Large-scale sparse logistic regression

By Jun Liu, Jianhui Chen, Jieping Ye. 2009

225 citations recorded by the source corpus.

Subfields: Artificial Intelligence, 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 Artificial Intelligence

In Computational Mechanics

In Computer Vision and Pattern Recognition

Citations in the corpus

Fast and Faster Convergence of SGD for Over-Parameterized Models and an Accelerated Perceptron (2018, methodology): “We denote the accelerated variant with and without line-search as Acc-SGD(LS) and Acc-SGD(T) respectively.”
Hyperspectral Image Classification Based on Structured Sparse Logistic Regression and Three-Dimensional Wavelet Texture Features (2013, methodology): “Here, adaptive backtracking line search [48] is used to select a proper step size t.”
An efficient algorithm for a class of fused lasso problems (2010, methodology): “For a given λ 2 , we can solve (40) by applying the solution corresponding to the large λ 1 as the “warm"-start to the smaller one; and this is the so-called “warm" start technique widely employed in the literature [8, 16].”
An improved GLMNET for l1-regularized logistic regression (2011, background): “Many optimization approaches have been proposed; for example, [2, 3, 4, 5, 6].”
GLMix: Generalized Linear Mixed Models For Large-Scale Response Prediction (2016, background): “Note that Equation (7) and (8) are both standard optimization problems for generalized linear model, for which many numerical optimizers are available for large data sets, such as L-BFGS, SGD, or TRON, and they can be easily parallelized for scalability [9, 14].”
Multi-Agent Distributed Optimization via Inexact Consensus ADMM (2014, unknown): “he computational burden of DC-ADMM. Conditions for global (linear) convergence are presented. September 12, 2014 DRAFT 5 Numerical examples for solving distributed sparse logistic regression problems [31] will show that the proposed IC-ADMM and IDC-ADMM methods converge much fasterthan the consensus”
A Stochastic Gradient Method with an Exponential Convergence Rate for Finite Training Sets (2012, methodology): “– AFG: Nesterov’s accelerated full gradient method [16], where iterations of (3) with a fixed step size are interleaved with an extrapolation step, and we use an adaptive line-search based on [27].”
Integrating low-rank and group-sparse structures for robust multi-task learning (2011, background): “APM has attracted extensive attentions in the machine learning and data mining communities [20, 19, 15, 8, 18, 21] due to its optimal convergence rate among all first-order techniques and its ability of dealing with large-scale non-smooth optimization problems.”
Addressing the straggler problem for iterative convergent parallel ML (2016, methodology): “Multinomial Logistic Regression (MLR) is a popular model for multi-way classification, such as used in the last layer of deep learning models for image classification [33] or text classification [37].”
Adaptive Consensus ADMM for Distributed Optimization (2017, methodology): “We also acquire large empirical datasets from the LIBSVM webpage (Liu et al., 2009), as well as MNIST digital images (LeCun et al.”
A Decentralized Proximal-Gradient Method With Network Independent Step-Sizes and Separated Convergence Rates (2017, methodology): “We consider a decentralized sparse logistic regression problem to classify the colon-cancer data [58].”
Face Liveness Detection from a Single Image with Sparse Low Rank Bilinear Discriminative Model (2010, methodology): “The sparse logistic regression [9,14] add a `1-norm regularization to the loss to avoid overfitting, i.”
Recent Advances of Large-Scale Linear Classification (2012, methodology): “, an interior-point method (l1 logreg [33]), (block) coordinate descent methods (BBR [34] and CGD [35]), a quasi-Newton method (OWL-QN [36]), Newton-type methods (GLMNET [37] and LIBLINEAR [19]), and a Nesterov’s method (SLEP [38]).”
Minimizing finite sums with the stochastic average gradient (2013, methodology): “We used an adaptive linesearch to estimate a local L based on the variant proposed for `2-regularized logistic regression by Liu et al. [2009]. • L-BFGS: A publicly-available limited-memory quasi-Newton method that has been tuned for log-linear models such as logistic regression [Schmidt, 2005]. Thi”
A Comparison of Optimization Methods and Software for Large-scale L1-regularized Linear Classification (2010, methodology): “In a manner similar to (15), Liu et al. (2009) proposed a gradient projection method called”
Active Sampler: Light-weight Accelerator for Complex Data Analytics at Scale (2015, methodology): “We shall evaluate the speedup of Active Sampler using a set of popular benchmark tasks, namely the hand-written digit classiﬁcation on MNIST using SVM [25], malicious URL detection in URL using feature selection [16], and image classiﬁcation on CIFAR10 using CNN [12].”
Multi-Task Feature Learning Via Efficient l2, 1-Norm Minimization (2009, methodology): “We plan to improve the practical performance of the proposed algorithms using the adaptive line search proposed in [11], and apply the proposed algorithms for solving the 2,1-norm regularized problems including the group Lasso [27], the logistic group-lasso [13], and the group-lasso for generalized ”
SLEP: Sparse Learning with Efficient Projections (2011, methodology): “The use of thel1 norm regularization has led to sparse models for linear regression [63], principle component analys is [75], linear discriminant analysis [70], canonical correlation analysis [59, 69], partial least squares [26], support vector machines [74], and logistic regression [25, 32, 58, 62]”