arXiv subjects

Sheng Chen

Publications and source records attributed to Sheng Chen.

At least 181 records · Page 10Linked to original sources

arXiv2017

Boundary Effects on Population Dynamics in Stochastic Lattice Lotka-Volterra Models

We investigate spatially inhomogeneous versions of the stochastic Lotka-Volterra model for predator-prey competition and coexistence by means of Monte Carlo simulations on a two-dimensional lattice with periodic boundary conditions. To study boundary effects for this paradigmatic population dynamics system, we employ a simulation domain split into two patches: Upon setting the predation rates at two distinct values, one half of the system resides in an absorbing state where only the prey survives, while the other half attains a stable coexistence state wherein both species remain active. At the domain boundary, we observe a marked enhancement of the predator population density. The predator correlation length displays a minimum at the boundary, before reaching its asymptotic constant value deep in the active region. The frequency of the population oscillations appears only very weakly affected by the existence of two distinct domains, in contrast to their attenuation rate, which assumes its largest value there. We also observe that boundary effects become less prominent as the system is successively divided into subdomains in a checkerboard pattern, with two different reaction rates assigned to neighboring patches. When the domain size becomes reduced to the scale of the correlation length, the mean population densities attain values that are very similar to those in a disordered system with randomly assigned reaction rates drawn from a bimodal distribution.

cond-mat.stat-mech↗

arXiv2017

DocTag2Vec: An Embedding Based Multi-label Learning Approach for Document Tagging

Tagging news articles or blog posts with relevant tags from a collection of predefined ones is coined as document tagging in this work. Accurate tagging of articles can benefit several downstream applications such as recommendation and search. In this work, we propose a novel yet simple approach called DocTag2Vec to accomplish this task. We substantially extend Word2Vec and Doc2Vec---two popular models for learning distributed representation of words and documents. In DocTag2Vec, we simultaneously learn the representation of words, documents, and tags in a joint vector space during training, and employ the simple $k$-nearest neighbor search to predict tags for unseen documents. In contrast to previous multi-label learning methods, DocTag2Vec directly deals with raw text instead of provided feature vector, and in addition, enjoys advantages like the learning of tag representation, and the ability of handling newly created tags. To demonstrate the effectiveness of our approach, we conduct experiments on several datasets and show promising results against state-of-the-art methods.

cs.CL↗

arXiv2017

Remote Channel Inference for Beamforming in Ultra-Dense Hyper-Cellular Network

In this paper, we propose a learning-based low-overhead channel estimation method for coordinated beamforming in ultra-dense networks. We first show through simulation that the channel state information (CSI) of geographically separated base stations (BSs) exhibits strong non-linear correlations in terms of mutual information. This finding enables us to adopt a novel learning-based approach to remotely infer the quality of different beamforming patterns at a dense-layer BS based on the CSI of an umbrella control-layer BS. The proposed scheme can reduce channel acquisition overhead by replacing pilot-aided channel estimation with the online inference from an artificial neural network, which is fitted offline. Moreover, we propose to exploit joint learning of multiple CBSs and involve more candidate beam patterns to obtain better performance. Simulation results based on stochastic ray-tracing channel models show that the proposed scheme can reach an accuracy of 99.74% in settings with 20 beamforming patterns.

cs.IT↗

arXiv2017

Laguerre Functions and Their Applications to Tempered Fractional Differential Equaitons on Infinite Intervals

Tempered fractional derivatives originated from the tempered fractional diffusion equations (TFDEs) modeled on the whole space R (see [23]). For numerically solving TFDEs, two kinds of generalized Laguerre functions were defined and some important properties were proposed to establish the approximate theory. The related prototype tempered fractional differ- ential problems was proposed and solved as the guidance. TFDEs are numerically solved by two domains Laguerre spectral method and the numerical experiments show some properties of the TFDEs and verify the efficiency of the spectral scheme.

math.NA↗

arXiv2017

Equation of state for random sphere packing with arbitrary adhesion and friction

We systematically generate a large set of random micro-particle packings over a wide range of adhesion and friction by means of adhesive contact dynamics simulation. The ensemble of generated packings covers a range of volume fraction $ϕ$ from $0.135 \pm 0.007$ to $0.639 \pm 0.004$, and of coordination number $Z$ from $2.11 \pm 0.03$ to $6.40 \pm 0.06$. We determine $ϕ$ and $Z$ at four limits (random close packing, random loose packing, adhesive close packing, and adhesive loose packing), and find a universal equation of state $ϕ(Z)$ to describe packings with arbitrary adhesion and friction. From a mechanical equilibrium analysis, we determine a critical friction coefficient $μ_{\rm f, c}$: when the friction coefficient $μ_{\rm f}$ is below $μ_{\rm f, c}$, particles' rearrangements are dominated by sliding, otherwise, they are dominated by rolling. Because of this reason, both $ϕ(μ_{\rm f})$ and $Z(μ_{\rm f})$ change sharply across $μ_{\rm f, c}$. Finally, we generalize the Maxwell counting argument to micro-particle packings, and show that the loosest packing, i.e., adhesive loose packing, satisfies the isostatic condition at $Z=2$.

cond-mat.soft↗

arXiv2016

Alternating Estimation for Structured High-Dimensional Multi-Response Models

We consider learning high-dimensional multi-response linear models with structured parameters. By exploiting the noise correlations among responses, we propose an alternating estimation (AltEst) procedure to estimate the model parameters based on the generalized Dantzig selector. Under suitable sample size and resampling assumptions, we show that the error of the estimates generated by AltEst, with high probability, converges linearly to certain minimum achievable level, which can be tersely expressed by a few geometric measures, such as Gaussian width of sets related to the parameter structure. To the best of our knowledge, this is the first non-asymptotic statistical guarantee for such AltEst-type algorithm applied to estimation problem with general structures.

stat.ML↗

arXiv2016

Compressive Sensing Based Multi-User Detector for the Large-Scale SM-MIMO Uplink

Conventional spatial modulation (SM) is typically considered for transmission in the downlink of small-scale MIMO systems, where a single one of a set of antenna elements (AEs) is activated for implicitly conveying extra bits. By contrast, inspired by the compelling benefits of large-scale MIMO (LS- MIMO) systems, here we propose a LS-SM-MIMO scheme for the uplink (UL), where each user having multiple AEs but only a single radio frequency (RF) chain invokes SM for increasing the UL-throughput. At the same time, by relying on hundreds of AEs but a small number of RF chains, the base station (BS) can simultaneously serve multiple users whilst reducing the power consumption. Due to the large number of AEs of the UL-users and the comparably small number of RF chains at the BS, the UL multi-user signal detection becomes a challenging large-scale under-determined problem. To solve this problem, we propose a joint SM transmission scheme and a carefully designed structured compressive sensing (SCS)-based multi-user detector (MUD) to be used at the users and BS, respectively. Additionally, the cyclic- prefix single-carrier (CPSC) is used to combat the multipath channels, and a simple receive AE selection is used for the improved performance over correlated Rayleigh-fading MIMO channels. We demonstrate that the aggregate SM signal consisting of SM signals of multiple UL-users in one CPSC block appears the distributed sparsity. Moreover, due to the joint SM transmission scheme, aggregate SM signals in the same transmission group exhibit the group sparsity. By exploiting these intrinsically sparse features, the proposed SCS-based MUD can reliably detect the resultant SM signals with low complexity. Simulation results demonstrate that the proposed SCS-based MUD achieves a better signal detection performance than its counterparts even with higher UL-throughtput.

cs.IT↗

arXiv2016

Structured Matrix Recovery via the Generalized Dantzig Selector

In recent years, structured matrix recovery problems have gained considerable attention for its real world applications, such as recommender systems and computer vision. Much of the existing work has focused on matrices with low-rank structure, and limited progress has been made matrices with other types of structure. In this paper we present non-asymptotic analysis for estimation of generally structured matrices via the generalized Dantzig selector under generic sub-Gaussian measurements. We show that the estimation error can always be succinctly expressed in terms of a few geometric measures of suitable sets which only depend on the structure of the underlying true matrix. In addition, we derive the general bounds on these geometric measures for structures characterized by unitarily invariant norms, which is a large family covering most matrix norms of practical interest. Examples are provided to illustrate the utility of our theoretical development.

stat.ML↗

arXiv2016

Non-equilibrium relaxation in a stochastic lattice Lotka-Volterra model

We employ Monte Carlo simulations to study a stochastic Lotka-Volterra model on a two-dimensional square lattice with periodic boundary conditions. If the (local) prey carrying capacity is finite, there exists an extinction threshold for the predator population that separates a stable active two-species coexistence phase from an inactive state wherein only prey survive. Holding all other rates fixed, we investigate the non-equilibrium relaxation of the predator density in the vicinity of the critical predation rate. As expected, we observe critical slowing-down, i.e., a power law dependence of the relaxation time on the predation rate, and algebraic decay of the predator density at the extinction critical point. The numerically determined critical exponents are in accord with the established values of the directed percolation universality class. Following a sudden predation rate change to its critical value, one finds critical aging for the predator density autocorrelation function that is also governed by universal scaling exponents. This aging scaling signature of the active-to-absorbing state phase transition emerges at significantly earlier times than the stationary critical power laws, and could thus serve as an advanced indicator of the (predator) population's proximity to its extinction threshold.

q-bio.PE↗

arXiv2015

Estimation with Norm Regularization

Analysis of non-asymptotic estimation error and structured statistical recovery based on norm regularized regression, such as Lasso, needs to consider four aspects: the norm, the loss function, the design matrix, and the noise model. This paper presents generalizations of such estimation error analysis on all four aspects compared to the existing literature. We characterize the restricted error set where the estimation error vector lies, establish relations between error sets for the constrained and regularized problems, and present an estimation error bound applicable to any norm. Precise characterizations of the bound is presented for isotropic as well as anisotropic subGaussian design matrices, subGaussian noise models, and convex loss functions, including least squares and generalized linear models. Generic chaining and associated results play an important role in the analysis. A key result from the analysis is that the sample complexity of all such estimators depends on the Gaussian width of a spherical cap corresponding to the restricted error set. Further, once the number of samples $n$ crosses the required sample complexity, the estimation error decreases as $\frac{c}{\sqrt{n}}$, where $c$ depends on the Gaussian width of the unit norm ball.

stat.ML↗

arXiv2015

Priori-Information Aided Iterative Hard Threshold: A Low-Complexity High-Accuracy Compressive Sensing Based Channel Estimation for TDS-OFDM

This paper develops a low-complexity channel estimation (CE) scheme based on compressive sensing (CS) for time-domain synchronous (TDS) orthogonal frequency-division multiplexing (OFDM) to overcome the performance loss under doubly selective fading channels. Specifically, an overlap-add method of the time-domain training sequence is first proposed to obtain the coarse estimates of the channel length, path delays and path gains of the wireless channel, by exploiting the channel's temporal correlation to improve the robustness of the coarse CE under the severe fading channel with long delay spread. We then propose the priori-information aided (PA) iterative hard threshold (IHT) algorithm, which utilizes the priori information of the acquired coarse estimate for the wireless channel and therefore is capable of obtaining an accurate channel estimate of the doubly selective fading channel. Compared with the classical IHT algorithm whose convergence requires the $l_2$ norm of the measurement matrix being less than 1, the proposed PA-IHT algorithm exploits the priori information acquired to remove such a limitation as well as to reduce the number of required iterations. Compared with the existing CS based CE method for TDS-OFDM, the proposed PA-IHT algorithm significantly reduces the computational complexity of CE as well as enhances the CE accuracy. Simulation results demonstrate that, without sacrificing spectral efficiency and changing the current TDS-OFDM signal structure, the proposed scheme performs better than the existing CE schemes for TDS-OFDM in various scenarios, especially under severely doubly selective fading channels.

cs.IT↗

arXiv2015

Computer simulation of random loose packings of micro-particles in presence of adhesion and friction

With a novel 3D discrete-element method specially developed with adhesive contact mechanics, random loose packings of uniform spherical micron-sized particles are fully investigated. The results show that large velocity, large size or weak adhesion can produce a relatively dense packing when other parameters are fixed, and these combined effects can be characterized by a dimensionless adhesion parameter ( $Ad=ω/2ρ_pU^2_0R$). Four regimes are identified based on the value of $Ad$: RCP regime with $Ad<\sim 0.01$; RLP regime with $\sim 0.01 20$. Force distribution of these adhesive loose packings follows $P(f)\sim f^θ$ for small forces and $P(f)\sim \exp^{-βf}$ for big forces, respectively, which shares a similar form with that in packings without adhesion but results in distinct exponents of $θ=0.879$, $β=0.839$. A local mechanical equilibrium analysis shows that adhesion enhances both sliding and rolling resistance so that fewer neighbours are needed to satisfy the force and torque balance.

cond-mat.soft↗

arXiv2015

Effect of long-range repulsive Coulomb interactions on packing structure of adhesive particles

The packing of charged micron-sized particles was investigated using discrete element simulations based on adhesive contact dynamic model. The formation process and the final obtained structures of ballistic packings are studied to show the effect of interparticle Coulomb force. It was found that increasing the charge on particles causes a remarkable decrease of the packing volume fraction ϕand the average coordination number Z, indicating a looser and chainlike structure. Force-scaling analysis shows that the long-range Coulomb interaction changes packing structures through its influence on particle inertia before they are bonded into the force networks. Once contact networks are formed, the expansion effect caused by repulsive Coulomb forces are dominated by short-range adhesion. Based on abundant results from simulations, a dimensionless adhesion parameter Ad* , which combines the effects of the particle inertia, the short-range adhesion and the long-range Coulomb interaction, is proposed and successfully scales the packing results for micron-sized particles within the latestly derived adhesive loose packing (ALP) regime. The structural properties of our packings follow well the recent theoretical prediction which is described by an ensemble approach based on a coarse-grained volume function, indicating some kind of universality in the low packing density regime of the phase diagram regardless of adhesion or particle charge. Based on the comprehensive consideration of the complicated inter-particle interactions, our findings provide insight into the roles of short-range adhesion and repulsive Coulomb force during packing formation and should be useful for further design of packings.

cond-mat.soft↗

arXiv2015

l1-norm Penalized Orthogonal Forward Regression

A l1-norm penalized orthogonal forward regression (l1-POFR) algorithm is proposed based on the concept of leaveone- out mean square error (LOOMSE). Firstly, a new l1-norm penalized cost function is defined in the constructed orthogonal space, and each orthogonal basis is associated with an individually tunable regularization parameter. Secondly, due to orthogonal computation, the LOOMSE can be analytically computed without actually splitting the data set, and moreover a closed form of the optimal regularization parameter in terms of minimal LOOMSE is derived. Thirdly, a lower bound for regularization parameters is proposed, which can be used for robust LOOMSE estimation by adaptively detecting and removing regressors to an inactive set so that the computational cost of the algorithm is significantly reduced. Illustrative examples are included to demonstrate the effectiveness of this new l1-POFR approach.

cs.LG↗

arXiv2015

Super Hom-Gel'fand-Dorfman bialgebras and Hom-Lie conformal superalgebras

The purpose of this paper is to introduce and study super Hom-Gel'fand-Dorfman bialgebras and Hom-Lie conformal superalgebras. In this paper, we provide different ways for constructing super Hom-Gel'fand-Dorfman bialgebras and obtain some infinite-dimensional Hom-Lie superalgebras from affinization of super Hom-Gel'fand-Dorfman bialgebras. Also, we give a general construction of Hom-Lie conformal superalgebras from Hom-Lie superalgebras and establish equivalence of quadratic Hom-Lie conformal superalgebras and super Hom-Gel'fand-Dorfman bialgebras. Finally, we characterize one-dimensional central extensions of quadratic Hom-Lie conformal superalgebras by using certain bilinear forms of super Hom-Gel'fand-Dorfman bialgebras.

math.QA↗

arXiv2015

Soft Pilot Reuse and Multi-Cell Block Diagonalization Precoding for Massive MIMO Systems

The users at cell edge of a massive multiple-input multiple-output (MIMO) system suffer from severe pilot contamination, which leads to poor quality of service (QoS). In order to enhance the QoS for these edge users, soft pilot reuse (SPR) combined with multi-cell block diagonalization (MBD) precoding are proposed. Specifically, the users are divided into two groups according to their large-scale fading coefficients, referred to as the center users, who only suffer from modest pilot contamination and the edge users, who suffer from severe pilot contamination. Based on this distinction, the SPR scheme is proposed for improving the QoS for the edge users, whereby a cell-center pilot group is reused for all cell-center users in all cells, while a cell-edge pilot group is applied for the edge users in the adjacent cells. By extending the classical block diagonalization precoding to a multi-cell scenario, the MBD precoding scheme projects the downlink transmit signal onto the null space of the subspace spanned by the inter-cell channels of the edge users in adjacent cells. Thus, the inter-cell interference contaminating the edge users' signals in the adjacent cells can be efficiently mitigated and hence the QoS of these edge users can be further enhanced. Our theoretical analysis and simulation results demonstrate that both the uplink and downlink rates of the edge users are significantly improved, albeit at the cost of the slightly decreased rate of center users.

cs.IT↗

arXiv2015

Generalized Dantzig Selector: Application to the k-support norm

We propose a Generalized Dantzig Selector (GDS) for linear models, in which any norm encoding the parameter structure can be leveraged for estimation. We investigate both computational and statistical aspects of the GDS. Based on conjugate proximal operator, a flexible inexact ADMM framework is designed for solving GDS, and non-asymptotic high-probability bounds are established on the estimation error, which rely on Gaussian width of unit norm ball and suitable set encompassing estimation error. Further, we consider a non-trivial example of the GDS using $k$-support norm. We derive an efficient method to compute the proximal operator for $k$-support norm since existing methods are inapplicable in this setting. For statistical analysis, we provide upper bounds for the Gaussian widths needed in the GDS analysis, yielding the first statistical recovery guarantee for estimation with the $k$-support norm. The experimental results confirm our theoretical analysis.

stat.ML↗

arXiv2015

A Survey of directed graphs invariants

In this paper, various kinds of invariants of directed graphs are summarized. In the first topic, the invariant w(G) for a directed graph G is introduced, which is primarily defined by S. Chen and X.M. Chen to solve a problem of weak connectedness of tensor product of two directed graphs. Further, we present our recent studies on the invariant w(G) in categorical view. In the second topic, Homology theory on directed graph is introduced, and we also cast on categorical view of the definition. The third topic mainly focuses on Laplacians on graphs, including traditional work and latest result of 1-laplacian by K.C.Chang. Finally, Zeta functions and Graded graphs are introduced, inclduing Bratteli-Vershik diagram, dual graded graphs and differential posets, with some applications in dynamic system.

math.CO↗