Understanding Label Bias in Single Positive Multi-Label Learning
This work addresses a practical issue in multi-label classification for researchers and practitioners, but it is incremental as it builds on existing SPML methods.
The paper tackles the problem of label bias in single positive multi-label learning, where realistic settings may not have uniformly random positive labels, and introduces protocols to study this bias with new empirical results.
Annotating data for multi-label classification is prohibitively expensive because every category of interest must be confirmed to be present or absent. Recent work on single positive multi-label (SPML) learning shows that it is possible to train effective multi-label classifiers using only one positive label per image. However, the standard benchmarks for SPML are derived from traditional multi-label classification datasets by retaining one positive label for each training example (chosen uniformly at random) and discarding all other labels. In realistic settings it is not likely that positive labels are chosen uniformly at random. This work introduces protocols for studying label bias in SPML and provides new empirical results.