Graph construction and b-matching for semi-supervised learning

By Tony Jebara, Jun Wang, Shih-Fu Chang. International Conference on Machine Learning, 2009

403 citations recorded by the source corpus.

Subfields: Artificial Intelligence, Computer Vision and Pattern Recognition

View Paper: https://www.semanticscholar.org/paper/ab0bbd0f8d2730398ab7a203fe4ba462496f41d5

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

Solving the multi-way matching problem by permutation synchronization (2013, methodology): “Instead of learning parameters [24, 26], we utilize supervision directly to provide the correct matches on a small subset of randomly picked image pairs (e.”
Discriminative multi-manifold analysis for face recognition from a single training sample per person (2011, methodology): “While many graph construction methods have been proposed in the machine learning and computer vision community recently [12], we apply the conventional k -nearest-neighbor method, for its high effectiveness and efﬁciency, to calculate the afﬁnity matrices A and B , as follows: where N k 1 inter ( x…”
Graph Signal Processing: Overview, Challenges, and Applications (2017, background): “, physical distance between sensors in the case of sensor networks or distance in feature space in the case of learning applications [5], [140], [141].”
A survey on semi-supervised learning (2019, methodology): “Extensive experiments have been conducted on different graph construction algorithms, and new methods have been introduced (de Sousa et al. 2013; Jebara et al. 2009; Subramanya and Talukdar 2014).”
Deep Co-Investment Network Learning for Financial Assets (2018, background): “This is a common issue in the graph learning tasks [16].”
Hubs in Space: Popular Nearest Neighbors in High-Dimensional Data (2010, background): “Also, Jebara et al. (2009) briefly mention hubness in the context of graph construction for semi-supervised learning.”
Non-negative low rank and sparse graph for semi-supervised learning (2012, methodology): “Traditional methods (such as k-nearest neighbors and Locally Linear Reconstruction [10]) mainly rely on pair-wise Euclidean distances and construct the graph by a family of overlapped local patches.”
Graph Learning Based on Total Variation Minimization (2018, background): “The authors of [23] construct a b-matched graph where each node has the same number of neighbors.”