Investigating Class-level Difficulty Factors in Multi-label Classification Problems
This work addresses multi-label classification challenges by introducing class-level difficulty factors, offering a novel approach to enhance performance without added computational cost, though it is incremental in its application of existing concepts to a new context.
The paper tackled the problem of multi-label classification by proposing four class-level difficulty factors and demonstrated their utility in predicting class-level performance and improving predictive performance through difficulty weighted optimization, achieving significant improvements in mAP and AUC on WWW Crowd and Visual Genome datasets.
This work investigates the use of class-level difficulty factors in multi-label classification problems for the first time. Four class-level difficulty factors are proposed: frequency, visual variation, semantic abstraction, and class co-occurrence. Once computed for a given multi-label classification dataset, these difficulty factors are shown to have several potential applications including the prediction of class-level performance across datasets and the improvement of predictive performance through difficulty weighted optimisation. Significant improvements to mAP and AUC performance are observed for two challenging multi-label datasets (WWW Crowd and Visual Genome) with the inclusion of difficulty weighted optimisation. The proposed technique does not require any additional computational complexity during training or inference and can be extended over time with inclusion of other class-level difficulty factors.