 Shapash: User-friendly Explainability and Interpretability to Develop Reliable and Transparent...
TeamHG-Memex/eli5
A library for debugging/inspecting machine learning classifiers and explaining their predictions
csinva/imodels
Interpretable ML package 🔍 for concise, transparent, and accurate predictive modeling...