Nonlinear Learning using Local Coordinate Coding

By Kai Yu, Tong Zhang, Yihong Gong. 2009

808 citations recorded by the source corpus.

Subfields: Representation & Self-Supervised Learning

How researchers see this paper

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

In Representation & Self-Supervised Learning

Citations in the corpus

Pylearn2: a machine learning research library (2013, methodology): “This includes autoencoders [6], RBMs [47] including RBMs with Ga ussian visible units [54], DBMs [45], MLPs [44], convolutional networks [33], and local coordina te coding [56].”
Representation Learning: A Review and New Perspectives (2012, background): “[223] is to show that that any Lipschitz-smooth function :M! IR defined on a smooth nonlinear manifold M embedded in IRx can be well approximated by a globally linear function with respect to the resulting coding scheme (i.”
Emergence of Complex-Like Cells in a Temporal Product Network with Local Receptive Fields (2010, background): “There are several system of this class (Lazebnik et al., 2006; Pinto et al., 2008, 2009; Serre et al., 2007; Yang et al., 2009; Lee et al., 2009; Yu et al., 2009) but speciﬁcally convolutional nets achieve 67% (Jarrett et al., 2009).”
Bag of visual words and fusion methods for action recognition: Comprehensive study and good practice (2014, neutral): “nonzero coefficients are often assigned to bases nearby to the encoded data [63].”
Pulling Things out of Perspective (2014, background): “Recent progress in computer vision and machine learning led to the development of algorithms [15, 27, 29, 28], that are able to successfully categorize images into hundreds [4] or even thousands [3] of different object classes.”
Object recognition with hierarchical kernel descriptors (2011, background): “recently introduced local coordinate coding [32] and its improved version [27, 31], which quickly computes sparse representations based on nearest neighbors and can model manifold geometric structures in high dimensions.”
Large-scale image classification: Fast feature extraction and SVM training (2011, methodology): “We apply two state-of-the-art ‘coding + pooling’ pipelines in our system, one is based on local coordinate coding (LCC) [26], and the other is based on super-vector coding (SVC) [28].”
The Importance of Encoding Versus Training with Sparse Coding and Vector Quantization (2011, background): “In particular, Yu et al. (2009) have argued for " locality preserving " encodings based on the idea that such encodings allow higher-level systems to learn functions across the data manifold more easily.”
Scattering Representations for Recognition (2013, background): “Local coordinate coding [YGZ09] is an encoding strategy half-way between vector quantization and sparse coding, which encodes a visual feature with a linear combination of its closest prototypes.”
Visual Classification by ℓ1-Hypergraph Modeling (2015, neutral): “Finally, our conclusions are presented in Section 5.”
In defense of soft-assignment coding (2011, background): “A fundamental assumption of LLC is that local features approximately reside on a lowerdimensional manifold in an ambient descriptor space.”
Adversarial Learning with Local Coordinate Coding (2018, positive): “Deﬁnition 3 (Latent Manifold (Yu et al., 2009)) A subset M embedded in the latent space R d B is called a smooth manifold with a intrinsic dimension d := d M , if there exists a constant c M , such that given any h ∈ M , there are where γ = [ γ 1 , . . . , γ d ] T is the local coding of a latent poi”
Image Classification Using Super-Vector Coding of Local Image Descriptors (2010, methodology): “Our approach is along the line of recent works on unsupervised feature learning for image classﬁication, especially, learning sparse representations e.g., [17][5][18] [19] [20].”
The generalized lasso is reducible to a subspace constrained lasso (2013, background): “Indeed, such representations have proven effective in applications ranging from image restoration [2, 3], face recognition [4, 5], object recognition [6], speech classification [7], speech recognition [8], music genre classification [9], and topic detection in text documents [10].”
The Manifold Tangent Classifier (2011, methodology): “More recently, the Local Coordinate Coding (LCC) algorithm (Yu et al., 2009) and its Local Tangent LCC variant (Yu and Zhang, 2010) were proposed to build a a local chart around each training example (with a local low-dimensional coordinate system around it) and use it to define a representation…”