Learning Maximally Predictive Prototypes in Multiple Instance Learning
This work addresses interpretability and classification challenges in multiple instance learning, offering a domain-specific solution that appears incremental.
The authors tackled the problem of interpretability and classification in multiple instance learning by proposing a model that generates maximally predictive prototypes and learns a linear classifier simultaneously, achieving accurate and efficient results on benchmark datasets.
In this work, we propose a simple model that provides permutation invariant maximally predictive prototype generator from a given dataset, which leads to interpretability of the solution and concrete insights to the nature and the solution of a problem. Our aim is to find out prototypes in the feature space to map the collection of instances (i.e. bags) to a distance feature space and simultaneously learn a linear classifier for multiple instance learning (MIL). Our experiments on classical MIL benchmark datasets demonstrate that proposed framework is an accurate and efficient classifier compared to the existing approaches.