Ensemble Methods for Multi-label Classification

DEEP-CARVING: Discovering visual attributes by carving deep neural nets

“…Again, such methods require complete supervision, and thus cannot be applied to our problem. Likewise for the various multi-label ranking methods such as [4,6,14,16,18,36] considered in the machine learning literature, which propose different types of feature models for efficient rank learning or label prediction, and the associated ensemble methods for multi-label classification such as [31,40]. Authors in [25] try to rank attributes in images in a completely unsupervised manner.…”