arXiv subjects

Xiao Zhang

Publications and source records attributed to Xiao Zhang.

At least 487 records · Page 27Linked to original sources

arXiv2018

Ellipticity dependence transition induced by dynamical Bloch oscillations

The dependence of high-harmonic generation (HHG) on laser ellipticity is investigated using a modified ZnO model. In the driving of relatively weak field, we reproduce qualitatively the ellipticity dependence as observed in the HHG experiment of wurtzite ZnO. When increasing the field strength, the HHG shows an anomalous ellipticity dependence, similar to that observed experimentally in the single-crystal MgO. With the help of a semiclassical analysis, it is found that the key mechanism inducing the change of ellipticity dependence is the interplay between the dynamical Bloch oscillation and the anisotropic band structure. The dynamical Bloch oscillation contributes additional quantum paths, which are less sensitive to ellipticity. The anisotropic band-structure make the driving pulse with finite ellipticity be able to drive the pairs to the band positions with larger gap, which extends the harmonic cutoff. The combination of these two effects leads to the anomalous ellipticity dependence. The result reveals the importance of dynamical Bloch oscillations for the ellipticity dependence of HHG from bulk ZnO.

physics.atom-ph↗

arXiv2018

Ferromagnetic van der Waals crystal VI3

We report structural, physical properties and electronic structure of van der Waals (vdW) crystal VI3. Detailed analysis reveals that VI3 exhibits a structural transition from monoclinic C2/m to rhombohedral R-3 at Ts ~ 79 K, similar to CrX3 (X = Cl, Br, I). Below Ts, a long-range ferromagnetic (FM) transition emerges at Tc ~ 50 K. The local moment of V in VI3 is close to the high-spin state V3+ ion (S = 1). Theoretical calculation suggests that VI3 may be a Mott insulator with the band gap of about 0.84 eV. In addition, VI3 has a relative small interlayer binding energy and can be exfoliated easily down to few layers experimentally. Therefore, VI3 is a candidate of two-dimensional FM semiconductor. It also provides a novel platform to explore 2D magnetism and vdW heterostructures in S = 1 system.

cond-mat.mtrl-sci↗

arXiv2018

Eigenvalue Estimate of the Dirac operator and Rigidity of Poincare-Einstein Metrics

We re-visit the eigenvalue estimate of the Dirac operator on spin manifolds with boundary in terms of the first eigenvalues of conformal Laplace operator as well as the conformal mean curvature operator. These problems were studied earlier by Hijazi-Montiel-Zhang and Raulot and we re-prove them under weaker assumption that a boundary chirality operator exists. Moreover, on these spin manifolds with boundary, we show that any $C^{3,α}$ conformal compactification of some Poincare-Einstein metric must be the standard hemisphere when the first nonzero eigenvalue of the Dirac operator achieves its lowest value, and any $C^{3,α}$ conformal compactification of some Poincare-Einstein metric must be the flat ball in Euclidean space when the first positive eigenvalue of the boundary Dirac operator achieves certain value relating to the second Yamabe invariant.In two cases the Poincare-Einstein metrics are standard hyperbolic metric.

math.DG↗

arXiv2018

Energy-momentum in general relativity

We briefly review of the definitions of the total energy, the total linear momentum and the angular momentum of gravitational field when the cosmological constant is zero. In particular, we show pseudo-tensor's definition of the energy and the momentum given by Prof. Duan in 1963 agree with the ADM total energy-momentum and the Bondi energy-momentum at spatial and null infinity respectively. We also review the relevant energy-momentum inequalities. Finally, we provide a short review of the positive energy theorem and the peeling property of the Newmann-Penrose quantities when the cosmological constant is positive.

gr-qc↗

arXiv2018

Dynamics of deformable straight and curved prolate capsules in simple shear flow

This work investigates the motion of neutrally-buoyant, slightly deformable straight and curved prolate capsules in unbounded simple shear flow at zero Reynolds number using direct simulations. The curved capsules serve as a model for the typical crescent-shaped sickle red blood cells in sickle cell disease (SCD). The effects of deformability and curvature on the dynamics are revealed. We show that with low deformability, straight prolate spheroidal capsules tumble in the shear plane as their unique asymptotically stable orbit, which contrasts with that for rigid spheroids where infinitely many neutrally stable Jeffery orbits exist. The dynamics of curved prolate capsules are more complicated due to a combined effect of deformability and curvature. At short times, depending on the initial orientation, slightly deformable curved prolate capsules exhibit either a Jeffery-like motion such as tumbling or kayaking, or a non-Jeffery-like behavior in which the end-to-end vector of the capsule crosses the shear-gradient plane back and forth. At long times, however, a Jeffery-like quasi-periodic orbit is taken regardless of the initial orientation. We further show that the average of the long-time trajectory can be well approximated using the analytical solution for Jeffery orbits with an effective orbit constant $C_{\textnormal{eff}}$ and aspect ratio $\ell_{\textnormal{eff}}$. As the capsule becomes more deformable or curved, $C_{\textnormal{eff}}$ decreases, indicating a shift of the orbit towards log-rolling motion, while $\ell_{\textnormal{eff}}$ increases weakly as the degree of curvature increases but shows negligible dependency on deformability. As cell deformability, cell shape, and cell-cell interactions are all pathologically altered in blood disorders such as SCD, these results will have clear implications on improving our understanding of the pathophysiology of hematologic disease.

physics.flu-dyn↗

arXiv2018

CMUT on Glass Substrate for Next-Generation Medical Imaging and Beyond

CMUT technology was invented and has been widely researched on silicon substrates. In recent years, fabrication of CMUT on glass substrates has raised significant interests because a number of advantages it can offer. First, parasitic capacitance could be easily reduced since glass is an insulating material. Second, insulation steps required in the silicon-based CMUT fabrication process could be avoided, leading to a reduced fabrication complexity. Third, anodic bonding can be used when the substrate is borosilicate glass, which is a low-temperature bonding technique and has a high tolerance to bonding surface area and roughness in addition to the advantages of silicon wafer bonding. Moreover, glass transparency can enable novel applications beyond pulse-echo ultrasound medical imaging and enlarge the potential markets of CMUTs.

physics.app-ph↗

arXiv2018

NetAdapt: Platform-Aware Neural Network Adaptation for Mobile Applications

This work proposes an algorithm, called NetAdapt, that automatically adapts a pre-trained deep neural network to a mobile platform given a resource budget. While many existing algorithms simplify networks based on the number of MACs or weights, optimizing those indirect metrics may not necessarily reduce the direct metrics, such as latency and energy consumption. To solve this problem, NetAdapt incorporates direct metrics into its adaptation algorithm. These direct metrics are evaluated using empirical measurements, so that detailed knowledge of the platform and toolchain is not required. NetAdapt automatically and progressively simplifies a pre-trained network until the resource budget is met while maximizing the accuracy. Experiment results show that NetAdapt achieves better accuracy versus latency trade-offs on both mobile CPU and mobile GPU, compared with the state-of-the-art automated network simplification algorithms. For image classification on the ImageNet dataset, NetAdapt achieves up to a 1.7$\times$ speedup in measured inference latency with equal or higher accuracy on MobileNets (V1&V2).

cs.CV↗

arXiv2018

Peeling Property of Bondi-Sachs metrics for nonzero Cosmological Constant

In this paper, we show that the peeling property still holds for Bondi-Sachs metrics with nonzero cosmological constant under the boundary condition given by Sommerfeld's radiation condition together with three nontrivial $Λ$-independent functions $B$, $a$, $b$. This should indicate the new boundary condition is natural. Moreover, we construct some nonstationary vacuum Bondi-Sachs metrics without Bondi news, which Newmann-Penrose quantities fall faster than usual. This provides a new feature of gravitational waves for nonzero cosmological constant.

gr-qc↗

arXiv2018

Entanglement-guided architectures of machine learning by quantum tensor network

It is a fundamental, but still elusive question whether the schemes based on quantum mechanics, in particular on quantum entanglement, can be used for classical information processing and machine learning. Even partial answer to this question would bring important insights to both fields of machine learning and quantum mechanics. In this work, we implement simple numerical experiments, related to pattern/images classification, in which we represent the classifiers by many-qubit quantum states written in the matrix product states (MPS). Classical machine learning algorithm is applied to these quantum states to learn the classical data. We explicitly show how quantum entanglement (i.e., single-site and bipartite entanglement) can emerge in such represented images. Entanglement characterizes here the importance of data, and such information are practically used to guide the architecture of MPS, and improve the efficiency. The number of needed qubits can be reduced to less than 1/10 of the original number, which is within the access of the state-of-the-art quantum computers. We expect such numerical experiments could open new paths in charactering classical machine learning algorithms, and at the same time shed lights on the generic quantum simulations/computations of machine learning tasks.

stat.ML↗

arXiv2018

Learning One-hidden-layer ReLU Networks via Gradient Descent

We study the problem of learning one-hidden-layer neural networks with Rectified Linear Unit (ReLU) activation function, where the inputs are sampled from standard Gaussian distribution and the outputs are generated from a noisy teacher network. We analyze the performance of gradient descent for training such kind of neural networks based on empirical risk minimization, and provide algorithm-dependent guarantees. In particular, we prove that tensor initialization followed by gradient descent can converge to the ground-truth parameters at a linear rate up to some statistical error. To the best of our knowledge, this is the first work characterizing the recovery guarantee for practical learning of one-hidden-layer ReLU networks with multiple neurons. Numerical experiments verify our theoretical findings.

stat.ML↗

arXiv2018

Bloch oscillations in arrays of helical waveguides

We study optical Bloch oscillations in the one- and two-dimensional arrays of helical waveguides with transverse refractive index gradient. Longitudinal rotation of waveguides may lead to notable variations of the width of the band of quasi-energies and even its complete collapse for certain radii of the helix. This drastically affects the amplitude and direction of Bloch oscillations. Thus, they can be completely arrested for certain helix radii or their direction can be reversed. If the array of helical waveguides is truncated and near-surface waveguide is excited, helix radius determines whether periodic Bloch oscillations persist or replaced by the irregular near-surface oscillations.

physics.optics↗

arXiv2018

NuSTAR Detection of a Hard X-ray Source in the Supernova Remnant - Molecular Cloud Interaction Site of IC 443

We report on a broadband study of a complex X-ray source (1SAX J0618.0+2227) associated with the interaction site of the supernova remnant (SNR) IC 443 and ambient molecular cloud (MC) using NuSTAR, XMM_Newton, and Chandra observations. Its X-ray spectrum is composed of both thermal and non-thermal components. The thermal component can be equally well represented by either a thin plasma model with kT=0.19 keV or a blackbody model with kT=0.11 keV. The non-thermal component can be fit with either a power-law with Gamma~1.7 or a cutoff power-law with Gamma~1.5 and a cutoff energy at E_cut~18 keV. Using the newly obtained NuSTAR dataset, we test three possible scenarios for isolated X-ray sources in the SNR-MC interaction site: 1) pulsar wind nebula (PWN); 2) SNR ejecta fragment; 3) shocked molecular clump. We conclude that this source is most likely composed of a SNR ejecta (or a PWN) and surrounding shocked molecular clumps. The nature of this hard X-ray source in the SNR-MC interaction site of IC 443 may shed light on unidentified X-ray sources with hard X-ray spectra in rich environments for star forming regions, such as the Galactic center.

astro-ph.HE↗

arXiv2018

Fast Deployment of UAV Networks for Optimal Wireless Coverage

Unmanned Aerial Vehicle (UAV) networks have emerged as a promising technique to rapidly provide wireless coverage to a geographical area, where a flying UAV can be fast deployed to serve as cell site. Existing work on UAV-enabled wireless networks overlook the fast UAV deployment for wireless coverage, and such deployment problems have only been studied recently in sensor networks. Unlike sensors, UAVs should be deployed to the air and they are generally different in flying speed, operating altitude and wireless coverage radius. By considering such UAV heterogeneity to cover the whole target area, this paper studies two fast UAV deployment problems: one is to minimize the maximum deployment delay among all UAVs (min-max) for fairness consideration, and the other is to minimize the total deployment delay (min-sum) for efficiency consideration. We prove both min-max and min-sum problems are NP-complete in general. When dispatching UAVs from the same location, we present an optimal algorithm of low computational complexity $O(n^2)$ for the min-max problem. When UAVs are dispatched from different locations, we propose to preserve their location order during deployment and successfully design a fully polynomial time approximation scheme (FPTAS) of computation complexity $O(n^2 \log \frac{1}ε)$ to arbitrarily approach the global optimum with relative error $ε$. The min-sum problem is more challenging. When UAVs are dispatched from the same initial location, we present an approximation algorithm of linear time. As for the general case, we further reformulate it as a dynamic program and propose a pseudo polynomial-time algorithm to solve it optimally.

cs.NI↗

arXiv2018

OMG - Emotion Challenge Solution

This short paper describes our solution to the 2018 IEEE World Congress on Computational Intelligence One-Minute Gradual-Emotional Behavior Challenge, whose goal was to estimate continuous arousal and valence values from short videos. We designed four base regression models using visual and audio features, and then used a spectral approach to fuse them to obtain improved performance.

cs.CV↗

arXiv2018

Electromagnetic response of quantum Hall systems in dimensions five and six and beyond

Quantum Hall (QH) states are arguably the most ubiquitous examples of nontrivial topological order, requiring no special symmetry and elegantly characterized by the first Chern number. Their higher dimension generalizations are particularly interesting from both mathematical and phenomenological perspectives, and have attracted recent attention due to a few high profile experimental realizations. In this work, we derive from first principles the electromagnetic response of QH systems in arbitrary number of dimensions, and elaborate on the crucial roles played by their modified phase space density of states under the simultaneous presence of magnetic field and Berry curvature. We provide new mathematical results relating this phase space modification to the non-commutativity of phase space, and show how they are manifested as a Hall conductivity quantized by a higher Chern number. When a Fermi surface is present, additional response currents unrelated to these Chern numbers also appear. This unconventional response can be directly investigated through a few minimal models with specially chosen fluxes. These models, together with more generic 6D QH systems, can be realized in realistic 3D experimental setups like cold atom systems through possibly entangled synthetic dimensions.

cond-mat.str-el↗

arXiv2018

Fast and Sample Efficient Inductive Matrix Completion via Multi-Phase Procrustes Flow

We revisit the inductive matrix completion problem that aims to recover a rank-$r$ matrix with ambient dimension $d$ given $n$ features as the side prior information. The goal is to make use of the known $n$ features to reduce sample and computational complexities. We present and analyze a new gradient-based non-convex optimization algorithm that converges to the true underlying matrix at a linear rate with sample complexity only linearly depending on $n$ and logarithmically depending on $d$. To the best of our knowledge, all previous algorithms either have a quadratic dependency on the number of features in sample complexity or a sub-linear computational convergence rate. In addition, we provide experiments on both synthetic and real world data to demonstrate the effectiveness of our proposed algorithm.

stat.ML↗

arXiv2018

Medical Exam Question Answering with Large-scale Reading Comprehension

Reading and understanding text is one important component in computer aided diagnosis in clinical medicine, also being a major research problem in the field of NLP. In this work, we introduce a question-answering task called MedQA to study answering questions in clinical medicine using knowledge in a large-scale document collection. The aim of MedQA is to answer real-world questions with large-scale reading comprehension. We propose our solution SeaReader--a modular end-to-end reading comprehension model based on LSTM networks and dual-path attention architecture. The novel dual-path attention models information flow from two perspectives and has the ability to simultaneously read individual documents and integrate information across multiple documents. In experiments our SeaReader achieved a large increase in accuracy on MedQA over competing models. Additionally, we develop a series of novel techniques to demonstrate the interpretation of the question answering process in SeaReader.

cs.CL↗

arXiv2018

A Unified Framework for Low-Rank plus Sparse Matrix Recovery

We propose a unified framework to solve general low-rank plus sparse matrix recovery problems based on matrix factorization, which covers a broad family of objective functions satisfying the restricted strong convexity and smoothness conditions. Based on projected gradient descent and the double thresholding operator, our proposed generic algorithm is guaranteed to converge to the unknown low-rank and sparse matrices at a locally linear rate, while matching the best-known robustness guarantee (i.e., tolerance for sparsity). At the core of our theory is a novel structural Lipschitz gradient condition for low-rank plus sparse matrices, which is essential for proving the linear convergence rate of our algorithm, and we believe is of independent interest to prove fast rates for general superposition-structured models. We illustrate the application of our framework through two concrete examples: robust matrix sensing and robust PCA. Experiments on both synthetic and real datasets corroborate our theory.

stat.ML↗