Affinity-Based Hierarchical Learning of Dependent Concepts for Human Activity Recognition
This addresses the challenge of inconsistent class separability in real-world multi-class classification tasks, such as human activity recognition, with incremental improvements.
The paper tackles the problem of overlapping classes in human activity recognition by organizing them into hierarchies based on transfer affinity, which improves classification performance and reduces the number of examples needed to learn, as demonstrated on the SHL dataset.
In multi-class classification tasks, like human activity recognition, it is often assumed that classes are separable. In real applications, this assumption becomes strong and generates inconsistencies. Besides, the most commonly used approach is to learn classes one-by-one against the others. This computational simplification principle introduces strong inductive biases on the learned theories. In fact, the natural connections among some classes, and not others, deserve to be taken into account. In this paper, we show that the organization of overlapping classes (multiple inheritances) into hierarchies considerably improves classification performances. This is particularly true in the case of activity recognition tasks featured in the SHL dataset. After theoretically showing the exponential complexity of possible class hierarchies, we propose an approach based on transfer affinity among the classes to determine an optimal hierarchy for the learning process. Extensive experiments show improved performances and a reduction in the number of examples needed to learn.