Hierarchical Transfer Learning for Multi-label Text Classification

no code implementations • ACL 2019 • Siddhartha Banerjee, Cem Akkaya, Francisco Perez-Sorrosal, Kostas Tsioutsiouliklis

Compared to binary classifiers trained from scratch, our HTrans approach results in significant improvements of 1{\%} on micro-F1 and 3{\%} on macro-F1 on the RCV1 dataset.

Binary Classification General Classification +4