Multi-Instance Multi-Label Learning

TL;DR

This paper introduces the MIML framework for learning from examples described by multiple instances and labels, proposing algorithms that improve performance on complex objects with multiple semantic meanings.

Contribution

It presents the MIML framework and new algorithms like MimlBoost, MimlSvm, and D-MimlSvm, addressing information loss and demonstrating effectiveness in various tasks.

Findings

MIML framework improves learning for complex objects.

Algorithms outperform traditional single-instance or single-label methods.

MIML remains useful even without access to real object data.

Abstract

In this paper, we propose the MIML (Multi-Instance Multi-Label learning) framework where an example is described by multiple instances and associated with multiple class labels. Compared to traditional learning frameworks, the MIML framework is more convenient and natural for representing complicated objects which have multiple semantic meanings. To learn from MIML examples, we propose the MimlBoost and MimlSvm algorithms based on a simple degeneration strategy, and experiments show that solving problems involving complicated objects with multiple semantic meanings in the MIML framework can lead to good performance. Considering that the degeneration process may lose information, we propose the D-MimlSvm algorithm which tackles MIML problems directly in a regularization framework. Moreover, we show that even when we do not have access to the real objects and thus cannot capture more information from real objects by using the MIML representation, MIML is still useful. We propose the InsDif and SubCod algorithms. InsDif works by transforming single-instances into the MIML representation for learning, while SubCod works by transforming single-label examples into the MIML representation for learning. Experiments show that in some tasks they are able to achieve better performance than learning the single-instances or single-label examples directly.