arXiv subjects

Lei Wang

Publications and source records attributed to Lei Wang.

At least 1,207 records · Page 67Linked to original sources

arXiv2015

An Efficient Dynamic Programming Algorithm for STR-IC-SEQ-EC-LCS Problem

In this paper, we consider a generalized longest common subsequence problem, in which a constraining sequence of length $s$ must be included as a substring and the other constraining sequence of length $t$ must be excluded as a subsequence of two main sequences and the length of the result must be maximal. For the two input sequences $X$ and $Y$ of lengths $n$ and $m$, and the given two constraining sequences of length $s$ and $t$, we present an $O(nmst)$ time dynamic programming algorithm for solving the new generalized longest common subsequence problem. The time complexity can be reduced further to cubic time in a more detailed analysis. The correctness of the new algorithm is proved.

cs.DS↗

arXiv2015

Fractional fractal quantum Hall effect in graphene superlattices

The Hofstadter energy spectrum provides a uniquely tunable system to study emergent topological order in the regime of strong interactions. Previous experiments, however, have been limited to the trivial case of low Bloch band filling where only the Landau level index plays a significant role. Here we report measurement of high mobility graphene superlattices where the complete unit cell of the Hofstadter spectrum is accessible. We observe coexistence of conventional fractional quantum Hall effect (QHE) states together with the integer QHE states associated with the fractal Hofstadter spectrum. At large magnetic field, a new series of states appear at fractional Bloch filling index. These fractional Bloch band QHE states are not anticipated by existing theoretical pictures and point towards a new type of many-body state.

cond-mat.mes-hall↗

arXiv2015

Boosting-like Deep Learning For Pedestrian Detection

This paper proposes boosting-like deep learning (BDL) framework for pedestrian detection. Due to overtraining on the limited training samples, overfitting is a major problem of deep learning. We incorporate a boosting-like technique into deep learning to weigh the training samples, and thus prevent overtraining in the iterative process. We theoretically give the details of derivation of our algorithm, and report the experimental results on open data sets showing that BDL achieves a better stable performance than the state-of-the-arts. Our approach achieves 15.85% and 3.81% reduction in the average miss rate compared with ACF and JointDeep on the largest Caltech benchmark dataset, respectively.

cs.CV↗

arXiv2015

Identifying Dwarfs Workloads in Big Data Analytics

Big data benchmarking is particularly important and provides applicable yardsticks for evaluating booming big data systems. However, wide coverage and great complexity of big data computing impose big challenges on big data benchmarking. How can we construct a benchmark suite using a minimum set of units of computation to represent diversity of big data analytics workloads? Big data dwarfs are abstractions of extracting frequently appearing operations in big data computing. One dwarf represents one unit of computation, and big data workloads are decomposed into one or more dwarfs. Furthermore, dwarfs workloads rather than vast real workloads are more cost-efficient and representative to evaluate big data systems. In this paper, we extensively investigate six most important or emerging application domains i.e. search engine, social network, e-commerce, multimedia, bioinformatics and astronomy. After analyzing forty representative algorithms, we single out eight dwarfs workloads in big data analytics other than OLAP, which are linear algebra, sampling, logic operations, transform operations, set operations, graph operations, statistic operations and sort.

cs.DB↗

arXiv2015

An efficient dynamic programming algorithm for the generalized LCS problem with multiple substring inclusive constraints

In this paper, we consider a generalized longest common subsequence problem with multiple substring inclusive constraints. For the two input sequences $X$ and $Y$ of lengths $n$ and $m$, and a set of $d$ constraints $P=\{P_1,\cdots,P_d\}$ of total length $r$, the problem is to find a common subsequence $Z$ of $X$ and $Y$ including each of constraint string in $P$ as a substring and the length of $Z$ is maximized. A new dynamic programming solution to this problem is presented in this paper. The correctness of the new algorithm is proved. The time complexity of our algorithm is $O(d2^dnmr)$. In the case of the number of constraint strings is fixed, our new algorithm for the generalized longest common subsequence problem with multiple substring inclusive constraints requires $O(nmr)$ time and space.

cs.DS↗

arXiv2015

Learning Discriminative Stein Kernel for SPD Matrices and Its Applications

Stein kernel has recently shown promising performance on classifying images represented by symmetric positive definite (SPD) matrices. It evaluates the similarity between two SPD matrices through their eigenvalues. In this paper, we argue that directly using the original eigenvalues may be problematic because: i) Eigenvalue estimation becomes biased when the number of samples is inadequate, which may lead to unreliable kernel evaluation; ii) More importantly, eigenvalues only reflect the property of an individual SPD matrix. They are not necessarily optimal for computing Stein kernel when the goal is to discriminate different sets of SPD matrices. To address the two issues in one shot, we propose a discriminative Stein kernel, in which an extra parameter vector is defined to adjust the eigenvalues of the input SPD matrices. The optimal parameter values are sought by optimizing a proxy of classification performance. To show the generality of the proposed method, three different kernel learning criteria that are commonly used in the literature are employed respectively as a proxy. A comprehensive experimental study is conducted on a variety of image classification tasks to compare our proposed discriminative Stein kernel with the original Stein kernel and other commonly used methods for evaluating the similarity between SPD matrices. The experimental results demonstrate that, the discriminative Stein kernel can attain greater discrimination and better align with classification tasks by altering the eigenvalues. This makes it produce higher classification performance than the original Stein kernel and other commonly used methods.

cs.CV↗

arXiv2015

HEp-2 Cell Image Classification with Deep Convolutional Neural Networks

Efficient Human Epithelial-2 (HEp-2) cell image classification can facilitate the diagnosis of many autoimmune diseases. This paper presents an automatic framework for this classification task, by utilizing the deep convolutional neural networks (CNNs) which have recently attracted intensive attention in visual recognition. This paper elaborates the important components of this framework, discusses multiple key factors that impact the efficiency of training a deep CNN, and systematically compares this framework with the well-established image classification models in the literature. Experiments on benchmark datasets show that i) the proposed framework can effectively outperform existing models by properly applying data augmentation; ii) our CNN-based framework demonstrates excellent adaptability across different datasets, which is highly desirable for classification under varying laboratory settings. Our system is ranked high in the cell image classification competition hosted by ICPR 2014.

cs.CV↗

arXiv2015

Cosmic reionization of hydrogen and helium: contribution from both mini-quasars and stars

Observations on the high-redshift galaxies at $z>6$ imply that their ionizing emissivity is unable to fully reionize the Universe at $z\sim 6$. Either a high escape fraction of ionizing photons from these galaxies or a large population of faint galaxies below the detection limit are required. However, these requirements are somewhat in tension with present observations. In this work, we explored the combined contribution of mini-quasars and stars to the reionization of cosmic hydrogen and helium. Our model is roughly consistent with: (1) the low escape fractions of ionizing photons from the observed galaxies, (2) the optical depth of Cosmic Microwave Background (CMB) measured by the WMAP-7, and (3) the redshift of the end of hydrogen and helium reionization at $z\approx 6$ and $z\approx 3$, respectively. Neither an extremely high escape fraction nor a large population of fainter galaxies is required in this scenario. In our most optimistic model, more than $\sim20\%$ of the cosmic helium is reionized by $z\sim6$, and the ionized fraction of cosmic helium rapidly climbs to more than $50\%$ by $z\sim5$. These results may imply that better measurements of helium reionization, especially at high redshifts, could be helpful in constraining the growth of intermediate-mass black holes (IMBHs) in the early Universe, which would shed some light on the puzzles concerning the formation of supermassive black holes (SMBHs).

astro-ph.CO↗

arXiv2015

A light pseudoscalar of 2HDM confronted with muon g-2 and experimental constraints

A light pseudoscalar of the lepton-specific 2HDM can enhance the muon g-2, but suffer from various constraints easily, such as the 125.5 GeV Higgs signals, non-observation of additional Higgs at the collider and even $B_s\to μ^+μ^-$. In this paper, we take the light CP-even Higgs as the 125.5 GeV Higgs, and examine the implications of those observables on a pseudoscalar with the mass below the half of 125.5 GeV. Also the other relevant theoretical and experimental constraints are considered. We find that the pseudoscalar can be allowed to be as low as 10 GeV, but the corresponding $\tanβ$, $\sin(β-α)$ and the mass of charged Higgs are strongly constrained. In addition, the surviving samples favor the wrong-sign Yukawa coupling region, namely that the 125.5 GeV Higgs couplings to leptons have opposite sign to the couplings to gauge bosons and quarks.

hep-ph↗

arXiv2015

Understanding Big Data Analytic Workloads on Modern Processors

Big data analytics applications play a significant role in data centers, and hence it has become increasingly important to understand their behaviors in order to further improve the performance of data center computer systems, in which characterizing representative workloads is a key practical problem. In this paper, after investigating three most impor- tant application domains in terms of page views and daily visitors, we chose 11 repre- sentative data analytics workloads and characterized their micro-architectural behaviors by using hardware performance counters, so as to understand the impacts and implications of data analytics workloads on the systems equipped with modern superscalar out-of-order processors. Our study reveals that big data analytics applications themselves share many inherent characteristics, which place them in a different class from traditional workloads and scale-out services. To further understand the characteristics of big data analytics work- loads we performed a correlation analysis of CPI (cycles per instruction) with other micro- architecture level characteristics and an investigation of the big data software stack impacts on application behaviors. Our correlation analysis showed that even though big data ana- lytics workloads own notable pipeline front end stalls, the main factors affecting the CPI performance are long latency data accesses rather than the front end stalls. Our software stack investigation found that the typical big data software stack significantly contributes to the front end stalls and incurs bigger working set. Finally we gave several recommen- dations for architects, programmers and big data system designers with the knowledge acquired from this paper.

cs.DC↗

arXiv2015

Fidelity susceptibility made simple: A unified quantum Monte Carlo approach

The fidelity susceptibility is a general purpose probe of phase transitions. With its origin in quantum information and in the differential geometry perspective of quantum states, the fidelity susceptibility can indicate the presence of a phase transition without prior knowledge of the local order parameter, as well as reveal the universal properties of a critical point. The wide applicability of the fidelity susceptibility to quantum many-body systems is, however, hindered by the limited computational tools to evaluate it. We present a generic, efficient, and elegant approach to compute the fidelity susceptibility of correlated fermions, bosons, and quantum spin systems in a broad range of quantum Monte Carlo methods. It can be applied both to the ground-state and non-zero temperature cases. The Monte Carlo estimator has a simple yet universal form, which can be efficiently evaluated in simulations. We demonstrate the power of this approach with applications to the Bose-Hubbard model, the spin-$1/2$ XXZ model, and use it to examine the hypothetical intermediate spin-liquid phase in the Hubbard model on the honeycomb lattice.

cond-mat.stat-mech↗

arXiv2015

Entanglement as a resource in adiabatic quantum optimization

We explore the role of entanglement in adiabatic quantum optimization by performing approximate simulations of the real-time evolution of a quantum system while limiting the amount of entanglement. To classically simulate the time evolution of the system with a limited amount of entanglement, we represent the quantum state using matrix-product states and projected entangled-pair states. We show that the probability of finding the ground state of an Ising spin glass on either a planar or non-planar two-dimensional graph increases rapidly as the amount of entanglement in the state is increased. Furthermore, we propose evolution in complex time as a way to improve simulated adiabatic evolution and mimic the effects of thermal cooling of the quantum annealer.

cond-mat.dis-nn↗

arXiv2015

The invariant tori of knot type and the interlinked invariant tori in the Nosé-Hoover system

We revisit the famous Nosé-Hoover system in this paper and show the existence of some averagely conservative regions which are filled with an infinite sequence of nested tori. Depending on initial conditions, some invariant tori are of trefoil knot type, while the others are of trivial knot type. Moreover, we present a variety of interlinked invariant tori whose initial conditions are chosen from different averagely conservative regions and give all the interlinking numbers of those interlinked tori, showing that this quadratic system possesses so rich dynamic properties.

nlin.CD↗

arXiv2015

A simplified 2HDM with a scalar dark matter and the galactic center gamma-ray excess

Due to the strong constrain from the LUX experiment, the scalar portal dark matter can not generally explain a gamma-ray excess in the galactic center by the annihilation of dark matter into $b\bar{b}$. With the motivation of eliminating the tension, we add a scalar dark matter to the aligned two-Higgs-doublet model, and focus on a simplified scenario, which has two main characteristics: (i) The heavy CP-even Higgs is the discovered 125 GeV Higgs boson, which has the same couplings to the gauge bosons and fermions as the SM Higgs. (ii) Only the light CP-even Higgs mediates the dark matter interactions with SM particles, which has no couplings to $WW$ and $ZZ$, but the independent couplings to the up-type quarks, down-type quarks and charged leptons. We find that the tension between $<σv>_{SS\to b\bar{b}}$ and the constraint from LUX induced by the scalar portal dark matter can go away for the isospin-violating dark matter-nucleon coupling with $-1.0< f^n/f^p<0.7$, and the constraints from the Higgs search experiments and the relic density of Planck are also satisfied.

hep-ph↗

arXiv2015

Efficient Continuous-time Quantum Monte Carlo Method for the Ground State of Correlated Fermions

We present the ground state extension of the efficient quantum Monte Carlo algorithm for lattice fermions of arXiv:1411.0683. Based on continuous-time expansion of imaginary-time projection operator, the algorithm is free of systematic error and scales \emph{linearly} with projection time and interaction strength. Compared to the conventional quantum Monte Carlo methods for lattice fermions, this approach has greater flexibility and is easier to combine with powerful machinery such as histogram reweighting and extended ensemble simulation techniques. We discuss the implementation of the continuous-time projection in detail using the spinless $t-V$ model as an example and compare the numerical results with exact diagonalization, density-matrix-renormalization-group and infinite projected entangled-pair states calculations. Finally we use the method to study the fermionic quantum critical point of spinless fermions on a honeycomb lattice and confirm previous results concerning its critical exponents.

cond-mat.str-el↗

arXiv2014

Multi-terminal electrical transport measurements of molybdenum disulphide using van der Waals heterostructure device platform

Atomically thin two-dimensional (2D) semiconductors such as molybdenum disulphide (MoS2) hold great promise in electrical, optical, and mechanical devices and display novel physical phenomena such as coupled spin-valley physics and the valley Hall effect. However, the electron mobility of mono- and few-layer MoS2 has so far been substantially below theoretically predicted limits, particularly at low temperature (T), which has hampered efforts to observe its intrinsic quantum transport behaviors. Potential sources of disorder and scattering include both defects such as sulfur vacancies in the MoS2 itself, and extrinsic sources such as charged impurities and remote optical phonons from oxide dielectrics. To reduce extrinsic scattering and approach the intrinsic limit, we developed a van der Waals (vdW) heterostructure device platform where MoS2 layers are fully encapsulated within hexagonal boron nitride (hBN), and electrically contacted in a multi-terminal geometry using gate-tunable graphene electrodes. Multi-terminal magneto-transport measurements show dramatic improvements in performance, including a record-high Hall mobility reaching 34,000 cm2/Vs for 6-layer MoS2 at low T. Comparison to theory shows a decrease of 1-2 orders of magnitude in the density of charged impurities, indicating that performance at low T in previous studies was limited by extrinsic factors rather than defects in the MoS2. We also observed Shubnikov-de Haas (SdH) oscillations for the first time in high-mobility monolayer and few-layer MoS2. This novel device platform therefore opens up a new way toward measurements of intrinsic properties and the study of quantum transport phenomena in 2D semiconducting materials.

cond-mat.mtrl-sci↗

arXiv2014

High-Speed Electro-Optic Modulator Integrated with Graphene-Boron Nitride Heterostructure and Photonic Crystal Nanocavity

Nanoscale and power-efficient electro-optic (EO) modulators are essential components for optical interconnects that are beginning to replace electrical wiring for intra- and inter-chip communications. Silicon-based EO modulators show sufficient figures of merits regarding device footprint, speed, power consumption and modulation depth. However, the weak electro-optic effect of silicon still sets a technical bottleneck for these devices, motivating the development of modulators based on new materials. Graphene, a two-dimensional carbon allotrope, has emerged as an alternative active material for optoelectronic applications owing to its exceptional optical and electronic properties. Here, we demonstrate a high-speed graphene electro-optic modulator based on a graphene-boron nitride (BN) heterostructure integrated with a silicon photonic crystal nanocavity. Strongly enhanced light-matter interaction of graphene in a submicron cavity enables efficient electrical tuning of the cavity reflection. We observe a modulation depth of 3.2 dB and a cut-off frequency of 1.2 GHz.

physics.optics↗

arXiv2014

Spin alignments of spiral galaxies within the large-scale structure from SDSS DR7

Using a sample of spiral galaxies selected from the Sloan Digital Sky Survey Data Release 7 (SDSS DR7) and Galaxy Zoo 2 (GZ2), we investigate the alignment of spin axes of spiral galaxies with their surrounding large scale structure, which is characterized by the large-scale tidal field reconstructed from the data using galaxy groups above a certain mass threshold. We find that the spin axes of only have weak tendency to be aligned with (or perpendicular to) the intermediate (or minor) axis of the local tidal tensor. The signal is the strongest in a \cluster environment where all the three eigenvalues of the local tidal tensor are positive. Compared to the alignments between halo spins and local tidal field obtained in N-body simulations, the above observational results are in best agreement with those for the spins of inner regions of halos, suggesting that the disk material traces the angular momentum of dark matter halos in the inner regions.

astro-ph.CO↗