A Survey on Multi-view Learning

TL;DR

This survey categorizes multi-view learning methods into co-training, multiple kernel learning, and subspace learning, highlighting how consensus and complementary principles drive performance. It details view generation and evaluation, view combination strategies, and a wide range of algorithms within each category, including theoretical analyses and practical applications. The paper emphasizes the importance of constructing informative views and a unified framework that leverages both shared structure and view-specific information to improve generalization over single-view approaches. Overall, it shows that exploiting multiple views yields robust, scalable improvements across tasks, while also identifying open challenges in view design and integration.

Abstract

In recent years, a great many methods of learning from multi-view data by considering the diversity of different views have been proposed. These views may be obtained from multiple sources or different feature subsets. In trying to organize and highlight similarities and differences between the variety of multi-view learning approaches, we review a number of representative multi-view learning algorithms in different areas and classify them into three groups: 1) co-training, 2) multiple kernel learning, and 3) subspace learning. Notably, co-training style algorithms train alternately to maximize the mutual agreement on two distinct views of the data; multiple kernel learning algorithms exploit kernels that naturally correspond to different views and combine kernels either linearly or non-linearly to improve learning performance; and subspace learning algorithms aim to obtain a latent subspace shared by multiple views by assuming that the input views are generated from this latent subspace. Though there is significant variance in the approaches to integrating multiple views to improve learning performance, they mainly exploit either the consensus principle or the complementary principle to ensure the success of multi-view learning. Since accessing multiple views is the fundament of multi-view learning, with the exception of study on learning a model from multiple views, it is also valuable to study how to construct multiple views and how to evaluate these views. Overall, by exploring the consistency and complementary properties of different views, multi-view learning is rendered more effective, more promising, and has better generalization ability than single-view learning.

Figures (6)

• Figure 1: Multi-view data: a) a web document can be represented by its url and words on the page, b) a web image can be depicted by its surrounding text separate to the visual information, c) images of a 3D object taken from different viewpoints, d) video clips are combinations of audio signals and visual frames, e) multilingual documents have one view in each language.
• Figure 2: The process of co-training style algorithms.
• Figure 3: Sketch map of multiple kernel learning.
• Figure 4: Sketch map of subspace learning for multi-view data.
• Figure 5: Flowchart of the semi-supervised multi-view subspace learning algorithm yu2012combining. The method first extracts multi-view features from cartoon characters. Then, by considering the constraints of each local patch and the complementary characteristics of multi-view features, the low dimensional representation $Y$ can be obtained through solving an alternating optimization problem. Finally, the cartoon character retrieval and clip synthesis can be conducted by measuring the dissimilarity in the subspace $Y$.