Learning Maximally Predictive Prototypes in Multiple Instance Learning

2019/10/02 by Mert Yüksekgönül, Yuksekgonul, Mert, Özgür Emre Sivrikaya +3
Computer Science · #Advanced Image and Video Retrieval Techniques #FOS: Computer and information sciences #Image Retrieval and Classification Techniques #Machine Learning (cs.LG) #Machine Learning (stat.ML) #Video Analysis and Summarization

paper · pdf · doi:10.48550/arxiv.1910.00965

openalex publication_date 2019/10/02 · openalex created_date 2025/10/10 · openalex updated_date 2026/07/28

Abstract

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.

Citations

• Batch Normalization: Accelerating Deep Network Training by Reducing Internal Covariate Shift
• Multiple instance classification: Review, taxonomy and comparative study
• Solving the multiple instance problem with axis-parallel rectangles
• Attention-based Deep Multiple Instance Learning
• PCL: Proposal Cluster Learning for Weakly Supervised Object Detection
• Layer Normalization

Related

• Multiple-Instance Learning by Boosting Infinitely Many Shapelet-based Classifiers
• Using Neural Network Formalism to Solve Multiple-Instance Problems
• On the Complexity of One-class SVM for Multiple Instance Learning
• A multi-instance learning algorithm based on a stacked ensemble of lazy learners
• Multi-Instance Learning by Treating Instances As Non-I.I.D. Samples