arXiv subjects

Jun Liu

Publications and source records attributed to Jun Liu.

At least 811 records · Page 45Linked to original sources

arXiv2016

Managing Deadline-constrained Bag-of-Tasks Jobs on Hybrid Clouds

Outsourcing jobs to a public cloud is a cost-effective way to address the problem of satisfying the peak resource demand when the local cloud has insufficient resources. In this paper, we study on managing deadline-constrained bag-of-tasks jobs on hybrid clouds. We present a binary nonlinear programming (BNP) problem to model the hybrid cloud management where the utilization of physical machines (PMs) in the local cloud/cluster is maximized when the local resources are enough to satisfy the deadline constraints of jobs, while when not, the rent cost from the public cloud is minimized. To solve this BNP problem in polynomial time, we proposed a heuristic algorithm. Its main idea is assigning the task closest to its deadline to current core until the core cannot finish any task within its deadline. When there is no available core, the algorithm adds an available PM with most capacity or rents a new VM with highest cost-performance ratio. Extensive experimental results show that our heuristic algorithm saves 16.2%-76% rent cost and improves 47.3%-182.8% resource utilizations satisfying deadline constraints, compared with first fit decreasing algorithm.

cs.DC↗

arXiv2016

NTU RGB+D: A Large Scale Dataset for 3D Human Activity Analysis

Recent approaches in depth-based human activity analysis achieved outstanding performance and proved the effectiveness of 3D representation for classification of action classes. Currently available depth-based and RGB+D-based action recognition benchmarks have a number of limitations, including the lack of training samples, distinct class labels, camera views and variety of subjects. In this paper we introduce a large-scale dataset for RGB+D human action recognition with more than 56 thousand video samples and 4 million frames, collected from 40 distinct subjects. Our dataset contains 60 different action classes including daily, mutual, and health-related actions. In addition, we propose a new recurrent neural network structure to model the long-term temporal correlation of the features for each body part, and utilize them for better action classification. Experimental results show the advantages of applying deep learning methods over state-of-the-art hand-crafted features on the suggested cross-subject and cross-view evaluation criteria for our dataset. The introduction of this large scale dataset will enable the community to apply, develop and adapt various data-hungry learning techniques for the task of depth-based and RGB+D-based human activity analysis.

cs.CV↗

arXiv2016

Improved Sufficient Conditions for Exact Convex Relaxation of Storage-Concerned ED

To avoid simultaneous charging and discharging of storages, complementarity constraints are introduced to storage-concerned economic dispatch (ED), which makes the problem non-convex. This letter concerns the conditions under which the convex relaxation of storage-concerned ED with complementarity constraints is exact. Two new sufficient conditions are proposed, proved and verified to significantly reduce the conservatism of recent results [3], [4].

math.OC↗

arXiv2016

Littlewood-Paley Characterizations of Anisotropic Hardy-Lorentz Spaces

Let $p\in(0,1]$, $q\in(0,\infty]$ and $A$ be a general expansive matrix on $\mathbb{R}^n$. Let $H^{p,q}_A(\mathbb{R}^n)$ be the anisotropic Hardy-Lorentz spaces associated with $A$ defined via the non-tangential grand maximal function. In this article, the authors characterize $H^{p,q}_A(\mathbb{R}^n)$ in terms of the Lusin-area function, the Littlewood-Paley $g$-function or the Littlewood-Paley $g_λ^*$-function via first establishing an anisotropic Fefferman-Stein vector-valued inequality in the Lorentz space $L^{p,q}(\mathbb{R}^n)$. All these characterizations are new even for the classical isotropic Hardy-Lorentz spaces on $\mathbb{R}^n$. Moreover, the range of $λ$ in the $g_λ^*$-function characterization of $H^{p,q}_A(\mathbb{R}^n)$ coincides with the best known one in the classical Hardy space $H^p(\mathbb{R}^n)$ or in the anisotropic Hardy space $H^p_A(\mathbb{R}^n)$.

math.CA↗

arXiv2016

Littlewood-Paley Characterizations of Hajłasz-Sobolev and Triebel-Lizorkin Spaces via Averages on Balls

Let $p\in(1,\infty)$ and $q\in[1,\infty)$. In this article, the authors characterize the Triebel-Lizorkin space ${F}^α_{p,q}(\mathbb{R}^n)$ with smoothness order $α\in(0,2)$ via the Lusin-area function and the $g_λ^*$-function in terms of difference between $f(x)$ and its average $B_tf(x):=\frac1{|B(x,t)|}\int_{B(x,t)}f(y)\,dy$ over a ball $B(x,t)$ centered at $x\in\mathbb{R}^n$ with radius $t\in(0,1)$. As an application, the authors obtain a series of characterizations of $F^α_{p,\infty}(\mathbb{R}^n)$ via pointwise inequalities, involving ball averages, in spirit close to Hajłasz gradients, here an interesting phenomena naturally appears that, in the end-point case when $α=2$, these pointwise inequalities characterize the Triebel-Lizorkin spaces $F^2_{p,2}(\mathbb{R}^n)$, while not $F^2_{p,\infty}(\mathbb{R}^n)$. In particular, some new pointwise characterizations of Hajłasz-Sobolev spaces via ball averages are obtained. Since these new characterizations only use ball averages, they can be used as starting points for developing a theory of Triebel-Lizorkin spaces with smoothness orders not less than $1$ on spaces of homogeneous type.

math.CA↗

arXiv2016

Forward model with space-variant of source size for reconstruction on x-ray radiographic image

Forward imaging technique is the base of combined method on density reconstruction with the forward calculation and inverse problem solution. In the paper, we introduced the projection equation for the radiographic system with areal source blur and detector blur, gained the projecting matrix from any point source to any detector pixel with x-ray trace technique, proposed the ideal on gridding the areal source as many point sources with different weights, and used the blurring window as the effect of the detector blur. We used the forward projection equation to gain the same deviation information about the object edge as the experimental image. Our forward projection equation is combined with Constrained Conjugate Gradient method to form a new method for density reconstruction, XTRACE-CCG. The new method worked on the simulated image of French Test Object and experimental image. The same results have been concluded the affecting range of the blur is decreased and can be controlled to one or two pixels. The method is also suitable for reconstruction of density-variant object. The capability of our method to handle blur effect is useful for all radiographic systems with larger source size comparing to pixel size.

physics.med-ph↗

arXiv2015

The Chiral Anomaly and Ultrahigh Mobility in Crystalline HfTe5

HfTe5 is predicted to be a promising platform for studying topological phases. Here through an electrical transport study, we present the first observation of chiral anomaly and ultrahigh mobility in HfTe5 crystals. Negative magneto-resistivity in HfTe5 is observed when the external magnetic and electrical fields are parallel (B//E) and quickly disappears once B deviates from the direction of E. Quantitative fitting further confirms the chiral anomaly as the underlying physics. Moreover, by analyzing the conductivity tensors of longitudinal and Hall traces, ultrahigh mobility and ultralow carrier density are revealed in HfTe5, which paves the way for potential electronic applications.

cond-mat.mtrl-sci↗

arXiv2015

Successive Ray Refinement and Its Application to Coordinate Descent for LASSO

Coordinate descent is one of the most popular approaches for solving Lasso and its extensions due to its simplicity and efficiency. When applying coordinate descent to solving Lasso, we update one coordinate at a time while fixing the remaining coordinates. Such an update, which is usually easy to compute, greedily decreases the objective function value. In this paper, we aim to improve its computational efficiency by reducing the number of coordinate descent iterations. To this end, we propose a novel technique called Successive Ray Refinement (SRR). SRR makes use of the following ray continuation property on the successive iterations: for a particular coordinate, the value obtained in the next iteration almost always lies on a ray that starts at its previous iteration and passes through the current iteration. Motivated by this ray-continuation property, we propose that coordinate descent be performed not directly on the previous iteration but on a refined search point that has the following properties: on one hand, it lies on a ray that starts at a history solution and passes through the previous iteration, and on the other hand, it achieves the minimum objective function value among all the points on the ray. We propose two schemes for defining the search point and show that the refined search point can be efficiently obtained. Empirical results for real and synthetic data sets show that the proposed SRR can significantly reduce the number of coordinate descent iterations, especially for small Lasso regularization parameters.

cs.LG↗

arXiv2015

Anisotropic Hardy-Lorentz Spaces and Their Applications

Let $p\in(0,1]$, $q\in(0,\infty]$ and $A$ be a general expansive matrix on $\mathbb{R}^n$. The authors introduce the anisotropic Hardy-Lorentz space $H^{p,q}_A(\mathbb{R}^n)$ associated with $A$ via the non-tangential grand maximal function and then establish its various real-variable characterizations in terms of the atomic or the molecular decompositions, the radial or the non-tangential maximal functions, or the finite atomic decompositions. All these characterizations except the $\infty$-atomic characterization are new even for the classical isotropic Hardy-Lorentz spaces on $\mathbb{R}^n$. As applications, the authors first prove that $H^{p,q}_A(\mathbb{R}^n)$ is an intermediate space between $H^{p_1,q_1}_A(\mathbb{R}^n)$ and $H^{p_2,q_2}_A(\mathbb{R}^n)$ with $0<p_1<p<p_2<\infty$ and $q_1,\,q,\,q_2\in(0,\infty]$, and also between $H^{p,q_1}_A(\mathbb{R}^n)$ and $H^{p,q_2}_A(\mathbb{R}^n)$ with $p\in(0,\infty)$ and $0<q_1<q<q_2\leq\infty$ in the real method of interpolation. The authors then establish a criterion on the boundedness of sublinear operators from $H^{p,q}_A(\mathbb{R}^n)$ into a quasi-Banach space; moreover, the authors obtain the boundedness of $δ$-type Calderón-Zygmund operators from $H^p_A(\mathbb{R}^n)$ to the weak Lebesgue space $L^{p,\infty}(\mathbb{R}^n)$ (or $H^{p,\infty}_A(\mathbb{R}^n)$) in the critical case, from $H_A^{p,q}(\mathbb{R}^n)$ to $L^{p,q}(\mathbb{R}^n)$ (or $H_A^{p,q}(\mathbb{R}^n)$) with $δ\in(0,\frac{\lnλ_-}{\ln b}]$, $p\in(\frac1{1+δ},1]$ and $q\in(0,\infty]$, as well as the boundedness of some Calderón-Zygmund operators from $H_A^{p,q}(\mathbb{R}^n)$ to $L^{p,\infty}(\mathbb{R}^n)$, where $b:=|\det A|$, $λ_-:=\min\{|λ|:\ λ\inσ(A)\}$ and $σ(A)$ denotes the set of all eigenvalues of $A$.

math.CA↗

arXiv2015

Correlation between the spin Hall angle and the structural phases of early 5d transition metals

We have studied the relationship between the structure and the spin Hall angle of the early 5d transition metals in X/CoFeB/MgO (X=Hf, Ta, W, Re) heterostructures. Spin Hall magnetoresistance (SMR) is used to characterize the spin Hall angle of the heavy metals. Transmission electron microscopy images show that all underlayers are amorphous-like when their thicknesses are small, however, crystalline phases emerge as the thickness is increased for certain elements. We find that the heavy metal layer thickness dependence of the SMR reflects these changes in structure. The spin Hall angle largest |θ$_{SH}$| of Hf, Ta, W and Re (~0.11, 0.10, 0.23 and 0.07, respectively) is found when the dominant phase is amorphous-like. We find that the amorphous-like phase not only possesses large resistivity but also exhibits sizeable spin Hall conductivity, which both contribute to the emergence of the large spin Hall angle.

cond-mat.mtrl-sci↗

arXiv2015

Littlewood-Paley Characterizations of Fractional Sobolev Spaces via Averages on Balls

In this paper, the authors characterize Sobolev spaces $W^{α,p}({\mathbb R}^n)$ with the smoothness order $α\in(0,2]$ and $p\in(\max\{1, \frac{2n}{2α+n}\},\infty)$, via the Lusin area function and the Littlewood-Paley $g_λ^\ast$-function in terms of centered ball averages. The authors also show that the condition $p\in(\max\{1, \frac{2n}{2α+n}\},\infty)$ is nearly sharp in the sense that these characterizations are no longer true when $p\in (1,\max\{1, \frac{2n}{2α+n}\})$. These characterizations provide a new possible way to introduce fractional Sobolev spaces with smoothness order in $(1,2]$ on metric measure spaces.

math.CA↗

arXiv2015

Electric-field control of spin-orbit torque in a magnetically doped topological insulator

Electric-field manipulation of magnetic order has proved of both fundamental and technological importance in spintronic devices. So far, electric-field control of ferromagnetism, magnetization and magnetic anisotropy has been explored in various magnetic materials, but the efficient electric-field control of spin-orbit torque (SOT) still remains elusive. Here, we report the effective electric-field control of a giant SOT in a Cr-doped topological insulator (TI) thin film using a top-gate FET structure. The SOT strength can be modulated by a factor of 4 within the accessible gate voltage range, and it shows strong correlation with the spin-polarized surface current in the film. Furthermore, we demonstrate the magnetization switching by scanning gate voltage with constant current and in-plane magnetic field applied in the film. The effective electric-field control of SOT and the giant spin-torque efficiency in Cr-doped TI may lead to the development of energy-efficient gate-controlled spin-torque devices compatible with modern field-effect semiconductor technologies.

cond-mat.str-el↗

arXiv2015

Computing finite abstractions with robustness margins via local reachable set over-approximation

This paper proposes a method to compute finite abstractions that can be used for synthesizing robust hybrid control strategies for nonlinear systems. Most existing methods for computing finite abstractions utilize some global, analytical function to provide bounds on the reachable sets of nonlinear systems, which can be conservative and lead to spurious transitions in the abstract systems. This problem is even more pronounced in the presence of imperfect measurements and modelling uncertainties, where control synthesis can easily become infeasible due to added spurious transitions. To mitigate this problem, we propose to compute finite abstractions with robustness margins by over-approximating the local reachable sets of nonlinear systems. We do so by linearizing the nonlinear dynamics into linear affine systems and keeping track of the linearization error. It is shown that this approach provides tighter approximations and several numerical examples are used to illustrate of effectiveness of the proposed methods.

eess.SY↗

arXiv2015

Lyapunov-based sufficient conditions for stability of hybrid systems with memory

Hybrid systems with memory are dynamical systems exhibiting both hybrid and delay phenomena. In this note, we study the asymptotic stability of hybrid systems with memory using generalized concepts of solutions. These generalized solutions, motivated by studying robustness and well-posedness of such systems, are defined on hybrid time domains and parameterized by both continuous and discrete time. We establish Lyapunov-based sufficient conditions for asymptotic stability using both Lyapunov-Razumikhin functions and Lyapunov-Krasovskii functionals. Examples are provided to illustrate these conditions.

math.DS↗

arXiv2015

A PCA-Based Convolutional Network

In this paper, we propose a novel unsupervised deep learning model, called PCA-based Convolutional Network (PCN). The architecture of PCN is composed of several feature extraction stages and a nonlinear output stage. Particularly, each feature extraction stage includes two layers: a convolutional layer and a feature pooling layer. In the convolutional layer, the filter banks are simply learned by PCA. In the nonlinear output stage, binary hashing is applied. For the higher convolutional layers, the filter banks are learned from the feature maps that were obtained in the previous stage. To test PCN, we conducted extensive experiments on some challenging tasks, including handwritten digits recognition, face recognition and texture classification. The results show that PCN performs competitive with or even better than state-of-the-art deep learning models. More importantly, since there is no back propagation for supervised finetuning, PCN is much more efficient than existing deep networks.

cs.LG↗

arXiv2015

Anisotropic Magnetotransport and Exotic Longitudinal Linear Magnetoresistance in WTe2 Crystals

WTe2 semimetal, as a typical layered transition-metal dichalcogenide, has recently attracted much attention due to the extremely large, non-saturating parabolic magnetoresistance in perpendicular field. Here, we report a systematic study of the angular dependence of the magnetoresistance in WTe2 single crystal. The violation of the Kohler rule and a significant anisotropic magnetotransport behavior in different magnetic field directions are observed. Surprisingly, when the applied field is parallel to the tungsten chains of WTe2, an exotic large longitudinal linear magnetoresistance as high as 1200% at 15 T and 2 K is identified. Violation of the Kohler rule in transverse magnetoresistance can be understood based on a dual effect of the excitons formation and thermal activation, while large longitudinal linear magnetoresistance reflects perfectly the scattering and nesting of quasi-1D nature of this balanced hole-electron system. Our work will stimulate studies of such double-carrier correlated material and corresponding quantum physics.

cond-mat.mtrl-sci↗

arXiv2015

A coefficient average approximation towards Gutzwiller wavefunction formalism

Gutzwiller wavefunction is a physically well motivated trial wavefunction for describing correlated electron systems. In this work, a new approximation is introduced to facilitate evaluation of the expectation value of any operator within the Gutzwiller wavefunction formalism. The basic idea is to make use of a specially designed average over Gutzwiller wavefunction coefficients expanded in the many-body Fock space to approximate the ratio of expectation values between a Gutzwiller wavefunction and its underlying noninteracting wavefunction. To check with the standard Gutzwiller approximation (GA), we test its performance on single band systems and find quite interesting properties. On finite systems, we noticed that it gives superior performance than GA, while on infinite systems it asymptotically approaches GA. Analytic analysis together with numerical tests are provided to support this claimed asymptotic behavior. At the end, possible improvements on the approximation and its generalization towards multiband systems are illustrated and discussed.

cond-mat.str-el↗

arXiv2014

Anisotropic Fermi Surface and Quantum Limit Transport in High Mobility 3D Dirac Semimetal Cd3As2

The three-dimensional (3D) topological Dirac semimetal is a new topological phase of matter, viewed as the 3D analogy of graphene with a linear dispersion in the 3D momentum space. Here, we report the angular dependent magnetotransport in Cd3As2 single crystal and clearly show how the Fermi surface evolves when tilting the magnetic field. Remarkably, when the magnetic field lies in [112] and [44-1] axis, only single oscillation period features present, however, the system shows double period oscillations when the field is applied along [1-10] direction. Moreover, tilting the magnetic field at certain direction also gives double period oscillations. We attribute the anomalous oscillation behavior to the sophisticated geometry of Fermi surface and illustrate a complete 3D Fermi surfaces with two nested anisotropic ellipsoids around the Dirac point. Additionally, a sub-millimeter mean free path at 6 K is observed in Cd3As2 crystal, indicating a large ballistic transport region in this material. Tracking the magnetoresistance oscillations to 60 T, we reach the quantum limit (n = 1 Landau Level) at about 43 T. These results improve the knowledge of the Dirac semimetal material Cd3As2, and also pave the way for proposing new electronic applications based on 3D Dirac materials.

cond-mat.mtrl-sci↗