Why state-of-the-art deep learning barely works as good as a linear classifier in extreme multi-label text classification.

Mohammadreza Qaraei, Sujay Khandagale, Rohit Babbar

Browse the full ESANN paper archive.