Incomplete Multi-View Weak-Label Learning with Noisy Features and Imbalanced Labels

TL;DR

This paper introduces a novel multi-view multi-label learning method that handles incomplete, noisy, and imbalanced data by joint embedding, adaptive view importance, and focal loss, showing superior performance on real datasets.

Contribution

It proposes an integrated approach combining low-dimensional embedding, adaptive view weighting, and imbalance mitigation for incomplete multi-view weak-label learning.

Findings

Effective in handling incomplete and noisy views.

Improves label imbalance with focal loss.

Outperforms existing methods on real datasets.

Abstract

A variety of modern applications exhibit multi-view multi-label learning, where each sample has multi-view features, and multiple labels are correlated via common views. Current methods usually fail to directly deal with the setting where only a subset of features and labels are observed for each sample, and ignore the presence of noisy views and imbalanced labels in real-world problems. In this paper, we propose a novel method to overcome the limitations. It jointly embeds incomplete views and weak labels into a low-dimensional subspace with adaptive weights, and facilitates the difference between embedding weight matrices via auto-weighted Hilbert-Schmidt Independence Criterion (HSIC) to reduce the redundancy. Moreover, it adaptively learns view-wise importance for embedding to detect noisy views, and mitigates the label imbalance problem by focal loss. Experimental results on four real-world multi-view multi-label datasets demonstrate the effectiveness of the proposed method.