PAC-Bayesian Analysis for a two-step Hierarchical Multiview Learning Approach
This work addresses the problem of improving generalization in multiview learning for researchers, but it is incremental as it builds on existing late fusion methods with a new theoretical analysis.
The paper tackles multiview learning with more than two views by proposing a two-step hierarchical approach under the PAC-Bayesian framework, deriving a generalization bound that highlights the importance of balancing diversity and accuracy in view-specific classifiers, and validates this on Reuters RCV1/RCV2 datasets.
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.