A theoretical and empirical analysis of support vector machine methods for multiple-instance classification

Abstract

The standard support vector machine (SVM) formulation, widely used for supervised learning, possesses several intuitive and desirable properties. In particular, it is convex and assigns zero loss to solutions if, and only if, they correspond to consistent classifying hyperplanes with some nonzero margin. The traditional SVM formulation has been heuristically extended to multiple-instance (MI) classification in various ways. In this work, we analyze several such algorithms and observe that all MI techniques lack at least one of the desirable properties above. Further, we show that this tradeoff is fundamental, stems from the topological properties of consistent classifying hyperplanes for MI data, and is related to the computational complexity of learning MI hyperplanes. We then study the empirical consequences of this three-way tradeoff in MI classification using a large group of algorithms and datasets. We find that the experimental observations generally support our theoretical results, and properties such as the labeling task (instance versus bag labeling) influence the effects of different tradeoffs.

Inferring User Gender from User Generated Visual Content on a Deep Semantic Space

“…Additionally, it allows one to explicitly consider all the user profile images for classification in an elegant manner. Concretely, extensions to the SVM classifiers for MIL have been proposed [13], [18], [19]. These classifiers are interesting because they keep the desirable properties of SVMs.…”

Learning to Find Common Objects Across Few Image Collections

“…In a standard MIL framework, instance labels in each positive bag are treated as hidden variables with the constraint that at least one of them should be positive. MI-SVM and mi-SVM [2] are two popular methods for MIL, and have been widely adapted for many weakly supervised computer vision problems, achieving state-of-the-art results in many different applications [7,13]. In these methods, images in each bag inherit the label of the bag and an SVM is trained to classify images.…”

Two-person interaction recognition via spatial multiple instance embedding

“…The k-NN classifier has been adapted for MIL by defining the distance between bags [49]. Later on, kernel methods have also been adapted to work with MI data such as [4,19]; a complete review on such approaches is available in [12]. More recently, algorithms that involve boosting [55], embedding the data into a different feature space [7], or treating the data in bags as graphs [57] have been proposed.…”