Label Propagation through Linear Neighborhoods

By Fei Wang, Changshui Zhang. 2006

911 citations recorded by the source corpus.

Subfields: Computer Vision and Pattern Recognition

How researchers see this paper

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

In Computer Vision and Pattern Recognition

Citations in the corpus

Scalable Semi-Supervised Learning by Efficient Anchor Graph Regularization (2016, neutral): “[37] developed a graph-based SSL approach based on a linear neighborhood model which assumes that each datapoint can”
Graph Learning Based on Total Variation Minimization (2018, background): “In [24] a graph with normalized node degree is obtained based on the assumption that each data point can be expressed as linear combination of its neighbors.”
A survey on semi-supervised learning (2019, methodology): “In particular, they compared Gaussian weighting (where edges are weighted using an isotropic Gaussian kernel), the locally normalized Gaussian weighting approach by Hein andMaier (2007), and LNP (Wang and Zhang 2008); somewhat surprisingly, their best results were obtained using Gaussian weighting.”
Learning to Propagate Labels: Transductive Propagation Network for Few-Shot Learning (2018, background): “The neighborhood structure behaves differently with respect to various σ, which means that it needs to carefully select the optimal σ for the best performance of label propagation (Wang and Zhang, 2006; Zhu and Ghahramani, 2002).”
An Adaptive Semisupervised Feature Analysis for Video Semantic Recognition (2018, background): “The variance parameter σ empirically tuned in real-world applications will affect the performance significantly [36].”
How to Learn a Graph from Smooth Signals (2016, methodology): “See for example the work by Zhang and Hancock (2008) for an application on image denoising by heat diffusion smoothing on the pixels graph.”
Semi-supervised learning by disagreement (2010, neutral): “Wang and Zhang [70] used the idea of LLE [60] that instances can be reconstructed by their neighbors to obtain weights over the edges in the graph.”
Adaptive Hypergraph Learning and its Application in Image Classification (2012, neutral): “, self-training [26], cotraining [27], transductive SVM [28], graph-based learning [29]–[31], and hypergraph learning [18].”
Discrete Signal Processing on Graphs (2012, methodology): “Our approach is based on the label propagation [66], [67], which is a simple, yet efficient technique that advances known labels from labeled graph nodes along edges to unlabeled nodes.”
Transductive Propagation Network for Few-shot Learning (2018, background): “tion: Wij= exp( d(xi;xj) 2˙2 ); (1) where d(;) is a distance measure (e.g., Euclidean distance) and ˙is the length scale parameter. However, it is difﬁcult to choose the optimal ˙in label propagation [25, 31]. In this paper, we propose a graph construction module built on the union set of support se”
Interactive shape co-segmentation via label propagation (2014, methodology): “Among them, Wang and Zhang [29] propose approximating the graph with a set of overlapped linear neighborhood patches (LNPs) and computing the edge weights in each patch using the neighborhood linear projection.”
Zero-Shot Recognition via Structured Prediction (2016, methodology): “[16]) and certain multi-class classification methods (e.”
Discrete Signal Processing on Graphs: Frequency Analysis (2013, methodology): “In signal processing on graphs, minimization problems similar to (39) and (41) formulated with the Laplacian quadratic form (see (51) in Appendix B) are used for data classiﬁcation [46], [47], [41], [48], characterization of graph signal smoothness [18] and recovery [8].”
Towards Personalized Medicine: Leveraging Patient Similarity and Drug Similarity Analytics (2014, background): “After label propagation, possibilistic label (i.e., the possibility when a drug is effective for a patient) matrix F can be obtained by a formula F =(1-µ )( I - µ W ) -1 Y (for details please refer to Wang and Zhang 23 ).”
Unifying Graph Convolutional Neural Networks and Label Propagation (2020, methodology): “…aggregated along edges in the graph, which is known as Label Propagation Algorithm (LPA) (Zhu et al., 2005; Zhou et al., 2004; Zhang & Lee, 2007; Wang & Zhang, 2008; Karasuyama & Mamitsuka, 2013; Gong et al., 2017; Liu et al., 2019a); On the node side V , node features are propagated along…”
Nonlinear Dimensionality Reduction with Local Spline Embedding (2009, background): “The utility of manifold learning has been demonstrated in many applications such as visualization [6], [50], [20], image analysis [11], image superresolution [8], denoising [25], motion analysis [26], data ranking [62], semisupervised learning [52], and so on.”
Graph construction and b-matching for semi-supervised learning (2009, methodology): “Second, the user selects an algorithm for .nding a sparse weighted subgraph from the fully connected weighted graph between all pairs of nodes.”