PAC-Bayesian Analysis for a two-step Hierarchical Multiview Learning Approach

TL;DR

This paper introduces a PAC-Bayesian framework for a two-step multiview learning approach, providing a theoretical risk bound that emphasizes the importance of balancing diversity and accuracy among view-specific classifiers.

Contribution

It presents a novel PAC-Bayesian generalization bound for multiview learning that accounts for diversity among classifiers, enhancing understanding of late fusion strategies.

Findings

Theoretical risk bound includes a diversity term.

Controlling diversity improves multiview learning performance.

Experimental validation on Reuters datasets supports the theoretical insights.

Abstract

We study a two-level multiview learning with more than two views under the PAC-Bayesian framework. This approach, sometimes referred as late fusion, consists in learning sequentially multiple view-specific classifiers at the first level, and then combining these view-specific classifiers at the second level. Our main theoretical result is a generalization bound on the risk of the majority vote which exhibits a term of diversity in the predictions of the view-specific classifiers. From this result it comes out that controlling the trade-off between diversity and accuracy is a key element for multiview learning, which complements other results in multiview learning. Finally, we experiment our principle on multiview datasets extracted from the Reuters RCV1/RCV2 collection.