search

arpeggiohq / explainable-ai

12 stars 4 forks source link

feedback - an interesting Interpretable ML opensource package #1

Open arrabi opened 1 year ago

arrabi commented 1 year ago

Hello Jay, A great video as usual. Explainable AI is a critical component to bring the AI revolution to many business areas.

I thought you might be interested in this opensource package: https://github.com/interpretml

It has many nice components. One of my favorites is Explainable Boosting Machines (EBM), a modern ML algorithm that is comparable to the perf of XGBoost, but fully transparent/interpretable by design. It's based on Generalized Additive Model (a statistical method from the 80s) but with modern ML additions.

The Github link has links to papers and videos. Hope this is useful, Muhammad

arrabi commented 1 year ago

FYI that Rich Caruana is working on Neural Additive Models with a number of collaborators including Hinton. I think we'll see very interesting new models in this space soon: https://arxiv.org/abs/2004.13912