arXiv subjects

Li Shen

Publications and source records attributed to Li Shen.

461 records · Page 26Linked to original sources

arXiv2017

Study of the Magnetizing Relationship of the Kickers for CSNS

The extraction system of CSNS mainly consists of two kinds of magnets: eight kickers and one lambertson magnet. In this paper, firstly, the magnetic test results of the eight kickers were introduced and then the filed uniformity and magnetizing relationship of the kickers were given. Secondly, during the beam commissioning in the future, in order to obtain more accurate magnetizing relationship, a new method to measure the magnetizing coefficients of the kickers by the real extraction beam was given and the data analysis would also be processed.

physics.acc-ph↗

arXiv2017

Contour Detection from Deep Patch-level Boundary Prediction

In this paper, we present a novel approach for contour detection with Convolutional Neural Networks. A multi-scale CNN learning framework is designed to automatically learn the most relevant features for contour patch detection. Our method uses patch-level measurements to create contour maps with overlapping patches. We show the proposed CNN is able to to detect large-scale contours in an image efficienly. We further propose a guided filtering method to refine the contour maps produced from large-scale contours. Experimental results on the major contour benchmark databases demonstrate the effectiveness of the proposed technique. We show our method can achieve good detection of both fine-scale and large-scale contours.

cs.CV↗

arXiv2016

Weighted iteration complexity of the sPADMM on the KKT residuals for convex composite optimization

In this paper we establish an $\mathcal{O}({1}/{k})$ weighted iteration complexity on the KKT residuals yielded by the sPADMM (semi-proximal alternating direction method of multiplier) for the convex composite optimization problem. This result, which is derived with the help of a novel generalized HPE (hybrid proximal extra-gradient) iteration formula, first fills the gap on the ergodic iteration complexity of the classic ADMM with a large step-size and its many proximal variants.

math.OC↗

arXiv2016

A Unified Model for Differential Expression Analysis of RNA-seq Data via L1-Penalized Linear Regression

The RNA-sequencing (RNA-seq) is becoming increasingly popular for quantifying gene expression levels. Since the RNA-seq measurements are relative in nature, between-sample normalization of counts is an essential step in differential expression (DE) analysis. The normalization of existing DE detection algorithms is ad hoc and performed once for all prior to DE detection, which may be suboptimal since ideally normalization should be based on non-DE genes only and thus coupled with DE detection. We propose a unified statistical model for joint normalization and DE detection of log-transformed RNA-seq data. Sample-specific normalization factors are modeled as unknown parameters in the gene-wise linear models and jointly estimated with the regression coefficients. By imposing sparsity-inducing L1 penalty (or mixed L1/L2-norm for multiple treatment conditions) on the regression coefficients, we formulate the problem as a penalized least-squares regression problem and apply the augmented lagrangian method to solve it. Simulation studies show that the proposed model and algorithms outperform existing methods in terms of detection power and false-positive rate when more than half of the genes are differentially expressed and/or when the up- and down-regulated genes among DE genes are unbalanced in amount.

stat.ME↗

arXiv2016

Linear convergence of the generalized PPA and several splitting methods for the composite inclusion problem

For the inclusion problem involving two maximal monotone operators, under the metric subregularity of the composite operator, we derive the linear convergence of the generalized proximal point algorithm and several splitting algorithms, which include the over-relaxed forward-backward splitting algorithm, the generalized Douglas-Rachford splitting algorithm and Davis' three-operator splitting algorithm. To the best of our knowledge, this linear convergence condition is weaker than the existing ones that almost all require the strong monotonicity of the composite operator. Withal, we give some sufficient conditions to ensure the metric subregularity of the composite operator. At last, the preliminary numerical performances on some toy examples support the theoretical results.

math.OC↗

arXiv2016

Study on the magnetic measurement results of the injection system for CSNS/RCS

A combination of the H- stripping and phase space painting method is used to accumulate a high intensity beam in the Rapid Cycling Synchrotron (RCS) of the China Spallation Neutron Source (CSNS). The injection system for CSNS/RCS consists of three kinds of magnets: four direct current magnets (BC1-BC4), eight alternating current magnets (BH1-BH4 and BV1-BV4), two septum magnets (ISEP1 and ISEP2). In this paper, the magnetic measurements of the injection system were introduced and the data analysis was processed. The field uniformity and magnetizing curves of these magnets were given, and then the magnetizing fitting equations were obtained.

physics.acc-ph↗

arXiv2016

Relay Backpropagation for Effective Learning of Deep Convolutional Neural Networks

Learning deeper convolutional neural networks becomes a tendency in recent years. However, many empirical evidences suggest that performance improvement cannot be gained by simply stacking more layers. In this paper, we consider the issue from an information theoretical perspective, and propose a novel method Relay Backpropagation, that encourages the propagation of effective information through the network in training stage. By virtue of the method, we achieved the first place in ILSVRC 2015 Scene Classification Challenge. Extensive experiments on two challenging large scale datasets demonstrate the effectiveness of our method is not restricted to a specific dataset or network architecture. Our models will be available to the research community later.

cs.CV↗

arXiv2016

Co-occurrence Feature Learning for Skeleton based Action Recognition using Regularized Deep LSTM Networks

Skeleton based action recognition distinguishes human actions using the trajectories of skeleton joints, which provide a very good representation for describing actions. Considering that recurrent neural networks (RNNs) with Long Short-Term Memory (LSTM) can learn feature representations and model long-term temporal dependencies automatically, we propose an end-to-end fully connected deep LSTM network for skeleton based action recognition. Inspired by the observation that the co-occurrences of the joints intrinsically characterize human actions, we take the skeleton as the input at each time slot and introduce a novel regularization scheme to learn the co-occurrence features of skeleton joints. To train the deep LSTM network effectively, we propose a new dropout algorithm which simultaneously operates on the gates, cells, and output responses of the LSTM neurons. Experimental results on three human action recognition datasets consistently demonstrate the effectiveness of the proposed model.

cs.CV↗

arXiv2015

Inexact indefinite proximal ADMMs for 2-block separable convex programs and applications to 4-block DNNSDPs

This paper is concerned with two-block separable convex minimization problems with linear constraints, for which it is either impossible or too expensive to obtain the exact solutions of the subproblems involved in the proximal ADMM (alternating direction method of multipliers). Such structured convex minimization problems often arise from the two-block regroup settlement of three or four-block separable convex optimization problems with linear constraints, or from the constrained total-variation superresolution image reconstruction problems in image processing. For them, we propose an inexact indefinite proximal ADMM of step-size $τ\in\!(0,\frac{\sqrt{5}+1}{2})$ with two easily implementable inexactness criteria to control the solution accuracy of subproblems, and establish the convergence under a mild assumption on indefinite proximal terms. We apply the proposed inexact indefinite proximal ADMMs to the three or four-block separable convex minimization problems with linear constraints, which are from the duality of the important class of doubly nonnegative semidefinite programming (DNNSDP) problems with many linear equality and/or inequality constraints. Numerical results indicate that the inexact indefinite proximal ADMM with the absolute error criterion has a comparable performance with the directly extended multi-block ADMM of step-size $τ=1.618$ without convergence guarantee, whether in terms of the number of iterations or the computation time.

math.OC↗

arXiv2015

Shadow Optimization from Structured Deep Edge Detection

Local structures of shadow boundaries as well as complex interactions of image regions remain largely unexploited by previous shadow detection approaches. In this paper, we present a novel learning-based framework for shadow region recovery from a single image. We exploit the local structures of shadow edges by using a structured CNN learning framework. We show that using the structured label information in the classification can improve the local consistency of the results and avoid spurious labelling. We further propose and formulate a shadow/bright measure to model the complex interactions among image regions. The shadow and bright measures of each patch are computed from the shadow edges detected in the image. Using the global interaction constraints on patches, we formulate a least-square optimization problem for shadow recovery that can be solved efficiently. Our shadow recovery method achieves state-of-the-art results on the major shadow benchmark databases collected under various conditions.

cs.CV↗

arXiv2015

A corrected semi-proximal ADMM for multi-block convex optimization and its application to DNN-SDPs

In this paper we propose a corrected semi-proximal ADMM (alternating direction method of multipliers) for the general $p$-block $(p\!\ge 3)$ convex optimization problems with linear constraints, aiming to resolve the dilemma that almost all the existing modified versions of the directly extended ADMM, although with convergent guarantee, often perform substantially worse than the directly extended ADMM itself with no convergent guarantee. Specifically, in each iteration, we use the multi-block semi-proximal ADMM with step-size at least $1$ as the prediction step to generate a good prediction point, and then make correction as small as possible for the middle $(p\!-\!2)$ blocks of the prediction point. Among others, the step-size of the multi-block semi-proximal ADMM is adaptively determined by the infeasibility ratio made up by the current semi-proximal ADMM step for the one yielded by the last correction step. For the proposed corrected semi-proximal ADMM, we establish the global convergence results under a mild assumption, and apply it to the important class of doubly nonnegative semidefinite programming (DNN-SDP) problems with many linear equality and/or inequality constraints. Our extensive numerical tests show that the corrected semi-proximal ADMM is superior to the directly extended ADMM with step-size $τ=1.618$ and the multi-block ADMM with Gaussian back substitution \cite{HTY12,HY13}. It requires the least number of iterations for $70\%$ test instances within the comparable computing time with that of the directly extended ADMM, and for about $40\%$ tested problems, its number of iterations is only $67\%$ that of the multi-block ADMM with Gaussian back substitution \cite{HTY12,HY13}.

math.OC↗