Information Recovery-Driven Deep Incomplete Multiview Clustering Network (2023-04-02T00:00:00.000000Z)

TL;DR

A two-stage autoencoder network with self-attention structure is built to synchronously extract high-level semantic representations of multiple views and recover the missing data, and a recurrent graph reconstruction mechanism is developed that cleverly leverages the restored views to promote representation learning and further data reconstruction.

Abstract

Incomplete multiview clustering (IMC) is a hot and emerging topic. It is well known that unavoidable data incompleteness greatly weakens the effective information of multiview data. To date, existing IMC methods usually bypass unavailable views according to prior missing information, which is considered a second-best scheme based on evasion. Other methods that attempt to recover missing information are mostly applicable to specific two-view datasets. To handle these problems, in this article, we propose an information-recovery-driven-deep IMC network, termed as RecFormer. Concretely, a two-stage autoencoder network with self-attention structure is built to synchronously extract high-level semantic representations of multiple views and recover the missing data. Besides, we develop a recurrent graph reconstruction mechanism that cleverly leverages the restored views to promote representation learning and further data reconstruction. Visualization of recovery results are given and sufficient experimental results confirm that our RecFormer has obvious advantages over other top methods.

Authors

Chengliang Liu

3 papers

Jie Wen

3 papers

Xiaoling Luo

2 papers

TL;DR

Abstract

Authors

References75 items

Self-supervised Dynamic Graph Embedding with evolutionary neighborhood and community

Deep Double Incomplete Multi-View Multi-Label Learning With Incomplete Labels and Missing Views

DICNet: Deep Instance-Level Contrastive Network for Double Incomplete Multi-View Multi-Label Classification

Self-Guided Partial Graph Propagation for Incomplete Multiview Clustering

Augmented Sparse Representation for Incomplete Multiview Clustering

A Survey on Incomplete Multiview Clustering

Localized Sparse Incomplete Multi-View Clustering

Refining Graph Structure for Incomplete Multi-View Clustering

Parameter-Free and Scalable Incomplete Multiview Clustering With Prototype Graph

High-Order Correlation Preserved Incomplete Multi-View Subspace Clustering

Fast Parameter-Free Multi-View Subspace Clustering With Consensus Anchor Guidance

AF: An Association-based Fusion Method for Multi-Modal Classification

Adaptive reverse graph learning for robust subspace learning

One-Stage Incomplete Multi-view Clustering via Late Fusion

Self-Representation Subspace Clustering for Incomplete Multi-view Data

Incomplete Multi-View Clustering With Joint Partition and Graph Learning

Consensus Graph Learning for Multi-View Clustering

Unified Tensor Framework for Incomplete Multi-view Clustering and Missing-view Inferring

Trusted Multi-View Classification

DMIB: Dual-Correlated Multivariate Information Bottleneck for Multiview Clustering

DIMC-net: Deep Incomplete Multi-view Clustering Network

Adaptive Graph Completion Based Incomplete Multi-View Clustering

Self-supervised Learning from a Multi-view Perspective

Generalized Incomplete Multiview Clustering With Flexible Locality Structure Diffusion

Efficient and Effective Regularized Incomplete Multi-View Clustering

Marginalized Multiview Ensemble Clustering

Unified Embedding Alignment with Missing Views Inferring for Incomplete Multi-View Clustering

One-Pass Incomplete Multi-view Clustering

Doubly Aligned Incomplete Multi-view Clustering

Attention is All you Need

Multiple Kernel k-Means with Incomplete Kernels

Incomplete Multi-Modal Visual Data Grouping

Deep Residual Learning for Image Recognition

Rethinking the Inception Architecture for Computer Vision

Multiple Incomplete Views Clustering via Weighted Nonnegative Matrix Factorization with L2, 1 Regularization

Constrained Multi-View Video Face Clustering

Robust Subspace Clustering for Multi-View Data by Exploiting Correlation Consensus

Very Deep Convolutional Networks for Large-Scale Image Recognition

Partial Multi-View Clustering

Multi-View K-Means Clustering on Big Data

Constrained Clustering and Its Application to Face Clustering in Videos

ImageNet classification with deep convolutional neural networks

Learning Generative Visual Models from Few Training Examples: An Incremental Bayesian Approach Tested on 101 Object Categories

I and i

Interpretable learning based Dynamic Graph Convolutional Networks for Alzheimer's Disease analysis

MVDRNet: Multi-view diabetic retinopathy detection by combining DCNNs and attention mechanisms

Multiple Kernel $k$k-Means with Incomplete Kernels

CPM-Nets: Cross Partial Multi-View Networks

and s

degree in computer science

Department of Computer Science, Harbin Engineering University, Harbin, China

received the B.S. degree in software engineering from South China Normal University (SCNU), Guangzhou, China

Deep Learning

I and J

Multiview Clustering with Incomplete Views

Agriculture and Natural Resources FSA 6146 A Guide to Fungicide Resistance in Turf Systems

Visualizing Data using t-SNE

“Uci machine learning repository

A. and Q

The Amsterdam Library of Object Images

Γ and B

Engineering in

Some methods for classification and analysis of multivariate observations

Some methods for classi cation and analysis of multivariate observations

2016 Ieee International Conference on Big Data (big Data) Online Multi-view Clustering with Incomplete Views

and as an in

Is a larger dataset with up to 10158 images and 50 categories, whose features are extracted by DECAF [36] and Vgg19 [37]

in computer science and technology at Harbin Institute of Technology, Shenzhen, China

the B.S. degree in Department of Computer Science from Harbin Engineering

Integrate the imputation of missing views and final clustering into one optimization procedure

“Multiple features,”

ACTIONS ON M ULTIMEDIA

Based on nonnegative matrix factorization technology, also introduces the prior missing information to help learn robust latent representations

Is another popular baseline method, which aligns all views with the same imputation strategy as BSV and simply concatenates them to conduct single-view clustering

, IEEE) received the Ph.D. degree in Computer Science and Technology

Field of Study