This is an introduction notebook to eXtreme Multi-Label Classification designed for beginners in Machine Learning.
It is designed to approach the deep XML problem in a simple way : we are using simple linear embedding and simple predictors, but we can visualize the results and understand why the method is interesting.
The method used in this notebook is based on the article Deep Extreme Multi-label Learning by Zhang et al.
If you want a quick look into XML, you can see our new video here:
The data come from The 9th annual MLSP competition: New methods for acoustic classification of multiple simultaneous bird species in a noisy environment and is used as an introduction example.