Multi-View Intact Space Learning

By Chang Xu, Dacheng Tao, Chao Xu. IEEE Transactions on Pattern Analysis and Machine Intelligence, 2015

385 citations recorded by the source corpus.

It is practical to assume that an individual view is unlikely to be sufficient for effective multi-view learning. Therefore, integration of multi-view information is both valuable and necessary. In this paper, we propose the Multi-view Intact Space Learning (MISL) algorithm, which integrates the encoded complementary information in multiple views to discover a latent intact representation of the data. Even though each view on its own is insufficient, we show theoretically that by combing multiple views we can obtain abundant information for latent intact space learning. Employing the Cauchy loss (a technique used in statistical learning) as the error measurement strengthens robustness to outliers. We propose a new definition of multi-view stability and then derive the generalization error bound based on multi-view stability and Rademacher complexity, and show that the complementarity between multiple views is beneficial for the stability and generalization. MISL is efficiently optimized using a novel Iteratively Reweight Residuals (IRR) technique, whose convergence is theoretically analyzed. Experiments on synthetic data and real-world datasets demonstrate that MISL is an effective and promising algorithm for practical applications.

Subfields: Artificial Intelligence, Computer Vision and Pattern Recognition

View Paper: https://www.semanticscholar.org/paper/280cd62f8830c80012c944a23703103641a636e6

How researchers see this paper

The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.

In Artificial Intelligence

In Computer Vision and Pattern Recognition

Citations in the corpus

Understanding Blooming Human Groups in Social Networks (2015, background): “studied the multi-view learning [18] [19] and scene classification [20].”
Feature Selection Based on Structured Sparsity: A Comprehensive Study (2017, background): “How to integrate these multiview features [120] and identify the important ones from them for specific tasks has become an increasingly important problem.”
GMC: Graph-Based Multi-View Clustering (2020, background): “Apart from clustering, multi-view learning methods and applications have been investigated in [45], [46], [47], [48], [49], [50], to list a few.”
Multi-view low-rank sparse subspace clustering (2017, methodology): “Multi-view learning method in [34] learns view generation matrices and representation matrix, relying on the assumption that data from all the views share the same representation matrix.”
Multi-Directional Multi-Level Dual-Cross Patterns for Robust Face Recognition (2014, background): “Index Terms—Face recognition, face image descriptors, face image representation F”
A Comprehensive Survey on Pose-Invariant Face Recognition (2015, methodology): “Of the existing works, the survey on face recognition across pose Zhang and Gao (2009) that summarizes PIFR approaches before 2009 is the most relevant to this paper.”
Learning Visual Semantic Relationships for Efficient Visual Retrieval (2015, neutral): “Index Terms—Visual concept relationship, hashing, image retrieval Ç”
NTIRE 2017 Challenge on Single Image Super-Resolution: Methods and Results (2017, methodology): “The training employed the cyclical learning rate strategy [29] and inspired by [37] used both train images and re-scaled images”
Structural Deep Network Embedding (2016, background): “Representation learning has long been an important problem of machine learning and many works aim at learning representations for samples [3, 35].”
Multiview Spectral Clustering via Structured Low-Rank Matrix Factorization (2017, background): “As implied by multiview research [7]–[9], an individual view is not capable of being faithful for effective multiview learning.”
Multi-Stream Deep Neural Networks for RGB-D Egocentric Action Recognition (2019, background): “Recent years have witnessed that multi-view learning has attracted growing attention in the research fields of machine learning and computer vision [49]–[59], which aims to jointly learn an optimal combination on multi-view (multi-modal) representations.”
Multi-view learning overview: Recent progress and new challenges (2017, background): “[48] proposed a multi-view intact space learning algorithm, which integrates the encoded complementary information in multiple views to discover a latent intact representation of data.”