Sparse Filtering

By Jiquan Ngiam, Pang Wei Koh, Zhenghao Chen, Zhenghao Chen, Sonia A. Bhaskar, Sonia A. Bhaskar, Andrew Y. Ng. 2011

276 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

Challenges in representation learning: A report on three machine learning contests (2013, background): “David Thaler and Lukasz Romaszko both observed that learning the sparse ﬁltering features on the combination of the labeled and unlabeled data worked worse than learning the features on just the labeled data.”
General Stochastic Networks for Classification (2014, background): “Most accomplishments originate from variants of restricted Boltzmann machines (RBMs) [1], auto-encoders (AE) [2, 3] and sparse-coding [4, 5, 6].”
Weakly-Shared Deep Transfer Networks for Heterogeneous-Domain Knowledge Propagation (2015, methodology): “, dropout technique for randomly omitting some hidden units with a constant probability [12], corrupted input [33], greedy layer-wise training with sparse filtering [17, 19], alternative convolution and pooling layers as previous layers [16].”
An Intelligent Fault Diagnosis Method Using Unsupervised Feature Learning Towards Mechanical Big Data (2016, methodology): “Sparse filtering can be extended to a nonlinear one using an activation function, and the activation function g(·) = |·| is recommended [29].”
Vehicle Type Classification Using a Semisupervised Convolutional Neural Network (2015, methodology): “…and C, we introduce the Kullback-Leibler (KL) divergence as the optimization principle min T ,C N∑ i=1 KL(d(i)‖y(i)) = min T ,C N∑ i=1 ( C∑ j=1 d (i) j ln 1 y (i) j − C∑ j=1 d (i) j ln 1 d (i) j ) = min T ,C − N∑ i=1 C∑ j=1 d (i) j lny (i) j , (23) where d(i)j and y (i) j represent the j-th…”
A Comprehensive Survey on Community Detection With Deep Learning (2021, background): “Sparse Filtering (SF) [134] is a two-layer learning model capable of handling high-dimensional graph data.”
Deep Learning of Invariant Features via Simulated Fixations in Video (2012, unknown): “9% Sparse Filtering [40] 53.”
Convolutional-Recursive Deep Learning for 3D Object Classification (2012, background): “Sparse coding is known to not scale well in terms of speed for large input dimensions [31].”
Unsupervised Deep Feature Extraction for Remote Sensing Image Classification (2015, methodology): “Sparsity is among the properties of a good feature representation [48], [50]–[53].”
A Digital-Twin-Assisted Fault Diagnosis Using Deep Transfer Learning (2019, background): “It could learn a compressed representation of input, even the number of hidden units is larger than input nodes [44].”
Unrolling Loopy Top-Down Semantic Feedback in Convolutional Deep Networks (2014, methodology): “Methods such as [17, 24, 22, 13, 5, 20] have been successfully used in the literature to extract sparse feature representations.”