Large Margin Component Analysis

By Lorenzo Torresani, Kuang-chih Lee. 2006

247 citations recorded by the source corpus.

Subfields: Computational Mechanics, 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 Computational Mechanics

In Computer Vision and Pattern Recognition

Citations in the corpus

Image-Based Recommendations on Styles and Substitutes (2015, background): “Similar ideas appear in [4, 35], which also consider the problem of metric learning via low-rank embeddings, albeit using a different objective than the one we consider here.”
A Survey on Metric Learning for Feature Vectors and Structured Data (2013, methodology): “Some metric learning approaches have been shown to be kernelizable (see for instance Schultz and Joachims, 2003; Shalev-Shwartz et al., 2004; Hoi et al., 2006; Torresani and Lee, 2006; Davis et al., 2007) using specific arguments, but in general kernelizing a particular metric algorithm is not trivi”
Transductive Episodic-Wise Adaptive Metric for Few-Shot Learning (2019, background): “(3) will degenerate into the popular Euclidean distance if we set Mt = I , that is, assuming all features are equally scaled and equally relevant [32].”
Learning Local Feature Descriptors Using Convex Optimisation (2014, background): “One approach to reducing dimension is to optimise directly over a projection matrix of the required size[8], [30], but this leads to non-convex objectives .”
Distance Metric Learning for Large Margin Nearest Neighbor Classification (2005, background): “The ideas behind LMNN classification have also been extended by others in various ways (Torresani and Lee, 2007; Kumar et al., 2007).”
Learning a Mahalanobis distance metric for data clustering and classification (2008, methodology): “Algorithms in this family include the neighboring component analysis [15], large margin nearest neighbor classification [16], large margin component analysis [17], class collapse [18], and other extension work [19,20].”
PCCA: A new approach for distance learning from sparse pairwise constraints (2012, methodology): “class labels must be given) [10, 11, 19, 27, 30], or they are too domain specific [1, 5, 33], or they suffer from significant loss in performance when the dimensionality of the input space is high or when the amount of available training data is low [7, 14].”
Metric Learning: A Survey (2013, methodology): “Finally, we also note the related work of Torresani and Lee [71], an earlier paper that considers the use of kernel PCA for solving LMNN in kernel space.”
Distance Metric Learning with Eigenvalue Optimization (2012, background): “Keywords: metric learning, convex optimization, semi-definite programming, first-order methods, eigenvalue optimization, matrix factorization, face verification”
Linear dimensionality reduction: survey, insights, and generalizations (2014, background): “Large margin nearest neighbors (LMNN; Weinberger et al., 2005; Torresani and Lee, 2006; Weinberger and Saul, 2009) assumes labeled data: (xi, zi), such that zi ∈ {1, .”
Regression on Fixed-Rank Positive Semidefinite Matrices: A Riemannian Approach (2010, methodology): “A simpler objective is pursued by the algorithms that optimize the Mahalanobis distance for the specific k-nearest neighbor classifier [Goldberger et al., 2004, Torresani and Lee, 2006, Weinberger and Saul, 2009].”
Large Margin Dimensionality Reduction for Action Similarity Labeling (2014, background): “This is closely related to CSML [14] and distance metric learning of large margin nearest neighbor (LMNN) [20], [21], [22].”
Fisher Vector Faces in the Wild (2013, methodology): “Another approach, which we use here, is to optimise directly the projection matrix W , as its size depends on the reduced dimensionality, although this results in a non-convex formulation [11, 33].”