Joint sparse graph and flexible embedding for graph-based semi-supervised learning

Dornaika, Fadi; Traboulsi, Y. El

doi:10.1016/j.neunet.2019.03.002

Cited by 30 publications

(9 citation statements)

References 18 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…It implicitly exploits a diffusion mapping that enhances the final discrimination. In [6], the authors proposed a framework that jointly estimates the non-linear projection, the graph structure, and the linear transform. In [32], the authors proposed a novel multiview learning model that can perform semi-supervised classification and graph estimation simultaneously.…”

Section: Graph-based Embedding Methodsmentioning

confidence: 99%

“…This method simultaneously provides a nonlinear embedding data and a linear regressor. The work described in [6] introduces a framework for graph-based flexible semi-supervised classification. The proposed framework jointly estimates the graph matrix, the nonlinear projection, and the linear regression model.…”

Section: Graph-based Embedding Methodsmentioning

confidence: 99%

“…We evaluate and compare our proposed method with the following methods: flexible graph-based semi-supervised embedding algorithm (FGSE) [4], manifold regularized deep learning architecture algorithm (MRDL) [50], kernel flexible model embedding (KFME) [8], joint embedding learning and sparse regression (JELSR) [12], supervised Laplacian egenmaps (SLE) [37], semi-supervised discriminant analysis (SDA) [3], semi-supervised discriminant embedding (SDE) [49], LLE [38], GraphSAGE [11], joint sparse graph and flexible embedding (JSGFE) [6], and semi-supervised projection with graph optimization (SPGO) [33].…”

Section: Methodsmentioning

confidence: 99%

“…In order to study the performance of the proposed method, we compared it with the following semi-supervised methods: Gaussian fields and harmonic functions (GFHF) [56], normalized GFHF, robust multi-class graph transduction (RMGT) [27], Laplacian regularized least square (LapRLS) [1], semi-supervised discriminant analysis (SDA) [3], semi-supervised discriminant embedding (SDE) [49], sparsity preserving discriminant analysis (SPDA) [35], flexible manifold embedding (FME) [30], flexible graph-based semi-supervised embedding (FGSE) [4], non-negative sparse graph (NNSG) [9], GraphSAGE [11], joint sparse graph and flexible embedding (JSGFE) [6], and semi-supervised projection with graph optimization (SPGO) [33]. Note that the GFHF and RMGT methods are transductive.…”

Section: Inductive Model Evaluationmentioning

confidence: 99%

See 3 more Smart Citations

Flexible data representation with graph convolution for semi-supervised learning

Dornaika

2020

Neural Comput & Applic

Self Cite

View full text Add to dashboard Cite

This paper introduces a scheme for semi-supervised data representation. It proposes a flexible nonlinear embedding model that imitates the principle of spectral graph convolutions. Structured data are exploited in order to determine nonlinear and linear models. The introduced scheme takes advantage of data graphs at two different levels. First, it incorporates manifold regularization that is naturally encoded by the graph itself. Second, the regression model is built on the convolved data samples that are obtained by the joint use of the data and their associated graph. The proposed semi-supervised embedding can tackle challenges related to over-fitting in image data spaces. The proposed graph convolution-based semi-supervised embedding paves the way to new theoretical and application perspectives related to the nonlinear embedding. Indeed, building flexible models that adopt convolved data samples can enhance both the data representation and the final performance of the learning system. Several experiments are conducted on six image datasets for comparing the introduced scheme with many state-of-art semi-supervised approaches. These experimental results show the effectiveness of the introduced data representation scheme.

show abstract

Section: Graph-based Embedding Methodsmentioning

confidence: 99%

Section: Graph-based Embedding Methodsmentioning

confidence: 99%

Section: Methodsmentioning

confidence: 99%

Section: Inductive Model Evaluationmentioning

confidence: 99%

See 2 more Smart Citations

Flexible data representation with graph convolution for semi-supervised learning

Dornaika

2020

Neural Comput & Applic

Self Cite

View full text Add to dashboard Cite

show abstract

“…However, only use linear regression to predict new sample class can not meet the actual demand, because most data has a non-linear relationship. Dornaika, F et.al [22] proposed margin-based discriminant embedding method, which process the data by non-linear embedding so that the linear regression model can be adapted to any data, and this method has been applied to FME NNSG [22] [23]. However, in the process of non-linear embedding, the dimension of non-linear subspace need to be set.…”

Section: Introductionmentioning

confidence: 99%

A Novel Adaptive Multi-View Non-Negative Graph Semi-Supervised ELM

et al. 2020

View full text Add to dashboard Cite

This paper represents a semi-supervised learning framework, which integrates multi-view learning, extreme learning machine (ELM) and graph-based semi-supervised learning. The aim is to expand the scope of adaptation of non-negative sparse graph (NNSG) framework, under a multi-view condition and a non-linear relationship. The proposed multi-view learning method will be adaptive since when data is single-view the framework will degenerate into an embedded framework for NNSG framework. The proposed ELM method also will be adaptive since the number of hidden layer neuron will change with different number of input and output layer neuron. The combination of both proposed methods outperforms traditional graph-based semisupervised learning, such as flexible manifold embedding (FME) and NNSG framework, which can not establish an affinity matrix for multi-view and can not establish a non-liner model for unknown data. Unlike traditional graph-based semi-supervised learning methods, which only can label propagation and build linear regression models for single or multi-view data, our proposed method has an obvious advantage that is applicable to any single or multi-view data, and builds linear or non-linear models. We provides extensive experiments on four public database in order to evaluate the performance of the proposed method. These experiments demonstrate significant improvement over the state-of-the-art algorithms in label propagation and processing of new data. INDEX TERMS Graph-based semi-supervised learning, Multi-view learning, Semi-supervised ELM, Graph learning.

show abstract