JonathanCrabbe/Label-Free-XAI
This repository contains the implementation of Label-Free XAI, a new framework to adapt explanation methods to unsupervised models. For more details, please read our ICML 2022 paper: 'Label-Free Explainability for Unsupervised Models'.
When you're working with unsupervised machine learning models, it's often hard to understand why the model arrived at a particular data representation. This framework helps you interpret those 'black-box' models by showing which specific input features or training examples were most important in shaping the model's internal data representation. Data scientists and machine learning researchers working with unlabelled datasets can use this to gain insight.
No commits in the last 6 months.
Use this if you need to understand why an unsupervised machine learning model, such as an autoencoder, is producing certain data representations from your unlabelled data.
Not ideal if your models are supervised and have clear output labels, as many existing explanation methods are already well-suited for that scenario.
Stars
25
Forks
9
Language
Python
License
MIT
Last pushed
Sep 21, 2022
Commits (30d)
0
Get this data via API
curl "https://pt-edge.onrender.com/api/v1/quality/ml-frameworks/JonathanCrabbe/Label-Free-XAI"
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...