Learning from labeled and unlabeled data on a directed graph

By Dengyong Zhou, Jiayuan Huang, Bernhard Schölkopf. International Conference on Machine Learning, 2005

474 citations recorded by the source corpus.

Subfields: Artificial Intelligence, Computational Theory and Mathematics, Media Technology

View Paper: https://www.semanticscholar.org/paper/df95ae968cb0b722143f6000fa0dc7ce21cc35e2

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 Computational Theory and Mathematics

In Media Technology

Citations in the corpus

Commute Times for a Directed Graph using an Asymmetric Laplacian (2011, methodology): “Symmetrized Laplacians and Symmetrized Graph The symmetrized Laplacians L = (L+L)/2 and L = [(L) +L]/2 correspond to those used in [9, 40], with various diagonal scalings.”
Feature Selection for Classification: A Review (2014, methodology): “4 Algorithms for Streaming Features Methods introduced above assume that all features are known in advance, and another interesting scenario is taken into account where candidate features are sequentially presented to the classifier for potential inclusion in the model [53, 78, 70, 67].”
Feature Selection (2016, methodology): “Alpha-investing (Zhou et al. 2005) is an adaptive complexity penalty method that dynamically changes the threshold of error reduction that is required to ACM Computing Surveys, Vol. 50, No. 6, Article 94.”
Learning with Hypergraphs: Clustering, Classification, and Embedding (2006, background): “Consequently, as shown in [20], with the least square loss function, the classification function is finally given by f = (I − ξΘ)−1y, where the elements of y denote the initial labels, and ξ is a parameter in (0, 1).”
An experimental investigation of kernels on graphs for collaborative recommendation and semisupervised classification (2012, methodology): “The regularized normalized Laplacian kernel matrix (I − αL̃)−1 also appears in the context of semisupervised classification of labeled nodes of a graph (Zhou et al., 2005), while the regularized Laplacian kernel matrix is used for directed graph embedding in Chen et al.”
Recommender Systems (2016, unknown): “Collective classification of directed graphs is discussed in [675].”
Introduction to Semi-Supervised Learning (2009, background): “In certain applications like the Web, the edges naturally are directed [27, 118, 200].”
Essential Tensor Learning for Multi-View Spectral Clustering (2018, background): “The work in [20], [21] shows that there is a natural connection between the spectral clustering and the Markov random walk.”
SemiBoost: Boosting for Semi-Supervised Learning (2009, background): “Asymmetric simila rity matrices arise when using directed graphs for modeling clas sification problems, and are shown to perform better in certa in applications related to text categorization [26].”
Clustering and Community Detection in Directed Networks: A Survey (2013, methodology): “That is, spectral clustering methods based on the directed Laplacian matrix [75, 61, 76, 105] are examples of approaches that identify density-based clusters.”
Efficient Sampling Set Selection for Bandlimited Graph Signals Using Graph Spectral Proxies (2015, background): “One can then define the following variation functional for signals on directed graphs [27], [28]:”
Robust Multi-View Spectral Clustering via Low-Rank and Sparse Decomposition (2014, unknown): “We refer the readers to (Zhou, Huang, and Schölkopf 2005) for more details of spectral clustering via Markov chains.”