arXiv subjects

Dong Zhang

Publications and source records attributed to Dong Zhang.

234 records · Page 13Linked to original sources

arXiv2016

The $1$-Laplacian Cheeger Cut: Theory and Algorithms

This paper presents a detailed review of both theory and algorithms for the Cheeger cut based on the graph $1$-Laplacian. In virtue of the cell structure of the feasible set, we propose a cell descend (CD) framework for achieving the Cheeger cut. While plugging the relaxation to guarantee the decrease of the objective value in the feasible set, from which both the inverse power (IP) method and the steepest descent (SD) method can also be recovered, we are able to get two specified CD methods. Comparisons of all these methods are conducted on several typical graphs.

math.SP↗

arXiv2016

Nodal Domains of Eigenvectors for $1$-Laplacian on Graphs

The eigenvectors for graph $1$-Laplacian possess some sort of localization property: On one hand, any nodal domain of an eigenvector is again an eigenvector with the same eigenvalue; on the other hand, one can pack up an eigenvector for a new graph by several fundamental eigencomponents and modules with the same eigenvalue via few special techniques. The Courant nodal domain theorem for graphs is extended to graph $1$-Laplacian for strong nodal domains, but for weak nodal domains it is false. The notion of algebraic multiplicity is introduced in order to provide a more precise estimate of the number of independent eigenvectors. A positive answer is given to a question raised in [{\sl K.~C. Chang, Spectrum of the $1$-Laplacian and Cheeger constant on graphs, J. Graph Theor., DOI: 10.1002/jgt.21871}], to confirm that the critical values obtained by the minimax principle may not cover all eigenvalues of graph $1$-Laplacian.

math.SP↗

arXiv2015

Robust Scene Text Recognition Using Sparse Coding based Features

In this paper, we propose an effective scene text recognition method using sparse coding based features, called Histograms of Sparse Codes (HSC) features. For character detection, we use the HSC features instead of using the Histograms of Oriented Gradients (HOG) features. The HSC features are extracted by computing sparse codes with dictionaries that are learned from data using K-SVD, and aggregating per-pixel sparse codes to form local histograms. For word recognition, we integrate multiple cues including character detection scores and geometric contexts in an objective function. The final recognition results are obtained by searching for the words which correspond to the maximum value of the objective function. The parameters in the objective function are learned using the Minimum Classification Error (MCE) training method. Experiments on several challenging datasets demonstrate that the proposed HSC-based scene text recognition method outperforms HOG-based methods significantly and outperforms most state-of-the-art methods.

cs.CV↗

arXiv2015

An Origin for Multi-Phase Gas in Galactic Winds and Halos

The physical origin of high velocity cool gas seen in galactic winds remains unknown. Following Wang (1995), we argue that radiative cooling in initially hot thermally-driven outflows can produce fast neutral atomic and photoionized cool gas. The inevitability of adiabatic cooling from the flow's initial 10^7-10^8K temperature and the shape of the cooling function for T<10^7K imply that outflows with hot gas mass-loss rate relative to star formation rate of beta=Mdot_hot/Mdot_star > 0.5 cool radiatively on scales ranging from the size of the energy injection region to tens of kpc. We highlight the beta and star formation rate surface density dependence of the column density, emission measure, radiative efficiency, and velocity. At r_cool, the gas produces X-ray and then UV/optical line emission with a total power bounded by 10^{-2} L_star if the flow is powered by steady-state star formation with luminosity L_star. The wind is thermally unstable at r_cool, potentially leading to a multi-phase medium. Cooled winds decelerate significantly in the extended gravitational potential of galaxies. The cool gas precipitated from hot outflows may explain its prevalence in galactic halos. We forward a picture of winds whereby cool clouds are initially accelerated by the ram pressure of the hot flow, but are rapidly shredded by hydrodynamical instabilities, thereby increasing beta, seeding radiative and thermal instability, and cool gas rebirth. If the cooled wind shocks as it sweeps up the circumgalactic medium, its cooling time is short, thus depositing cool gas far out into the halo. Finally, conduction can dominate energy transport in low-beta hot winds, leading to flatter temperature profiles than otherwise expected, potentially consistent with X-ray observations of some starbursts.

astro-ph.GA↗

arXiv2014

Half metallic and insulating phases in BN/Graphene lateral heterostructures

We investigate theoretically the electronic structure of graphene and boron nitride (BN) lateral heterostructures, which were fabricated in recent experiments. The first-principles density functional calculation demonstrates that a huge intrinsic transverse electric field can be induced in the graphene nanoribbon region, and depends sensitively on the edge configuration of the lateral heterostructure. The polarized electric field originates from the charge mismatch at the BN-graphene interfaces. This huge electric field can open a significant bang gap in graphene nanoribbon, and lead to fully spinpolarized edge states and induce half-metallic phase in the lateral BN/Graphene/BN heterostructure with proper edge configurations.

cond-mat.mes-hall↗

arXiv2014

Hot Galactic Winds Constrained by the X-Ray Luminosities of Galaxies

Galactic superwinds may be driven by very hot outflows generated by overlapping supernovae within the host galaxy. We use the Chevalier & Clegg (CC85) wind model and the observed correlation between X-ray luminosities of galaxies and their SFRs to constrain the mass loss rates (\dot{M}_hot) across a wide range of star formation rates (SFRs), from dwarf starbursts to ultra-luminous infrared galaxies. We show that for fixed thermalization efficiency and mass loading rate, the X-ray luminosity of the hot wind scales as L_X ~ SFR^2, significantly steeper than is observed for star-forming galaxies: L_X ~ SFR. Using this difference we constrain the mass-loading and thermalization efficiency of hot galactic winds. For reasonable values of the thermalization efficiency (<~ 1) and for SFR >~ 10 M_sun/yr we find that \dot{M}_hot/SFR <~ 1, significantly lower than required by integrated constraints on the efficiency of stellar feedback in galaxies, and potentially too low to explain observations of winds from rapidly star-forming galaxies. In addition, we highlight the fact that heavily mass-loaded winds cannot be described by the adiabatic CC85 model because they become strongly radiative.

astro-ph.GA↗

arXiv2013

Face Verification Using Boosted Cross-Image Features

This paper proposes a new approach for face verification, where a pair of images needs to be classified as belonging to the same person or not. This problem is relatively new and not well-explored in the literature. Current methods mostly adopt techniques borrowed from face recognition, and process each of the images in the pair independently, which is counter intuitive. In contrast, we propose to extract cross-image features, i.e. features across the pair of images, which, as we demonstrate, is more discriminative to the similarity and the dissimilarity of faces. Our features are derived from the popular Haar-like features, however, extended to handle the face verification problem instead of face detection. We collect a large bank of cross-image features using filters of different sizes, locations, and orientations. Consequently, we use AdaBoost to select and weight the most discriminative features. We carried out extensive experiments on the proposed ideas using three standard face verification datasets, and obtained promising results outperforming state-of-the-art.

cs.CV↗

arXiv2013

Interface-induced Topological Insulator Transition in GaAs/Ge/GaAs Quantum Wells

We demonstrate theoretically that interface engineering can drive Germanium, one of the most commonly-used semiconductors, into topological insulating phase. Utilizing giant electric fields generated by charge accumulation at GaAs/Ge/GaAs opposite semiconductor interfaces and band folding, the new design can reduce the sizable gap in Ge and induce large spin-orbit interaction, which lead to a topological insulator transition. Our work provides a new method on realizing TI in commonly-used semiconductors and suggests a promising approach to integrate it in well developed semiconductor electronic devices.

cond-mat.mes-hall↗

arXiv2012

Radiation Pressure Driven Galactic Winds from Self-Gravitating Discs

(Abridged) We study large-scale winds driven from uniformly bright self-gravitating discs radiating near the Eddington limit. We show that the ratio of the radiation pressure force to the gravitational force increases with height above the disc surface to a maximum of twice the value of the ratio at the disc surface. Thus, uniformly bright self-gravitating discs radiating at the Eddington limit are fundamentally unstable to driving large-scale winds. These results contrast with the spherically symmetric case, where super-Eddington luminosities are required for wind formation. We apply this theory to galactic winds from rapidly star-forming galaxies that approach the Eddington limit for dust. For hydrodynamically coupled gas and dust, we find that the asymptotic velocity of the wind is v_\infty ~ 1.5 v_rot and that v_\infty SFR^{0.36}, where v_rot is the disc rotation velocity and SFR is the star formation rate, both of which are in agreement with observations. However, these results of the model neglect the gravitational potential of the surrounding dark matter halo and an old passive stellar bulge or extended disc, which act to decrease v_\infty. A more realistic treatment shows that the flow can either be unbound, or bound, forming a "fountain flow" with a typical turning timescale of t_turn ~ 0.1-1 Gyr. We provide quantitative criteria and scaling relations for assessing whether or not a rapidly star-forming galaxy of given properties can drive unbound flows via the mechanism described in this paper. Importantly, we note that because t_turn is longer than the star formation timescale in the rapidly star-forming galaxies and ULIRGs for which our theory is most applicable, if rapidly star-forming galaxies are selected as such, they may be observed to have strong outflows, even though their winds are eventually bound on large scales.

astro-ph.CO↗

arXiv2012

The Very Massive and Hot LMC Star VFTS 682: Progenitor of a Future Dark Gamma-Ray Burst?

VFTS 682, a very massive and very hot Wolf-Rayet (WR) star recently discovered in the Large Magellanic Cloud near the famous star cluster R136, might be providing us with a glimpse of a missing link in our understanding of Long Gamma-Ray Bursts (LGRBs), including dark GRBs. It is likely its properties result from chemically homogeneous evolution (CHE), believed to be a key process for a massive star to become a GRB. It is also heavily obscured by dust extinction, which could make it a dark GRB upon explosion. Using Spitzer data we investigate the properties of interstellar dust in the vicinity of R136, and argue that its high obscuration is not unusual for its environment and that it could indeed be a slow runaway ("walkaway") from R136. Unfortunately, based on its current mass loss rate, VFTS 682 is unlikely to become a GRB, because it will lose too much angular momentum at its death. If it were to become a GRB, it probably would also not be dark, either escaping or destroying its surrounding dusty region. Nevertheless, it is a very interesting star, deserving further studies, and being one of only three presently identified WR stars (two others in the Small Magellanic Cloud) that seems to be undergoing CHE.

astro-ph.GA↗

arXiv2011

Impact of Primordial Ultracompact Minihaloes on the Intergalactic Medium and First Structure Formation

(Abridged) Ultracompact Minihaloes (UCMHs), which formed by dark matter accretion onto primordial black holes (PBHs) or initial dark matter overdensity produced by the primordial density perturbation, provide a new type of compact dark matter structure to ionize and heat the IGM after matter-radiation equality z_eq, which is much earlier than the formation of the first cosmological dark halo structure and later first stars. We show that dark matter annihilation density contributed by UCMHs can totally dominated over the homogenous dark matter annihilation background even for a tiny UCMH abundance, and provide a new gamma-ray background in the early Universe. The IGM ionization fraction x_ion and gas temperature T_m can be increased from the recombination residual and adiabatically cooling in the absence of energy injection to the highest value of x_ ion 0.1 and T_m ~ 5000 K at z>10 for the upper bound UCMH abundance constrained by the CMB optical depth. A small fraction of UCMHs are seeded by PBHs. The X-ray emission from gas accretion onto PBHs may totally dominated over dark matter annihilation and become the main cosmic ionization source, but the constraints of gas accretion rate and X-ray absorption by the baryon accumulation within the UCMHs and accretion feedback show that X-ray emission can only be a promising source much later than UCMH annihilation at z<z_m<1000, where z_m depends on the masses of PBHs, their host UCMHs, and the dark matter particles. Also, UCMH radiation including both annihilation and X-ray emission can significantly suppress the low mass first baryonic structure formation. The effects of UCMHs radiation on the baryonic structure evolution are quite small for the gas temperature after virialization, but more significant to enhance the gas chemical quantities such as the ionization fraction and molecular hydrogen abundance in the baryonic objects.

astro-ph.CO↗

arXiv2010

Hyperaccreting Disks around Neutrons Stars and Magnetars for GRBs: Neutrino Annihilation and Strong Magnetic Fields

Hyperaccreting disks around neutron stars or magnetars cooled via neutrino emission can be the potential central engine of GRBs. The neutron-star disk can cool more efficiently, produce much higher neutrino luminosity and neutrino annihilation luminosity than its black hole counterpart with the same accretion rate. The neutron star surface boundary layer could increase the annihilation luminosity as well. An ultra relativistic jet via neutrino annihilation can be produced along the stellar poles. Moreover, we investigate the effects of strong fields on the disks around magnetars. In general, stronger fields give higher disk densities, pressures, temperatures and neutrino luminosity; the neutrino annihilation mechanism and the magnetically-driven pulsar wind which extracts the stellar rotational energy can work together to generate and feed an even stronger ultra-relativistic jet along the stellar magnetic poles.

astro-ph.HE↗

arXiv2010

Hyperaccreting Disks around Magnetars for Gamma-Ray Bursts: Effects of Strong Magnetic Fields

(Abridged) The hyperaccreting neutron star or magnetar disks cooled via neutrino emission can be a candidate of gamma-ray burst (GRB) central engines. The strong field $\geq10^{15}-10^{16}$ G of the magnetar can play a significant role in affecting the disk properties and even lead to the funnel accretion process. We investigate the effects of strong fields on the disks around magnetars, and discuss implications of such accreting magnetar systems for GRB and GRB-like events. We discuss quantum effects of the strong fields on the disk, and use the MHD conservation equations to describe the behavior of the disk flow coupled with a large scale field, which is generated by the star-disk interaction. In general, stronger fields give higher disk densities, pressures, temperatures and neutrino luminosity, and change the electron fraction and degeneracy state significantly. A magnetized disk is always viscously stable outside the Alfvén radius, but will be thermally unstable near the Alfvén radius where the magnetic field plays a more important role in transferring the angular momentum and heating the disk than the viscous stress. The funnel accretion process will be only important for an extremely strong field, which creates a magnetosphere inside the Alfvén radius and truncates the plane disk. Because of higher temperature and more concentrated neutrino emission of the magnetar surface ring-like belt region covered by funnel accretion, the neutrino annihilation rate from the accreting magnetars can be much higher than that from accreting neutron stars without fields. Furthermore, the neutrino annihilation mechanism and the magnetically-driven pulsar wind from the magnetar surface can work together to generate and feed an ultra-relativistic jet along the stellar magnetic poles.

astro-ph.HE↗

arXiv2009

Hyperaccreting Neutron-Star Disks, Magnetized Disks and Gamma-Ray Bursts

This thesis focuses on the study of the hyperaccreting neutron-star disks and magnetized accretion flows. It is usually proposed that hyperaccreting disks surrounding stellar-mass black holes with a huge accretion rate are central engines of gamma-ray bursts (GRBs). However, hyperaccretion disks around neutron stars may exist in some GRB formation scenarios. We study the structure and neutrino emission of a hyperaccretion disk around a neutron star. We consider a steady-state hyperaccretion disk, and as a reasonable approximation, divide it into two regions, the inner and outer disks. The outer disk is similar to that of a black hole. The inner disk has a self-similar structure, such as the entropy-conservation or the advection structure, depending on the energy transfer and emission in the disk. We see that the neutron star disk can cool more efficiently, produce much higher neutrino luminosity and neutrino annihilation luminosity than a black hole disk. The neutrino emission from the neutron star surface boundary layer could increase the neutrino annihilation luminosity. Moreover, we study the effects of a global magnetic field on viscously-rotating and vertically-integrated accretion disks around compact objects using a self-similar treatment, and use two methods to study magnetized flows with convection. We also give a review on GRB progenitors, central engines and an outlook in this thesis.

astro-ph.HE↗

arXiv2009

Hyperaccreting Neutron-Star Disks and Neutrino Annihilation

Newborn neutron stars surrounded by hyperaccreting and neutrino-cooled disks may exist in some gamma-ray bursts (GRBs) and/or supernovae (SNe). In this paper we further study the structure of such a neutron-star disk based on the two-region (i.e., inner & outer) disk scenario following our previous work, and calculate the neutrino annihilation luminosity from the disk in various cases. We investigate the effects of the viscosity parameter, energy parameter (measuring the neutrino cooling efficiency of the inner disk) and outflow strength on the structure of the entire disk as well as the effect of emission from the neutron star surface boundary emission on the total neutrino annihilation rate. The inner disk satisfies the entropy-conservation or the advection-dominated self-similar structure depending on the energy parameter. An outflow from the disk decreases the density and pressure but increases the thickness of the disk. Moreover, compared with the black-hole disk, the neutrino annihilation luminosity above the neutron-star disk is higher, and the neutrino emission from the boundary layer could increase the neutrino annihilation luminosity by about one order of magnitude higher than the disk without boundary emission. The neutron-star disk with the advection-dominated inner disk could produce the highest neutrino luminosity while the disk with an outflow has the lowest. Although a heavily mass-loaded outflow from the neutron star surface at early times of neutron star formation prevents the outflow material from being accelerated to a high bulk Lorentz factor, an energetic ultrarelativistic jet via neutrino annihilation can be produced above the stellar polar region at late times if the disk accretion rate and the neutrino emission luminosity from the surface boundary layer are sufficiently high.

astro-ph.GA↗

arXiv2008

Self-similar structure of magnetized ADAFs and CDAFs

(Abridged) We study the effects of a global magnetic field on viscously-rotating and vertically-integrated accretion disks around compact objects using a self-similar treatment. We extend Akizuki & Fukue's work (2006) by discussing a general magnetic field with three components ($r, ϕ, z$) in advection-dominated accretion flows (ADAFs). We also investigate the effects of a global magnetic field on flows with convection. For these purposes, we first adopt a simple form of the kinematic viscosity $ν=αc_{s}^{2}/Ω_{K}$ to study magnetized ADAFs. Then we consider a more realistic model of the kinematic viscosity $ν=αc_{s}H$, which makes the infall velocity increase but the sound speed and toroidal velocity decrease. We next use two methods to study magnetized flows with convection, i.e., we take the convective coefficient $α_{c}$ as a free parameter to discuss the effects of convection for simplicity. We establish the $α_{c}-α$ relation for magnetized flows using the mixing-length theory and compare this relation with the non-magnetized case. If $α_{c}$ is set as a free parameter, then $|v_{r}|$ and $c_{s}$ increase for a large toroidal magnetic field, while $|v_{r}|$ decreases but $|v_ϕ|$ increases (or decreases) for a strong and dominated radial (or vertical) magnetic field with increasing $α_{c}$. In addition, the magnetic field makes the $α_{c}-α$ relation be distinct from that of non-magnetized flows, and allows the $ρ\propto r^{-1}$ or $ρ\propto r^{-2}$ structure for magnetized non-accreting convection-dominated accretion flows with $α+gα_{c}< 0$ (where $g$ is the parameter to determine the condition of convective angular momentum transport).

astro-ph↗

arXiv2008

Hyperaccretion Disks around Neutron Stars

(Abridged) We here study the structure of a hyperaccretion disk around a neutron star. We consider a steady-state hyperaccretion disk around a neutron star, and as a reasonable approximation, divide the disk into two regions, which are called inner and outer disks. The outer disk is similar to that of a black hole and the inner disk has a self-similar structure. In order to study physical properties of the entire disk clearly, we first adopt a simple model, in which some microphysical processes in the disk are simplified, following Popham et al. and Narayan et al. Based on these simplifications, we analytically and numerically investigate the size of the inner disk, the efficiency of neutrino cooling, and the radial distributions of the disk density, temperature and pressure. We see that, compared with the black-hole disk, the neutron star disk can cool more efficiently and produce a much higher neutrino luminosity. Finally, we consider an elaborate model with more physical considerations about the thermodynamics and microphysics in the neutron star disk (as recently developed in studying the neutrino-cooled disk of a black hole), and compare this elaborate model with our simple model. We find that most of the results from these two models are basically consistent with each other.

astro-ph↗

arXiv2001

Branching ratios and CP-violating asymmetries of $B_s \to h_1 h_2 $ decays in the general two-Higgs doublet models

Based on the low-energy effective Hamiltonian with the generalized factorization, we calculate the new physics contributions to branching ratios and CP-violating asymmetries of the charmless hadronic decays $B_s \to h_1 h_2$ in the standard model and the general two-Higgs doublet models (models I, II, and III). Within the considered paramter space, we find the following. (a) In models I and II, the new physics corrections are always small in size and will be masked by other larger known theoretical uncertainties. (b) In model III, the new physics corrections to the branching ratios of those QCD penguin-dominated decays $\ov B_s \to K^0\etapp, K^+ K^{-*}, etc.$, are large in size and insensitive to the variations of $\mhp$ and $\nceff$. For tree- or electroweak penguin-dominated decay modes, however, the new physics corrections are very small in size. (c) For $\ov B_s \to K^+ K^{-*}$ and other seven decay modes, the branching ratios are at the level of $(1-3)\times 10^{-5}$ and will be measurable at the future hadron colliders with large $b$ production. (d) Among the studied thirty nine $B_s$ meson decay modes, seven of them can have a CP-violating asymmetry ${\cal A}_{CP}$ larger than 20% in magnitude. The new physics corrections are small or moderate in magnitude. (e) Because of its large and $\nceff$ stable branching ratio and CP violating asymmetry, the decay $\ov B_s \to K^+ K^{-*}$ seems to be the "best" channel to find CP violation of $B_s$ system through studies of two-body charmless decays of $B_s$ meson.

hep-ph↗