arXiv subjects

Chong Li

Publications and source records attributed to Chong Li.

At least 109 records · Page 6Linked to original sources

arXiv2021

Impact of Distributed Rate Limiting on Load Distribution in a Latency-sensitive Messaging Service

The cloud's flexibility and promise of seamless auto-scaling notwithstanding, its ability to meet service level objectives (SLOs) typically calls for some form of control in resource usage. This seemingly traditional problem gives rise to new challenges in a cloud setting, and in particular a subtle yet significant trade-off involving load-distribution decisions (the distribution of workload across available cloud resources to optimize performance), and rate limiting (the capping of individual workloads to prevent global over-commitment). This paper investigates that trade-off through the design and implementation of a real-time messaging system motivated by Internet-of-Things (IoT) applications, and demonstrates a solution capable of realizing an effective compromise. The paper's contributions are in both explicating the source of this trade-off, and in demonstrating a possible solution.

cs.NI↗

arXiv2020

Accurate phonon blockade detector composed of a quadratically coupled optomechanical system

The observation of phonon blockade in a nanomechanical oscillator is clear evidence of its quantum nature. However, it is still a severe challenge to measure the strong phonon blockade in an optomechanical system with effective nonlinear coupling. In this paper, we propose a theoretical proposal for detecting the phonon blockade effect in a quadratically coupled optomechanical system by exploiting supervised machine learning. The detected optical signals are injected into the neural network as the input, while the output is the mechanical equal-time second-order correlation. Our results show our scheme performs superior performance on detecting phonon blockade. Specifically, it is efficient for nonlinear coupling systems; it performs a high precision for strong photon blockade; it is robust against the small disturbance of system parameters. Our work opens a promising way to build a phonon blockade detector.

quant-ph↗

arXiv2020

Accurate Retinal Vessel Segmentation via Octave Convolution Neural Network

Retinal vessel segmentation is a crucial step in diagnosing and screening various diseases, including diabetes, ophthalmologic diseases, and cardiovascular diseases. In this paper, we propose an effective and efficient method for vessel segmentation in color fundus images using encoder-decoder based octave convolution networks. Compared with other convolution networks utilizing standard convolution for feature extraction, the proposed method utilizes octave convolutions and octave transposed convolutions for learning multiple-spatial-frequency features, thus can better capture retinal vasculatures with varying sizes and shapes. To provide the network the capability of learning how to decode multifrequency features, we extend octave convolution and propose a new operation named octave transposed convolution. A novel architecture of convolutional neural network, named as Octave UNet integrating both octave convolutions and octave transposed convolutions is proposed based on the encoder-decoder architecture of UNet, which can generate high resolution vessel segmentation in one single forward feeding without post-processing steps. Comprehensive experimental results demonstrate that the proposed Octave UNet outperforms the baseline UNet achieving better or comparable performance to the state-of-the-art methods with fast processing speed. Specifically, the proposed method achieves 0.9664 / 0.9713 / 0.9759 / 0.9698 accuracy, 0.8374 / 0.8664 / 0.8670 / 0.8076 sensitivity, 0.9790 / 0.9798 / 0.9840 / 0.9831 specificity, 0.8127 / 0.8191 / 0.8313 / 0.7963 F1 score, and 0.9835 / 0.9875 / 0.9905 / 0.9845 Area Under Receiver Operating Characteristic curve, on DRIVE, STARE, CHASE_DB1, and HRF datasets, respectively.

eess.IV↗

arXiv2020

Simultaneous blockade of a photon phonon, and magnon induced by a two-level atom

The hybrid microwave optomechanical-magnetic system has recently emerged as a promising candidate for coherent information processing because of the ultrastrong microwave photon-magnon coupling and the longlife of the magnon and phonon. As a quantum information processing device, the realization of a single excitation holds special meaning for the hybrid system. In this paper, we introduce a single two-level atom into the optomechanical-magnetic system and show that an unconventional blockade due to destructive interference cannot offer a blockade of both the photon and magnon. Meanwhile, under the condition of single excitation resonance, the blockade of the photon, phonon, and magnon can be achieved simultaneously even in a weak optomechanical region, but the phonon blockade still requires the cryogenic temperature condition.

quant-ph↗

arXiv2020

Entanglement enhanced and one-way steering in PT -symmetric cavity magnomechanics

We study creation of entanglement and quantum steering in a parity-time- (PT -) symmetric cavity magnomechanical system. There is magnetic dipole interaction between the cavity and photon-magnon, and there is also magnetostrictive interaction which is induced by the phononmagnon coupling in this system. By introducing blue-detuned driving microwave field to the system, the bipartite entanglement of the system with PT -symmetry is significantly enhanced versus the case in the conventional cavity magnomechanical systems (loss-loss systems). Moreover, the one-way quantum steering between magnon-phonon and photon-phonon modes can be obtained in the unbroken-PT -symmetric regime. The boundary of stability is demonstrated and this show that the steady-state solutions are more stable in the gain and loss systems. This work opens up a route to explore the characteristics of quantum entanglement and steering in magnomechanical systems, which might have potential applications in quantum state engineering and quantum information.

quant-ph↗

arXiv2020

The Molecular Clouds associated with the H II Regions/Candidates between $l=207.7^\circ$ and $l=211.7^\circ$

Using the PMO-13.7 m millimeter telescope at Delingha in China, we have conducted a large-scale simultaneous survey of $^{12}$CO, $^{13}$CO, and C$^{18}$O $J=1-0$ emission toward the sky region centered at $l$=$209.7^\circ$, $b$=$-$2.25$^\circ$ with a coverage of $4.0^\circ \times 4.5^\circ$. The majority of the emission in the region comes from the clouds with velocities lying in the range from $-$3 km s$^{-1}$ to 55 km s$^{-1}$, at kinematic distances from 0.5 kpc to 7.0 kpc. The molecular clouds in the region are concentrated into three velocity ranges. The molecular clouds associated with the ten H II regions/candidates are identified and their physical properties are presented. Massive stars are found within Sh2-280, Sh2-282, Sh2-283, and BFS54, and we suggest them to be the candidate excitation sources of the H II regions. The distributions of excitation temperature and line width with the projected distance from the center of H II region/candidate suggest that the majority of the ten H II regions/candidates and their associated molecular gas are three-dimensional structures, rather than two-dimensional structures.

astro-ph.GA↗

arXiv2020

Automatic Crack Detection on Road Pavements Using Encoder Decoder Architecture

Inspired by the development of deep learning in computer vision and object detection, the proposed algorithm considers an encoder-decoder architecture with hierarchical feature learning and dilated convolution, named U-Hierarchical Dilated Network (U-HDN), to perform crack detection in an end-to-end method. Crack characteristics with multiple context information are automatically able to learn and perform end-to-end crack detection. Then, a multi-dilation module embedded in an encoder-decoder architecture is proposed. The crack features of multiple context sizes can be integrated into the multi-dilation module by dilation convolution with different dilatation rates, which can obtain much more cracks information. Finally, the hierarchical feature learning module is designed to obtain a multi-scale features from the high to low-level convolutional layers, which are integrated to predict pixel-wise crack detection. Some experiments on public crack databases using 118 images were performed and the results were compared with those obtained with other methods on the same images. The results show that the proposed U-HDN method achieves high performance because it can extract and fuse different context sizes and different levels of feature maps than other algorithms.

cs.CV↗

arXiv2020

A Cold and Diffuse Giant Molecular Filament in the Region of $l=41^\circ$, $b=-1^\circ$

Data of $^{12}$CO/$^{13}$CO/C$^{18}$O $J=1\to0$ emission toward the Galactic plane region of $l=35^\circ$ to $45^\circ$ and $b= -5^\circ$ to $+5^\circ$ are available with the Milky Way Imaging Scroll Painting (MWISP) project. Using the data, we found a giant molecular filament (GMF) around $l\approx38\sim42^\circ$, $b\approx-3.5\sim0^\circ$, $V_{LSR} \approx 27 \sim 40$ km~s$^{-1}$, named the GMF MWISP G041-01. At a distance of 1.7 kpc, the GMF is about 160 pc long. With a median excitation temperature about 7.5 K and a median column density about $10^{21}$ cm$^{-2}$, this GMF is very cold and very diffuse compared to known GMFs. Using the morphology in the data cube, the GMF is divided into four components among which three show filamentary structure. Masses of the components are $ 10^3 \sim 10^4 M_\odot$, with a total mass for the whole filament being about $7\times10^4 M_\odot$ from the LTE method. $^{13}$CO cores inside each component are searched. Virial parameters are about 2.5 for these cores and have a power-law index of -0.34 against the mass. The mass fraction of dense cores traced by $^{13}$CO to the diffuse clouds traced by $^{12}$CO are about 7% for all components of the GMF. We found signatures of possible large scale filament-filament collision in the GMF.

astro-ph.GA↗

arXiv2020

Ensemble of Deep Convolutional Neural Networks for Automatic Pavement Crack Detection and Measurement

Automated pavement crack detection and measurement are important road issues. Agencies have to guarantee the improvement of road safety. Conventional crack detection and measurement algorithms can be extremely time-consuming and low efficiency. Therefore, recently, innovative algorithms have received increased attention from researchers. In this paper, we propose an ensemble of convolutional neural networks (without a pooling layer) based on probability fusion for automated pavement crack detection and measurement. Specifically, an ensemble of convolutional neural networks was employed to identify the structure of small cracks with raw images. Secondly, outputs of the individual convolutional neural network model for the ensemble were averaged to produce the final crack probability value of each pixel, which can obtain a predicted probability map. Finally, the predicted morphological features of the cracks were measured by using the skeleton extraction algorithm. To validate the proposed method, some experiments were performed on two public crack databases (CFD and AigleRN) and the results of the different state-of-the-art methods were compared. The experimental results show that the proposed method outperforms the other methods. For crack measurement, the crack length and width can be measure based on different crack types (complex, common, thin, and intersecting cracks.). The results show that the proposed algorithm can be effectively applied for crack measurement.

cs.CV↗

arXiv2019

The Column Density Structure of Orion A Depicted by N-PDF

We have conducted a large-field simultaneous survey of $^{12}$CO, $^{13}$CO, and C$^{18}$O $J=1-0$ emission toward the Orion A giant molecular cloud (GMC) with a sky coverage of $\sim$ 4.4 deg$^2$ using the PMO-13.7 m millimeter-wavelength telescope. We use the probability distribution function of the column density (N-PDF) to investigate the distribution of molecular hydrogen in the Orion A GMC. The H$_2$ column density, derived from the $^{13}$CO emission, of the GMC is dominated by log-normal distribution in the range from $\sim$4$\times10^{21}$ to $\sim$1.5$\times10^{23}$ cm$^{-2}$ with excesses both at the low-density and high-density ends. The excess of the low-density end is possibly caused by an extended and low-temperature ($\sim$10 K) component with velocities in the range of 5$-$8 km s$^{-1}$. Compared with the northern sub-regions, the southern sub-regions of the Orion A GMC contain less gas with column density in $N_{H_2} > 1.25\times 10^{22}\ \rm{cm}^{-2}$. The dispersions of the N-PDFs of the sub-regions are found to correlate with the evolutionary stages of the clouds across the Orion A GMC. The structure hierarchy of Orion A GMC is explored with the DENDROGRAM algorithm, and it is found that the GMC is composed of two branches. All structures except one in the tree have virial parameters less than 2, indicating self-gravity is important on the spatial scales from $\sim$0.3 to $\sim$4 pc. Although power-laws and departures from log-normal distributions are found at the high-density end of N-PDFs of active star-forming regions, the N-PDFs of structures in the Orion A GMC are predominantly log-normal on scales from R$\sim$0.4 to 4 pc.

astro-ph.GA↗

arXiv2019

Sparse recovery via nonconvex regularized $M$-estimators over $\ell_q$-balls

In this paper, we analyse the recovery properties of nonconvex regularized $M$-estimators, under the assumption that the true parameter is of soft sparsity. In the statistical aspect, we establish the recovery bound for any stationary point of the nonconvex regularized $M$-estimator, under restricted strong convexity and some regularity conditions on the loss function and the regularizer, respectively. In the algorithmic aspect, we slightly decompose the objective function and then solve the nonconvex optimization problem via the proximal gradient method, which is proved to achieve a linear convergence rate. In particular, we note that for commonly-used regularizers such as SCAD and MCP, a simpler decomposition is applicable thanks to our assumption on the regularizer, which helps to construct the estimator with better recovery performance. Finally, we demonstrate our theoretical consequences and the advantage of the assumption by several numerical experiments on the corrupted errors-in-variables linear regression model. Simulation results show remarkable consistency with our theory under high-dimensional scaling.

math.ST↗

arXiv2019

Sparse estimation via $\ell_q$ optimization method in high-dimensional linear regression

In this paper, we discuss the statistical properties of the $\ell_q$ optimization methods $(0<q\leq 1)$, including the $\ell_q$ minimization method and the $\ell_q$ regularization method, for estimating a sparse parameter from noisy observations in high-dimensional linear regression with either a deterministic or random design. For this purpose, we introduce a general $q$-restricted eigenvalue condition (REC) and provide its sufficient conditions in terms of several widely-used regularity conditions such as sparse eigenvalue condition, restricted isometry property, and mutual incoherence property. By virtue of the $q$-REC, we exhibit the stable recovery property of the $\ell_q$ optimization methods for either deterministic or random designs by showing that the $\ell_2$ recovery bound $O(ε^2)$ for the $\ell_q$ minimization method and the oracle inequality and $\ell_2$ recovery bound $O(λ^{\frac{2}{2-q}}s)$ for the $\ell_q$ regularization method hold respectively with high probability. The results in this paper are nonasymptotic and only assume the weak $q$-REC. The preliminary numerical results verify the established statistical property and demonstrate the advantages of the $\ell_q$ regularization method over some existing sparse optimization methods.

stat.ML↗

arXiv2019

Assessing the Performance of Molecular Gas Clump Identification Algorithms

The detection of clumps(cores) in molecular clouds is an important issue in sub-millimetre astronomy. However, the completeness of the identification and the accuracy of the returned parameters of the automated clump identification algorithms are still not clear by now. In this work, we test the performance and bias of the GaussClumps, ClumpFind, Fellwalker, Reinhold, and Dendrograms algorithms in identifying simulated clumps. By designing the simulated clumps with various sizes, peak brightness, and crowdedness, we investigate the characteristics of the algorithms and their performance. In the aspect of detection completeness, Fellwalker, Dendrograms, and Gaussclumps are the first, second, and third best algorithms, respectively. The numbers of correct identifications of the six algorithms gradually increase as the size and SNR of the simulated clumps increase and they decrease as the crowdedness increases. In the aspect of the accuracy of retrieved parameters, Fellwalker and Dendrograms exhibit better performance than the other algorithms. The average deviations in clump parameters for all algorithms gradually increase as the size and SNR of clumps increase. Most of the algorithms except Fellwalker exhibit significant deviation in extracting the total flux of clumps. Taken altogether, Fellwalker, Gaussclumps, and Dendrograms exhibit the best performance in detection completeness and extracting parameters. The deviation in virial parameter for the six algorithms is relatively low. When applying the six algorithms to the clump identification for the Rosette molecular cloud, ClumpFind1994, ClumpFind2006, Gaussclumps, Fellwalker, and Reinhold exhibit performance that is consistent with the results from the simulated test.

astro-ph.GA↗

arXiv2019

Convergence Analysis of Gradient Algorithms on Riemannian Manifolds Without Curvature Constraints and Application to Riemannian Mass

We study the convergence issue for the gradient algorithm (employing general step sizes) for optimization problems on general Riemannian manifolds (without curvature constraints). Under the assumption of the local convexity/quasi-convexity (resp. weak sharp minima), local/global convergence (resp. linear convergence) results are established. As an application, the linear convergence properties of the gradient algorithm employing the constant step sizes and the Armijo step sizes for finding the Riemannian $L^p$ ($p\in[1,+\infty)$) centers of mass are explored, respectively, which in particular extend and/or improve the corresponding results in \cite{Afsari2013}.

math.OC↗

arXiv2019

Phonon laser in a cavity magnomechanical system

The phonon analog of an optical laser has become the focus of research. We theoretically study phonon laser in a cavity magnomechanical system, which consist of a microwave cavity, a small ferromagnetic sphere and an uniform external bias magnetic field. This system can realize the phonon-magnon coupling and the cavity photon-magnon coupling via magnetostrictive interaction and magnetic dipole interaction respectively, the magnons are driven directly by a strong microwave field simultaneously. Frist, the intensity of driving magnetic field which can reach the threshold condition of phonon laser is given. Then, we demonstrate that the phonon laser can be well controlled by an adjustable external magnetic field without changing other parameters, which provides an additional degree of freedom compared to the phonon laser in optomechanical systems. Finally, with the experimentally feasible parameters, threshold power in our scheme is close to the case of optomechanical systems. Our results provide a theoretical basis for the realization of phonon lasers in magnomechanical systems.

cond-mat.mes-hall↗

arXiv2019

Ground-state cooling of an magnomechanical resonator induced by magnetic damping

Quantum manipulation of mechanical resonators has been widely applied in fundamental physics and quantum information processing. Among them, cooling the mechanical system to its quantum ground state is regarded as a key step. In this work, we propose a scheme which one can realize ground-state cooling of resonator in a cavity magnomechanical system. The system consists of a microwave cavity and a small ferromagnetic sphere, in which phonon-magnon coupling and cavity photon-magnon coupling can be achieved via magnetostrictive interaction and magnetic dipole interaction, respectively. After adiabatically eliminating the cavity mode, an effective Hamiltonian which consists of magnon and mechanical modes is obtained. Within experimentally feasible parameters, we demonstrate that the ground-state cooling of the magnomechanical resonator can be achieved by extra magnetic damping. Unlike optomechanical cooling, magnomechanical interaction is utilized to realize the cooling of resonators. We further illustrate the ground-state cooling can be effectively controlled by the external magnetic field.

quant-ph↗

arXiv2019

A Large-field J=1-0 Survey of CO and Its Isotopologues Toward the Cassiopeia A Supernova Remnant

We have conducted a large-field simultaneous survey of $^{12}$CO, $^{13}$CO, and C$^{18}$O $J=1-0$ emission toward the Cassiopeia A (Cas A) supernova remnant (SNR), which covers a sky area of $3.5^{\circ}\times3.1^{\circ}$. The Cas giant molecular cloud (GMC) mainly consists of three individual clouds with masses on the order of $10^4-10^5\ M_{\odot}$. The total mass derived from the $\rm{^{13}CO}$ emission of the GMC is 2.1$\times10^{5}\ M_{\odot}$ and is 9.5$\times10^5\ M_{\odot}$ from the $\rm{^{12}CO}$ emission. Two regions with broadened (6$-$7 km s$^{-1}$) or asymmetric $^{12}$CO line profiles are found in the vicinity (within a 10$'\times10'$ region) of the Cas A SNR, indicating possible interactions between the SNR and the GMC. Using the GAUSSCLUMPS algorithm, 547 $^{13}$CO clumps are identified in the GMC, 54$\%$ of which are supercritical (i.e. $α_{\rm{vir}}<2$). The mass spectrum of the molecular clumps follows a power-law distribution with an exponent of $-2.20$. The pixel-by-pixel column density of the GMC can be fitted with a log-normal probability distribution function (N-PDF). The median column density of molecular hydrogen in the GMC is $1.6\times10^{21}$ cm$^{-2}$ and half the mass of the GMC is contained in regions with H$_2$ column density lower than $3\times10^{21}$ cm$^{-2}$, which is well below the threshold of star formation. The distribution of the YSO candidates in the region shows no agglomeration.

astro-ph.GA↗

arXiv2019

Band gap and band alignment prediction of nitride based semiconductors using machine learning

Nitride has been drawing much attention due to its wide range of applications in optoelectronics and remains plenty of room for materials design and discovery. Here, a large set of nitrides have been designed, with their band gap and alignment being studied by first-principles calculations combined with machine learning. Band gap and band offset against wurtzite GaN accurately calculated by the combination of screened hybrid functional of HSE and DFT-PBE were used to train and test machine learning models. After comparison among different techniques of machine learning, when elemental properties are taken as features, support vector regression (SVR) with radial kernel performs best for predicting both band gap and band offset with prediction root mean square error (RMSE) of 0.298 eV and 0.183 eV, respectively. The former is within HSE calculation uncertainty and the latter is small enough to provide reliable predictions. Additionally, when band gap calculated by DFT-PBE was added into the feature space, band gap prediction RMSE decreases to 0.099 eV. Through a feature engineering algorithm, elemental feature space based band gap prediction RMSE further drops by around 0.005 eV and the relative importance of elemental properties for band gap prediction was revealed. Finally, band gap and band offset of all designed nitrides were predicted and two trends were noticed that as the number of cation types increases, band gap tends to narrow down while band offset tends to go up. The predicted results will be a useful guidance for precise investigation on nitride engineering.

cond-mat.mtrl-sci↗